Help with Wharfedale SPX815 Pro please

[jstbrowsin]

Hi there,

I have a SPX815 powered mixer on the bench at the moment and I'm totally stumped. It arrived with a low distorted signal output on the left side and running it up on the scope we had the wave below.


While I was testing the transistors I shorted out a transistor (Q805) which in turn took out all the output transistors and few other components (circled on schematic)

Wharfdale SPX815.pdf
I replaced all the parts and while doing so noted the input circuit was underneath the board (all SMD). So I tested what I could and replaced a burnt out transistor (Q802A) and later (U801A)

Then I ran it up on the scope again and got a clean sine wave yay! But... and here in lies the problem when I plug the in the speaker I get output for a couple of seconds then the protection circuit kicks in for 5 secs and so on. I know the speaker is fine and there is 0V on the output before the relays so now I'm stumped??? Having never worked with this type of circuit before ( I normally repair guitar amps etc)

Can anybody help me please?

Many thanks in advance,

Trevor

[soundguruman]

You still have shorted or leaky parts in the circuit.

When you rebuilt the circuit, you missed something.

All the components should be rechecked.

The protection is shutting the circuit down, because when you load it with a speaker, something is drawing too much current...

You must still have a bad transistor, diode, blown resistor, etc...
Maybe the bias is shut off?

You have to check all the parts, not just some.

[diydidi]

Circuit looks very similar to Behringer EP2500 circuit. Class-H design.
Check all of the things sound guru mentioned.

[jstbrowsin]

Thanks,
Firstly the bias is working fine which is 2V same as the right side of circuit and when I turn the pot the voltage changes on the bias transistor also. . . so I think that means the bias part if working right yes?

So are you saying that just any faulty component on the board could cause this issue?
. Is there any way I can isolate the fault to a certain part of the circuit other than testing all of the components?
Is there any part of this circuit that wouldn't affect this please?

Cheers,
Trevor

You still have shorted or leaky parts in the circuit.

When you rebuilt the circuit, you missed something.

All the components should be rechecked.

The protection is shutting the circuit down, because when you load it with a speaker, something is drawing too much current...

You must still have a bad transistor, diode, blown resistor, etc...
Maybe the bias is shut off?

You have to check all the parts, not just some.

[km6xz]

I have never seen one of that model before but it should not be terrible hard to figure out. Drive is OK since you can get signal with no load so that limits the problem to the current boost of the switching rails output section. The protection circuit is triggered by several source, temperature, power switch, and LD1 and LD2 which trigger is base drive is higher than 5.6 volts above the emitters of the power transistors.
So, that reduces the number of possible causes quite a bit. Does the loaded output appear to be normal for those few seconds before the protection triggers?
Monitor the base drive to see if the base emitter voltage rises high enough to trigger the protection circuit. Use a scope with two channels. Attach one channel's probe to the base rail and the other to the output line, both on 2 or 5 volts per division. Then switch it to Channel 1 and 2 Summing, and then switch the polarity invert switch. Most decent 2 channel analog scopes have those opention. The resulting trace should display only the difference between the bases and emitters regardless of signal level so it should be easy to see if there is a sudden shift in the difference. If there isn't, good, move on to testing the thermal switching by monitoring the anode of ZD903. Does it shift in level right before the relays trigger protection? No? OK move on. If it does, note whether the fan control amp is working, does the fan start up during any of this?
I would be most suspect of anything replaced or damaged when testing since these are not the same symptoms. Q802 controls the offset between positive and negative drive needed for turning on the power transistors just enough to keep it in Class AB instead of Class B with very low or not bias. That is not the problem however since your loaded output looks OK for a few seconds without a ton of cross over distortion.

Does anything seem to be pulling more current from the wall than normal? If there is nothing suspicious there, it could be the protection circuit itself, like the temperature tracking resistors R942 and 944. Open emitter resistors could trigger the protection and give a good output with no load, but not with a load.
You should not use speakers as loads until the amp is stable and trustworthy. Use a power resistor to test without risking an expensive driver, besides testing a power amp this size with a live speaker would not make you popular with your neighbors.
Three or four logically test points could narrow it down to a very few possible causes. Don't bother checking each part passively, that is a waste of a lot of time. A mechanic would not do that, they would test a subsystem for expected output and conclude that all the individual parts that make up the subsystem are OK. So if the engine starts, idles and accelerates, it tells him hundreds of parts are functioning without testing them individually.
Same with an amp. Figure out what you should be getting from each subsystem and make one test that would show it working as a subsystem or not. Seeing that the circuit could swing a wide voltage range with no load is the only test you needed to show that everything prior to the level shift circuit is fine, for example.
Oh, that power output 80 and 50 volt rail switching should not be the culprit if you are trying to drive the load with less than 98 volts peak to peak. If it only shuts down with a wider swing voltage, you again have isolated the problem to the rail switches Q819 or Q818 or their level detectors U805 a and be section. Ignore that portion of the circuit if the protection triggers at lower output voltage.

Good luck

[jstbrowsin]

Thanks heaps km6xz.

I'll go through your suggestions systematically and come back to you when I'm done and (hopefully) have found the offender(s)

[km6xz]

Hope you get it.
As an aside....I am probably one of the few on this forum who been in your neighborhood, or even know where Tauranga is. In the 80s, had a beautful girlfriend from there. She was an actress, serious drama, and model and came to the US for a Pepsi commercial when I met her. She overstayed her visa and it took years to get it straightened out. We visited her home town a few times after getting legal residency in the US so she could travel. We spent most of our visits around her home town, a few miles south of you along the coast of Bay of Plenty. We tramped around South Island for a month during our second trip back. Her little sister was a model also. My last visit there was a surprise. Her town was really a small ocean side village but 15 years later when I went back it was completely built up with homes and shopping centers.

[jstbrowsin]

Hi "Does the loaded output appear to be normal for those few seconds before the protection triggers?"

Funny but the output wave resembles that of the first wave with cross over distortion but also the wave is inverted. Does this mean the fault lies on the negative rail somewhere?? Possibly an RC network??
I'm not too sure wether to carry on with your diagnosis procedure so just wait for your reply.

Heres a pic of the output before protection kicks in.

Thanks

Trevor

[jstbrowsin]

Hi km6xz,

Was her hometown a place called Whakatane? Anyhow yes we count ourselves lucky to be living here I just love this place. It sounds like you may have been a bit of a rock star in your day surrounding yourself with models. . . you lucky bugger
Anyhow if I throw a couple of diodes and capacitors into this conversation then we're back on topic eh LOL

Anytime you over this way look us up ok


Cheers,
Trevor

Hope you get it.
As an aside....I am probably one of the few on this forum who been in your neighborhood, or even know where Tauranga is. In the 80s, had a beautful girlfriend from there. She was an actress, serious drama, and model and came to the US for a Pepsi commercial when I met her. She overstayed her visa and it took years to get it straightened out. We visited her home town a few times after getting legal residency in the US so she could travel. We spent most of our visits around her home town, a few miles south of you along the coast of Bay of Plenty. We tramped around South Island for a month during our second trip back. Her little sister was a model also. My last visit there was a surprise. Her town was really a small ocean side village but 15 years later when I went back it was completely built up with homes and shopping centers.

[km6xz]

The wave form shown suggests a device breaking down, reaching a zener point and avalanche occurs until the drive forces it off. That might be one transistor failing. The wave form looks normal without a load? You isolate the device by using the scope in differential mode, attaching one probe to each side of an emitter resistor and invert one channel. If a single device is causing this, you will see a low level version of this sudden spike while the others will see a lower level with inverted waveform due to responding to the negative feedback trying to correct it.
The waveform shows 150 peck negative volts so it the MOSFET on the negative rail switch is turning on. That is probably a symptom of the higher negative swing than cause however.

[jstbrowsin]

HI km6xz, thanks for getting back to me. . . yes the wave form is normal without a load, but I just noted something else. . . with a load just before the protection circuit switches on the input shows a small amount of cross over distortion also. Please note that both the input and output are blown up to 50V per div to make it easier to see.




Cheers,
Trevor

The wave form shown suggests a device breaking down, reaching a zener point and avalanche occurs until the drive forces it off. That might be one transistor failing. The wave form looks normal without a load? You isolate the device by using the scope in differential mode, attaching one probe to each side of an emitter resistor and invert one channel. If a single device is causing this, you will see a low level version of this sudden spike while the others will see a lower level with inverted waveform due to responding to the negative feedback trying to correct it.
The waveform shows 150 peck negative volts so it the MOSFET on the negative rail switch is turning on. That is probably a symptom of the higher negative swing than cause however.

[jstbrowsin]

Latest update. . .

It seems that after testing the output transistors via a scope (thanks km6xz) it appears to me that I have 3 bad transistors compared to all the others. It took over 3 weeks for me to get these last time so can anyone recommend a reliable seller with fast international shipping please? I got these from Polida2008 on Ebay its funny that the negative rail transistors are all running as expected though yes?

Anyway here's the link to the video footage. . . http://youtu.be/ZjTr27KyJ2w

Comments welcomed

Cheers,
Trevor