What voltage does LED load simulator go up to?

[treez]

Hello,
Do you know what voltage the M98 LED load simulator goes up to? That is, can we simulate 48V, 3.5A led array loads?

The page says 150V, but this seems very high for a led string, does that really mean it can simulate strings of voltage up to 150V?...

M9811
http://www.maynuo.com/english/pro.asp?tid=96

M9811 manual
http://www.maynuo.com/downloadfile/2012050437183497.pdf

Also, why does the manual say nothing of what dynamic impedance settings exist for the "led load"?

 

[Audioguru]

The dynamic impedance of the LEDs does not matter when you set it for constant current.

 

[treez]

https://www.ti.com/lit/an/slyt308/slyt308.pdf

..As the above shows, the dynamic impedance is crucial for led driver stability testing.
Its essential to be able to set the dynamic impedance to the same as the leds (at the led load current level) in order to check for stability.

 

[Audioguru]

We do not know if the Chinese LED driver has an unstable DC to DC boost converter. Since it talks about using a fan for cooling then I doubt it and it is linear.

 

[treez]

the circuit we are testing is a buck led driver.
Do you agree with me that in order to test for instability in a 3.5A led driver, using a 3.5A constant current load is of no use....it needs to be a 3.5A load with the same dynamic impedance as the leds.

 

[chuckey]

Suppose you switch on your LED string, it will present a capacitance and a resistance (48/3.5 =~ 14 ohms). As the LEDs warm up the voltage across them will fall to say 46V at the same current. So what is "dynamic" about that. The load that you show does not quote an output resistance so you can work out how much the current changes for the change in volt drop. You must not have a slow start load, as the LEDs take that current from the start and a bit more to charge up their capacitance.
Use a resistive load set for 46 (or what ever) volt drop at 3.5 A and switch in another 2V worth of volt drop at 3.5A and see how your PSU reacts. Make sure you know where the on/off switch is because if its going to be between your PSU and the LEDs, then you have to actuate this at a fairly high speed to make sure that the capacitive inrush current does not upset the PSU. If the on/off switch is before the PSU then the PSU will be slower at powering and the inrush would be zero if your PSU is well designed with out any ringing.
Frank

 

[treez]

sorry but that is not relevant to the question of this thread.
As already stated in post #3 and again in post #5, the only way to test for stability in a switch mode , current regulated led driver is to use a load which has the same dynamic impedance as the leds that will be used.
Obviously it would be best to use the leds themselves as a test load, but you cannot guarantee getting the highest vf voltage led string for your led test load, so you need to use a led load simulator, so that you can set the total vf, and the dynamic impedance.

Do you agree that to test for stability of such a led driver, it is essential to use a test load that has the same "dynamic impedance" as the leds that will be used? It is not good enough to simply choose a constant current load, or a resistive load with value V(load)/(I(LEDs).....Do you agree?

 

[chuckey]

"dynamic impedance" Can you define this and say why it is useful in the case of a LED string that is run at a constant current? As far as I can see its only Dv/D(temperature) that changes as by definition the LEDs are running at a constant current.
Frank

 

[treez]

if you see the article in post #3 it explains all about led dynamic impedance. small signal feedback loop calculations refer to the dynamics of the loop......which means the loads dynamic impedance is needed to be known. Just because its run at a constant current doesn't change this...its constant current within a dynamic feedback loop.
For stability testing of LED drivers...the test load must, and we mean "must", have the same dynamic impedance as the actual led load...you agree?
You cannot use a "same power and voltage" resistive load to test the feedback loop of a switch mode led driver for stability..you agree?

 

[chuckey]

No I disagree, the LED, if fed with a proper DC constant current has no dynamic impedance , because if the current is constant then the voltage is constant, only subject to thermal drift. If your PSU has not got proper decoupling and exhibits some ripple voltage, then the dynamic characteristics of the LEDs should be taken into account. But then its not a constant current source, it is a stabilised current source of some waveform or other.
Frank

 

[treez]

I don't know if we are talking at crossed purposes here?...are you speaking about a regulated current source?.....one that uses a negative feedback loop to regulate it to give a constant current?
If so, then the dynamic impedance of the load must be taken into account, as per the article of post #3.
Do you disagree with that article?
That article is central to the initial question.

If testing a current regulated switch mode led driver (regulated with a negative feedback loop) then the only way to test for stability is to use a test load that has the same dynamic impedance as the leds that will eventually be used as the load (or of course, you can use the leds themselves as the test load)...do you agree?

 

[treez]

so do you believe that in order to test a negative feedback loop , current regulated, Switch mode led driver, its ok to use an output test load without any regard to what is the dynamic impedance of the test load?

 

[FvM]

I suppose there are misunderstandings in some responses. I agree that the I/V characteristic of a LED simulator and also it's frequency dependant behaviour, which hasn't been even addressed in your question, can matter for the stability of the DUT (tested LED driver).

Obviously the instrument manufacturer doesn't give much information about it, so which answers do you expect in this thread?

 

[treez]

Thanks, basically we need to test a 500W, 3 channel buck led driver, and the majority of our company just want to use a resistive load of the same power and voltage as the actual led load that the customer will use. (we can use the actual leds for some of the tests, but for many of the tests , its just not practical to use the actual led load that the customer will use).

So the decision is ....
1..whether to use a resistive load of the same power and voltage as the actual customer led load...OR...
2..Use any, off-the-shelf constant current load at the led current and voltage level...OR...
3..Use a led simulator which has the same dynamic impedance as the actual customer led load.

The testing needs to show if our negative feedback loop buck led driver will go unstable or not.

May i ask which of the above three options should we use.?

I have asked the company who make the led load simulator, and they say they will get back to me, but that was three weeks ago...they have quoted us 4700 Euros. (as you can see, number three is the expensive option)
I believe we need to use option 3 above, but I hope upon hope that I am wrong, as its much more expensive.
(Please I hope someone can prove us wrong.)

Good point about the frequency dependent behaviour of the led simulator, I was just hoping that it would be the same as that of the leds.....than again, maybe this is why the led simulator manufacturer is taking so long to get back to us.

Here's an example of what we mean by led simulator...
https://www.chromausa.com/pdf/63110a-instruction-demo-ver-4.pdf

 

[chuckey]

If this box did want you wanted it to do, it would have a menu of LED types, how many you were using, and operating current and their operating temperature. It would also give some sort of tolerance on how accurately it compares with them.
so. . LED types .... no
number of LEDs ...no
operating current ..yes (no accuracy specified over voltage range)
operating temperature of LED...no
and no mention of capacitance of the load
Err, I make that about 1/2 out of 4 so I would not buy it for this job.
Frank

 

[treez]

we don't need the current to be accurate ....the current accuracy is determined by our led driver.
The chroma device is just supposed to simulate a led string....of whatever dynamic impedance one types in to it.
4700 euros is a lot of money for a led simulator that doesnt truly simulate a led string.
So are we saying that the only way we can test for instability is to use the actual led load?...that's a problem, because if we make an actual led load up, it will take time and money , and we cant be sure we'll get leds of the maximum vf.
The maximum vf led string is the worst case for stability, so we need to test the led driver out with maximum vf led string....but since its going to be very time-consuming to make this, we thought of using the simulator?

 

[chuckey]

This secret LED driver off yours - is it mean't to be a DC supply? - low ripple. Then if it fires up into an open circuit and the full load without any major ringing, then that would seem to be a good first step. it also would mean that its output reservoir capacitor would likely swamp the capacitance of the LEDs AND wiring. But you have to find out what your LED capacitance is going to be (+_!!).
If its some high frequency pulse unit, then you will have to take into effect the wiring as well, as the ringing of the wiring and the capacitance of the LEDs could be catastrophic.
You still have not said where the power on/off switch is. For production testing you need to spend your money on gadgets that differentiate between a fault free PSU and a faulty one. If the design is so marginal that it cannot work into LEDs but can work into an equivalent load resistor, then its the design that needs looking at and not the load simulation.
So I reckon you need:- a ripple window detector, a DC current window detector, for a load of max Vf ( cold working LEDs) and min Vf (reduced for hot working LEDs). If the PSU is prone to overshoots when powered, an automatic on/off switch with a window detector on the overshoots (AC coupled peak detector). And a way of repeating these test at the extremes of mains voltages. Three seconds to perform all these tests?
Things like long term stability will be built into the PSU by the choice of components and quality of construction.
Frank

 

[treez]

I do appreciate your answer, but the discussion is about the stability of led drivers, as in the gain and phase margin.
-And whether a led simulator load (eg chroma 63110A) is better than a dummy resistive load in precipitating any instability that may be prevalent in the led driver.