Frustrating Intermittent Fault Triggers Amp's Protection...

[thermionic]

Hi,

I’ve now completely refurbished around 10 Bozak mixers with new pots / caps etc, and have fixed a few faults in my time. You can see the schematic here: http://djdalelee.tripod.com/bozak.html

Anyway, one particular mixer has got me completely stumped… Of course, the fault is intermittent and I am unable to get it to occur on my bench or in my listening room… I’ve had it over here twice now, and on both occasions it was fine – no DC offset, no oscillation.

Originally, my friend who owns the mixer told me his amp would trip out when he swept the volume control… I traced this to a leaky tant. The electrolytic at the output was new, but when the volume pot was swept fast, you could see an instant DC offset of up to 100mV or so – I reckon this swept DC got through the 220u output cap as it would have the appearance of low-frequency AC, but low enough to trigger the amp’s DC protection. I decided to replace all tants as par for the course. Result: no DC offset however fast the pot was swept.

He’s subbed other amps btw.

I get a call this evening telling me that his amp is cutting out again… (This is after I listened to the mixer for several hours before returning it, turning it off and on, and noting that DC offset never goes above 5mV). He says the bass driver jumps out (suggesting DC) and the right channel goes into ‘protect’ mode on his Audio Research amp.

I’ve now had the mixer on my bench twice and I can’t trigger the fault, let alone get a ‘scope on it to get an idea of frequency and shape…

All tants and electrolytics are new. If this were a PSU fault, i.e. intermittent cutting out, I can’t see why just the right channel's protection would trigger on the amp. Out of the many CMA10-2-DL-series mixers I’ve worked on, I’ve never seen a fault like this – I have always had the design down as inherently stable.

I wonder if any of the brains here can see something I might have missed in the schematic that would cause such a fault? I would really like to put this one to bed…this is losing me time I cannot afford. Many thanks in advance for any comments – please excuse my lack of data, but I cannot trigger the fault!


Justin

 

[JohnRoberts]

Single supply design so it is cap coupled at every stage. DC on output is most likely caused in final output stage.

If something is oscillating, that could be in PS or any stage with gain.

Did customer hear unusual noises, or just get DC and amp protect?

Problem in both outputs suggests possible common fault. power supply is probably only thing common to both outputs. Perhaps power supply fault high voltage or oscillating?

Good Luck.

JR

 

[thermionic]

Thanks, John.

The problem has always been on the right side. Last time I spoke to Felix, I told him to reverse L/R and see if the amp's protection was as sharp on its left bank.

I have never been able to get more than 5mV from the output, regardless of what I do in terms of abuse. The only thing I can think of now is an intermittent short somewhere internally. Felix seems to think that when he had the problem before, it was totally random, whereas now it's been re-capped, the issue starts when it's warm... He only got the mixer back a few days ago, so I'll take this with a pinch of salt as he won't have had time to reliably detail this kind of thing.

He claims the bass driver pops out (as if the amp's on and a DC offset has occurred) and then the amp shuts down into 'get off my back' mode.

One noteworthy observation is that the RCA 2N6291 / 2N6108 transistors on the output appear to have been replaced at some point... Do transistors have intermittent faults, such as when a junction reaches a temperature? The OPS are unobtainium here in the UK...it'll be a PITA to sub them to find out. Whereas the rest of the joints are free of flux traces, the OPS' pins are pretty messy with a rusty-coloured flux... Something tells me someone's been here before... The mixer has always caused the amp to trip, ever since its acquisition via Ebay... When I first saw it, the first 2,200u filter cap preceding the regulation was hanging off...that would cause an issue I thought... I then managed to stop the DC issue when the Master pot was swept - fault found I thought...




Justin

 

[JohnRoberts]

I vaguely recall something about transistor betas causing a factory problem in 10_2DL, but have no idea what module or any details. Transistor problems should not cause DC issues that would get past a blocking capacitor. If one channel has been repaired, maybe compare voltages and waveforms between two channels.
------
At the risk of asking a stupid question have you checked output blocking capacitors for polarity..? In rare cases I've even heard of caps mislabeled, but I've never seen that myself. With DC blocking, there should be no way to pass DC.

Here is another long shot. What is the DC input impedance of the following amplifier? All electrolytic capacitors have a tiny amount of DC leakage current. If there is no path to ground or very high impedance it might build up voltage.

I don't see any DC path on the schematics I looked at. Is there any resistance to ground on the Bozak output lines (say 50k-100k) to drain off any leakage current? I am not aware of significant temperature effect on leakage current but it is proportional with amount of capacitance and voltage.

Most amplifiers will have some dc resistance to ground in their input stage but who knows?

Of course it seem all of this should show up on the bench.

JR

 

[PRR]

I had a low-price phono preamp kick-out a receiver when turned to near live-piano level. It had large 1/f subsonic noise and extended bass out its output coupling. The VTVM on 1.5VDC would show ~~20mV flicks, apparently just large/long enuff at the speaker to upset the protection. And yeah, the woofer cone was bopping.

That Bozak stuff is even less sophisticated tham my $49 chip amp.

You may "need" a good stiff 20Hz low-cut to feed modern over-sensitive power amps.

 

[JohnRoberts]

Warped records can generate infrasonic energy. The Bozak RIAA looks like it has a HPF pole at 5 Hz, line inputs go even lower. A bad transistor could introduce LF noise depending on failure mode.

This also depends on how twitchy the amp is...

JR

 

[thermionic]

Last night I had another thought. I failed to mention that, when I was working on the ramped DC issue that I thought was causing the problem, during testing afterwards, I found a peculiar anomaly with the Tone Control.

When touching the Right (problem channel) channel’s Treble pot, I noticed all kinds of random DC offsets…I couldn’t get a steady waveform or offset, so it was difficult to figure out the exact nature of the issue, so I put it down to a bad transistor and replaced both BC550B on the Tone PCB for good measure (as well as the 10pF cap going between Gate and Source on the input FET – for the couple of pennies it cost me, I thought it made sense).

I put everything back together and the mixer was as good as gold. I played it into my mono-blocks with no issues (my amps don’t have protection btw). I played it a couple of times to myself and to Felix with no issues. I spent a few minutes abusing the Treble pot to see if the problem would come back, but it appeared ok, with no DC to speak of appearing anywhere.

The Tone control uses AB modpots which have an unusual construction. I recently found one that had a short on the wiper connection, but it was so sensitive you could almost blow it and it’d make contact – any slight vibration would make it connect intermittently, in a kind of pulsing, switching manner. I wonder if, by re-centering the Treble pot after I removed it for inspection, I gave it a temporary reprieve, making me think that the transistor replacement had done the job, only for the problem to return once the mixer had travelled back to Felix’s…???

Does the above make sense?

The Bozak specifies BC550B / 560B (it’s identical in topology to the schematic in the link, but subs transistors as the 550/560 have better Hfe – as John knows), and for the 5cents the C versions now cost, I almost feel tempted to replace the whole lot in the Tone / Summing and Amp cards. I can’t see an issue if I fit the ‘C’ versions – they have better Hfe, so it could be an improvement.

The RCA 2N6291 / 2N6108 output drivers which have been replaced are a bit ominous… But surely, if the engineer who replaced them really knew what he was doing, he wouldn’t have left such messy flux traces after the work… What about the 2N3822 FETs? Could they cause an intermittent fault of this nature?

If I can trigger the fault and get a probe in there to look I have a chance. If I can’t, I’m looking at pretty crude, scatter-gun techniques…

Thanks again!

Justin

 

[bcarso]

I'd suspect the pot.