Carvin Legacy 3 bias issue or design?

[DeepBias]

What I don't understand is that the source of the problem IOW what caused R55 to burn out, was turning the bias pot all the way up. So what does increasing the bias have to do with the D rail ? Would someone be kind to explain this.

However, it is obvious that there's an over current issue on D rail persisting even since R55 (and R56?) were changed. He removed all preamp tube (minus the inverter), the over current issue was still there, he checked Q3-Q6 and none were shorted. Could it be that R98, R108, R34, or R104 overheated and that the resistance of one or more of them is out of spec?
 

[EddyMunster]

Here's a pic of the transistors. They are all the same model.

[DeepBias]

Is this a cold solder I see on the picture (circled in red)?

[EddyMunster]

I noticed C56 and C167 are 22uf 50v caps. On the drawing it shows *10uF N.P.

Anybody know what this means?

[EddyMunster]

Deepbias,
   Could be. I've seen some like that around on the boards but they look real good on the back side. I will try cleaning them up but they are pretty solid.

[DeepBias]

I noticed C56 and C167 are 22uf 50v caps. On the drawing it shows *10uF N.P.

Anybody know what this means?

Those are coupling capacitors for the audio signal

Edit: N.P. means Non Polarised
 

[EddyMunster]

Okay thanks, Just wondering if the caps are supposed to be 10uf instead of 22uf and is 50v enough?  Would it even be a problem?

Both amps have the 22uf caps

[DeepBias]

okay thanks, Just wondering if the caps are supposed to be 10uf instead of 22uf and is 50v enough?  Would it even be a problem?

The capacitance doesn't matter that much, more bass is passing through 22uF than 10uF ... check the voltage with you DMM and you will see if 50V is enough
 

[SEL49]

I noticed C56 and C167 are 22uf 50v caps. On the drawing it shows *10uF N.P.

Anybody know what this means?

10µF N.P. is a 10µF non polarized electrolytic.

[AlNewman]

What I don't understand is that the source of the problem IOW what caused R55 to burn out, was turning the bias pot all the way up. So what does increasing the bias have to do with the D rail ? Would someone be kind to explain this.

I would think that what could cause a fault in the transistors is if the bias was turned too low, causing the voltages to rise above maximum specs.  It's also possible if one was testing their amp with their multimeter settings wrong, it could short the connections.  Transistors aren't very forgiving.

[AlNewman]

You could check for DC voltage at R100 or R101, that should indicate if the 10uf caps are leaking.

[DeepBias]

What I don't understand is that the source of the problem IOW what caused R55 to burn out, was turning the bias pot all the way up. So what does increasing the bias have to do with the D rail ? Would someone be kind to explain this.

I would think that what could cause a fault in the transistors is if the bias was turned too low, causing the voltages to rise above maximum specs.  It's also possible if one was testing their amp with their multimeter settings wrong, it could short the connections.  Transistors aren't very forgiving.

Right, he measured 8ma per tube using the bias probe so indeed on the low side, it makes sense.. Thanks a lot for the explanation!
 

[EddyMunster]

You could check for DC voltage at R100 or R101, that should indicate if the 10uf caps are leaking.

I checked them both. No DC voltage either one. :)

[EddyMunster]

What I don't understand is that the source of the problem IOW what caused R55 to burn out, was turning the bias pot all the way up. So what does increasing the bias have to do with the D rail ? Would someone be kind to explain this.

I would think that what could cause a fault in the transistors is if the bias was turned too low, causing the voltages to rise above maximum specs.  It's also possible if one was testing their amp with their multimeter settings wrong, it could short the connections.  Transistors aren't very forgiving.

Hi,
  I double checked all 4 transistors with the diode setting. From the center leg with the + lead of the fluke I get .6 vdc to each of the outer legs on all 4 transistors. When I reverse the - lead to the center I get nothing on the outer legs. Does that sound right?

Here's a pretty cool pic I took with a cheap IR camera. The red plus automatically finds the hottest spot the the white I just pointed at the transistor.

[AlNewman]

Hi,
  I double checked all 4 transistors with the diode setting. From the center leg with the + lead of the fluke I get .6 vdc to each of the outer legs on all 4 transistors. When I reverse the - lead to the center I get nothing on the outer legs. Does that sound right?

Yes, that would indicate the transistors are working ok.  Sorry, I'm all out of ideas, besides something's drawing too much current.

[astronomicum]

I plugged the numbers into a transistor bias calculator https://www.omnicalculator.com/physics/transistor-biasing and it appears the transistors are fine. Each should draw and are drawing about 3mA. Four transistors at 3mA plus five triodes at lets say 1mA each is 17mA. Through R55, that is an 80V drop with a dissipation of 1.3W. According to the schematic, there should be a 30V drop which is 6mA at 0.2W. For 5 triodes and 4 transistors, just 6mA does not make sense to me (assuming I am not missing something). Adding another lets say 3mA for the PI, that would be 20mA through R56 yielding a voltage drop of 44V with a dissipation of 0.9W. R55 at 1W is under rated while R56 at 1W is near its spec. I cannot see how a bias adjustment has had anything to do with the issues at R55 and R56. IMO, they are unrelated. IF the Carvin bias setting of 100mA is based on the math from the schematic, then we can assume that that setting is lower than desired due to the higher actual current draw in the D rail meaning less current going to the power tubes, about 11mA less. Not a whole lot but would warm up the bias a little. I would use a 3W rating for R55 and a 2W rating for R56, increase the bias across the standby switch to 111mA (using the correct meter settings), and go through and recheck everything after burn in. This would put the amp operation where I believe the designers intended. And of course make some corrections to your schematic.

[SEL49]

Odd question for Eddy... HOW DO THE AMPS SOUND?

[EddyMunster]

Hi, short answer is they sound pretty good. Loud as hell too!

There's an attempted explanation of why the "rabbit hole" I put on reply #20.

There is a 79vdc voltage drop on both amps across R55 but per the drawing should be 30vdc (rabbit hole).

The main issue is R55 gets hot enough to discolor. It's getting over 300 degrees F. R56 gets hot too and is discolored on both amps.

Please see attachment on reply #20. In his procedure he shows 29ma on top of the current draw of the 4 power tubes. Then he cranks up the bias up and ends up around 187ma across the standby. Eurotubes video on the V3 amp gets somewhere around 160ma.

I know it's not a legacy 3 that's shown in either example but the designs are similar. Based on that I am assuming it is normal to have some extra current besides what the power tubes draw.

I'm getting about 22ma extra on top of the 20ma per power tube while running bias across the standby at 102ma.

I'm pretty weak when it comes to electronics but I'm learning and that's pretty rewarding.

Any input is greatly appreciated.

 

[AlNewman]

Yes, here's the original schematic.

https://el34world.com/charts/Schematics/files/Carvin/Carvin_legacy_vl100_vl212.pdf

It had tube reverb and only one high voltage transistor.  So they copy pasted the power section and voltages, but added another 8ish mA of current to the D rail.
I guess it must have been a Friday design.

[EddyMunster]

I plugged the numbers into a transistor bias calculator https://www.omnicalculator.com/physics/transistor-biasing and it appears the transistors are fine. Each should draw and are drawing about 3mA. Four transistors at 3mA plus five triodes at lets say 1mA each is 17mA. Through R55, that is an 80V drop with a dissipation of 1.3W. According to the schematic, there should be a 30V drop which is 6mA at 0.2W. For 5 triodes and 4 transistors, just 6mA does not make sense to me (assuming I am not missing something). Adding another lets say 3mA for the PI, that would be 20mA through R56 yielding a voltage drop of 44V with a dissipation of 0.9W. R55 at 1W is under rated while R56 at 1W is near its spec. I cannot see how a bias adjustment has had anything to do with the issues at R55 and R56. IMO, they are unrelated. IF the Carvin bias setting of 100mA is based on the math from the schematic, then we can assume that that setting is lower than desired due to the higher actual current draw in the D rail meaning less current going to the power tubes, about 11mA less. Not a whole lot but would warm up the bias a little. I would use a 3W rating for R55 and a 2W rating for R56, increase the bias across the standby switch to 111mA (using the correct meter settings), and go through and recheck everything after burn in. This would put the amp operation where I believe the designers intended. And of course make some corrections to your schematic.

I don't know how I skipped over your eval this morning but thanks so much for your time! Makes sense to me.

Yes, here's the original schematic.

https://el34world.com/charts/Schematics/files/Carvin/Carvin_legacy_vl100_vl212.pdf

It had tube reverb and only one high voltage transistor.  So they copy pasted the power section and voltages, but added another 8ish mA of current to the D rail.
I guess it must have been a Friday design.

Thanks again!! I can't thank all of you enough!

I will chew on the schematic, change out the resistors, and put the amp back together. I'll let you guys know what happens. Might take a while if I need to order parts.

-Eddy

[astronomicum]

I will... change out the resistors...

-Eddy

Set those new resistors off the board a bit to provide good airflow around them.

[EddyMunster]

All,
 
Replaced R55 and R56 with new 3 watt resistors (couldn't find a 2 watt). Mounted them a little off the board to stay cooler.

Took some voltage readings after setting the bias to 111mA & 176mA DC across the standby and marked the schematic. Red is @ 176mA (176mA is based on ETBP reading of 15 watts which is 60% dissipation).

Also took some readings with the Eurotubes bias probe. ETBP (I only measured 1 tube. They are a matched set but are  slightly different).

see pics

Is it possible and would it be a good idea to change the resistance values of R55 and R56 to achieve voltages closer to the schematic?

[astronomicum]

Is it possible and would it be a good idea to change the resistance values of R55 and R56 to achieve voltages closer to the schematic?

IMO, no. Your voltages at A and B look good for biasing at 111mA. Your voltages at C and D are where they should be. The C and D schematic voltages are not correct assuming they did not intend to use different value resistors. I find that unlikely. You build amps based on the component specified on the schematic, not to the voltages specified, and they did that. Voltages will fall where they will. Make a note of your voltages for future reference. I would leave the amp at 111mA unless some day you find Carvin put out a service bulletin to change the bias target. You did say the amp sounded fine, and I would think as intended. I assume you will do both amps. Make sure both sets of voltages are comparable.

[EddyMunster]

Thanks again. Good to hear no changes are needed! I will check the voltages on the other amp soon.

[EddyMunster]

All,

Got the other amp done. Resistors replaced and voltages are the same within a volt or two between the two amps.

I spent some time with the bias set at 111mA and 176mA for about an hour or so.

I think the higher bias sounds better and more natural to my ears but either way sounds good.

I'd say at this point SUCCESS!

Thoughts? Questions?

Here's some pics of the resistors.

Thanks again

-Eddy

[astronomicum]

Got the other amp done. Resistors replaced and voltages are the same within a volt or two between the two amps.

Good to know. YAW.