120vac led driver 30w; Changing Current level?

[magudaman]

I purchased a led driver for my 30w led back in December and has been working great up till up burned up my led. I have a new led on the way but would like to drop the current of my driver from 1000ma to around 750ma max. I've opened up the driver and it is not exactly obvious of what to change. Has anyone successfully adjusted a driver like this?

The primary chip is a gc9910 which seems to reference an chinese led driving circuit but I can't seem to find any diagrams or datasheets for it? It maybe just another generic for something like the HV9910B which does have datasheets. I don't see any feedback lines for the DC side so it must be regulating current going into the isolation transformer. There is one capacitor connected to both sides

I also wrote the seller to see if they would reply but I doubt they will get back to me.


Outside:


Close Up of chip:


Bottom Side:


Top Side:


HV9910 PDF:

http://www.premier-electric.com/files/supertex/pdf/hv9910.pdf

 

[SemiMan]

Best I can do is suggest looking (probing) around for a resistor of about 0.22 - 0.25 ohm. That should be the sense resistor for the 1A current output.

I like the "SELV Equivalent" on the sticker. Never cease to be amused by the crap that Chinese suppliers will put out there to make a buck.


Semiman

 

[magudaman]

Thanks for the response! I did see that in the data sheet of the unit I linked to. I need to figure out how the chip is truly orientated and if the pins match the ones I in the datasheet. Once I can identify the pin then I should be able to backtrack to figure out where to adjust. Looking over the device itself I did find one pin that may corispond correctly and is rated at 220ohm, so maybe that is is it! I get back to you when I figure it out a bit more.

 

[Denjava]

Hi! Resistor for 1A current must be rated at 2-5Watts so it is one of those grey ones. Probably R1 on the PCB near that transistor. You should write here the name of that transistor which is probably our output MOSFET from the scheme. And one thing about measuring the resistance... you need to take out the resistor from the PCB because other components may affect on the result.

Greetings!!!

 

[SemiMan]

Thanks for the response! I did see that in the data sheet of the unit I linked to. I need to figure out how the chip is truly orientated and if the pins match the ones I in the datasheet. Once I can identify the pin then I should be able to backtrack to figure out where to adjust. Looking over the device itself I did find one pin that may corispond correctly and is rated at 220ohm, so maybe that is is it! I get back to you when I figure it out a bit more.

Not 220 ohm, 0.22 ohms.

Semiman

 

[SemiMan]

Hi! Resistor for 1A current must be rated at 2-5Watts so it is one of those grey ones. Probably R1 on the PCB near that transistor. You should write here the name of that transistor which is probably our output MOSFET from the scheme. And one thing about measuring the resistance... you need to take out the resistor from the PCB because other components may affect on the result.

Greetings!!!

It's likely the only thing remotely close to 0.22ohms, so you can probably measure it in the circuit without any issues.

Why would the resistor be rated 2-5 watts? ..... The power in the resistor is I * I * R ... or about 0.22 - 0.25 watts. Quite possibly a 1/2 watt resistor, but may have gone 1W if really concerned about heat sensitivity of the current setting.

Semiman

 

[SemiMan]

There is a very good chance that these 3 resistors in parallel are setting the current level. I can't read that top resistor though. The other two in parallel would give 0.5ohm, higher than I expected.

The GC9910 is a clone or somewhat dropin for the Supertex HV9910.

 

[Denjava]

It's likely the only thing remotely close to 0.22ohms, so you can probably measure it in the circuit without any issues.

Why would the resistor be rated 2-5 watts? ..... The power in the resistor is I * I * R ... or about 0.22 - 0.25 watts. Quite possibly a 1/2 watt resistor, but may have gone 1W if really concerned about heat sensitivity of the current setting.

Semiman

Hello, I was pointing at resistors on the PCB which are about 2W as possibility for the right choice. In theory this is correct but these are on the current path and can be exposed to some voltage peaks, current peaks and if nothing else small cross section of connection wires can cause voltage drop that could be the problem for the output part of the circuit so heat is not everything to be considered about. In the Audio, MOSFET Source resistors are at least 5Watts. But maybe I am wrong anyway, thanks for the opinion.

 

[Denjava]

I don't know if anyone has noticed but this PCB has the current control mechanism (Page 4.). It is the CS pin on the U1 (pin 2), which is connected on the MOSFET source pin and the resistor we are looking for. This can be help for finding it, but also maybe nothing should be done about this whole subject because of this "smart" addition. And also the first LED setup was maybe burned up because of something else if this option also shuts down output completely and so the current flow.

 

[easyledlighting]

Should you be temping that?
wouldn't it be easier to use a Lab power supply which can adjust current setting to test it out?
then buy one that has correct output

 

[easyledlighting]

And the other thing about high power LED is: Not just because it says 30W on paper, so you can drive it up to 30W. The LED and PCB must be carefully designed to handle the waste heat come out of it.
So the other equipment you need is thermo meter to monitor the LED, to connect it as close as to its case.
<img src="http://easyledlighting.com/_files/images/overdrive.jpg" alt="my set up">

 

[SemiMan]

Hello, I was pointing at resistors on the PCB which are about 2W as possibility for the right choice. In theory this is correct but these are on the current path and can be exposed to some voltage peaks, current peaks and if nothing else small cross section of connection wires can cause voltage drop that could be the problem for the output part of the circuit so heat is not everything to be considered about. In the Audio, MOSFET Source resistors are at least 5Watts. But maybe I am wrong anyway, thanks for the opinion.

Voltage peaks, current peaks, etc. don't impact dissipation, only average voltage and average current. When a 1/4 watt resistor is all that is needed, you can bet a low cost power supply is not going to use anything bigger than that.

In Audio, the current in the output stage can be many times what this power supply provides, hence why a 5W resistor, but they are not "always" this big .... they may not even be there. It all comes down to topology.

Semiman

 

[magudaman]

Not 220 ohm, 0.22 ohms.

Semiman

Ah ha good catch!

Voltage peaks, current peaks, etc. don't impact dissipation, only average voltage and average current. When a 1/4 watt resistor is all that is needed, you can bet a low cost power supply is not going to use anything bigger than that.

In Audio, the current in the output stage can be many times what this power supply provides, hence why a 5W resistor, but they are not "always" this big .... they may not even be there. It all comes down to topology.

Semiman

Dude you are the man, I totally over looked those at being paralleled! Sure enough a 1 ohm, 1 ohm, 1.2ohm makes .35 ohm. I took off the 1.2 ohm one and I was at a maximum current of 512ma at the led! Perfect! The led runs much cooler now and is still bright enough for the kitchen table

Just for reference the resistors that control current on this unit are these:



Thanks guys for the information and replies!

 

[X_Marine]

Great detective work. ')
I have a similar looking driver from a cpl yrs ago and would love to be able to make it variable. Already swiped the image and printed it out for when I have time.

I also have one, could be this one, that I lost my description for and was going to scan and post to see if anyone recognized. Man you got to label everything you order from plum&nelly. lol

Many thanks
X/BillyD..