Fault finding on a DC coupled amplifier. - UK Vintage Radio Repair and Restoration Discussion Forum

vinrads · 28th Sep 2020, 9:25 am

Having repaired the fault on the Grundig R35, I was wondering because all the voltages will be wrong the mid point will be high or low, could you make an artificial mid point then you can fault find as normal, you could isolate the outputs to do this, any thoughts? Mick.

Refugee · 28th Sep 2020, 9:57 am

You could remove the output pair and fit two power resistors in the 220 to 1K range.
Use a third resistor in the same range to simulate the speaker.

vinrads · 28th Sep 2020, 10:20 am

Good idea! . I used a 24volt bulb in series with the speaker when the fault appeared the bulb lit saving burning the speaker out , but could not fault find checking voltages this way so I just swapped the transistors from the good to the faulty side one at a time , I know a bit scary you could end up with two duff channels

Mick.

Radio Wrangler · 28th Sep 2020, 10:50 am

If you are developing an amplifier circuit, not just repairing one, you've lost the support of knowing it once worked, or having a good channel to compare against. So you really do have to find ways of testing a DC coupled amplifier in stages.

The difficulty arises in the first two stages having so much DC gain that the offset voltage of the input could try to put the DC quiescent point of the output anywhere between the rails.

As the output sections of the usual architecture amount to real or pseudo emitter followers, everythng after the VAS is unity voltage gain. So the above statement also relates to the voltage on the VAS collector.

So, one thing that can be done in a typical amplifier is to remove the VAS transistor and replace it with a resistor equal to the load resistor(s) at the top side of the Vbe multiplier/diode string.

So the VAS collector string should now sit around half way, and the quiescent current in the string should be right. The vbe multiplier should be working, or diode string, depending on design.

The power stage should be giving roughly the halfway voltage on the output. Voltage checks on the emitter resistors should show the right output transistor quiescent current. And in theis condition, you can test, debug, and replace output transistors quite easily with reduced consequences and things a lot easier to understand.

To run the early stages in isolation:

Disconnect the signals to the bases of the transistors after the VAS. Leave the collecto load above the Vbe multiplier and leave the vbe multiplier in position. (Put back the VAS transistor if yo had it out, of course)

Disconnect the feedback resistor at the end going to the output. Link the freed end of the resistor temporarily to the VAS collector.

You now have the input stage and VAS with DC feedback. Normally the feedback resistor is large enough for the VAS to drive it if there's nothing else. You can now have a prod around these two stages without risking the output section. You can measure gain, frequency response etc.

Going a stage further you can split the feedback (the series one) resistor into two resistors in series and put a decoupling capacitor to ground from their midpoint. This kills the AC feedback but leaves the DC. You can not only test DC conditions, but you can also measure the AC gain open loop and check the compensation roll-off and phase shift. The results will be affected a bit by the loss of the loading the power stages would give but you can get a useful indication, but if the VAS load is bootstrapped with a capacitor from the output, the gains and frequency response will be changed. You can calculate or simulate to find what to expect.

Welcome to the deep end of solid state amplifiers! Don't swim under the diving-board, there are people diving in blindly.

David

Diabolical Artificer · 29th Sep 2020, 7:10 am

Make sure you use a CC type bench PSU/s in place of it's normal PSU, saves tears and hair pulling. If there's handy jumpers or coupling resistors you could take those out whilst you sort each stage.

Andy.

vinrads · 29th Sep 2020, 8:43 am

Here is the cct of the Grundig r35 the fault was T2004 not sure what it does! Mick.

Radio Wrangler · 29th Sep 2020, 9:18 am

T2004 is the VAS... Voltage amplifier stage.

The input differential pair provides substantially less voltage gain than the overall amp needs. The power stages, those power darlingtons only do unity gain as emitter followers, so another gain stage is necessary.

Also, the output of the input stage cannot swing very far. The drive to those Darlingtons needs to swing almost between the rails, so the added stage needs an output capable of dramatic voltage swings. T2004 is it. Size for size, the transistor with the hardest job in the amplifier.

Note also the weird bodge-capacitor symbol from base to collector. Yuk! this is the place where the dominant pole lowpass filter is implemented, that sets the stability of the whole amplifier within its feedback loop. Ground zero for a whole pile of difficulties, that transistor. What they have to do makes them prime contenders for second breakdown mode failures, as well.

Notice also C2006 the bootstrap capacitor driving the tapping on what amounts to the collector load resistor for T2004. This can drive the effective collector voltage T2004 far above the positive supply voltage, while the emitter and base live down by the negative supply voltage.... did I say that transistor had a hard life?

The replacement part needs to be rated well above the total power supply voltages added together.

David

stuarth · 29th Sep 2020, 10:42 am

I’ve never liked the quiescent current setting pot between the base and collector of the vbe multiplier (T2005). If the pot goes open circuit, the vbe multiplier turns off, and the output Darlingtons turn hard on. Ouch.

Much better for the pot to be between the base and emitter, then if it goes open circuit, the vbe multiplier turns hard on and all you get is (non destructive) increased crossover distortion.

Stuart

stuarth · 29th Sep 2020, 10:50 am

Ah, just re read your other thread on the R53. If all transistors check ok, your intermittent fault causing DC onto the speaker could be a dirty contact between the track and moving contact on the pot (R2016)

Stuart

Radio Wrangler · 29th Sep 2020, 10:55 am

Quote:

Originally Posted by stuarth

I’ve never liked the quiescent current setting pot between the base and collector of the vbe multiplier (T2005). If the pot goes open circuit, the vbe multiplier turns off, and the output Darlingtons turn hard on. Ouch.

It's a good indicator that there may be other incompetent bits of design lurking around.

David

vinrads · 29th Sep 2020, 3:44 pm

That it why it's a good reason for fitting a speaker protection circuit.me thinks , Stuart I am aware of the pre set being suspect , I did give it a clean and twiddle ,then re set the idle current , the fault was very intermittent ,so whilst the fault was on I swapped all the transistors one at a time , it wasn't until I swapped the last one did the lamp go out , ie the fault disappeared this was the veb multiplier transistor , still on soak it's looking good . Mick.

28th Sep 2020, 9:25 am	#1
vinrads Dekatron Join Date: Feb 2009 Location: Chesterfield, Derbyshire, UK. Posts: 3,756	Fault finding on a DC coupled amplifier. Having repaired the fault on the Grundig R35, I was wondering because all the voltages will be wrong the mid point will be high or low, could you make an artificial mid point then you can fault find as normal, you could isolate the outputs to do this, any thoughts? Mick.

28th Sep 2020, 9:57 am	#2
Refugee Dekatron Join Date: Apr 2012 Location: Worksop, Nottinghamshire, UK. Posts: 5,553	Re: Fault finding on a dc coupled amplifier. You could remove the output pair and fit two power resistors in the 220 to 1K range. Use a third resistor in the same range to simulate the speaker.

28th Sep 2020, 10:20 am	#3
vinrads Dekatron Join Date: Feb 2009 Location: Chesterfield, Derbyshire, UK. Posts: 3,756	Re: Fault finding on a dc coupled amplifier. Good idea! . I used a 24volt bulb in series with the speaker when the fault appeared the bulb lit saving burning the speaker out , but could not fault find checking voltages this way so I just swapped the transistors from the good to the faulty side one at a time , I know a bit scary you could end up with two duff channels Mick.

28th Sep 2020, 10:50 am	#4
Radio Wrangler Moderator Join Date: Mar 2012 Location: Fife, Scotland, UK. Posts: 22,868	Re: Fault finding on a dc coupled amplifier. If you are developing an amplifier circuit, not just repairing one, you've lost the support of knowing it once worked, or having a good channel to compare against. So you really do have to find ways of testing a DC coupled amplifier in stages. The difficulty arises in the first two stages having so much DC gain that the offset voltage of the input could try to put the DC quiescent point of the output anywhere between the rails. As the output sections of the usual architecture amount to real or pseudo emitter followers, everythng after the VAS is unity voltage gain. So the above statement also relates to the voltage on the VAS collector. So, one thing that can be done in a typical amplifier is to remove the VAS transistor and replace it with a resistor equal to the load resistor(s) at the top side of the Vbe multiplier/diode string. So the VAS collector string should now sit around half way, and the quiescent current in the string should be right. The vbe multiplier should be working, or diode string, depending on design. The power stage should be giving roughly the halfway voltage on the output. Voltage checks on the emitter resistors should show the right output transistor quiescent current. And in theis condition, you can test, debug, and replace output transistors quite easily with reduced consequences and things a lot easier to understand. To run the early stages in isolation: Disconnect the signals to the bases of the transistors after the VAS. Leave the collecto load above the Vbe multiplier and leave the vbe multiplier in position. (Put back the VAS transistor if yo had it out, of course) Disconnect the feedback resistor at the end going to the output. Link the freed end of the resistor temporarily to the VAS collector. You now have the input stage and VAS with DC feedback. Normally the feedback resistor is large enough for the VAS to drive it if there's nothing else. You can now have a prod around these two stages without risking the output section. You can measure gain, frequency response etc. Going a stage further you can split the feedback (the series one) resistor into two resistors in series and put a decoupling capacitor to ground from their midpoint. This kills the AC feedback but leaves the DC. You can not only test DC conditions, but you can also measure the AC gain open loop and check the compensation roll-off and phase shift. The results will be affected a bit by the loss of the loading the power stages would give but you can get a useful indication, but if the VAS load is bootstrapped with a capacitor from the output, the gains and frequency response will be changed. You can calculate or simulate to find what to expect. Welcome to the deep end of solid state amplifiers! Don't swim under the diving-board, there are people diving in blindly. David __________________ Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done

29th Sep 2020, 7:10 am	#5
Diabolical Artificer Dekatron Join Date: Jul 2009 Location: Sleaford, Lincs. UK. Posts: 7,658	Re: Fault finding on a dc coupled amplifier. Make sure you use a CC type bench PSU/s in place of it's normal PSU, saves tears and hair pulling. If there's handy jumpers or coupling resistors you could take those out whilst you sort each stage. Andy. __________________ Curiosity hasn't killed this cat...so far.

29th Sep 2020, 8:43 am	#6
vinrads Dekatron Join Date: Feb 2009 Location: Chesterfield, Derbyshire, UK. Posts: 3,756	Re: Fault finding on a dc coupled amplifier. Here is the cct of the Grundig r35 the fault was T2004 not sure what it does! Mick. Attached Thumbnails Last edited by vinrads; 29th Sep 2020 at 8:57 am.

29th Sep 2020, 9:18 am	#7
Radio Wrangler Moderator Join Date: Mar 2012 Location: Fife, Scotland, UK. Posts: 22,868	Re: Fault finding on a dc coupled amplifier. T2004 is the VAS... Voltage amplifier stage. The input differential pair provides substantially less voltage gain than the overall amp needs. The power stages, those power darlingtons only do unity gain as emitter followers, so another gain stage is necessary. Also, the output of the input stage cannot swing very far. The drive to those Darlingtons needs to swing almost between the rails, so the added stage needs an output capable of dramatic voltage swings. T2004 is it. Size for size, the transistor with the hardest job in the amplifier. Note also the weird bodge-capacitor symbol from base to collector. Yuk! this is the place where the dominant pole lowpass filter is implemented, that sets the stability of the whole amplifier within its feedback loop. Ground zero for a whole pile of difficulties, that transistor. What they have to do makes them prime contenders for second breakdown mode failures, as well. Notice also C2006 the bootstrap capacitor driving the tapping on what amounts to the collector load resistor for T2004. This can drive the effective collector voltage T2004 far above the positive supply voltage, while the emitter and base live down by the negative supply voltage.... did I say that transistor had a hard life? The replacement part needs to be rated well above the total power supply voltages added together. David __________________ Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done Last edited by Radio Wrangler; 29th Sep 2020 at 9:24 am.

29th Sep 2020, 10:42 am	#8
stuarth Heptode Join Date: Feb 2010 Location: Heysham, Lancashire, UK. Posts: 668	Re: Fault finding on a dc coupled amplifier. I’ve never liked the quiescent current setting pot between the base and collector of the vbe multiplier (T2005). If the pot goes open circuit, the vbe multiplier turns off, and the output Darlingtons turn hard on. Ouch. Much better for the pot to be between the base and emitter, then if it goes open circuit, the vbe multiplier turns hard on and all you get is (non destructive) increased crossover distortion. Stuart

29th Sep 2020, 10:50 am	#9
stuarth Heptode Join Date: Feb 2010 Location: Heysham, Lancashire, UK. Posts: 668	Re: Fault finding on a dc coupled amplifier. Ah, just re read your other thread on the R53. If all transistors check ok, your intermittent fault causing DC onto the speaker could be a dirty contact between the track and moving contact on the pot (R2016) Stuart

29th Sep 2020, 3:44 pm	#11
vinrads Dekatron Join Date: Feb 2009 Location: Chesterfield, Derbyshire, UK. Posts: 3,756	Re: Fault finding on a dc coupled amplifier. That it why it's a good reason for fitting a speaker protection circuit.me thinks , Stuart I am aware of the pre set being suspect , I did give it a clean and twiddle ,then re set the idle current , the fault was very intermittent ,so whilst the fault was on I swapped all the transistors one at a time , it wasn't until I swapped the last one did the lamp go out , ie the fault disappeared this was the veb multiplier transistor , still on soak it's looking good . Mick.