GT120 - troublehsooting

[plokij88]

Hello everyone, thanks in advance for any guidance.

I'm building a Matamp style GT120. I am building it from the Hillbish chassis and circuit board. I am in the UK 240v. This is my 2nd amp build and today was first power up and the HT fuse was starting to smoke when I got to around 60v on the variac going in to the power transformer, with the power tubes in. And I haven't figured out why.

I'll go through what I have done so far.

• I powered the amp up on the variac with no tubes installed, starting at 10v, moving up in 10v increments

• My voltage readings were a little high but i expected them to decrease when i put the tubes in

• I put the tubes in

• I power the amp up on the variac with all tubes installed, starting at 10v moving up in 10v increments

• When I got to 60v on the variac I noticed smoke coming from the HT fuse, I powered off

• The HT fuse is still intact, i checked it was the correct rating and reinstalled it

• All tubes are brand new, I checked them for a short and didn't find one

• I gradually powered up again, with no tubes installed, no issues

• I gradually powered up again, with only preamp tubes installed, no issues

• I took voltage readings with the preamp tubes in, I set the variac so I would get 470A+. See attached. Many are high, V1-1 is very high

• I went over all the wiring again and couldn't spot a problem

• I came here and made this post

Any advice greatly appreciated thank you

[johnnyreece]

I don't use a variac, so excuse me if this is a stupid question...Is it possible that bringing it up slowly does something wonky to the bias, such that the tubes are way overdrawing current? Seems as though you've got it narrowed down to the power amp, anyhow. With your power section not conducting, it makes sense that all your preamp voltages will be high.

[Helmholtz]

Before re-inserting power tubes make sure you have sufficient negative bias voltage (e.g. -40V) at EL34 socket pins 5.

[Stevem]

If your HT fuse was starting to go then there’s no way you can trust it now even if you had the amp working normal at just an idle condition.

Whenever I fire up a new or rebuilt amp that has a HT fuse I install one with 1/2 the needed amperage rating because if all is well the amp should idle fine like that without popping that fuse.

This can potentially save you from burned up new output tubes and shorted output transformers.

[plokij88]

Is it possible that bringing it up slowly does something wonky to the bias...

I do not know how i would test this asside from powering up at full voltage which seems dangerous

Before re-inserting power tubes make sure you have sufficient negative bias voltage (e.g. -40V) at EL34 socket pins 5.

I have around -35v on each el34 pin 5

Whenever I fire up a new or rebuilt amp that has a HT fuse I install one with 1/2 the needed amperage rating...

I swapped in a 0.5amp fuse and it smoked at 160v a+/C30

[Stevem]

There's only 3 main reasons a HT fuse would blow.
1)
The output tubes do not have enough or any negative bias.

2) you do have a shorted output tube.

3) a output tube socket is arcing.

How did you test these tubes for shorts, meter or tube tester?

[plokij88]

3) a output tube socket is arcing.

This was the issue, I have found it and fixed it.

New Problem

• I plugged the amp in going through a passive attenuator (spl reducer) into my speaker cab,
  I powered it up with the tubes in, set the bias, left it for like 10 minutes, all was well.
  I plugged my guitar in, amp volume set at 0.
  I strum a string.
  The tubes whistled like a kettle, then rumbled, then there was an arc on one of the tubes.
  I power it off.
  I went through the tubes with the continuity tester, they all only had continuity between 2 and 7

I did some solid googling but everything I could find about tube rattle was about microphonic tubes and all references to whistling were coming out of the speakers.

I made a lightbulb current limiter.

• I power it up with preamp tubes and without power tubes, the bulb glows bright then dims, which i believe is what i want to see. https://youtu.be/9aZQ42KZZ9w
  Power it up with power tubes, bulb stays bright, which I believe means that something is drawing too much current. https://youtu.be/Rkt9uMcyl5c

Videos attached, Voltage is c30+ (A+) to ground.
In the video I have the bias tester sockets in but disconnected from the tester. I tried it again without those sockets in and the same result.

I keep going over the tube socket wiring and the output transformer and it all seems correct.
I've considered that my output transformer impedances are miswired, and i had the incorrect impedance for the cab, but i don't know how to test for that.

Any help appreciated, Thanks again,

Nick

[martin manning]

I don't like the look of the first video. The bulb is too bright, and stays bright too long. It's difficult to judge from a video because the camera is continuously adjusting the exposure, but with no tubes installed it should appear more like a brief flash, and then settle to a very dim glow.

See video link in this post: https://ampgarage.com/forum/viewtopic.p ... 10#p462410

Is that a 150W flood lamp bulb? I typically start with a bulb that is less than or equal to the amp's output power.

After the bulb dims, can you see any glow at all?

Can you link or post a schematic?

Start with no tubes installed and verify the B+ voltages and the bias voltage are reasonable. The whistling and rumbling indicates an oscillation, which may be what caused the arcing.

[Stevem]

As said above, it seems that the amp is oscillating.

The first way to confirm this and attack it at the same time is to do this and try the easiest of these 3.

1) swap the blue and brown wires on your output tubes pin 3 if there long enough.

If not.

2) swap the PI tube wires going to pin 5 on your output tubes.

3) swap the ground and hot wires to your speaker jack from the secondary of your output transformer.


In doing this 95% of the time the issue goes away ( you have the needed proper feedback phase), or the issue stays the same or gets worse.

Report back on how you make out.

[nuke]

Ditto on the oscillation. Flipping the phase of the OPT leads will oftentimes fix it, if that is the source of oscillation. The easy quick check is to simply disconnect the feedback lead from the output transformer secondary (R15/C13 in your schematic).

The schematic is kind of odd, being that the HT fuse, is between A+ and the choke. It isn't protecting the plate circuit, and the only thing that can draw substantial current is the screens of your EL-34's, through the choke. That would affirm output section oscillation.

I can't really see the build, but I noticed a lot of wiring hanging out around the side of the chassis. Lead dress and routing can be a source of oscillation, particularly around the output section.

Maybe disconnect the feedback loop, and then just pull V2 out and power it up.

[plokij88]

This is the 3rd time im typing this out as TAG keeps making me log in again and deleting what i have written. Sorry If I am curt.

I disconnected the NF wire from the 16ohm lug of my impedance selector. Lighbulb limiter flashed bright and then dimmed.

I then reattached the NF wire, and swapped the OT wires on pin 3. Lighbulb limiter stayed bright.

I swapped the OT wires on pin 3 back the way they were, and swapped the wires going into pin 6. Lightbulb limiter flashed bright and dimmed.

I removed the lightbulb limiter and powered up. Bias readings were around 60ma, i was flicking through them when a tube redplated. I switched it off

I replaced the tube, and brought the lightbulb limited back into the circuit. The video shows the current situation with all tubes in. I switch the amp off and on during the video

With the lightbulb limiter and a 100W bulb I have 300v at C30+ but -0.2v at c19/c12 etc

I checked all the filter capacitors with my lcr meter and they all read good.

picture and my wiring diagram attached. I've tried to keep it neat and stick to the diagram but some of it I did not, most of the wiring is single core, all of the flexible wires are things i find myself having to keep detaching for testing. The green and yellow wire leaving the board near the power tubes would be going to pin 6 on the sockets, but are detached and the flying grey wires are making that connection since I swapped them.

Once again any advice appreciated,

[R.G.]

I disconnected the NF wire from the 16ohm lug of my impedance selector. Lighbulb limiter flashed bright and then dimmed.

I then reattached the NF wire, and swapped the OT wires on pin 3. Lighbulb limiter stayed bright.

This is a pretty solid confirmation that the problem is oscillation in the output stage, if there was any doubt.

With the lightbulb limiter and a 100W bulb I have 300v at C30+ but -0.2v at c19/c12 etc

Weird. Maybe the oscillation is fooling your meter on C19/C12.
I don't'suppose you have access to an oscilloscope, do you?

picture and my wiring diagram attached. I've tried to keep it neat and stick to the diagram but some of it I did not, most of the wiring is single core, all of the flexible wires are things i find myself having to keep detaching for testing. The green and yellow wire leaving the board near the power tubes would be going to pin 6 on the sockets, but are detached and the flying grey wires are making that connection since I swapped them.

I noticed that the wiring diagram shows a not on the NFB wire that it has to be kept close to chassis to prevent oscillation. If that was from the original designer/manufacturer, it indicates that there is a sensitivity to oscillation in that wire placement. That being the case, it might be a good idea to look at exactly how that wire is routed, or to change something to eliminate the sensitivity.

The output winding floats, and is only earthed by an explicit earth wire. In this wiring diagram, that is connected to the earthed cathodes of the output tubes. In an ideal world, that earthing wire would run not to the output tubes' cathodes, but to the closest possible earth/ground point of the phase inverter tube itself. This prevents any earth/ground wire voltage caused by the cathode currents from causing a feedback signal to the earth/ground of the phase inverter. Happens sometimes.

If it were me, I would be very tempted to move the earth/ground connection wire from the speaker jack to the tube cathodes to terminate instead on the board earth/ground nearest the phase inverter - the earth/ground end of R14. If that didn't help, I'd try substituting a shielded/screened wire for the feedback wire, and tie the shield/screen to either chassis, or to the earth/ground at the bottom of R14.

[nuke]

Might be dumb thought, but which output transformer are you using and what are the wire color codes?

Some years ago, I went in circles on an OPT that the color codes of the wires didn’t match the spec sheet. I guessed wrong, instead of carefully checking my assumptions.

Maybe do a quick DCR reading of the primary coils?

[plokij88]

I don't'suppose you have access to an oscilloscope, do you?

Don't tempt me

I noticed that the wiring diagram shows a not on the NFB wire that it has to be kept close to chassis to prevent oscillation

I believe I have it routed where it should be

If it were me, I would be very tempted to move the earth/ground connection wire from the speaker jack to the tube cathodes to terminate instead on the board earth/ground nearest the phase inverter - the earth/ground end of R14.

I've done this, and it does not appear to have had any effect

I'd try substituting a shielded/screened wire for the feedback wire, and tie the shield/screen to either chassis, or to the earth/ground at the bottom of R14

I will do this, but right now I do not think it is a cause of my problems so I will save it for later.

Maybe do a quick DCR reading of the primary coils?

The sheet from the factory says I/P 2.5KΩ C.T. Brown, Red, Yellow. I have brown and yellow going to pin3's on the power tubes, and I have red going to the fuse holder/choke. I disconnected the primary side from the amp and my readings are.

Brown to Red = 14.3ohms
Yellow to Red = 16.5ohms
Yellow to Brown = 30ohms

Thanks as always for your advice

[plokij88]

I started going through everything with the dcr meter, this is the grid screen resistor to socket 2. I have replaced it.

Now interestingly, when I switch off power to the amp, the capacitors are discharging themselves, my A+ reading drops quite quickly.