Preamp broken

[billybob]

My friend has an ART MPA Gold Dual tube preamp. It wasn’t working when he bought it. I replaced a few of the op amps by the tubes and it worked again. Now he’s saying it doesn’t work again. I don’t thing replacing the op amps would help this time. Would could possibly be the issue? When I opened it I noticed the heat dissipation was kind of lossy, but thought it could withstand it. Would running it for hours at a time burn the amps out? Would it be the tubes this time?
Thank you, here are some pics...

 

[Nigel Goodwin]

You don't repair things by randomly replacing components - you've got a schematic, fault find it - and if it's killing opamps, find out why.

 

[billybob]

You don't repair things by randomly replacing components - you've got a schematic, fault find it - and if it's killing opamps, find out why.

It’s hard for me to find exactly what the problem is just by looking at the schematic. I would love to learn this troubleshooting technique better, but not sure how as my conclusions are often wrong. I also don’t have an osciliscope so that posses another issue. The transistors probably to power the tube are really abnormally hot, my first reaction would be to find the source of that, but there are so many possibilities.
Here is the tube section

 

[Nigel Goodwin]

It’s hard for me to find exactly what the problem is just by looking at the schematic. I would love to learn this troubleshooting technique better, but not sure how as my conclusions are often wrong. I also don’t have an osciliscope so that posses another issue.

There are essentially two methods of fault finding this type of thing - signal tracing (using a scope, or a simple audio amplifier, or even a crystal earpiece - so pretty cheap and easy), or signal injecting (using a signal generator, a simple multivibrator would do, or even buzzing using a screwdriver with your finger on it). Either method is dirt cheap, unless you go to the expense of a scope.

A multivibrator also works for RF signal injection as well, due to the squarewave and the many harmonics - and was the first thing I built when I started work, and used it for ages on huge numbers of repairs.

The transistors probably to power the tube are really abnormally hot, my first reaction would be to find the source of that, but there are so many possibilities.

Which transistors are getting hot? - although it's very possible it's designed like that.

 

[ljcox]

Have you measured the voltages at various relevant points & checked if that are correct?
Also, look for dry joints. Sometimes they are not obvious.
I had a fault in a TV years ago & eventually found by, using a magnifying glass, a hair line crack in one of the soldered joints.

 

[KeepItSimpleStupid]

Generally check the power supplies. (typo fixed)

You can get a $20.00 USD scope. e.g. The DSO150 Sometimes they even come with a probe. I added a 9V battery holder with a barrel connector. Mine didn't come with an adapter, but came with a x1, x10 probe. there is no place for a battery bummer).

 

[unclejed613]

for the op amps that have the noninverting inputs (marked +) grounded, you should also have 0V DC on the inverting inputs (marked -)... if you have more than a few millivolts DC showing on those inputs, you have a problem that can damage op amps..... most likely cause would be one of the transistors in the cascade of current mirrors connected to the plates of the tubes...

 

[billybob]

for the op amps that have the noninverting inputs (marked +) grounded, you should also have 0V DC on the inverting inputs (marked -)... if you have more than a few millivolts DC showing on those inputs, you have a problem that can damage op amps..... most likely cause would be one of the transistors in the cascade of current mirrors connected to the plates of the tubes...

What usually causes components to fail? Such as transistors?

 

[KeepItSimpleStupid]

What usually causes components to fail? Such as transistors?

Let's say that nothing is perfect. An OP amp is not a perfect OP-amp. let's also say that stuff can be Automotive Qualified and Space qualified. They need to be put through tests before they are used.

Cosmic rays were found to interfere with the function of computer memory as initially designed.

I worked on a couple of really high voltage 100kV and current power supplies. Another was 15kV at 1.5A or so.. In both cases, failures were caused by loose connections.

A problem that annoyed me the most was having to spend about $1500 to have a board repaired. Power line spikes damaged the board. once I put a surge suppressor on the AC line, the problem went away. The manufacturer said we specify 120 V 60 Hz and they didn;t design against normal power line spikes.

Heat is a problem. Dust stops a fan. The components heat up past their limit and you have failure.

You have design issues, you have assembly issues, you have process issues.

I conditioned the power to a computer. That computer was a Macintosh and it lasted 17 years and was still running. the The hard drive was SCSI which is a better drive and it lasted 17 years. The floppy drive failed mechanically. The other problem was dust.. The power protection probably cost as much as the computer did. The time lost IF that computer went south easily exceeded the cost of power line protection.

 

[rjenkinsgb]

What usually causes components to fail? Such as transistors?

In a lot of gear, failures can simply be from heat and thermal cycling.

Anything that runs hot when on, and is switched on and off regularly, means the parts are expanding and contracting with every power cycle.

That can eventually start to cause internal cracks and bad connections, especially in semiconductors, as the plastic encapsulation and silicon die expand slightly different amounts and cause stress.

 

[throbscottle]

In the face of not having much in the way of fault finding skills or test equipment, and with the given that you replaced some components after which it worked for a while, then stopped again, I think it's not a bad idea to re-work all the soldered joints on the board. You only need to melt each one with your iron and add a tiny bit of fresh solder (keep the solder sucker handy, as these things occasionally go sideways). Blanket coverage like this is crude but usually effective if the problem is simply a bad joint - trying to track one down individually can be like trying to catch pixies. Or elves.

Running it for hours at a time won't burn the op-amps out, or any other part. It is designed to run indefinitely. Thermal faults can occur - ie ones that only show when it's warmed up past a certain point. Hard to trace. However if it's not working from power-up, this is quite unlikely.

You haven't said in what way it doesn't work. Is it no power at all, or no sound on both channels, or on just one channel, or buzzing, or very quiet, affected by the controls, do the VU meters show any activity, or what? Is the new fault the same as the original fault? Symptoms like these can give a big clue as to where the fault lies.

There are some voltages given on the schematic. Check all of these. I'd take a punt on any more than 10% different to the drawing as needing further investigation.

A quick and dirty check for op-amp functionality is, does the voltage on the inverting input, match (or is very close to) the output? If it's different by more than 100mV (let's say) it probably indicates a fault associated with that op-amp - use the schematic to figure out candidates for testing. It being different by design is possible, but unlikey (eg, if it's used as a comparator, or to only kick in under certain conditions)

Check transistor voltages - for linear circuits the base should be around 0.6v higher (for NPN) or lower (for PNP) than the emitter. If it's around 0.7v then something is likely causing the transistor to saturate. If it's much more than that you could have a faulty transistor. Much less than 0.6 and the transistor will be off, which doesn't necessarily indicate the fault to be directly associated with it, but could explain screwy voltages elsewhere. You should try to find why it's low, in this case (again, it could be by design, but unlikely) Transistor used as switches however will operate with high or low base voltages - something to bear in mind

The valves should have a negative grid voltage in relation to their cathodes. The grids are connected to ground via grid-leak resistors. The anodes are supplied via current mirrors which are balanced up by a further current mirror, supplied by 48v. The cathodes are supplied by I'm not sure what kind of circuit, some kind of gain control possibly, which runs off the -15v supply. So anyway, check these +48 and -15 volt supplies. You can confirm the current mirrors are ok by checking the voltage across each of the 10k emitter resistors. They should be about the same.

If the whole thing is simply not giving any output it's most likely that power is not getting to one of the stages. The first thing you should check is that each stage has power. Check the op-amp power pins individually too - a cracked trace will stop the juice getting through.

 

[billybob]

In the face of not having much in the way of fault finding skills or test equipment, and with the given that you replaced some components after which it worked for a while, then stopped again, I think it's not a bad idea to re-work all the soldered joints on the board. You only need to melt each one with your iron and add a tiny bit of fresh solder (keep the solder sucker handy, as these things occasionally go sideways). Blanket coverage like this is crude but usually effective if the problem is simply a bad joint - trying to track one down individually can be like trying to catch pixies. Or elves.

Running it for hours at a time won't burn the op-amps out, or any other part. It is designed to run indefinitely. Thermal faults can occur - ie ones that only show when it's warmed up past a certain point. Hard to trace. However if it's not working from power-up, this is quite unlikely.

You haven't said in what way it doesn't work. Is it no power at all, or no sound on both channels, or on just one channel, or buzzing, or very quiet, affected by the controls, do the VU meters show any activity, or what? Is the new fault the same as the original fault? Symptoms like these can give a big clue as to where the fault lies.

There are some voltages given on the schematic. Check all of these. I'd take a punt on any more than 10% different to the drawing as needing further investigation.

A quick and dirty check for op-amp functionality is, does the voltage on the inverting input, match (or is very close to) the output? If it's different by more than 100mV (let's say) it probably indicates a fault associated with that op-amp - use the schematic to figure out candidates for testing. It being different by design is possible, but unlikey (eg, if it's used as a comparator, or to only kick in under certain conditions)

Check transistor voltages - for linear circuits the base should be around 0.6v higher (for NPN) or lower (for PNP) than the emitter. If it's around 0.7v then something is likely causing the transistor to saturate. If it's much more than that you could have a faulty transistor. Much less than 0.6 and the transistor will be off, which doesn't necessarily indicate the fault to be directly associated with it, but could explain screwy voltages elsewhere. You should try to find why it's low, in this case (again, it could be by design, but unlikely) Transistor used as switches however will operate with high or low base voltages - something to bear in mind

The valves should have a negative grid voltage in relation to their cathodes. The grids are connected to ground via grid-leak resistors. The anodes are supplied via current mirrors which are balanced up by a further current mirror, supplied by 48v. The cathodes are supplied by I'm not sure what kind of circuit, some kind of gain control possibly, which runs off the -15v supply. So anyway, check these +48 and -15 volt supplies. You can confirm the current mirrors are ok by checking the voltage across each of the 10k emitter resistors. They should be about the same.

If the whole thing is simply not giving any output it's most likely that power is not getting to one of the stages. The first thing you should check is that each stage has power. Check the op-amp power pins individually too - a cracked trace will stop the juice getting through.

He says the tubes are cold. And he ruled out the main power supply board.

 

[throbscottle]

I found a much better copy of the schematic here: https://elektrotanya.com/art_mpa-gold_schematic.pdf/download.html#dl

If the tubes are cold, then they have no heater supply. The heaters look to be in series and powered from a +12v regulator, which shares it's input with the +15v regulator. So I'd check to see if that 12v supply is present - trace it back from the heaters to the PSU. The 12v regulator is attached to a heatsink. I can see 5 regulators with heatsinks in the schematic, but only 4 heatsinks are visible on the PCB photo. Or trace it out from the PSU and see where it disappears.
The supplies should be +5v, +12v, +15v, -15v, -15v(lights) and +48v which is provided by a discrete regulator and +150v I missed before, which is provided by a voltage multiplier.

I'm puzzled by the phrase "main power supply board" - there is no separate power supply board that I can see, it's all on the main board.

 

[billybob]

I found a much better copy of the schematic here: https://elektrotanya.com/art_mpa-gold_schematic.pdf/download.html#dl

If the tubes are cold, then they have no heater supply. The heaters look to be in series and powered from a +12v regulator, which shares it's input with the +15v regulator. So I'd check to see if that 12v supply is present - trace it back from the heaters to the PSU. The 12v regulator is attached to a heatsink. I can see 5 regulators with heatsinks in the schematic, but only 4 heatsinks are visible on the PCB photo. Or trace it out from the PSU and see where it disappears.
The supplies should be +5v, +12v, +15v, -15v, -15v(lights) and +48v which is provided by a discrete regulator and +150v I missed before, which is provided by a voltage multiplier.

I'm puzzled by the phrase "main power supply board" - there is no separate power supply board that I can see, it's all on the main board.

Thank you so much for the help. I will check all of this very thoroughly whenever he brings the preamp over. Sorry for the confusion I was quoting him when he said the “main power supply board.” I think he meant the first few stages such as the rectifier diodes and filter caps.

 

[throbscottle]

Ah, I wondered if you were quoting your friend. Good luck with it!

 

[billybob]

What transistors would possibly be driving the tubes? Most of the transistors in the schematic are rated around 0.05 Amps would that be enough to light the tube?

 

[Nigel Goodwin]

What transistors would possibly be driving the tubes? Most of the transistors in the schematic are rated around 0.05 Amps would that be enough to light the tube?

No the heaters are 300mA at 12.6V, so 600mA for the two - presumably they are simply fed from the 12V regulator U105.

 

[throbscottle]

It shows quite clearly the heaters are series connected and powered by +12v. There's only one +12v regulator.

 

[billybob]

After a while of putting this thing off. I was board and decided to tackle it again. Is this what tubes are supposed to look like? The glow is Bailey visible and I didn’t do anything besides test the output voltages of the voltage regulators.
This thing is weird.

 

[rjenkinsgb]

They look perfect to me.
The heater filaments are inside the cathodes, so you only really see them in the end of the cathode "tube" assembly.