Troubleshooting a hifi system, speakers got a complex filter circuit.

[Yamin]

Hi guys,
I was working on this Onkyo amp which had a shorted channel. One of the power NPN transistor had shorted out and one half of the noble resistor was also open. (I made a previous post about transistor replacement). Anyway I took the two power transistors and the noble resistor from a working channel that is not used. Amp was working perfectly when I tested it. I should ask is it ok to leave the unused channel without the two transistors and the resistor?
Ran it for sometime I heard a relay click not sure which one it was - no error message nothing and the amp kept on working. Turned off and on again no problem was working.
So I took it to its owners place and I wanted to check whether the speakers had shorted as this was my suspicion for the original failure.
He had two tower speakers - I measured across the terminal (it has got a filter circuit) and here is something peculiar I found:
1.Left speaker measured in the mega ohms range right away.
2.Right speaker showed the numbers crawling up in other words I saw the capacitor charging, and I heard the woofer move when touched with the DMM.

The owner was using it for sometime and he called me back saying the amp had died again. From what he described I know that the fuses on the B+ voltage rails had blown and the amp didn’t have time to go into protection mode.

So guys am I right in thinking the speakers might at be at fault here. I am going there tomorrow to see whether the same channel had shorted out.

Any adivce/tips on the matter would be greatly appreciated.

[_Wim_]

1.Left speaker measured in the mega ohms range right away.
2.Right speaker showed the numbers crawling up in other words I saw the capacitor charging, and I heard the woofer move when touched with the DMM.

Both measurements seem strange to me, unless you mean with crawling up it goes from very low impedance to something like 8 ohms, but it should never go to a high value for a tower speaker (no high pass filter on the woofer, so only the coil(s) and woofer should result in some max low DC resistance). But if the speakers sound normal, then it must have something todo with how you measured them, because a mega ohms impedance with not give you a lot of sound on a normal amplifier...

To check the speakers you can easily measure there impedance of the full frequency range using a normal sound card and free software: https://www.roomeqwizard.com/help/help_en-GB/html/impedancemeasurement.html
The impedance should not drop below 3-4 ohms over the entire frequency range, it is does, than they do indeed have a problem.
However, it could be impedance looks ok at this low test voltage generated by the sound card, but the speaker can still have a problem at high voltage (volume). Did the amp always die when the owner was listening to loud music?

As to the channel without transistors, without a schematic it is hard to tell for sure, but if it is the final output transistors and the emitter resistor it should be ok I think.

[Yamin]

Thanks, the speaker measurement seems very strange to me as well - the values crawls up and does not stop at 8 ohms or any familiar value, it goes into the kiloohms. I'm thinking that there might be a series capacitor, but still two speakers showing different values is a bit odd.
With regards to fidelity I haven't heard the speakers before so was not sure of how they should sound. It was kind of difficult to gauge the lower end as the owner also had a subwoofer connected. He seemed happy with the sound.
With regards to the transistor removed - it very much like any other amplifier schematic wise. I had removed the two power transistors and the emitter resistor. Shouldn't be an issue right leaving it like that.

[_Wim_]

I'm thinking that there might be a series capacitor, but still two speakers showing different values is a bit odd.

That would be very strange for a tower speaker. If the speaker were satellites used together with a subwoofer I would understand, but for a full range tower speaker I do not see why a high pass filter would be installed on the complete speaker.

[Yamin]

Yes my thought exactly, however I came across a online post and someone was saying the same thing as me.

Update:
So I had a look today. Checked out the speakers and they both does in fact behave the same way numbers crawl up like a cap charging and stops at around 0.7Mohm and when I short the two speaker leads together I can hear the woofer moving (possibly the cap discharging).
This time too the same channel had shorted out, but for a change the two power transistors (NPN,PNP) has shorted. The noble resistors remains ok. Looks like something upstream from the power transistors had shorted damn  .
I did not swap the driver transistors from the good channel I should have done so!  I didn't do that as it test ok. Even now its testing ok. I shall order them and change all when I'm getting the new spares.
Could temperature detection circuit cause this kind of failure? I noticed that the thermal compound has become very crusty with the sensors attached to the transistors.
What else should I keep in mind to further troubleshoot?
Something else I should remark is that I noticed that B voltage rails are around 54 to 55V. Schematic says it should be 51V, I'm assuming that this is ok.
Appreciate the help.

[Yamin]

Ive been trying to upload the schematic of the power amp section and its not going through. The amp model is onkyo SR508

[_Wim_]

Is this by any change a DIY speaker or a modified (tweaked) cross-over in the speaker? Could be that at certain frequencies the impedance goes to zero creating problems with the amplifier. High capacitance to ground could for example also make the amp oscillate. I have heard stories of audiophiles thinking they tweaked the system perfectly because the high tone are "so much clearer" now, but what actually happened was the amplifier started oscillating 

I would really measure the impedance of the speaker. It’s easy, only a laptop with headphone out and a 100 ohm resistor or similar is needed.

Edit: corrected 10 ohm to 100 ohm

[Yamin]

Is this by any change a DIY speaker or a modified (tweaked) cross-over in the speaker? Could be that at certain frequencies the impedance goes to zero creating problems with the amplifier. High capacitance to ground could for example also make the amp oscillate. I have heard stories of audiophiles thinking they tweaked the system perfectly because the high tone are "so much clearer" now, but what actually happened was the amplifier started oscillating 

I would really measure the impedance of the speaker. It’s easy, only a laptop with headphone out and a 10 ohm resistor or similar is needed.

Wow very interesting thing to be aware of, thanks for that. As per to the owner this is the first time he had any problems in six years - it was bought  straight from the shop and no one had tinkered with it.
But still a fault developed in the crossover circuitry could in fact cause similar kind of  problems right?
Could you please explain the setup a bit  on how you go on about measuring the impedance across the frequency ranges. I have got a external sound card. How should the 10 ohms resistor be connected, would the sound card be able to drive the tower speaker?
Very much appreciated.

[_Wim_]

Right channel of the headphone is connected to 2 places:
- the 100 ohm resistor (forgot a zero in the previous post, now corrected)
- the left input channel (line in, not microphone input) of the sound card

The other end of the 100 ohm resistor is also connected to 2 places
- the speaker system that is being tested (DUT)
- the right input channel (line in, not microphone input) of the sound card

This is how it works:
- because of the series resistor, the headphone output sees now a resistance of >100 ohm and can drive the speaker (of course not very loud)
- the 100ohm resistor and the speaker form a "voltage divider"
- the left soundcard input measures the "full" signal
- the right soundcard input measures the divided signal

Because the software knows you connected a 100 ohm resistor (you have to enter the actual value used), it can calculate the speaker impendence by comparing the left and right input channels. The software will generate a sweep tone from 20Hz to 20kHz and measure the speaker impedance of each frequency. This way you get a plot of the complete impedance and you can check if it behaves abnormal somewhere.

[_Wim_]

But still a fault developed in the crossover circuitry could in fact cause similar kind of  problems right?

Hard to say for sure, but for me this seems less likely because the speaker still sounds "normal" to the owner, and I would expect that a commercially build cross-over has no extreme impedance dips that could destroy an amplifier.

[_Wim_]

Software can be downloaded here:
https://www.roomeqwizard.com/