Sunday 1 July 2018

Electronics - High frequency metal vapour arc lamp power supply

Changed; 22-04-2024, 19-04-2024

A reliable high-frequency ballast for a metal vapour arc lamp was made long ago.  It works similarly to a conventional 50Hz mains power source and choke but works at a high-frequency regulated AC current source instead.  It uses variable frequency to tune the L.C resonant circuit to create the high voltage necessary to strike the arc lamp.  This power supply is rated at 1A for a Spectral Lamp made by Osram. 

Images in Blogger are limited resolution.  Click or right-click to view the image with a good definition. 

Summary
I have given myself a design objective to improve the power supply efficiently and to update the parts used.  The Arc-lamp power supply design involves a lot of iterations of various models.  The requirement is unusual in that starting up voltage and power circulating in the power supply and the lamp is very high for a few seconds, then drops and runs at 150mA until the voltage falls further and the current rises to 1A until it fully warms up after several minutes.  After starting and warming up the unit needs to run efficiently at 10-60W depending on the spectral lamp type chosen. 


In this case, it was useful to get a functioning though inefficient, robust design to work to prove the design idea and the understanding of the lamp.  There were many design iterations that I have not included.  By using a newer MOSFET transistor type and the switch mode power supply type to a resonant mode the transistor losses should improve.  The power supply shuts down and restart if the lamp fails or is not fitted.

Conventional mains frequency ballast;
The metal vapour arc lamp arc is struck with a high voltage of 900V peak (GE fluorescent tube starter data sheet) and in the high-frequency lamp, a limit was set at 1.5KV peak which works reliably.  The noble gases ionize and begin conducting current.  In the case of Sodium (Na), lamp voltage quickly drops to 500V and then operates at 100V for a short time, heating up the metal until it becomes a vapour within the lamp.  During this short period, it is likely that the lamp unit will behave like a spark-gap transmitter.  The lamp voltage then falls to about 15V at 1A as the metal vapour conducts current and the lamp appears to have a resistive characteristic.

The mains lamp unit appears to re-strike each mains cycle 2x 50Hz so could cause emissions all the running time that can be observed with an oscilloscope or a spectrum analyser.  You can also observe some blue and flickering light the re-striking causes in a low-frequency ballast system.

I found this video on YouTube of a sodium vapour arc lamp starting.  You will see 
some blue and other colours around the electrodes plus flicker and fluctuation in this 
conventional mains power ballast when it is fully warmed up and running normally. 

When warmed up and running normally blue and other colours plus flicker and fluctuation are much less evident in the fully warmed up high-frequency metal vapour-arc-lamp power supply.  The circuit below is different from modern electronic ballast lighting in that the lamp's light is constant and regulated, that is, the current waveform is not modulated by rectified AC mains voltage but by 2x 50 KHz or whatever the running frequency which makes this sodium arc lamp in the video behave as a resistive load.  So it will display no doubt flicker still but with fewer other colours because other noble gases used to start the lamp are evidently not activated.

Starting with guesswork then developing a sound theory leads to a robust solution;
The starting and running characteristics were all discovered by observation and experimentation, as described above, but with a range of different metal arc lamps.  The data sheet on the lamps has little information.  Because the lamp voltages range from 10 to 60V the current range is at least +-12% with other tolerances = 100% * 50V/240V + other tolerances.  I set the starting current low but high enough to cause the lamp to warm up. 

The concluded design is a simpler solution.  This is often the way the thing turns out that the final simplicity conceals the effort necessary to achieve it.

Heath Robinson - drew cartoons of elaborate machines but engineers should make machines 
to function well in an uncomplicated way from a range of imprecise materials and parts.  I 
think Heath Robinson showed in fiction a desire for complicated machines as toys. 

It is reasonable to start a project and expect it to first be overly complicated but this does not have to be so and I have constrained my designs from the start.  In due course, the design and the function should become simpler and plainer for the user to use.  I have observed designs that present too many complications that are not liked by the customers and the company eventually simplifies the design so it is more like an earlier model but the problem is for the company to recognise the problem.

It is not an uncommon problem for small companies and individuals to have difficulty in getting support from big companies on products those companies make.  Many parts made by big companies may not be available to other customers but it can be worth asking about them.  But also many very big companies are particularly supportive of small companies and individuals.  There has been a positive change since the 1990s with big companies helping small companies and individuals much more although some never stopped and have always worked that way.

Spectral-lamp
Laboratory spectral lamp unit

Half-bridge - switch mode vapour arc-lamp power supply;

Based on a successful design but with changes and using newer parts than were fitted in the actual lamp power supply in the early 1990s.  The lamp and power supply are enclosed within a metal box with metallic contact on all edges to the lid.  There is an aperture for the light to exit - this provided the lowest emissions in conducted EMC test up to 30 MHz at that time.

Since working on this project I have worked for automotive contractors one of which had a test chamber, LISN (line impedance stabilisation network) and a large field test site.  It is possible to use a LISN and analyser up to 150 MHz - although uncalibrated this will make initial testing more practical.  Using an oscilloscope is an essential first step to look at the diode overshoot which is the diode turn-on time and may cause EMI.  Measure the width of the overshoot if it is say 10 ns (half a cycle) then any related to the diode emission is likely about 50 MHz.

The most significant thing about using a LISN is that you do not just measure conducted emissions but you also get a good indication of potential radiated emissions without needing to leave your workspace and equipment. 

Some of the spectral lamps could behave as a spark gap transmitter.  This is why the output choke should be a low Q filter type but R13 also achieves a similarly low Q.
  The mains power supply AL-0026-01D (6).  The power supply has components rated for 200W 
or more briefly although the lamp power is much less when it is warmed up.  C4 was fitted across
 the source current sense resistor on the PCB but on reflection, this may be counterproductive to 
fit if it extends the switching moment pulse outside any dead-band filter included in the controller? 
  • C4 - slightly compromises the dead-band turn-on noise filter but beneficially should reduce the chance of high voltage transients in the Source pin compromising the gate driver.   In any case, the driver chosen is much more robust and can withstand 100V <300nS than the original driver IR2111 which had lower over-voltage ratings.  Even so, it was adequately robust.  
This circuit is a little different and is not finalised but the original was working well and reliably for a decade of daily use. 
  • The circuit above does not work over the full voltage range and lamp power rating.  But the regulation will cease if the mains voltage drops to 150Vac depending on the maximum power 60V Helium lamp.  But from the models below the circuit may continue to work down to 100 Vac if the power supply were to run at close to but just a little above the resonant frequency.  That is although the Q of the resonant circuit would be low it would still create a higher voltage than the incoming supply voltage.  Operating slightly above resonant frequency so that the transistors switch on at zero inductor voltage and therefore more efficiently.  That is the high currents seen in the models below do not occur.
  • The Status LEDs for starting could be supplemented with a status LED for in-regulation.  See controller diagram sheet 5 in pdf; AL-0026-06A.pdf
C16 & L2 form a tuned circuit to create the high voltage necessary to strike the vapour arc lamp's arc.  The frequency first reduces to below resonant frequency of 50 KHz to maintain about 150mA until the frequency drops to 50 KHz.  The circuit then runs at 50 KHz and the current in the lamp current increases until it is controlled at 1A by pulse width modulation.  A change that needs to be made is that the frequency needs to settle operating just above resonance similarly to the ballast controller ICs further below.

Revised for 50 KHz operation a smaller inductor L2 (100uH) has been selected.  The resonant frequency for striking the arc is now 87 KHz (C12=33nF) and L1 protects the circuit from damage when the arc is struck.  Note that although L2 is physically smaller and the originals got hot, the manufacturer's models are very accurate, although the working voltage of 2,000V may be too high for the part it should be possible to evaluate with that part then have a special choke with better interwinding insulation.  The change to this strategy would be to reduce the frequency ramping or not use the frequency ramping for starting but keep the starting circuit in place to accommodate choke inductance changed with power throughput. 
 
A fairly large difference between the starting high frequency and lower running frequency was designed in, to moderate the starting up power in the lamp which in any case was still high at roughly 75W = 150mA * 500V or 150W = 1A * 150V for a 15W rated sodium lamp.
 
The controller and half-bridge have an added high lamp voltage sense to reduce the power as the lamp voltage increases.  There is also a Zener diode added to allow 10% current when the lamp strikes and first starts to warm up.  This still might be an excessive current? See controller and half-bridge sheets; AL-0026-06A.pdf

The current mode switch mode power supply controller works well in this simple current output design providing some of the current regulation, particularly for mains 100 Hz ripple on the power supply.  There is a conventional slower outer control to set the lamp current accurately that responds as fast as is practical.  The operation is unsymmetrical the negative pulse width is shorter, higher and set by the controller, and the longer positive pulse completes the remainder of the period of oscillation.  The controller also includes over-voltage shutdown in case the lamp is an open circuit such as if it is not fitted.  The over-voltage is sampled on the positive peak and the limit is set at about 1,500V peak, the peak voltage will be a little higher than this, so the capacitors are rated at 2,000V but the inductor's voltage rating would need to be confirmed.  The lamp's light output and life appear to be very good is more uniform light noticeably at each electrode where you see fluctuation and ran reliably daily for at least a decade.

The circuit operating frequency ramps through L2 & C16's resonance which causes the high voltage necessary for the lamp to strike which is more like how a leakage transformer rather than a fluorescent tube starter operates.  Otherwise, the circuit operates in a similar way to how the conventional 50Hz choke ballast above works.  When the lamp is at full temperature C12's purpose is mostly complete although it may provide some filtering of RF emissions.  L2 provides a 1A average current source for the lamp.  Spectral Lamps require 10 to 60V at 1A RMS depending on the lamp metal.  The lamp was plugged into a ceramic B9A valve base, called PICO 9.

Using average current instead of true RMS current introduces an error.  The current waveforms are a mixture of rounded sine, triangle, sawtooth and square waves but using a multiplication factor of 1.1 would be about right and is accurate as is required.

So for an; (RMS/Average)
Sine wave; 1.11 = 0.707/0.635
Square wave; 1.0   = 1.0/1.0
Triangle or sawtooth; 1.15 = 0.577/0.5

Multiply the average value by 1.1 to approximate the RMS value required.


Low Voltage Power Supply

AL-0026-01D (4) low voltage power supply.  Just enough power will be supplied from the current sensing when the lamp is drawing the rated 1A to run the low-voltage circuit. 

A mains voltage 2-3mA PTC could not be found now so a higher power rating transistor has been selected instead.  
  • Raising the value of R23 to 47K so that if the lamp does not start the power supply shuts down may be practical but the value will need to be experimented to be found? The gate driver's shutdown pin is wired in but even so, the operation needs to be checked?
  • R23, 22K should provide 2mA at 100V, enough to start for 50 seconds but not run.
This part of the design is not fully optimised.  N2 is the neutral half-wave filtered mains voltage input therefore the power wasted is better than halved at <1.4W in the start-up condition. 


Assessment of the original arc-lamp power supply;
The power consumption of the whole power supply was fairly high but what mattered was that the lamp was an excellent source of flicker-free and low-ripple light.  The main inductor which is about 40 x 40 x 40mm depending on which was used ran hot or very hot depending on which one was used.  The transistors were each mounted to PCB mounting heat sinks rated at about 15'C/W each and they ran cool.
  • Increase the size of the inductor or increase the frequency of operation to reduce power loss or at least to reduce the inductor's temperature rise.
  • But regulations were coming in that required electrical equipment to have power factor correction if the power was greater than 75W.  In addition, sodium street lamps if only sodium light was required are better priced and lamp life with this option is longer but street lamps use more power and produce more light than a sodium Na Spectral lamp. {STM data sheet says this amounts to 25W maximum lamp - which varies with other advice}.
  • Consequently, if the full range of Spectral Lamps are supported then provided the power supply's efficiency is improved then no power factor correction circuit needs to be added.  That is because the highest power-rated spectral lamp is Helium (HE) which is rated at 60W.
  • The operating frequency has been increased to 50KHz and the starting frequency resonance of 70KHz to start at up to 100KHz. 

Earlier designs often destroyed transistors at start-up;
When the arc was struck the output transistors were often destroyed.  My design assumption was that this would not happen because there would be enough stray inductance for the C16 discharge current to be limited the current adequately was wrong.  Secondly, the circuit for the arc-lamp current path was far from the power transistors and I could not see a path on the PCB that was nearby either to explain this.  Although the current sense transformer did carry the high discharge current nothing was damaged in that circuit's path.  L1 was added although in that location about 4u7H or 220mR wire wound resistor adequately resolved the issue -  for such a small impedance to be significant then evidently the current pulse was very high and fast.

Op-amp unstable;
The dual op-amp in the controller section could be unstable due to the electromagnetic emissions from the power inductor.  A workaround is to add a 100uF or more likely lower capacitance to the output - so that it is overcompensating.  The data sheet recommends loading the output with a pull-up resistor to draw more current.  A capacitor is shown on the circuit diagram.

Another option is to also place the Op-amp some distance from the power parts and under the PCB.  Hopefully, the screening will be adequate so that the amplifier does not need to be placed on the bottom layer of the PCB?  The design has been changed to an SMT design which may also allow us to separate the input inductive filter from the output choke by distance plus better screening using 0V and power planes 5V1+, or Vc+.
 
Op-amps and other circuits within switch mode power supply controllers intentionally have greater charge storage at the pins I read recently. 

Power Input filter;
The input filter circuit with the safe X and Y class capacitors has changed a little since about 1985 other than to become required standard practice to fit rather than being optional.  Some of the parts shown will also need to be reviewed.
Input filtering AL-0026-01D (3) - there is only a single point connection 0V to Earth connection although bonding at 
many points would be better - this cannot be done for safety reasons one single point RF  connection is achieved by 
commonly available safety-approved type Y-class capacitor.  In any case, the star point strategy works fine with care.

The input filter circuit is compromised for safety in that there is a single bond 0V to Earth via the Y-class capacitor.  I have considered a more efficient design that includes an output transformer but I have not pursued that further.  With that design one of the lamp supply voltage connections is connected to Earth for potentially lower Electromagnetic emissions (EMI).

The Earth and Neutral connections have two resistors in a series to discharge C3, the discharge would be safe but is likely to cause complaint.   The addition of these resistors does compromise safety and the particular types chosen should be checked for their suitability.   Without those resistors, the discharge energy is safely within the standards and those resistors could be omitted. 

AL-0026-01D (5) control circuit.  This circuit is wrong or has incorrect component values. 
The Red LED flashes during starting and is stead on when warming up and also running.
  • The left circuit creates the slow voltage ramps that modulate the Switch mode power supply oscillator frequency.  When the lamp is struck and 150mA of lamp current is reached then the ramp stops and the frequency drops until the lowest frequency about 50KHz is reached and then the current rises to 1A and the pulse width modulator starts to control the current.  When the lamp current has risen to 1A there is enough current from the current transformer to run the low-voltage circuit so the wastage from the mains ceases. 
  • The right-hand circuit is the switch mode power supply circuit.  The IC chosen is a low-power version of a 1980s Unitrode Current mode controller that was stable in this application even without a small element of Voltage mode control added. 

ALTERNATIVELY;  (simplification)

  • Select components that will resonate at about the operating frequency.  The supply voltage can be set high enough that the Q can be low and the LC mis-tuned but still attain 1,500V to strike the arc.  
    • If the chokes value reduces by, say, 20% then the frequency could be set 10% higher in future the variable frequency starting circuit might be removed?
  • Modelled below with L2 = 100uH then C12 = 100nF for 50KHz.  This circuit can run with 100Vdc supply even though the lamp voltage is 120V - 150V pk-pk.
  • This alternative could be tried by changing C12 and removing D12.  This will likely work even with a lower voltage supply but in any case, PFC that will boost the supply voltage is required.
A different controller with many more functions integrated could be used such as; TEA19161T/2,  TEA2016AAT these use a half-bridge driving into a transformer with leakage inductance so a PFC section may not be required. 
  • Burst mode would need to be deactivated.  I have not investigated if these ICs will work, they work in voltage mode but they are efficient.
AL-0026-06A.pdf does not yet incorporate all the changes but does show a different method of current reduction when the lamp is starting.  Another change might be to start the lamp at slightly lower frequency than the running frequency.  A better strategy would be to use a fluorescent tube/arc lamp high frequency ballast controller.
AL-0026-01D (2) Block diagram. 

It is most likely that the running power consumption will be up to 65W but if it turns out to be 70W such as with a He. lamp then a PFC will need to be added in case 75W were exceeded.  Most lamps though are lower powered and the whole unit's power should be below 20W with many lamps such as Sodium Na.   It depends on how much better efficiency newer chokes turn out to be compared to what was available or could be wound in the 1990s?  Some of the chokes modelled further down in this blog are much better.

Notes Mods and ToDos AL-0026-01D.  PDF of all pages of the circuit design.
 
------------------------------------------------------------------------- 
  
Developing and refining this 1990's design

The circuit is very similar to the working original circuit.  There are several incorrect parts such as connectors put down and develop the ideas without being distracted by detail.  Polypropylene and Polycarbonate power capacitors will need to be changed to Greener RoHS-compliant types.
  • There was an earlier stage that amounted to play that was very important.
  • A step-up high leakage, high frequency transformer did not work well I struggled with the design which a Mullard field applications engineer suggested and supported me with. 
  • An earlier working design included a full-bridge 24V DC, which worked well. 
  • A mains voltage full bridge also worked well but was unnecessarily complicated.  I did a lot of work with this design.

  Revision;
 
Output transistors - first look;
The original circuit used IRF841 transistors plus PCB-mounted heat sinks each for the half-bridge output.  Although the newer transistors below also include an avalanche diode and the diode has a reasonable switching time the switching losses could be a significant factor.  That original circuit had a transistor current limit of 4A.
  • A rough estimate for the candidate's new transistor power can be done using Ron.
  • If the transistors are soldered to PCB with no additional heat-sinking then the package dissipation is 62'C/W.  That figure depends a lot on the mounting orientation of the PCB.
The top transistor's Drain is soldered to an area of copper for additional heat sinking.  I have included a lot of decoupling to minimise EMI radiated by this pad which is connected to the positive supply.  As discussed below a PTC thermistor to sense over temperature and shut down the controller could be placed in this area.

The bottom transistor's Drain is soldered to an area of copper for additional heat sinking.  However, the copper area will cause EMI to radiate and it will be necessary to experiment with snubbing and switching time to minimise EMI adequately.  The bottom transistor conducts high and shorter current pulses (although the maximum current is not as high as might seem at first).

https://electronics.stackexchange.com/questions/26783/how-do-i-determine-the-area-of-copper-needed-on-a-pcb-to-provide-adequate-heatsi?utm_medium=organic&utm_source=google_rich_qa&utm_campaign=google_rich_qa 

This example with 50mm x 50mm on one side of the board but with a smaller transistor would improve the thermal cost-efficiency to better than 35 'C/W.

----- Return to the transistor power selection later -----

Simulation to show basic circuit operation;
Simulation adds a step but gives me more insight into the operation of the circuit before proceeding to have a PCB made.  This model will confirm or remind me of assumptions and calculations made previously. 
AL-0028-07B, SiMetrix 9.1 simulations of the output choke/filter and the lamp represented by the 
bridge rectifier and resistor in this case and a voltage clamp in other cases.  Top right is the current
transformer monitors average current as an approximation to RMS current. 
  • The current transformer is; 1:150, therefore R2 = 150R gives a 1V per amp average but the waveform is not square, triangular, sawtooth or other but complex and I have not been able to estimate the RMS current using the power measurement probe in SiMetrix.  However, I am hoping to find a value that is near enough using the average value scaled.  The mains frequency choke ballast would have no regulator but for a 110VAC and 230VAC switch, so the higher power lamps probably should be running at a lower current.
AL-0028-07B lamp; 15V 1A into 1 ohm.
    • The Voltage offset is due to the duty cycle 96% was adjusted until the power looked about right at 15W.
Waveforms are ripply because the circuit includes capacitors with significant reactance. 

Purple - Filter voltage.  The coupling capacitor's voltage is 200V and a 500V part may be substituted.
I-sens = current transformer rectified voltage (across 150R), 0.8V.
Brown - I-Probe Drive Transistors, +1A to -2.5A
Blue - Lamp Voltage, ~11V.
Red - Rectified Average lamp voltage (12V).
Green - Power in load resistor R1, difficult to determine ~15W? 
---------------------------------------------------------

AL-0028-08B is a similar circuit as -07? but the lamp voltage is; 
60V and R1 = 60R, the highest power lamp.

Purple - Filter Voltage.
I-sens = current transformer rectified voltage across 150R, 1V/Amp average.
Brown/Green Probe 1 Mauve = Transistors Current, +2.5A -4.5A.
 Green - Lamp Voltage, 60V?
Green - Power in output load R1.  Looks very high >100W but that looks wrong.
Red = Rectified smoothed Lamp-V R1, 60 ohms, 60V. 
      • The duty cycle is 87% 
  • Simulation of an arc lamp warming up assuming the lamp is behaving resistively.
-----------------------------------------------------------------------------

AL-0028-09B 60V 1A from 110VDC minimum V+, duty cycle 50%.  Same circuit as version -07?
With a 220uF reservoir capacitor, 90Vac, 60 Hz minimum is required.

Purple - Filter Voltage.
I-sens = current transformer rectified voltage. 0.9V
Probe 1 Brown = Transistors Current, +2A -2A. 
 Green - Lamp Voltage.
Green - Power in load R1,  >60W?
Red = Rectified smoothed Lamp-V,  60 ohms.
  • The power supply is running flat out therefore there is no regulation.  So anticipate double mains frequency fluctuation in the light output.  So double mains frequency fluctuation in the light output is to be anticipated. 
    • This issue can be resolved by adding a Power Factor Controller (PFC) or a voltage doubler rectifier switch.
A resistor may need to be added across L2 to stop it from ringing, the part should not be fitted unless it is found necessary.  C1 when reduced caused the peak current to increase from 4-5A to 7-8A, that is its reactance is significant and C1 was increased from 220nF to 470nF. 
  • AL-0028-07A Graph  Old version C1=220nF, 15V 1A, 400Vdc - peak current is okay.
  • AL-0028-08A Graph  Old version C1=220nF, 60V 1A, 400Vdc - high peak current.
  • AL-0028-09A Graph  Old version C1=220nF, 60V 1A, 100Vdc - peak current is okay.
  • Status LEDs have been added to this revised power supply particularly relevant is the controller page See pdf; AL -0026-06A
Q of L2 an RF filter choke should be quite low.  In any case, a soft magnetic should be chosen which has a low Q because when the lamp starts and is warming up the lamp is likely to behave like a spark-gap transmitter.  But when it is warmed up it should behave resistively and tests tend to confirm this.
  • Using average current rather than the RMS current worked well and is about right for an approximate triangle wave.


If the higher-power lamps operate as an arc when fully warmed up;

If the higher voltage lamps operate as an arc rather than a conducting metal as the Sodium lamp does, which is most likely,  then the circuit will work regulated at low voltage, with high circulating current but without PFC.


L1=100uH, C1=470nF, 60V 1A;
  • And for a 15V lamp at 400Vdc +1A to -2.5A (AL-0028-07B above)

Consider adding Power Factor correction to improve regulation and efficiency over the working voltage;

AL-0053-01A High-Frequency Arc Lamp power supply with PFC
Power Factor correction function added. 

This PFC would always provide a higher DC voltage than the rectified mains peak-to-peak voltage whereas the configuration below can provide more than or less than the peak-to-peak mains input voltage.  The voltage out could be variable based on the measured lamp voltage plus the reservoir capacitor minimum voltage but always less than 450Vdc. 
AL-0053-01A (3) Power Factor Correction buck-boost (Full Bridge).  Arbitrary value components are shown.
Provides a high enough voltage to power a lamp without excessive voltage that would be inefficient.  About 30Vdc to 150Vdc + Reservoir capacitor ripple amplitude.  The drawback is that I have not found a suitable controller and using logic to switch modes between buck and boost would put a little glitch in the current drawn.  The glitch could possibly be minimised by using a current mode controller.

The risk is that current mode controllers often have a dead period to prevent current glitches from upsetting the control.

A current mode PFC may satisfactorily transition between buck and boost more smoothly; L4984D, L6562A, TEA19162.

Buck-boost PFC control strategy draft;
If the input voltage is greater than the output voltage then the PFC works in buck mode. Otherwise the PFC works in boost mode.

  • Buck mode - the top transistor is switched by the controller.  The bottom transistor is switched off.
  • Boost mode - the bottom transistor is switched by the controller.  Ideally, the top transistor should be switched on but it will need to be switched off and then on at the beginning of the PWM on cycle to recharge the gate driver boost power supply.


The lamp power supply - considering the modelling further above;
  • Consider removing the power from the current transformer and using power from PFC for low-voltage sections. 
  • Changed output L, C & C to 470uH, 100nF + 6n8F.  As, AL-0028-07A,  AL-0028-08A, AL-0028-09A.
  • Reduce the size of the reservoir capacitor to; 50uF.  Remove the choke that is between the rectifier and the half-bridge output. 
  • This change reduces the peak current in the transistors and improves the light output regulation. 
  • The poor regulation warning circuit and RED LED is not required.  There is always adequate power for the regulation to be good, when everything is in order, with the PFC option included. 

AL-0028-10A 60V 1A from >=400VDC PFC,  duty cycle 72%.  Same basic circuit as -07? 
The inductor is larger and more expensive but should result in less power loss. 470uH, 100nF & 6n8F.  Starting Resonance 87KHz.

Purple - Filter Voltage.
I-sens = current transformer rectified voltage. 1V
Probe 1 Brown = Transistors Current, +2A -2.5A. 
Green - Lamp Voltage. 
Green - Power in load R1,  >60W?
Red = Rectified smoothed Lamp-V,  60 ohms.


Modelling a switch-mode power supply without using simulation tools is also relevant;
An inverter needs to be placed between the output and the gate driver because the controlled current is sensed in the bottom transistor.  There is an inverted output controller UC3847 that was used before but it consumes more power at up to 21mA.   

Consequently using an estimation of minimum on-time based on square law to consider increasing the operating frequency is a useful modelling exercise.  The next output transition (minimum PWM on time) is determined by the time between the current flowing into the substrate diode in that transistor (current flowing out of the inductor but not conducted by the MOSFET) must be at least tRR (~200nS).

Conclusions for the simulation of general operation above;

Lowering the Q of L2 by adding R3 to reduce ringing when the lamp is behaving as an arc should reduce RF emissions.

It is unlikely that this extra design step would be assisted in the design process application notes generally do a good job of explaining standard topologies.  But modelling the power in the transistors is likely to be beneficial - I had been advised that would be the case!

      Simulation for power supply running at near L.C resonance;

      The simulation below now considers the switching power but does not include changes with temperature increase.  It will, therefore, underestimate power consumption because Ron increases with temperature rise so, therefore, the power dissipated increases with temperature. 
       
      This simulation does not take account of the change in resonant frequency due to inductor L1's inductance reduction with power throughput.  The operating frequency probably needs to be increased by 10% if the inductor value drops by 20%.
       
      AL-0054-09A - Lower Ron transistors and a higher current gate 
      driver are used.  400Vdc minimum from a PFC boost section.

      LTspice's available parts were used mainly because making them is fiddly hence I've used the library 20:1 current transformer.  This is a new more comprehensive simulation using LTspice 24.x - replacing the old SiMetrix model does not run now.  The power MOSFETs were created using the LTspice nmos part and changing the model to the manufacturer's model. 

      LTspice alternate solver was used and the simulation was 800uS to 1mS for measurements, with the maximum time step 2-5ns.  Ton is adjusted until Vav = ~0.9V

      Gate resistors increased to 100R, diode gate clamp was added this value is almost certainly too high for efficient operation.  A snubber is required even so the peak currents and power losses without one are okay.

      The .meas directive should give 100% accurate results I am told.  But have not found how to use .meas but I do not need to use it.  I have made more mistakes, resulting in wrong results, as a consequence of reading LTspice instructions.  They are clear but often short of the required detail.
       
      The new snubber values 220pF and 22R have not been modelled below but there would be a small increase in the peak currents in some cases.

      AL-0054-05B 400Vdc, L1 = 100uH, C3 = 33nF, C2 =470nF, 60V Zener load

      Power measurements; 

      Isense 20R = 80mW,  Vav = 935mV,  Pulse width; 1.55us in 20us. 

      Gate driver; U1 = 22mW,  Top Transistor; M1 = 628mW,  Bottom Transistor; M2 = 2.6W. 

      Currents; IL1 = 4.6A to 2.1A,  MOSFET currents are 12.2A to 12.9A 

      AL-0054-06B 15 Ohm resistor load.  400Vdc,  L1 = 100uH, C3 = 33nF, C2 =470nF.

      Power measurements; 

      Isense 20R = 46.9mW, Vav = 882mV.  Pulse width; 1.0us in 20us. Lamp 15W

      Power; gate driver; U1 =  21mW,  Top Transistor; M1 = 59mW,  Bottom Transistor; M2 = 301mW. 

       Currents; IL1 = 2.9A to 1.16A,  MOSFET currents are;  +-2.9A

      AL-0054-07B  150Vdc. L1 = 100uH, C3 = 33nF, C2 =470nF, 60V Zenner load

      Power measurements; 

      Isense 20R = 69mW and 59mWrms, Vav = 886mV.  Pulse width; 3.75us in 20us.

      Power; gate driver; U1 =  21mW,  Top Transistor; M1 = 1.0W,  Bottom Transistor; M2 = 421W. 

      Currents; IL1 = 3.3A to 2.0A,  MOSFET currents are 11A and -12A. 

      AL-0054-08B Into 60V zener diode load, 220uH + 470nF + 15nF, 400Vdc.

      Power measurements; 

      Isense 20R = 71mW, Vav = 923mV.  Pulse width; 2.5us in 20us.

      Power; gate driver; U1 =  21mW,  Top Transistor; M1 = 134mW,  Bottom Transistor; M2 = 400mW. 

      Currents; IL1 = 1.1A to 3.2A,  MOSFET currents +-3.1A. 

      Maximum power (50% duty model below);

      • With 47uH at 100V and 60R does not achieve the required power predictably and actually. but with 100uH and 60V Zener diode load does achieve the power.  It may be possible to run this power level without a boost PFC.
      • With 470uH at 400V and 60R does achieve the required power and the inductor could be reduced to 330uH to ensure there is adequate power.

      AL-0054-09A Into 60W resistive load, 470uH + 470nF + 6.8nF, 400Vdc.

      Power measurements;  

      Isense 20R = 56mW, Vav = 942mV.  Pulse width; 2.5us in 20us. 

      Power; gate driver; U1 =  21mW,  Top Transistor; M1 = 160mW,  Bottom Transistor; M2 = 228mW.  

      Currents; IL1 = 1.6A to 2.0A,  MOSFET currents -2.0A. 

      The required Vav = 900mV is achieved with a pulse width on time of; 7us.

      Conclusion power measurement and transistor efficiency;

      LTspice at least gives a power estimate but it looks like it is wrong because the average value Vav varies with the power in R4.  The PWM was set by measuring the Vav repeatedly until I get about 900mV.

      On the other hand, SiMetrix models AL-0026-04b and -05b do not work since SiMetrix has been updated since I made those models.  In any case, those models did not show the mean of the squares or products. 

      I have not taken the precautions mentioned in Infineon; DT-97 3 art "Managing Transients in Control IC Driven Power Stages", Method B. In any case, the original circuit was fine and the newer drivers are more robust.

      By using a lower Ron MOSFET the half-bridge becomes more efficient, confirming previous measurements and estimates.

      Because LTspice is not limited I used the gate driver model rather than a simplified model using current and voltage sources.
      • The model for the metal vapour arc lamp is changed to simply a resistor or two Zener diodes in series back to back. 
      • The objective is for the RMS current in the lamp to be 1A (+-13% I guess) which occurs when 50mW is dissipated in the sense resistor.
      • From above the average current value is a rounded mixture of sine, triangle, sawtooth and square waves and using a multiplication factor of 1.1 would be about as accurate as is required. Unfortunately, modelling does not confirm this and there is a big discrepancy in measuring the power in the sense resistor and using the Vav to set the lamp Irms.
            The lower Ron IPD60R280P7 MOSFET gives better efficiency.  But the higher current driver 2ED2184 improvement was marginal.  Both have been incorporated into the circuits above.  IPD50R280P7 would introduce greater switching losses but should be an acceptable alternative.

            Adding Schottky diodes to the gate drive did not reduce the exceptionally high gate pulses in some cases, of 20A (which is okay) but with the increase in Rgate to 100R the peak current was reduced to an acceptable level though sacrifices efficiency.  A snubber is still required though. 

            -----------------------------------------------------------------
            Evaluation of reducing the start-up frequency ramping frequency so that the ramp is enough to  bring about resonance over the range of inductor change with power through put.

            Although 150mA lamp current during starting was found adequate the method of current limiting is changed to sensing the higher lamp voltage and reducing the current proportionally but not down to zero.  This is similar to how a traditional ballast works and will surely result in shorter lamp life similar to what is quoted for the lamps.  See alternative circuit AL-0026-06A, ramp frequency for starting is still interoperated but it needs to be changed.

            I have run such a circuit from a 48V power supply and it started a sodium spectral lamp and ran fine but that was at 20KHz with a starting frequency of 50 or 70KHz?  That was using a full bridge stepper motor BIFET driver IC.  In other words, this circuit should work. 

             The circuit below is the same as above but for component values which are for operating at about resonant frequency.
             
            AL-0054-04F The operating frequency is the same as the tuned frequency of L1, C3, 
            (100uH, and 100nF, 50KHz). 

            LTspice alternate engine was used and the simulation results were taken between; 800uS to 1mS.  And the maximum time step is set to 2ns to 4ns.

            AL-0054-01E 100Vdc, L1 = 100uH, C3 = 100nF, C2 =470nF, 60V resistor load
            Power Measurements; Note no change from Zener load to a resistor was applied.

            Isense 20R = 59mW, Vav = 979mV.  Pulse width; 6.0us in 20us.

            Power; gate driver; U1 = 21mW,  Top Transistor; M1 = 754mW, Bottom Transistor; M2 = 972mW.

            Currents; IL1 = 3.2A to 4.2A,  MOSFET currents 4.2A and 4.9A

            AL-0054-02E  Into 15W resistive load. 100uH 100nF, 100Vdc.
            Power Measurements;

            Isense 20R = 58mW, Vav = 966mV.  Pulse width; 5.6u in 20uS. Lamp 19W

            Power; gate driver; U1 = 24mW,  Top Transistor; M1 = 200mW, Bottom Transistor; M2 = 414mW.

            Currents; IL1 = 2.8A to 1.9A,  MOSFET currents; 2.7A and 2.7A. 

              AL-0054-03F  Into 60W resistive load. 100uH 100nF, 100Vdc. 

            Power Measurements;

            Isense 20R = 100mW, Vav = 1.0V.  Pulse width; 4.5u in 20uS.

            Power; gate driver; U1 = 21mW,  Top Transistor; M1 = 631mW, Bottom Transistor; M2 = 1.6W

             Currents; IL1 = 3.2A to 5.5A,  MOSFET currents are -5.3A -5.5A. 

            AL-0054-04F  Into 60W Zener clamp load. 100Vdc {L.C = 100uH 100nF}
            Power Measurements;

            Isense 20R = 80mW, Vav = 941mV.  Pulse width; 6u in 20uS.

            Power; gate driver; U1 = 21mW,  Top Transistor; M1 = 612mW, Bottom Transistor; M2 = 1.1W

            Currents; IL1 = 4.3A and 3.2A,  MOSFET currents -4.8A and -4.2A. 

            Conclusion for operating at about  LC resonate frequency;

            As before, LTspice at least gives power estimates that look wrong.  The power in R4 and the average voltage vary widely, they should agree well when the arithmetic is done. 
            • The objective is for the RMS current in the lamp to be 1A (at least +-13%) which occurs when 50mW is dissipated in the sense resistor. 
            • The high circulating current could be reduced therefore efficiency improved by fitting a snubber.  The snubber should in any case be fitted.  The gate resistor 100R is almost certainly too high although it works well.
            • The efficiency could be improved by using a buck-boost PFC controller.   Although some of the chokes modelled below look much better than the working earlier power supply.  That is a good efficiency that should be expected. 
            • The current mode controller will not work with this tuned mode power supply because the current flow is not increasing during the whole of the top transistor on time at full power output. {Bottom transistor in the circuits at the top of this blog}.
            • The circuit efficiency could almost certainly be improved by using one of the zero voltage or zero current variable frequency controllers.  But I do not know if such a controller would work with a wide input voltage or if it generates the resonant frequency high voltage for starting?  That is though these controllers are usually called resonant mode controllers.

            LTspice 24 is the Analog Devices version that supersedes XVII, the last Linear Technology version
            • Where there is a part for example nmos use that part rather than the auto-generated part.
              • Change the prefix to X (for the external part) rather than NM.
              • Change the value to the exact name in the model library.
              • Remove the path and ensure the library is placed in the working directory.
              • add a .inc with the added model file name to the diagram.
            • Avoid using more than one net name for a net e.g. placing 0V on the reference ground plane (has a down-pointing triangle symbol). 
            • In this case, it was necessary to set the maximum time step as low as possible.
              • It is also recommended to turn the compression off.
              • If a different solver is used then add a note in big text on the diagram. Eg Alternative solver.
            • In addition power is measured by pressing alt and clicking on parts. RMS voltage current or average power over an exact number of periods press control and click on the graph output title waveform. 
            -----------------------------------------------------------------

            Modelling the power output transistors;

            Advice from PCB makers and assemblers on the exposed metal area for heat-sinking;
            • Gardner Osborn advises that it is possible to have a copper area exposed so that there is no green solder resist mask but it is not possible to have this area black oxidised.  The inner layers are black oxidised so it would be necessary to cut an aperture in the top and bottom layers so that the inner layer is exposed.
            • Wilson Process Systems says if it is possible to have a black oxide exposed copper surface it is likely that that would only etch back a few micrometres of the 70um copper thickness but they gave me the PCB's supplier name to check with.
            • I received a similar answer from Minnitron Ltd. which is the oldest PCB manufacturer in the UK.  Black Oxide used to be used but it probably would not fare well exposed.  Generally, transistors are mounted to a Lead-free hasl or Ni/Au finish on exposed copper lands.
              In the 1990's I found that a PCB with an area of copper for heat-sinking was improved by having the green resist left off of the heat dissipation area of a PCB.  The area of PCB with the exposed tinned copper quickly changed from silver to tarnished mat grey improving the power dissipation.  A variation like this is still possible.  Particularly leaving the black oxidised copper exposed - doing this could increase the cooling by 50% provided a thick enough area of copper can be used such as 70um thickness rather than the standard 35um thickness.
              ---------------------------

              Low Voltage power supply current;                                    Before Start      Running          Running
                                                                                                              AL-0026-??       AL-0026-01D   AL-0026-06?
              IPD60R600P7 - Gate charge 9nC typical at 10uS intervals,   0mA                   1mA
              IPD60R280P7 - Gate charge 18nC typical at 10uS intervals, 0mA                   2mA
              2ED2104S06F - Gate driver max 245uA + 650uA,                 900uA                900uA
              2ED2184S06F - Gate driver max 170uA + 550uA,                 720uA                720uA
              UCC3806 - Current mode SMP max 1.4mA,                           0.1mA                1.4mA
              TSV712 - OP-AMP    Max 2x 16uA,                                          32uA                 32uA
              Power Status LEDs;                                                                 0                        0-2.2mA           2-4mA
              Significant factors total;                                                          1mA                6.5mA             8.3mA
              Status LED not fitted;                                                                                 4.3mA

              --------------------------

              The snubber and switching - new draft
               
              IPD60R280P7 Infineon

              Robustness; 80V/ns

              Maximum current; 5A due to load L1.

              Substrate diode maximum; 50V/ns

              Maximum Vc; 400V

              I = C.dv/dt therefore;
                  C = 5A * 1ns / 80V. (MOSFET)
              62.5pF would be the minimum value; 100pF

              C = 5A * 1ns / 50V. (substrate diode)
              100pF would be the minimum value, therefore; 220pF.

              Maximum Current pulse 36A. try 20A
                  Therefore; 22R = 400V / 20A.

              Snubber 220pF & 22R added to; AL-0054-???
               
              If a ballast controller is used which operates in quasi resonant mode then the snubber requirement becomes different, the capacitor becomes part of the tuning circuit and the resistor which had been a problem is redundant.
               
              The gate pull-down diode;
              • Maximum gate current that will cause power losses due to miller capacitance is;

                • The gate current that will not turn on the MOSFET is less than; 285mA = 3V-1V / (7R + 1R) // Because only typical values are given for Rgate.

              • Maximum dv/dt is that can be handled without gate current exceeding 285mA is;
                • Using Qc=18nC and Q = i * t, rearranged for t. then t = Q / i.

                • 18 ns  = 5nC  / 285mA  // This looks like the wrong method are the figure below
              • But the datasheet gives a turn-off time of 9 ns.  // But that is with Vgs = 0V rather than 1V and Rg =10R instead of 0R.
               i = C.dv/dt
               5A = C * 400V / 9ns
               14pF = 5 * 9 n / 400
              The snubber capacitor is much smaller when calculated this way.  
               
              However the different mode of operation under consideration in these calculations may not be relevant.
              ---------------
               High voltage reservoir capacitor;
              Now that the output transformer has been chosen a large electrolytic capacitor needs to be selected along with a supply voltage at <100W running and warming up.  Looking at prices 400V to 450V at >33uF and > 250mA ripple. 47uF 450V is reasonably priced. The assumption is that the starting current of <5A is only brief and not significant.

              Location; common to the output of the PFC and the HF light source Half-bridge output.
              F=100Hz, C=33uF (400V or 450V)
              Impedance = 1 / 2.pi.f.c
              50 ohms = 1 / (2 x 3.14 x 100 x 33E-6)

              The ripple voltage when running will be ~25V. (100W lamp 400V).
              This capacitor can handle the maximum ripple current required at the lowest price.  May be formed by two capacitors for lower EMI.
              ------------------------
              Gate driver selection;

              Infineon part number    Pins   Dead-Time    Iout               Voltage                     Feature                In
              2ED2103S06F              8        520nS        0.29A/0.7A   -5V +650V (-100V)   Internal diode    HIN & -LIN
              2ED2104S06F              8        520nS        0.29A/0.5A   -5V +650V  (-100V)  Internal diode    IN & -SD
              2ED2108                       8        540nS        0.45A/0.7A   -5V +650V (-100V)   Internal diode    HIN & -LIN
              2ED2109                       8        540nS        0.45A/0.7A   -5V +650V (-100V)   Internal diode    IN & -SD
              2ED2183                       8        400nS        2.5A/2.5A     -5V +650V (-100V)   Internal diode    HIN & -LIN
              2ED2184                       8        400nS       2.5A/2.5A     -5V +650V (-100V)   Internal diode    IN & -SD
              IR2182                          8         400ns   1.4A/1.8A               +600V
              IR2111S                        8         650 ns       0.2A / 0.4A             +600V                                        IN
              IR2108S                        8         540 ns      0.12A / 0.25A         +600V                                        HIN & -LIN
              IR2109S                        8         540 ns      0.12A / 0.25A         +600V                                         IN & -SD
              IR2302S                        8         540 ns      0.2A/0.35A                                                                 IN & -SD
              IRS21844MPBF           16       400 ns      1.9A & 2.3A             +600V
              Texas Instruments - no special protection required.
              • UCC20255 TI both drivers are isolated and one PWM input. dead-time 8nS (no resistor), 200nS (20K) 4/6A. price about $2.00 ----- This driver does not require protection against negative voltage caused by tRR in the commutation current that the IRxxxx requires.
              Analogue Devices - not suitable.
              • ADuM3223C/ADuM4223C ADI 5V supply and digital inputs also require an inverter to one side input. - not convenient?
              • ADuM7223C 5V supply and digital inputs also require an inverter to one side input. - not convenient?
              ST - good single driver.
              • STCAP2SCM_ ST - This is a single gate driver with negative and positive input options so the top transistor can be driven from the +in and the bottom transistor driven from the -in. Additionally, there is a gate pull-down clamp which will be more efficient than the diode used in the model giving a slightly faster turn-off time and a little better anti-miller capacitance clamping. $1.50 although requires 2 off and other parts this part is included because the PCB layout may be better consequently.
              Toshiba - Good single driver, fast for opt isolator type.
              • TLP5832 Toshiba - fast single opt isolator, min 5mA in, >1A >1.6A output. 
              Conclusion gate driver;

              To do later, Consider other transistors such as;
              • STD8N60DM2 - ST MDmesh DM2 MOSFETs which are specifically designed for half-bridge circuits with commutation currents and claim to have a good low tRR diode integrated.

              Magnetics

              Using CoilCraft web coil-loss modeller;
              • It is necessary to enter a small amount of DC current. I set the value to 0.1A the minimum.
              • AGP4233-470ME CoilCraft is 42x36x28mm
                • Others, MSS1278, MSS1583 are smaller and have lower power.
              • CoilCraft web-model for; 470uH at 20KHz running, 3A pk-pk.
                • MSS1278, 410mW 54'C 
                • MSS1583, 760mW 68'C
                • AGP4233, 4W 76'C
                • Starting 50KHz (f x 2.5) 14A pk-pk,
                  • MSS1278, 24W
                  • MSS1583, 46W
                  • AGP4233, 160W
                • Starting 71KHz (f x 3.5) 10A pk-pk,
                  • MSS1278, 23W
                  • MSS1583, 43W
                  • AGP4233, 98W
                • Starting 100KHz (f x 5) 7A pk,
                  • MSS1278,15W off the graph.
                  • MSS1583, 28W inductance drops to 100uH
                  • AGP4233, 59W inductance drops to 400uH
                • Starting 150KHz (f x 7.5) 5A pk, 
                  • MSS1278, 13W inductance drops to 150uH
                  • MSS1583, 13W inductance drops to 250uH
                  • AGP4233, 40W inductance drops to 450uH
              • CoilCraft web-model for; 220uH at 50KHz running.
                • MSS1278, 250mW 37'C
                • MSS1583, 450mW 52'C 
                • AGP4233, 1.1W 39'C
                • Starting 250KHz (f x 5) 7A pk-pk,
                  • MSS1278, 23W L reduced to 80uH 
                  • MSS1583, 40W L reduced to 120uH. 
                  • AGP4233, 66W, L unchanged. 
                • Starting 350KHz (f x 7.5) 5A pk-pk,
                  • MSS1278, 17W inductance reduced to 165uH.
                  • MSS1583, 31Winductance reduced to 195uH.
                  • AGP4233, 44W, inductance unchanged.
              • CoilCraft web-model for;100uH at 100KHz running. 
                • MSS1278, 550mW, 50'C 
                • MSS1583, 1W, 125'C
                • AGP4233, 1.8W, 50'C
                • Starting 250KHz (f x 2.5) 14A pk-pk, 
                  • MSS1278, 35W off the graph
                  • MSS1583, 71W off the graph
                  • AGP4233, 90W, L reduced to 60%, least 'C rise.
                • Starting 500KHz (f x 5) 7A pk-pk,
                  • MSS1278, 36W L reduced to 80uH
                  • MSS1583, 68W L reduced to 90uH
                  • AGP4233, 45W, L unchanged
                • Starting 750KHz (f x 7.5) 5A pk-pk, 
                  • MSS1278, 28W inductance reduced to 95uH
                  • MSS1583, 53Winductance reduced to 95uH.
                  • AGP4233, 75W, inductance unchanged.
              • DMT2-380-2.4L CoilCraft is an input or output filter choke and may not be suitable little is said about whether is it a low Q type. 380uH, 2.4A Toroidal, 36 x 36 x 23mm
              • DMT3-402-3.7L CoilCraft but as DM2 above. 402uH 3.7A, 41 x 41 x 23mm
              MSS1278-474KL_
              470 ±10%
              707.5
              786.2
              2.2
              1.34
              1.54
              1.64
              0.66
              0.90
              It is unnecessary to increase the frequency due to the availability of a suitable power choke. Increasing the operating frequency to 100KHz with 500KHz starting looks very good though but the frequency will need to ramp to 750KHz and this may not be possible. This should be simulated.

              Current Transformer;
              20KHz, 15V. Gives V.us; 750 V.us = 50uS x 15V (would be duty cycle near 100%).

              •  Pulse Electronics - some arithmetic required.
                • PA10005.xxx - both types will only produce a low voltage. (8.4x7.2x5.5mm)
                • PA820xNL  
                • 1:125 would only produce 2.7V at 20kHz duty 98%.
              • PA1005.125NL Is small and will need extra circuit components. Will provide  (~4V) with a minimum operating frequency is 30KHz. 
                • The PCB clearance between primary and secondary windings is only >1mm which may not be adequate?
                • The PCB clearance between primary and secondary is not adequate for Earth but a connection to Neutral is better anyway. Hipot voltage is 500Vrms but is okay for 220Vrms.
              •  Coil Craft
                • CST2010-100L_ is only; 254 V.us
                • SCS-100L_ is only; 160 V.us
                • CS4100V-01L is only 298 V.us
                • CST2020-100L is only 395 V.us
                  • 200:1 sense no power
                    • CST2010-200L_ ; 508 V.us $1.32 (15x20x10mm)
                    • SCS-200L_ ; 320 V.us $2.50 (15x15x10mm)
                    • CS4200V-01L ; 596 V.us
                    • CST2020-200L  791 V.us
                    • CST2020-300L  1186 V.us (300:1)
                • CTS3015-100ED is 638 V.us, 100:1
              • Wurth Electronics - none suitable. 
                • MID-SNS Sense Transformers (15x20x10mm)
                • 1:200 sense only no power
                • Part No. 750316796
                • 496 V.us
              Controller IC
              UC3846 TI - Used originally and is still available. This is still a good choice
              UC3856 TI - Has a higher output drive version.
              UCC3806 TI - Is it a lower-power version and is a good candidate.
              -------------------------------

              Comment;

              Potentially the running frequency could be increased to 500KHz if the emissions turn out to be modest, which is unlikely.  The Inductors and capacitors could be much smaller saving cost.    The gate drive power would be modest <200mW, this is unlikely and the best strategy may be to fit a lower Ron MOSFET?
              ---------------

              Review circuit topology and alternative strategies
              • A variant of this used a 24V full-bridge STM BIFET stepper motor driver worked successfully. The new drivers have thermal protection built in which would be an advantage.
              • Add a step-down transformer so that the current in the transistors is reduced, and the power in the inductor and all other power circulating be further reduced by operating with a smaller input-output difference.
              • Spectral lamps are expensive but the following sodium lamps are often used instead they are also higher voltage, lower current conveniently; NAV-T 50 W SUPER 4Y or NAV-T 50 W SUPER 6Y OSRAM, 50W, 86V.
              • A 390VDC out from the PFC;

                • IC type; ucc28056 TI looks interesting with a low start-up current few pins and low component count. I suspect that the inductor must not conduct continually even at maximum load so that there is no current period for the double-function input pin can sense the input voltage. Such an operation is likely to cause EMI but it is worth reading the data sheet and considering the input filtering despite what I have said the part will work and should meet all international standards without undue cost.
              • IC type; L6562AT ST, is a conventional continuous mode PFC controller but uses a few more resistors. So it may not be such cheap a solution?
              • A single transistor steps-up forward converter would require the design of a leakage transformer (loosely coupled) or use a transformer and choke.  This option was tried but it is likely to be less optimal than the circuit chosen. What has been developed subsequently resonant to start up works well so the forward step-up solution need not be developed further at this stage?
                • The circuit was tried but not refined using a  GTO and LT1070 but these parts are now mostly superseded.  The circuit was robust but did not function fully just proving that this type of thyristor is also very robust.
              • Operate at or near resonance frequency but the startup frequency probably needs not to be varied to tune for inductor value change with the power level.
                • Some units will not start lamps due to component tolerances resulting in the frequency being too far away from the resonant frequency to reach the arc-lamp striking voltage. 
                  • The losses are higher than optimal at starting because of the higher circulating current. This may be offset by not having a ramping frequency which may or may not cause a higher EMI than with a fixed frequency. 
              ------------------------ 

              Revision of the controller IC;

              Modern IC controllers have many features integrated that are useful,
               
              NXP IC controllers have;
              • Lower power startup and some sort of high voltage regulator.
              • X capacitor discharge so the user is less likely to be shocked by touching the mains plug. 
              • Voltage mode operation which should work better than current mode if running in with the LC resonate is used. 
              • Adaptive gate drive for efficient switching.
              • Resonate mode may be good but I have asked about noise and regulation.  Particularly with promoted LLC converters such as TEA19161.
              • These LLC have a narrow voltage range but I am hoping for a narrow current range and wide voltage range operation.
              Recommended;

              A link describing various AC-AC power supplies;  https://www.monolithicpower.com/en/power-electronics/ac-ac-converters/introduction-to-ac-ac-converters 

              • LCC types are not efficient at such low power.
              • The nearest controllers that might be suitable are; TEA1755T for up to 250W and TEA2017 (LLC) for 90W to 1KW. 
              • Flyback types;
              • TEA1733LT up to 75W flyback no PFC.  Current mode, CCM, 66KHz (62 to 71KHz).
                • 72% duty cycle limit that can be programmed lower with a resistor to CTRL pin.
                • OVERPOWER pin sets soft start.
                • This is a current mode controller.
                • TEA1755T is a flyback controller with a PFC controller.  It would be used with a transformer therefore so presumably the PFC would not be required.  The advice is very similar to what I was given 40 years ago by the same company but called Mullard at that time. 
                • TEA18363T/2 Flyback. up to 75KHz.
                  • 132.5KHz (125 - 140KHz) or 25KHz (23 - 27KHz)
                  • Warning It can go into burst mode at low power DCM.
                  • At medium power, the frequency is reduced.
                  • At high power quasi resonate mode operates.
                •  TEA19363LT Fixed Frequency DCM, QR. flyback
                  • Warning It can go into burst mode at low power.
                  • At medium power, the frequency is reduced.
                  • At high power quasi resonate mode operates.
                  • 128KHz (120 - 136KHz) or 25.5KHz (23 -28KHz)
                  • Powered by HV pin also does the X capacitor discharge up to 100nF.
                  • For PSU's up to 75W. Discontinuous and, Burst modes could be problems. It works in Quasi-Resonant mode.
                    • Isense must be >140mV to prevent these low power modes.  As well as frequency must be a maximum of 132KHz.
                  • It uses on AUX pin for demagnetisation sense.
                  • This part probably would be difficult to use.
              • LLC types for >90W These all feature 
                • High power to low power mode can transition as low as 10% power level.
                • In Low power mode, 1 in 4 cycles are missed the low driver stays switched on for a half period.
                • Low power mode is lower lower-frequency operation it is unfortunate that it seems to be a switch rather than a transition.  The default in 20 and 30%.
                • Burst mode can be set as low as 1% - this mode must never occur.  The device would need to be programmed to reduce the default 10%. 
                • In Bust mode the low driver is held on for a long period. This mode must be avoided.
                • quasi-resonant mode, discontinuous conduction mode, or burst mode. And demagnetization sense.
                • Variable frequency 25-132.5KHz 
                • The protection pin operates by low voltage detection.
                • Over-voltage is detected with the AUX pin that comes from the transformer.  So it won't detect the lamp Over Voltage.
                • Burst mode be avoided by ensuring CTRL pin >0.5V.
                • Isense pin also manages soft start.
                • TEA2016 LLC Half bridge with PFC 90-500W.  In high power mode, the circuit runs at variable frequency.
                • TEA2017 (LLC) Half bridge and also has a PFC controller.  Although the application circuit includes a transformer it looks like it might do what is required with a simple choke and feedback for switching control by capacitor coupling. 90-1000W.
                • TEA2226 LLC Half bridge has low power burst modes which are not advantageous.  But is designed for much bigger 90W-1KW power supplies.
                • TEA6017 LLC with PFC. 90 - 1000W.
                • TEA18361LT/2
                • TEA19161 LLC Half bridge. 90 - 500W.  This does not have a PFC but works with a TEA19162 PFC controller. 
              • Other quasi-resonant mode power supply ICs;
               AL-0026-02C XDPS2201 controller.  Incomplete for discussion.

              PDF   Infineon community question

                • Hybrid-flyback Controller XDPS2201 - Infineon
                  • This controller is very similar to the working circuit at the top of this blog but for;
                  • It is resonant mode so the switching and general efficiency should be improved.
                  • Because it works in resonant mode frequency ramping may not be required for arc starting and to run the higher power lamps from 90Vac input.  On the other hand, there is no oscillator frequency control.
                  • The part has PWM for control rather than running 50% duty.  In other words, it runs in the same way as the UC3806 or UC3847 used previously.  I do not know if the controller can be configured to run with a PWM limited to a 50% duty cycle?
                  • L2 could be replaced by a 1:2 step-up transformer if the resonant mode can not be relied on to boost the supply voltage for higher-power lamps.  The transistors would therefore need to handle higher current. 
              • Unfortunately, the Infineon community reply is that I need to do more work but in any case, the controller is untested without a transformer. 
              • Infineon recommend using one of the florescent ballast controllers because the lamp life would be shorted by 5-10% using a PWM type which is an unsymmetrical drive.  Such as IRS2158D or ICB2FL03.

              Arc lamp controllers;

              • NXP used to make many high-frequency operating controllers and ballast units they now do not make them and they are not recommended for new design. 
              • These either run fixed frequency or regulated and dimmed by frequency control.  That means that there is a delay in the control between mains frequency fluctuation and rectifier capacitor ripple and the correction applied.
                • Perhaps some feed-forward compensation could be added to address this issue?
                • There is a conflict in reducing power by increasing the frequency and the power increasing as the frequency approaches starting resonance. In the models above this may not be an issue provided the inductors value is not too low.  But to achieve this the control range needs to be limited by using PFC to boost the supply voltage. 
                • A more expensive pre-regulator solution such as UC1872 would resolve all these issues. Or use any of the controllers with a high voltage 150V 200Vdc regulated power supply pre-regulator.


              UC1872 works in resonate mode so models AL-0054-01? to AL-0054-04? 
               may be relevant.  An advantage also is that there is no pre-heat support. 
               
              The potential drawback could be that the control loop response is slow so mains ripple on the supply will modulate the light a little.  This circuit is similar to a thyristor sine wave inverter but for C5 which is an extra component.  The higher voltage could be created by adding series diodes in A and B MOSFET drain leads, thereby reducing the required transformer step up ratio and reducing the transistors peak current. 

               AL-0026-04C output section adapted for high voltage supply.  Components have arbitrary values.  UC1872 has a high operating power at 500mW more or less compared to newer IC types.
               
              The output circuit is adapted to mains voltage operation.  The MOSFETs substrate diodes are used to limit the voltage but the substrate diode current could be high.

              Other ICs for various arc lamps that are suitable for new design;

               AL-0026-03B, L6574 STM fluorescent lamp controller with current regulation by frequency control.  This circuit does not work in resonant mode so either a step up output transformer need to be added or PFC be added.  PDF 
               
              But the circuit might work unreliably starting by design and chance near enough to resonance. There is a conflict in the control in that the power is reduced by increasing the frequency but if the resonance is higher the power will increase then reduce.  This concern is true of all of this type of controller.
              • L6571 - has no regulation, fixed high frequency.  This is simply an oscillator and a half-bridge driver. 
                • This offers no regulation and the Lamp current will have a tolerance due to the oscillator frequency and inductor value tolerances plus the supply voltage and the lamp rating.  That is the spectral lamps would be run more or less the same as a traditional mains frequency leakage transformer type ballast.
                   
              • L6574 - Fixed frequency + variable frequency for starting.  This controller has no switching loss protection modes such as zero voltage switching.  And operates with a delay in the control loop that will introduce ripple into the light output. 
                • The lamp strike resonance is passed through quickly so there is no long resonate high voltage phase where the noble gas can heat up and vaporise the metal.  It may work though because much of the warming is within the high-voltage input or created by a PFC.  In any case the lowest running frequency and the starting higher resonant frequency do not need to be more than 30% different.  But the drawback is that the warming up power would have been high without the high voltage sense that should reduce the current to 150mA.
                • The PFC st.com says is required for lamps greater than 25W this is something to do with a rectified and smoothed mains draw current at the peaks of the wave.  The advice seem different to the 75W limit published elsewhere Is it probably a typo? 
                • The basic current setting is mostly by selection of the inductor, but  trimming the lamp current by frequency control using the op-amp integrated. 
                  • So there is potential for supply voltage feed-forward control (not part of the application note) which could reduce supply ripple appearing in the light output.  This option has not been included.
              • L6574 and L6561 - As above plus PFC. AN993 Application note.

              • IRS2158D - Starting by variable frequency resonance.  The data sheet does not mention power saving zero voltage or current modes.  Pins;
                • RFMIN 10K to 300K Minimum frequency set.
                • CT >330pF
                • VCC 10.5V - 12.5V clamp 14.6 - 16.6V 5mA
                • CPH preheat. Ends rising to 8.8V - 9.8V.  Preheat time set
                • CPH Ignition. Starts falling to 4.4V - 5.0V
                • CPH fault 0V
                  • Rph and Cph set the preheat time but as no preheat is required Cph = 0 and Rph = 22K?
                     
                • VCO 0V preheat mode. 0.65V ignition mode. 215uA open V run mode. 0V = fault.  Ignition ramp generator.
                • CS over current sense 1.0 - 1.4V, EOL when 60 events occur.
                • SD/EOL shut down.  rising 4.5 - 5.5V falling 2.7 - 3.3V not latching.  End of life rising 1.8V - 2.2V, falling 0.9V - 1.1V
                • VDC enable 4.5V - 5.5V, Disable 2.7V - 3.4V.  Brown out detect.
                • FMIN 0V = fault. Normally 5V
                  • Run mode 45.5KHz
                  • Preheat 68KHz 
                  • CT 2V - 5V
                • Op-amp NINV INV inputs 10mV, 100uA, 0-11.5V. Output 12-14mA,
                • VB High side gate driver boot strap has an internal diode.
              • IRS2530 - Simpler IC has dimming and current regulation.  This has current feedback/dimming pin which operates by adjusting the frequency.   There is also start up frequency ramping.  Not zero voltage switching or high MOSFET current they call crest factor detection when running to protect when the circuit is operating just above resonance when high current pulses can occur (these high MOSFET currents can be seen in some of the models above).  The data sheet explains that this protection features are turned off during lamp starting otherwise detection of not ZVS causes the operating frequency to be increased.
              • ICB2FL03 - With PFC which had zero current detection.  Variable frequency 20 to 120KHz. Preheat is above this frequency to 150KHz.  The warm-up period needs to be set short or not at all.  One of the protections is the detection of capacitive mode when the frequency is below optimum but like other ICs, they all start with high frequency and then reduce the frequency.  When started the frequency adjusts to about 40KHz.  Over-voltage is detected as a fault this could be a problem with some spectral lamps that run high voltage whilst the metal warms up in the ionised noble gas running period. 
                • Power Factor Controller pins; PFCGD, PFCCS, PFCZCD, PFCVS.
                • LVS pin - Over voltage detect for lamp missing but perhaps be used for high resonant voltage detection? 
                • RES pin - Filament detect feature can be turned off by connecting pin to 0V. 
                • RFRun pin - resistor sets the running frequency 20KHz to 120KHz. presumably could be used to adjust the lamp current? 
                • RFPH pin - resistor sets pre-heat frequency up to 150KHz.  This feature needs to be turned off or at least minimised.
                • RTPH pin - resistor set pre-heat time.  For zero time required; R=0.
                • RES pin - 
                • LSGD pin - MOSFET current sense  
                  • Over-current at 800mV for 500ns
                    • And 205mV/uS - ignition control. Puts controller into ignition control by holding the frequency. The LSCS pin now handles ignition control.  This saves from choke saturation.
                  • Shutdown at 1.6V for 500ns at start-up or soft start ignition and pre-run. 
                  • +-50mV detects other conditions to detect capacitive mode inefficiency during running.
                  • 2V and other comparators
                  • -50mV to set dead time.
                • Power and gate drive pins; LSGD, Vcc, GND.
              • ICB2FL03 or IRS2158D are recommended for this application by Infineon community.  They advise that the lamp's life would be reduced 5-10% by using unsymmetrical drive.
              ICB2FL03 Infineon fluorescent, cold cathode lamp controller.
              Many of these controllers are very simple or are less well developed
              than the NXP parts now not made.
              • A fluorescent lamp controllers operate 40KHz to 100KHz and meet emissions standards so a spectral lamp should be expected to also meet EMI standards.  Note the only broadcast in this band I believe is NPL's time standard at 60KHz.
              • Spectral lamps do not have a heater.  Fluorescent lamps can also start up in high-frequency high voltage without requiring the heater.  Therefore the heater windings are not required and the capacitor shown must be fitted across the lamp, not between the heater terminals and the series bead EMI filter choke therefore included.
              • Because of the use of zero voltage switching the snubber is not an issue and high currents should therefore be avoided.
              • The heat-up cycle boost is best avoided and the lamp is running at a much higher power than running power unless a limiting circuit is added.  It would be better if the current were reduced similarly to my original circuit that runs at only 150mA in the early stage of starting before rising to 1A.
              ---------------------------------------------------------
              Blogger.com has not been working properly since March 2024 and does not display pictures from drive.google.com. The pictures recently added have been uploaded to Blogger instead.