LED driver with current source bias supply?

[eem2am]

Hello,

The following PDF (&LTspice .txt simulation) schematic is a buckboost LED driver with Vin = 100V.

Please can you confirm that the bias supply for the high side optocoupler feedback, MUST be from a constant current source as shown (LT3092 is a constant current source IC).
If it is not a constant current bias supply , then the varying draw of current by the optocoupler feedback circuit interferes with the feedback loop?

Do you agree with this, (that the current source high side bias supply must indeed be a current source)?


Spec:
Buckboost
Vin = 12-100V
Vout = 40 to 64V
Switching frequency = 200KHz
LED current = 100mA
Discontinuous Mode.

 

[KrisBlueNZ]

That's an interesting circuit! There are some ideas there that I haven't seen before. Not too surprising I guess, since this type of LED driver is outside my experience.

I don't see anything wrong with the general idea, but that's no guarantee of anything. As I said, I'm not experienced with this type of design.

Re powering the current measurement and feedback circuit, I guess the other option would be to connect it directly across the bottom LEDs in the string, without a current regulator and without a zener. Then, as you say, the optocoupler LED current will affect the current drawn by the circuit, which will subtract from the current measured across the current shunt, R8. And that might make the feedback loop unstable. Is that what you were considering?

I'm not knowledgeable or experienced enough to say whether it would be a problem or not. I can suggest one possibility: modify the optocoupler driving circuit so that the optocoupler current flows through the shunt as well. This would keep that current in parallel with the bottom five LEDs; it would affect their brightness (slightly) but would not affect the current measured by the shunt.

That is just a suggestion for you to consider. I don't know for sure whether it would fix any instability problems.

I think perhaps the only way to do this would be to move the feedback circuit's 0V rail to the top of the shunt resistor and have it measure the voltage on the bottom end of the shunt resistor using a voltage divider between that point and the reference voltage; the centre of that voltage divider would then be compared to the 0V rail.

Do you think that would work?

I have two other comments. I don't consider myself knowledgeable or experienced enough to know all of the issues that the design might have; these are just the ones I can see.

You should have a good quality decoupler between the left end of the inductor and 0V.

Re your calculations to ensure adequate CTR in the optocoupler, what is the maximum current you expect the LT1243 to source from pin 1? The data sheet says it's 0.75 mA typical, but it doesn't state a maximum value.

Out of interest, why do you need such a wide input voltage range?

 

[eem2am]

OK thanks very much , i am coming to your other points.....the wide input voltage range...you wouldnt believe me if i told you......some customers believe.....that they must protect against automotive load dump and so they insist the product can run off up to 100V, -even though the auto battery is just 24V....they dont want TVS protection as they say it'll just blow the TVS.
Also, Other customers use forklift truck with 80V battery, and some customers just 12V auto battery.
I am now happily investigating your other comments.

 

[eem2am]

Then, as you say, the optocoupler LED current will affect the current drawn by the circuit, which will subtract from the current measured across the current shunt, R8. And that might make the feedback loop unstable. Is that what you were considering?

..yes, thats exactly my fear

what is the maximum current you expect the LT1243 to source from pin 1?

i beleive its 1mA max.....the same as in the UC3843......this is a good point.....it'd be worth grounding the FB pin and measuring the current out of the comp pin....tough there will be a tolerance, as you allude to.

I see what you are trying to do with the feedback as per your suggestion.....ill think more on that one.

 

[KrisBlueNZ]

AFAIK load dump is a real phenomenon. I'm surprised that your customer expects your circuit to operate as normal during load dump... I would have thought it would be enough for the circuit to just disconnect itself and stop working during the overvoltage.