Behringer PMP4000 half working

[DaedalusWren]

Question... T2 & T3 require 30+ volts to Gate to open. Since it is in circuit, would measuring Gate to ground show the gate voltage or would I have to pull them to check if the 30+vdc was present? Seeing as though the Secondary has no voltage, I can only assume that with the transformer ringing through, there is no power to the primary windings. Therefore, following logic of standard non-switching supplies, I look to the source. In that thinking I checked the 110vac through to the windings. D3 has supply and output. Looking to the other side of the winding I see the circuit controlled by IC4. As T2 and T3 have been replaced, I assume they are good and would move onto testing T4, T5, D5 and D7. Would this logic be right or would you suggest I probe or test some other area?

[nickb]

Question... T2 & T3 require 30+ volts to Gate to open. Since it is in circuit, would measuring Gate to ground show the gate voltage or would I have to pull them to check if the 30+vdc was present? Seeing as though the Secondary has no voltage, I can only assume that with the transformer ringing through, there is no power to the primary windings. Therefore, following logic of standard non-switching supplies, I look to the source. In that thinking I checked the 110vac through to the windings. D3 has supply and output. Looking to the other side of the winding I see the circuit controlled by IC4. As T2 and T3 have been replaced, I assume they are good and would move onto testing T4, T5, D5 and D7. Would this logic be right or would you suggest I probe or test some other area?

I like the way your are logically thinking it through. The FETs only need about 5v to turn on. They are fed with a pulsing signal so the average level will be a bit less than half of the peak level, whatever that it.

Anyway...

The symbol for the KA3525 shows the shutdown pin 10 as active low (SHDN with a line over it means the function is active when the signal is low). THis is why earlier I said the pin should be low if shutdown. This is wrong. That pin has to be pulled high to shutdown. I have no idea why they got the convention wrong on this particular chip. At first I couldn't believe it, but it's true.

So if you are seeing +3.7V on pin 10 then it's shutdown. So follow that through - what is pin 6 of IC3 doing? Is Vref 5V?

[Enzo]

T2 T3 are not referenced to ground, they are referenced to the V- rail, approx -170vDC. That is why we use isolation transformers to work on SMPS, and why they are so dangerous, the mains are directly rectified and filtered.


Are we scoping or metering? Unless it is pretty special, a hand meter will not measure anything in teh transformer circuits, they work at WAy too high frequency for a meter.

[nickb]

T2 T3 are not referenced to ground, they are referenced to the V- rail, approx -170vDC. That is why we use isolation transformers to work on SMPS, and why they are so dangerous, the mains are directly rectified and filtered.


Are we scoping or metering? Unless it is pretty special, a hand meter will not measure anything in teh transformer circuits, they work at WAy too high frequency for a meter.

It's true that to really look at dynamic operation, a meter is almost useless, but we haven't got to that stage yet. That said, measuring average DC across the gate-source that is greater than say 0.1V at least says some drive is going on. But, since the PWM is still is shutdown, we have to look earlier in the system to stand a chance.

BTW, the transformers are not both referenced to -170V only the low side driver is. They are both referenced to their respective source terminals.

PS: Just in case you were think of breaking out a scope you will need an isolating transformer.

[Enzo]

My concern was that he reported finding no signal on the transformer windings, and I doubt a meter could detect such even in a working unit.

he also asked if he should measure the gates to ground. Regardless what they are referenced to really, none is to ground.

I didn't want to see him basing measurements on a faulty premise.


The shutdown pin on the IC4 has a bar over it, to me that means active low, so I would expect above the threshold to be NOT shutdown. Does the data sheet say otherwise?

[nickb]

My concern was that he reported finding no signal on the transformer windings, and I doubt a meter could detect such even in a working unit.

he also asked if he should measure the gates to ground. Regardless what they are referenced to really, none is to ground.

I didn't want to see him basing measurements on a faulty premise.


The shutdown pin on the IC4 has a bar over it, to me that means active low, so I would expect above the threshold to be NOT shutdown. Does the data sheet say otherwise?

Gotcha When he said 'ground' I interpreted it as 'local ground i.e the source of each FET.

I made the same assumption as you about active low. It's just not true as far as I can see. When that signal is high the PWM inv input is held low, the soft start cap is discharged, the PWM flip flop is held SET and both drivers are set to pull down via the OR gates. I also found a reference elsewhere that said the same thing. ALso, the lab test fixture shows a switch pulling the signal high to turn off. It just seems to be a documentation quirk and something of a timewaster for us.





Go figure...

[DaedalusWren]

Looking at the Lab Test pic, I would also assume the same. When Vref is applied to pin 10, bringing the pin high, the shutdown is in effect.
IC3 has 0.64vdc at pin 6 and 5.04Vdc at pin 4 or Vref.

[nickb]

Looking at the Lab Test pic, I would also assume the same. When Vref is applied to pin 10, bringing the pin high, the shutdown is in effect.
IC3 has 0.64vdc at pin 6 and 5.04Vdc at pin 4 or Vref.

Sorry been busy..

There are 3 ways for IC3 pin 6 to be low:
(1) Overcurrent via IC2c - since you don't see an output I don't think that is it.
(2) Amplifier error via IC7. the amp is disconnected so that can't be it and
(3) Undervoltage via IC2a. Check pin 7 it should be about 8.5V

[DaedalusWren]

As far as you being busy, you have helped me a tremendous amount with this and i appreciate the time you have spared to aid me in this issue. So I do not mind you being busy at all.
As far as this project goes, I think I will put all probed readings thus far here to avoid having to re-trace the progress in previous posts.
Also, since you mentioned the amp being disconnected in the last post I thought I would ring through all the voltages again with the amp connected just to see if there was any difference. I found that the only difference was that Vcc increased by only o.o2Vdc with the amp connected while the rest stayed the same.
The following are all probes taken thus far.

IC4 pin 10 (shutdown) =3.7Vdc
IC4 pin 15 (Vcc) =15.3Vdc
IC3 pin 4 (Vref) =5.04Vdc
IC3 pin 6 =0.45Vdc
IC3 pin 12 =0.01Vdc
IC2a pin 7 =7.56Vdc
IC2a,d,c pins 6.9.11 =5.04Vdc

[DaedalusWren]

as a side note. In regards to the function of pin 10 of IC4. The following was found regarding the SG3525, which from what I can gather is similar to the KA3525A (IC4) except for some different values. However the Shutdown is also labeled the same on this chip with the line over the shutdown name.

Understanding SG3525 IC Pin Outs - Electronic Circuit Projects

Pin#10 (Shutdown): As the name suggest this pinout may be used for shutting down the outputs of the IC in an event of a circuit malfunction or some drastic conditions. A logic high at this pin out will instantly narrow down te PWM pulses to the maximum possible level making the output device's current go down to minimal levels. However if the logic high persists for longer period of time, the IC prompts the slow start capacitor to discharge, initiating a slow turn ON and release. This pin out should not be kept unconnected for avoiding stray signal pick ups.

[nickb]

As far as you being busy, you have helped me a tremendous amount with this and i appreciate the time you have spared to aid me in this issue. So I do not mind you being busy at all.
As far as this project goes, I think I will put all probed readings thus far here to avoid having to re-trace the progress in previous posts.
Also, since you mentioned the amp being disconnected in the last post I thought I would ring through all the voltages again with the amp connected just to see if there was any difference. I found that the only difference was that Vcc increased by only o.o2Vdc with the amp connected while the rest stayed the same.
The following are all probes taken thus far.

IC4 pin 10 (shutdown) =3.7Vdc
IC4 pin 15 (Vcc) =15.3Vdc
IC3 pin 4 (Vref) =5.04Vdc
IC3 pin 6 =0.45Vdc
IC3 pin 12 =0.01Vdc
IC2a pin 7 =7.56Vdc
IC2a,d,c pins 6.9.11 =5.04Vdc

Nice table

Well since IC2 pin 7 is > IC2 pin 6 then undervoltage is not the reason that IC2 pin 1 is low. So, we keep looking. Ignoring the possibility that IC3 is bad until other possibilities are checked, measure IC2 pin 10: I'd expect that to be zero. On the other hand it may be retrying. If that is the case then IC2 pin 10 will be greater than zero and we will have learned something interesting, namely that we have a overcurrent problem. This is an example point where a scope would give a quick unambiguously answer.

The other possibility is there is enough current flowing through the DC shutdown optoisolator LED IC7 to turn the phototransistor on and so pull down the signal of interest. To check that measure the voltage across pin terminals +15V and DC_SD.

Lastly IC7 could be bad, I suppose. And the other things I might have missed, of course...

On the SHUTDOWN polarity, the data sheet for the pin compatible UC3525 from Texas Insts shows the shutdown as active high and gives some descriptive text the says the same thing. Bizarrely, in the their lab fixture, they have the shutdown switch pulling the signal low

[DaedalusWren]

Ok. since we have both found indicators of the shutdown being active when Pin 10 of IC4 is logic high, The most solid I could find supporting this in the last sentence of the pin description. " This pin out should not be kept unconnected for avoiding stray signal pick ups." Meaning to me that the IC will function properly with this pin unconnected (logic low), but should not be applied as such. Assumption follows, that something is pulling this high and shutting down the circuit. Also seeing as these voltages are consistent, whether the amp it connected or not, can I tentatively assume that the amp board is not the cause of this state?

The probes you mentioned in your last post were as follows.

IC2 pin 10 = 0.00Vdc
voltage across +15 and DS SD of the IC7 Optocoupler = 0.07Vdc
voltage across the Emitter and collector of IC7 = 0.44Vdc

[nickb]

With a reading of zero on IC2 pin 10 we can be fairly confident that it's not retrying. The main outputs will also be zero if so. We also now know the DC_SD signal is not the cause. That leaves R39, IC2, IC3, IC7, T7, R52 and C32. With the power off you should be able to measure the resistance across R52 - this will test R52 & C32. Do the same with R39. Try meter both ways round and pick the higher for these measurements. IC3 would be my next target since it's connected to IC4 which was damaged.

With the others you can remove the chips one at a time or, if you are very careful, you can lift the pin connected to that signal of interest. Since this will be defeating the built-in protections and there still might be another underlying fault, it might be a good idea to disable the main switching path - I think the easiest way is to remove C21 temporarily.

[DaedalusWren]

Well I have been dealing with a cold for the past few days so I have not done too much. I did manage to get out and check the resistors though. The readings were not what I expecting. Neither rung through as 47kohms, but as they are in circuit I imagine they would not.
R52/C32 = .424 tested both directions with the same value.
R39 =1.830 tested both directions with same value.

I will get further into testing hopefully tomorrow.

[nickb]

0.4242 and 1.83? What are the units? It may not be possible to measure these in circuit. That's often the case but sometimes you get lucky so it's worth a try. BTW R39 is 4.7K.