Chauvet LED 300 LED Driver Problems

[alistermorton]

Perhaps there's a ramp generator and a comparator to convert the analogue voltage to PWM?

[posty07]

1 hour ago, alistermorton said:

Perhaps there's a ramp generator and a comparator to convert the analogue voltage to PWM?

There is a Shunt regulator.. not sure if that's the same thing - but no ramp generator. As far as I can tell this isn't using pwm to dim the LED.

Seems to be current limiting instead, perhaps to ensure flicker free? I'm no expert so I'm not exactly sure but that's how I've understood the circuit.

[DrV]

Have had a dig through some stuff and this schematic looks as though it is what you have. Ignore the circuit references as they won't line up with your board but the principle will be the same I think. If you can't access the pdf I'll stick it somewhere public. I haven't tried posting a picture with the new forum software since it changed

Dave

 

Current source.pdf

[DrV]

Principle of operation: Voltage from the reference is compared with the voltage across the current sense resistors to establish a constant current through the LEDs.

U2 (the LM393) is used to short out the drive to the MOSFET. PWM from the controller is fed in through an overtemperature switch so the output is either off or at the current set by the LM358.

[posty07]

So if the 393 is outputting high, then the 358 also outputs a voltage to drive the MOSFET?

I really need to do some more research on these circuits as they're confusing me slightly!

[DrV]

The LM393 is a comparator and has an open collector output so it doesn't output a voltage as such. It can only short the gate of the MOSFET to the 0V rail. The LM358, in contrast, is a conventional op-amp with a push-pull output so it does output a voltage. This op-amp forms a negative feedback loop with the MOSFET so, provided the LM393 is off, a constant current flows through the current sense resistors and hence the LED.

[posty07]

AHH I think I get it. 

So when the 393 is 'off' it is floating but when on, the output shorts to ground..? Because of this, the current sense should be low, so the 358 turns off? 

Have I understood that correctly?

[DrV]

25 minutes ago, posty07 said:

AHH I think I get it. 

So when the 393 is 'off' it is floating but when on, the output shorts to ground..? Because of this, the current sense should be low, so the 358 turns off? 

Have I understood that correctly?

Actually when the 393 is on the voltage from the current sense resistors will be zero (because no current is flowing) so the 358 output will go high but it still won't be able to turn on the MOSFET because the 393 is preventing it. The 393 "overrides" the 358.

When the 393 is off the 358 is allowed to function normally and sets the current as required.

[posty07]

13 minutes ago, DrV said:

Actually when the 393 is on the voltage from the current sense resistors will be zero (because no current is flowing) so the 358 output will go high but it still won't be able to turn on the MOSFET because the 393 is preventing it. The 393 "overrides" the 358.

When the 393 is off the 358 is allowed to function normally and sets the current as required.

I finally get it! 

Thank you for your excellent explanation, that's genuinely really helped me.

So to test this I'm going to need a scope aren't I..?

[DrV]

Actually you could do a lot with a decent voltmeter - when the led is full on you should at least be able to measure the voltage across the current sense resistors and the reference voltage. Maybe compare with a working PCB.

[posty07]

18 hours ago, DrV said:

Actually you could do a lot with a decent voltmeter - when the led is full on you should at least be able to measure the voltage across the current sense resistors and the reference voltage. Maybe compare with a working PCB.

Yep that worked perfectly, the dim one is only getting 23.5mv whereas a good one is showing 100mv.

However the dimmer one is the one that I've replaced the MOSFETs with the wrong or at least different ones. Is this likely the issue?

I'd guess this will happen if the gate doesn't get enough voltage to turn on enough?

[posty07]

So, I was doing a bit more measuring now I'm familiar with the proper operation of those 2 chips and turns out even the "working" one had a problem.

Dry joint on the drain of one of the MOSFETs was causing it to dim slightly too because the current sense resisters weren't getting anything.

Fixed that and I can now see that those replacement MOSFETs are just not letting through enough current at all. I'll update once the new ones are in!

Thanks for all your help so far

[posty07]

On 11/2/2022 at 10:12 PM, DrV said:

Actually you could do a lot with a decent voltmeter - when the led is full on you should at least be able to measure the voltage across the current sense resistors and the reference voltage. Maybe compare with a working PCB.

Hi @DrV

I've now replaced with the new exact replacements and it's working perfectly. Both sets of current sense resistors are getting ~100mv.

One thing I'm a little concerned about is the MOSFET driving voltage is different for each one and my thought is maybe the LM358 is damaged and not outputting the full voltage.. is this likely?

So measuring the LM358 we have:

Vcc: ~5v

In1+: ~0.9mv

In1-: ~0.9mv

Out1: ~4.8v

In2+: ~0.9mv

In2-: ~0.9mv

Out2: ~3.6v

 

Am I wrong to assume there should be a much lower voltage drop? 

[DrV]

Ignore the In+ and In- pins - you can't glean any information from their voltages as they are current inputs rather than voltage inputs. What you are seeing is simply the LM358s doing their job - their outputs will vary in order to get the correct current through the sense resistors. The difference between the drive voltages is to adjust for slight differences in component values. These include, but are not restricted to, the value of the sense resistors, the gains of the MOSFETs, the slight temperature difference between the MOSFETs, the Vgsth of the MOSFETS and whether there is an R in the month! If you were to squirt some freezer spray on one of the MOSFETs while monitoring the output voltage of its associated LM358 you'd probably see the voltage swing quite markedly while the LED current remained constant.

So, in summary - don't panic! It's working correctly!

Dave

[posty07]

On 11/7/2022 at 10:22 PM, DrV said:

Ignore the In+ and In- pins - you can't glean any information from their voltages as they are current inputs rather than voltage inputs. What you are seeing is simply the LM358s doing their job - their outputs will vary in order to get the correct current through the sense resistors. The difference between the drive voltages is to adjust for slight differences in component values. These include, but are not restricted to, the value of the sense resistors, the gains of the MOSFETs, the slight temperature difference between the MOSFETs, the Vgsth of the MOSFETS and whether there is an R in the month! If you were to squirt some freezer spray on one of the MOSFETs while monitoring the output voltage of its associated LM358 you'd probably see the voltage swing quite markedly while the LED current remained constant.

So, in summary - don't panic! It's working correctly!

Dave

Ah that makes perfect sense, thanks! In that case they're all working as they should after mosfet (and thermal pad) replacement surgery.

Thanks again for your help on this!