Peavey XR696F crowbar issue

[Mick Bailey]

I have a problem with the power amp that's got me beat at the moment. It came to me with intermittent serious noise issues. I did eventually find one of the ribbon cables chafing against a component leg. I'd got the mixer section suspended and tapping the board flashed the bulb limiter.

When I initially got the amp I noticed the emitter resistors looked cooked on the RH channel and the protection SCR had heat discoloration to the tab. So I temporarily left the SCR out and checked the amp on the limiter. It was running fine, so connected a load and it still looked good. With the amp powered directly it was still OK on load thought the +/-90v rails read +/-103v and the 40v rails read 46v. With a signal it easily reached full output. I checked the diac to see if it was leaking and that's fine, so installed a fresh SCR.

The problem is, the amp will crowbar occasionally at switch-on. Maybe one in 50 or more. It does not do this with a load and with the bulb limiter connected. Neither does it do this if I reduce the input voltage so that the rails read correctly. Once the amp is powered up OK it continues to run fine for hours on end.

The amp must have been running with the high voltage rails all its life, but I'm wondering if component aging means that the power rails aren't symmetric enough at switch-on to prevent the protection circuit from operating.

Any thoughts?

[J M Fahey]

+/-90v rails read +/-103v and the 40v rails read 46v.
the amp will crowbar occasionally at switch-on. Maybe one in 50 or more. It does not do this with a load and with the bulb limiter connected. Neither does it do this if I reduce the input voltage so that the rails read correctly.

I BET that PT primary expects some 230VAC (2 x 115V windings in series) and you are feeding juicy healthy around 250VAC John Bull mains.
So not surprising.

Doubt the PT has a 250V tap so you might try to brown it down.
Even without crowbar triggering I'd feel uncomfortable with such high rail voltages.

At least 12VAC bucking would be needed.

The auxiliary transformer won't be small by any means, it must stand the full mains current needed.

Ouch!!! , just checked: 1200WRMS?
You'll need a 180VA 12VAC secondary transformer, which will *not* fit the overcrowded cabinet, mount it into some convenient box and tell owners it's a needed "power adapter" ... which it is.

Tell them about very high UK mains voltage and that a lot of stuff was designed for the old 220/230V standard.

They *may* not use this, at their own risk, but then it's their choice.

The crowbar circuit is not malfunctioning, just doing its watchdog job, and I wouldn't mess with it.

OK, OK, so there's no way they will do the proper thing and you want to brush them off?

Crowbar circuits in general protect from DC at the speaker output, typically there is an RC circuit delaying triggering.
Either increase capacitance (better) or resistance by 50% , it should take care of the erratic self triggering.

It will do nothing about the way too high rails of course

For a (possibly) better answer, please post the schematic.

[Jazz P Bass]

http://music-electronics-forum.com/a...pvy-xr696f.zip

[mozwell]

an alternative is a 240V to 220V auto transformer (like a variac but with a single fixed output setting)
A 10A 220V output auto transformer would be the same size & cost as a 200VA isolation transformer.
A 15A 220V output auto transformer would be the same size & cost as a 300VA isolation transformer.
Just get the auto transformer in a box, with a socket for your amp to plug into & a lead to plug the auto transformer into the wall outlet.
There should be transformer manufacturers in the UK who can help you

Note an auto transformer does NOT provide an isolated output, so you should treat the output socket exactly as you would your normal 240V wall outlet

If you were in Australia, i could help you with this, i don't know who can do this in the UK, but a google search shouldn't take too long.

Your rails will reduce to +/-94V and +/-42V with a 420-220V auto transformer

[Mick Bailey]

They run to around £100 over here. It's a good idea, though I would have liked to have got it back running how it was, though I think for the age and condition of it there's a limit on cost. I've had it running all week on 220v and it hasn't failed. I can't find anything wrong, but the over-voltage is clearly having an effect at startup. There could be an accumulation of component stress, but I'm suspecting the main cap charge time could be causing a differential that's large enough to trigger the diac/scr protection. It's only doing this one one channel - I think I'll swap over the filter caps and see if the fault moves.

[mozwell]

I would just remove the diac & scr, but inform the customer why you are doing this. With the parts gone, there is no possibility a false trigger can hurt the amp.
The reason they are there, is if there is a large DC on the output (failed transistor) the crowbar activates & blows a power amp fuse. This to protect the speaker.
The vast majority of amps have no protection like this & work fine.

One thing you could do, is add in a relay & control circuit to do the same job. You should be able to find a circuit easily enough, and it will eliminate any turn off thump and disconnect the speaker if there is a large DC on the amp output.

I have always thought the diac & triac to be a very crude circuit, which can be subject to false trips through noise injection (RF etc).
The other problem with this circuit, is at turn on, or turn off, if the power amp rails don't come up / down at the same rate, the power amp can have DC on its output (heard as a turn on / turn off thump), and this thump can activate the diac / triac....

So my 2c (or 2p for the UK) is just to leave the diac & triac out altogether, or try Juan's excellent suggestion, double the ohms of R193 to 100K, and double the capacitance of C336 to 4u7, or leave R193 as it is & use 10u 50V NON POLARISED for C336
It may be the 2u2 has dried out & you are now getting "nuisance trips". Those small value electros can not always be reliable long term.
The only thing i would add to Juan's suggestion, is to put a 1k 1/4W between the triac gate & GND (the connection of triac terminal 1, C336, C146, WJ102). The triacs can sometimes have "sensitive" gates and adding a 1k (or even 470 ohm) will help bleed off any leakage currents so the triac doesn't trigger falsely.