Screaming amp6 basic?!

[chris661]

Hi all.

Some of you guys might recognise me from the speaker area that I usually lurk in.

Anyway, I've had my little amp6 basic for a while now, and it seems to have developed a fault.

When I turn up the input (both straight out of my PC and out of a mixing desk), past a certain point, it puts out maybe 3kHz at seemingly full power, but it gets louder and louder (some kind of feedback loop?). I've never had this problem before, but trying it with the mixing desk with the PC turned down confirmed that it's the amplifier itself that's the problem.

The modifications I have made to the original design are:
Additional power capacitors (4,000uF added in total - some recommend more, others don't). When I did this, there were very high frequency pulses that came out around 6kHz, which was the SMPS (60w, so definitely enough), not enjoying driving a capacitive load. To sort this, I added an air-core inductor, sufficient to filter out the noise. It sounded fantastic (and has done for months) until just now, where I tried actively crossing over my speakers via foobar2000 (also tried windows media player to double check, same problem).

The volume at which the problem becomes apparent has been surpassed before now, with no complaints, so I'm struggling here.

For now, I'm going to get the scope out, and try it with a 10ohm resistor, see exactly what's going on here.

If anyone has heard of anything similar, or knows what to check, I'm happy to have a poke around. There's no apparent damage to any of the components, and heatsinking is sufficient for a design like this.

Will report back when I've played around with the scope.

Thanks for reading,
Chris

[chris661]

Well, had a poke around with the 'scope.

Nothing out of the ordinary to report. There was +6vDC on each side of the resistor, resulting in no current flow, so I figure that's Ok.

When turned up, it clips hard at around supply voltage. The power supply, under heavy load, seems to emit a high frequency sine wave, so I guess that needs to be investigated further, but I suspect it would be caught by the capacitors and inductor.

The input doesn't exceed 2.5v, and supply is less than the 14.4v maximum.

I'm going to introduce a reactive element, in the form of a rubbish Wharfedale speaker, that's survived various amplifier tested stages previously. I'll also take some pics before I proceed, see if you guys notice anything I've missed...

Chris

[chris661]

Took the resistor out of circuit (it was in series with the speaker, to reduce volume a bit).

6 ohm load, turned it louder and louder. It clipped, but it didn't do the high pitched noise. What's going on here...? The only thing changed is the source (MP3 player, instead of PC, as the whole thing is down stairs, where there's more room).

Gonna try 'scoping the PC output now. I can't see that having any problems though...

[chris661]

Nothing showed on the PC output, checked both channels. No inconsistency there.

I'm beginning to question the drivers in use at the time. Visaton FRS8M, cones coated and phase plugged. In the stock impedance curves, there's no hideous impedance spikes or anything like that which might cause problems to some amps (I know valve amps tend to prefer smooth impedance curves, though it's very dependant on design).

Took out the extra capacitors and inductor. If I'm going to bi-amp using this for the mid-high, bass isn't an issue.

[chris661]

Well, I'm going to take everything back upstairs, see if there's a repeat performance.

[Pano]

That's a new one on me, Chris. Never had that problem with an AMP6. I would have suspected oscillation at the input stage, not output. You do see the 3Khz on the scope, right?

Can you get it to "scream" on a purely resistive load?

[chris661]

I couldn't, and I've just confirmed what you suggested - something wrong with the input.

Using a cheap USB DAC, it's fine no matter how loud it is on the mixer. Ohh, I didn't notice before (assumed it was just the music) - the "scream" shows up on the LED VU meter on the mixer. However, when I 'scoped the PC's internal sound card's output, nothing showed up, so I moved on.

Gonna try again with it, but I suspect the sound card is to blame...

[chris661]

The mystery deepens.

The mixer has 2 outputs (identical, one for a recorder, one for an amp), so I put the scope on the rec. and the amp6b on the amp output. It still oscillated. Unplugged the amp, and it went away, no matter how far up I turned it.

It's difficult to get a reading of the frequency: it starts off as a constant, but, after half a second or so, flies off the vertical scale. I don't want the amp/speakers to cook, so I turn it back down.

[Pano]

So the AMP6 and the sound card are not playing nice together?
What's the front end of the AMP6 like? What sort of pots, caps connectors, wire have you got in there?

I assume that the mixer does not do this with any other amp, yeah? And the AMP6 does not do this with any other source?

[Saturnus]

Could be a ground loop.

Try a much different input cable length and see if that change the pitch.