Testing a High-Output Driver PCB

I have a 2" diameter PCB driver with a very large coil on it and 4 output leads. It is actually intended to drive 12 LEDs (pulled from flashlight).

Now, I would like to test this driver for its maximum output on the Vf side. The old technique was a series of 4001 diodes from RShack, with a 10W/1ohm resistor in the series, then millivolts measured across the circuit = mA of current.

Since I suspect this driver to deliver 12+ amps at 3.4V+/- (it's either 1 or two channels of diodes, will figure this out shortly), what type of test load method would you recommend for this power range?

I was thinking, similar configuration, but the circuit splits on the test load so that two parallel 10W/1ohm resistors are wired in series (4, 10W-1ohm resistors total) to make a 1ohm load again, then test mV across the resistor bank again like usual. I figured this would be enough resistor mass to perform the quick test without cooking the load mass in 5 seconds. Is this how you might do it? Or, double impedance to 2 ohms, then multiply mV reading x2.

I also have an oscilloscope (besides a usual 10A multimeter), if that can be of any use here. I am attempting to use this driver type in a build and it's chinese, so I'm trying to obtain some sort of quality measurements on it and its ability to hold a setting wattage.

I don't know yet how it actually regulates, whether on voltage setting or on amperage draw. I can get some pics tomorrow if needed.

Thanks

A 1ohm resistor + diodes will not work imo.

to flow 12A in a 1ohm res, you need 12v, plus Vf of each diode, so it won't work, and maybe could damage your driver (dunno).
Maybe with a 0.1ohm resistor and beefy diodes (1n540x, or heatsinked diodes).

You will need some better advice there, don't test that with a conventional test load. It may raise the voltage so 12A current could flow and BANG, who knows. I'd like to help but I can't

Knowing its an LED driver, you can reasonably suspect it can drop ~3 V on the output.

In this case, I would wire 8 1ohm 10W resistors all in a single parallel line.

Then at say 12A out of the multimeter, a total of 1.5V is dropped by the huge resistor, and each 1ohm unit sinks 1.5A. This puts 2.25W in each resistor - so they will still get warm.

So that leaves around (Vfmax - 1.5V) of headroom to experiment with. You simply keep adding 20A diodes in the series chain until the driver fails to regulate, or noticeably struggles or gets hot.

That should give you a pretty good idea to within +- .7V.

Either that or find yourself a 50W pot and just turn up the resistance until the driver can't push 12A through anymore.

There are many correct ways to do this. Here's how I would do it.

Get a few cheap and beefy rectifiers. Use your 10A meter as the shunt (12A is probably not high enough to damage anything). We'll assume the shunt is 750µΩ for the sake of illustration, but you can work that out later. Wire it as such:



Then you can put your oscilloscope across the meter to check the current waveform. Be sure to use 18awg or bigger for the wiring so you don't have too much drop on the leads.

I like Cyparagon's diagram because I see clearly: 4, 6, and 8, for 2V, 3V, and 4V.

Where can I get some rectifiers quick? Heh...I have an audio amplifier board a friend gave me, anything useful on those?

12 Cree XM-Ls. That's what's powered. It could be two banks of 6A, I simply suspect each LED is getting 1A by the massive heat output of the lumens on your hand. It feels like 30W or so, but I don't know yet the actual Vf of the driver. Let me pry the light head apart (it's epoxied), and I will be certain how the LEDs are wired. I will post back, but do appreciate the schematic--thanks.

Or you could do this. Your choice, though it doesn't look like the guy can count. One for the fail thread, maybe

Is that 750 micro ohms?
That's a piece of wire isn't it?
HMike

Prettymuch. I had to make one for a 100A analog meter once, the kind that goes in a dashboard. The original one got lost. It came out to be a couple inches of 12AWG copper wire. (For that particular meter, anyway) Then the meter movement then goes across it. On smaller meters it is inside the meter. That's why they call them shunt ammeters.

Hemlock_Mike said:

Is that 750 micro ohms? That's a piece of wire isn't it?

Click to expand...

Precisely. You see that large wire to the left of R29? That's your typical multimeter shunt.

Manganin - Wikipedia, the free encyclopedia apparently this is the metal of choice for ammeter shunts

Been meaning to reply to this. A few things briefly:

One; I'm partially confused. :undecided:

When it is talked about placing the multimeter in the load, you guys are talking about the shunts, etc. I have a Klein, MM1000 multimeter. Probably the shunt type like mentioned, I assume.

So what's all this mention of the shunts? Are you saying, I need to find it's resistance before I can accomplish what's been recommended?

Also, I no longer deal with eBay. I was ripped off, never went back again.

I know how to do a google search for those rectifiers of course, but, do you guys have a preference of a seller here in the US--when it comes to electronic parts like those?

Multiple rectifiers are being shown, what makes you guys believe one is better than the other, or etc? I just know they look very different, one type is 4-prong the other linked set is a single. What would the difference there be in wiring? I may in fact go up above 10-12 amps. There is a driver I need to test which will push 22A. I'm not sure how I would take this test one step higher to that amperage range without buying a bigger more expensive multimeter, *when* the time comes that I need to begin working with that 22A driver. For now, it is the lower range mentioned which really needs driver testing.

Now, Cyparagon, you stated "R29", did you mean that heavy solid wire by R39 that has roughly an "M" shape is the shunt? I do not see an R29.

Thanks a lot for the help guys. I'm trying to tackle a lot at once here and cutting corners seems to be my only option, as the stuff I've been finding for testing is $$$. I'm about to spend over $600, something I didn't want to do, because the power supply needed that was 0-10V/0-20A is over $300, and now that I know I'll need over 20A for the diode testing, I need a 0-30A lab supply. It's funny because I have what's called an EFuel 15-24V, 0-60A AC-to-DC supply for lithium high current battery chargers (link: http://www.quadrocopter.com/product.asp?itemid=729&gclid=CK7b_aHjursCFSUOOgodVBEAOw ), and it was only $200. I was like, "man if I could just modify the thing to go down to 1V (so I can test 1V-4.5V), I would be set on amperage ability", in my mind, as I thought about these expensive and no doubt chinese, lab supply prices. These big diodes I'm working with don't really need the super lab precision to drive them, they just need the proper current to the 1/3 of an amp I would say, for the proper amount of time--without the driver burning up during that time.

Digi-Key
Digi-Key

You can see that by avoiding eBay there is a price hit. They have a good buyer protection
policy now and if you pay with PayPal you can get a full refund when a seller tries to
scam.

I have been doing a little testing of my own and am working on a driver torture device. It
simulates switching the driver on and off very quickly at a regular interval. This makes it
much easier to view start and stop transients on an oscilloscope. It is just basically a
PIC sending PWM pulses to a logic level MOSFET. There is a button to vary the pulse
width. I can provide a schematic and source code if anyone is interesting.

Well I'd be interested in knowing basically, what the difference is, it looks like the body is ground, and the center post is positive? So each would use a ring-terminal pigtail soldered to the next in series to follow Cyparagons schematic?

I'm just wondering why you are saying get this type Lightning Stalker, and the other is saying get a totally different looking type who posted first...I end up with confusion as to why a second rectifier is being shown to me with a link to it. What are you saying is more useful? And, source code? For a simple load device? Heh, that's all I'm trying to do is simulate the load of 1 single, big LED. I don't really want programming to be part of soldering together a simple load, no offense to your project.

You have offended my project! j/k.

Those are a different type from what Cy showed, just a single rectifier in each package. Get
these.

Jameco

I tried searching for those first on Digi-Key, but they were more than twice the cost of the
single diodes.

Steer clear of the Lamp Lighter.

Hah!

Awesome, though. Thank you, both. I will order them, and from there, once they arrive, I will be back with results from my experiment on the driver. I'm about to go mill and turn me a nice set of adjustable spanners to get the actual driver out. Darn grounding/holding ring in there on top the driver is very tight.

...And oh yeah, about that shunt^ What do I need to do, make sure it is 750 micro-ohms? It is the "M" shaped solid metal?

Yeah, that's it. Don't worry about it. Just attach the probes the way it specifies in the manual and take a
reading.

Bob's Kart Shop!