10 ohm R's going high - problem.

[Terry]

To my surprise you can get 1% wirewounds for not too much money ...

Thinking about it, it must be much easier to make precision wirewound resistors than any other type because the winding machine can be arranged to measure the total resistance of the wire as it is wound and stop at exactly the right point.

[Refugee]

I have seen resistors burn out in the IF stages of an AR88 comms receiver.
There was a load of char coal in the mains selector so that it was intermittent.
The repeated power cycling caused the grids to act like diodes so that the resistors in the anodes were burned and split open due to pulses of anode current.
The circuit in question appears to have a number of power switching events going on using more than one device.

[Murphyv310]

I'd be getting the scope on and seeing what is going on, especially with signal applied and the amp screwed right up, might be a bit of a shock, especially if these resistors are not decoupled.
Let the capacitors deal with spikes and high AC content from the audio not just the resistors.

One other thought. If the Amp has NFB its possible you have too much and the amp is subsonically oscillating that could stress the resistors.

[Diabolical Artificer]

Yes Jeffrey those R's look just the job, unfortunately as Mike pointed out there's a £15 surcharge. This is a big problem with Farnell, yesterday I looked at 10r MF R's, the first page of results all had the £15 surcharge ( why it costs £15 bog knows. It's £15 from Lille in Franch and the US. Digikey is only £3). So I ordered some bigger Vishay 3w's from RS and some bigger (physically) 3W's from rapid. Someone did suggest using W21's. I'll see how I go and report back.

I scoped as well as monitored voltage and current yesterday at turn on and off and at full whack. Apart from a dip of several volts at Sw off - nada. The start up is gentle, V rises over about 6 seconds due to the heaters warming up.

There is 34dB of NFB. (Hope that's right, 1v on the OP - 0.02 applied to the cathode of the IP valve so 20 log V1/V2 = 33.9.) NFB R is two 10k's in parallel = 5k and an 820p C. This isn't quite right, but ok for now. Amp is stable with a speaker load, but at some point I'm going to have to improve closed loop stability and test with a proper inductive/capacitive test load.

I'm also going to put diodes reversed biased from anode to ground to stop big inductive spikes if someone left the dummy load off by mistake sort of thing (whistling whilst looking skywards) and spikes when over driven. It can kick out 160w (36v RMS - OPT rated 120w)) without the protection circuit and has whilst fettling, gone into oscillation, but the R's in question survived this abuse. The R's have gone high whilst everything is bolted down so to speak.

A.

[Murphyv310]

Very odd then Andy...... Thanks for your previous post.
Perhaps a bad batch of resistors, were they all from the same lot?

[Diabolical Artificer]

Most duds were from the same lot Trev, though I recollect another from a different order went the same way.

Replacements arrived ( Rapid really are Rapid, ordered dinner time day 1, arrived early AM day 2 ... and cheap), will report back when I have anything new.

Andy.