Welcome to Solid State Guitar Amp Forum | DIY Guitar Amplifiers. Please login or sign up.

April 23, 2024, 04:14:40 PM

Login with username, password and session length

Recent Posts


Fault finding

Started by Geoffb, October 30, 2012, 06:26:42 AM

Previous topic - Next topic


Hi everyone,

I've worked on a number of valve and solid sate amplifiers over the years but never had much success with fault finding on DC coupled power output stages. For example, on occasions I've found the power output transistors blow, replaced them and checked all other transistor/diode junctions and all looks O/K. Switched on and the expensive power transistors have blow again.
I was just wondering if anyone can suggest a clever (and safe) way of fault finding on these DC coupled circuits to avoid having to keep replace expensive parts.

J M Fahey

Yes, the lamp bulb current limiter.
It was invented exactly for that.


I was thinking the same thing but I never go full voltage until I go through dim bulb test stage first... However, I think there are times that full voltage will blow something up that a lamp bulb tester might not...? In that case maybe those parts are now out of the way once replaced, but if those parts are ones you just replaced... Well then S happens...

Geoffb have you been using a light bulb limiter at all?
"A person who never made a mistake never tried anything new." -Albert Einstein


Valve/tube amps tend to be rather more forgiving than solid-state ones.

When you find a solid-state output stage with dead silicon in it anywhere, and it's generally one or more of the output devices, it is vital that all the other devices in the DC loop be also checked to be sure than nothing else has been damaged as well - or you get the "galloping silicon cancer" effect at next switch on.  This does not mean a cursory in-circuit check with an ohm-meter, it means removing the drivers etc and checking them out of circuit to be quite sure they are functional.

I can't remember ever having a situation where a device that tested okay out of circuit was unable to perform in circuit.

Exhaustive checking is a PITA, but if you do this you will be spared that "oh bugger" moment when a set of expensive new output devices instantly go to the Great Sandpit in the Sky (particularly if there are half a dozen in each side of a large rack amp).

Even so, if you are wise you will bring the renewed output stage up gently.  I normally use a pair of 150 ohm 10 watt resistors soldered to a couple of blown fuses (for use where clip-in fuse holders are fitted) or the same across the rear of the panel fuse holders with the fuses removed (which have generally blown anyway), or otherwise in series with the supply/supplies.

Where this is not suitable I fall back on the trusty limiting lamp, and have a selection of globes from 5 watts up to 100 watts for different situations, and to bring the power up in stages if needs be.

I do not use a Variac or variable mains auto-transformer because these are actually transformers and can still deliver considerable power at reduced voltage, whereas a conventional lamp has a resistance that rises sharply as it gets hot and so acts to strictly limit the amount of power that can be dissipated in the amplifier under test.

If you encounter an amp that has a switch-mode power supply you may have to fit your load-limiting between the supply and the output stage because the SMPS a) won't start, and then b) try and compensate for the low supply voltage.

When an amp balances the half rail to half whatever the reduced supply is you can normally increase  the lamp wattage without fear and check again, but if it is not maintaining half-rail balance you need to have a closer look before applying more voltage/power.
If you say theory and practice don't agree you haven't applied enough theory.


Hi All,

Many thanks for the great advice. Althouh I do use the old light bulb techneque, I guess I've been unlucky a few times or not used a small enough wattage bulb. Do you think I should go down to a 25w bulb for these DC coupled amps?
Also on the subject of silicon devices checking O/K out of circuit and then failing when fitted, I've had two occasions recently when the audio sounds like electric arcing going on and it has been a diode breaking down under load. At least replacing the affending diode has solved the problem.



If anything I'd consider a 25W globe to be starting a bit high.  As I said, my lamp collection goes down to 5W pilot lights, and you can find low wattage globes intended for fridges and microwave ovens &c.

As far as intermittent diodes go, if they were zeners I wouldn't be surprised, but ordinary diodes?  That would surprise me a bit.  I've had faults where it has been days, weeks, months and sometimes years later that I realise I had misinterpreted what was going on.

I've also had a few really strange ones.  A transistor I could turn on and off by applying left or right torque to one leg; another that seemed to be acting like a humidity sensor until I discovered a conductive layer across the bottom between the leads - washing cured it.  Asked the owner if the amp had spent any time at the seaside.  "Oh yes, I lived right on the beach for a couple of years; how did you know that?".  More commonly it's corrosion, but in this case it was salt film.

Just when you think you've seen everything, the mud wasps, the skinks, the spiders, the mice living in the reverb tank or speaker cab, you pick up an amp and it has a python inside.  I charge extra for python removal.   8|
If you say theory and practice don't agree you haven't applied enough theory.

J M Fahey

You should charge by the yard ;)



The first yard is the worst.
If you say theory and practice don't agree you haven't applied enough theory.

J M Fahey

Now I know the reason for that *biting* sound !!   ;)

And, of course, that could not happen in China, where the python would quickly be turned into a delicious, spicy soup.   :loco


Oi!  Hands off my Python!

When I was living in the bush I had a juvenile (1.5 - 2m) co-habit with me off and on for a number of years.  Very placid (except were mice are concerned) it would sit on top of my nice warm CRT screen and watch me work.  Did you know that (soup notwithstanding) they live to 50-60 years, and I've seen one over 10m long and a good 20cm in diameter, head the size of a dogs'.
If you say theory and practice don't agree you haven't applied enough theory.

J M Fahey

Up to 60 years??  :tu:


Many thanks for the great advice. I'm off to the shops tomorrow to buy a 5w bulb. Both the diodes I had the arcing problem with were 1n4148.

Thanks once again.


I know it is an old thread now, but a couple thoughts.

First, you may check all the semiconductors, but it is equally important to check every resistor associated with the bad parts.  They tend to burn out when transistors short.  An open resistor may LOOK OK, but you have to test it with a meter.  AN open one is just as dangerous as a shorted part.   Same thing with connections two perfectly good parts can still allow the amp to blow up if the connection between them is open.  Two diodes in series in your bias circuit for example.

And second, your dim bulb may say OK, but you always always work with NO speaker load until you are certain the amp does not put out DC when fully powered.  Always check the output for voltage, we want to see none.  If your amp is sitting there, limit bulb and all, with 40v DC on the output, it will do that all day without a speaker, but as soon as you connect a load, that 40 volts will start burning things out.