18 Watt -- Problem voltages

[Lectroid]

I've been completing an 18W amp sort of like a Marshall 1974.  A recent thread helped me reduce the power to a more useful range. It was sounding great after some here suggested certain tube swaps.

But I still had a hum in the reverb circuit, and I read a small DC voltage (60mV) on the downstream side.  I replaced the reverb recovery coupling cap.  The sound changed immediately; distorted and lacking a lot of the power it had before.

Of course I suspected I messed up the replacement, so I replaced it again but no change: still low-power, distorted sound.  I wondered what else might be going on.

At quitting time last night, I took voltage readings for the whole amp--attached below.   The preamp, reverb, and power tube sections seem healthy, based on voltages.  I know the first preamp stage is working fine, and will continue to trace the signal through with the scope this morning. 

But right away the PI voltages in the "0624" version look screwy.  Compared to the PreviousVersion (also posted below) it's like one side of the PI is getting way less voltage while the other side's voltage increased.  That tells me to check the PI plate resistors and the PS nodes, and try a couple of new tubes in the PI socket, to rule out a bad tube.  Otherwise, I'll just keep tracing down the signal with the scope, try to find where things go south.

Any suggestions, ideas, things to check, or troubleshooting tips?

Thanks!

[sluckey]

V3A is not conducting. Concentrate on pins 1, 2, and 3 and associated components.

[Lectroid]

V3A is not conducting. Concentrate on pins 1, 2, and 3 and associated components.

Thanks, sluckey.  The problem is that the plate voltages are so unequal, right?  I can't figure out why it's happening electrically, so let me ask:

Can you explain how you knew which side of the PI tube is not conducting, just from looking at the schematic?

Today I tried three different 12AU7 tubes in the PI. With each tube, voltages at each position changed slightly up or down, but otherwise just the expected random-variation-within-limits for versions of the same tube.
I've been testing every part from in the signal path from the reverb return point RV through the PI and MV. Tested continuity on wire runs.  I lifted some caps to get a reading but one or two read okay in place.  I found some issues in the tone stack and reverb mixing section so I rebuilt that area. 

Only weird thing: on V3b, I can't measure resistance on the the 1M grid leak; the meter numbers keep wandering up and down.  If I lift one end, it measures  998K instantly. Soldered in, the voltage wanders.  I've never seen that before.

One question: when I'm testing coupling caps that follow after a plate, should I measure absolutely no DC voltage downstream of the cap?  Or is some small level of leakage ever acceptable?

[tdvt]

Just a bad tube socket?


Maybe re-tighten/clean the connectors on the PI tube socket?


Seems like if V3B is conducting, then the shared cathode network is OK, which wouldn't leave too much else to go wrong, if the issue persists with other tubes.

[sluckey]

V3A is not conducting. Concentrate on pins 1, 2, and 3 and associated components.

Can you explain how you knew which side of the PI tube is not conducting, just from looking at the schematic?

Node C B+ is 283V and V3 pin 1 is 280. That means only 3V is dropped across that 82K plate resistor. 3V/82000Ω=36µA. That's almost no current.

One question: when I'm testing coupling caps that follow after a plate, should I measure absolutely no DC voltage downstream of the cap?  Or is some small level of leakage ever acceptable?

That depends. Sometimes there should be voltage on the downstream side. You must disconnect the downstream side of the cap to test for leakage.

You seem to be helter-skelter jumping around the amp. Stop doing that and just concentrate on why V3A is not conducting. Not much involved. Could be a bad socket as tdvt suggested.

We need to see some hi-rez pics.

[Lectroid]

We need to see some hi-rez pics.

Sure, I'll post some pics tomorrow. 

Node C B+ is 283V and V3 pin 1 is 280. That means only 3V is dropped across that 82K plate resistor. 3V/82000Ω=36µA. That's almost no current.

That makes perfect sense.  But then, if the 82K side has an anode voltage too high, I suspect the 82k resistor has drifted lower in resistance. I measured it at 81.2K actual, pretty close.  Then when I measure across it with power on, I get voltage drops of around 3V, like you did. 

Let me understand the scenario.  Because no current is flowing in the upper tube V3a, then only one phase is getting amplified through a limping V3b.  Does this explain why the output power is so low?

My confusion is that I don't see how an anode, supplied with a signal at the grid, with plenty of voltage available, and properly biased, could fail to conduct.  I can't see an answer yet.  Unless the tube is bad.  I'll try a few more tubes?  I have lots of 12AU7s. 

Anyway, I'll keeping working on this part of the amp.  Thanks for your advice, everything helps.

[sluckey]

Resolder all the pins on V3 socket. Any joy?

[Lectroid]

Will do. I’ll tighten the pins also.

[tubeswell]

... replaced the reverb recovery coupling cap. 

which coupling cap do you mean?

[Lectroid]

The .01uF cap right after the reverb recovery tube, the high-mu side of the 12DW7.  Since hum was coming into the dry circuit as the reverb Intensity knob is turned up, it seemed worthwhile to look at that cap first, esp. since I measured DC (mV) on the downstream side of the cap while operating.  Hence my question about leakage to sluckey:  How much is too much?

[sluckey]

Measure V3 voltages on pins 1, 2, 3, 6,  7, and 8. Also measure resistance to ground on pins 2. 3, 7, and 8.

Where are the pics?

[Lectroid]

sluckey,

You were right--it was the socket wiring.  I rewired the whole socket and the amp came right back to life.  Even the slight hum in the reverb is gone, so double-win.  Then I ran out of time, have to go out tonight.  Tomorrow I'll measure voltages again, and post them here.

Thanks for the instant diagnosis, and all the troubleshooting tips.  You saved me a lot of time.

[tdvt]

Now that it is up & running, are you happy with the reverb?

Thinking I might lift that 1-tube circuit for a 6G2-ish build I have been plodding along on.

I have many of Tubenit's 1-tube schematics saved, enough of them to be confused about which is which, so thinking it's good to confirm an in-service version.

[Lectroid]

I do like this reverb.  It's stronger than I expected from coming out of just one tube.  The tone is different from Fender's--not so metallic like surf music or acid rock.  It's got depth but it's softer in feel, more feathery.  Maybe part of its sound is that I used a short tank?  I don't know.

I can recommend the 12DW7 tube; its high-mu/low-mu sides are perfect for the different needs of driver vs. recovery.  I swapped in a 12AX7 in that spot two days ago when troubleshooting, just to rule out a bad tube.  It works okay with the 12AX7, but the 12DW7 puts out a noticeably better reverb and I'll be putting it back in.

I'll post tubenit's original drawing below and will also post a final schematic of the 18 watt later today with voltages in case anyone's still interested.

Edit: Forgot to mention that I left off the mosfet cathode follower tubenit put in his design.  Maybe it degrades the sound somewhat, but the amp seems loud and clear to me.  So far it isn't a problem. 

[Lectroid]

Here is the latest schematic with voltages for anyone interested.  I changed the 10K resistor in the power supply down to 4.7K to boost the preamp voltages.  Worked out okay; I gained 10-12 volts on C & D nodes.  Changed the coupling cap after V1a from .01 to .022.

Bias is warm, 87% max dissipation, so I may get than down a little.  Otherwise, I'm pretty happy with it.

[tdvt]

I do like this reverb.  It's stronger than I expected from coming out of just one tube.  The tone is different from Fender's--not so metallic like surf music or acid rock.  It's got depth but it's softer in feel, more feathery. 

That sounds like just what I am looking for, something for some depth but doesn't need to be over the top.  I also already have a couple 12DW7s on hand, so I just need to get back to that project.


I have a 6G15 Reverb I built a while ago to get surfy if the need arises.

[Lectroid]

I've continued to work on this problem and made enough progress to make it worth writing down for posterity.   Schematic attached below.

Swapping tubes had the biggest effect. The PI works very well with a 12AU7.  It works great; removes a lot of the amp's fizz and distortion w/o losing much output power. 

I tried a 12AU7, and 12AY7 to replace the original 5751 in V1.  The 12AU7 has more clean headroom, but the OD is not as smooth.  The 12AY7 sounds a little grittier and punchier and has a smoother OD tone. Life is compromise.  Also on V1, I changed the cathode bypass caps to: 10uf for V1a  and 2.2uf on V1b, to smooth out even more distortion.

I like the amp.  I still have some voicing to do but I like it already.  My hope was that the amp would sound clean up to about 5 on the Volume knob, then add increasing OD after that.  But now the cleans end around 3 on the Volume knob and increasing distortion begins to come in. 

Does anyone know if changing over to a single pre-PI master volume would open up some more clean headroom?  Or is there another way to make that clean-to-distorted point happen at a higher Volume setting?  Is this just how early  Marshalls sound?

[tubeswell]

I've continued to work on this problem and made enough progress to make it worth writing down for posterity.   Schematic attached below.

Swapping tubes had the biggest effect. The PI works very well with a 12AU7.  It works great; removes a lot of the amp's fizz and distortion w/o losing much output power. 

I tried a 12AU7, and 12AY7 to replace the original 5751 in V1.  The 12AU7 has more clean headroom, but the OD is not as smooth.  The 12AY7 sounds a little grittier and punchier and has a smoother OD tone. Life is compromise.  Also on V1, I changed the cathode bypass caps to: 10uf for V1a  and 2.2uf on V1b, to smooth out even more distortion.

I like the amp.  I still have some voicing to do but I like it already.  My hope was that the amp would sound clean up to about 5 on the Volume knob, then add increasing OD after that.  But now the cleans end around 3 on the Volume knob and increasing distortion begins to come in. 

Does anyone know if changing over to a single pre-PI master volume would open up some more clean headroom?  Or is there another way to make that clean-to-distorted point happen at a higher Volume setting?  Is this just how early  Marshalls sound?

Try going to a 12AT7* in the LTP, and removing the cathode bypass cap** for the EL84s.


*Will make a better load line than a 12AU7.


** will take some hair off the gain and sound more bluesy.

[Lectroid]

tubeswell,

Thank you. I'll try both suggestions today.

[Lectroid]

tubeswell,

Changing to a 12AT7 in the PI was most definitely the secret sauce.  Harshness is mellowed.  Cleans are crisp and bright up to about halfway on the Volume knob, then the OD increases from there.  No appreciable hum. (A lot of luck there, but I'll take it.)  Tone controls work tolerably well. Good range on the  master volume.  Reverb is good enough for anything short of covering "Pipeline". 

What does a 'better' load line mean in this context?

I also lifted the 6BQ5s' cathode bypass cap--that also helped a lot.  I thought about switching it, but I finally just left it out.  Much better without it.  Blues tones are smoother, more rounded.  My goal was something like a Marshall 1974 and I think this gets close.

I'll also post my layout doc.  It's ugly, but it's built on a Hoffman Stout Reverb board and that saved some time and cash.

Thanks again to everyone who offered advice; this amp would be a sad thing without all your help.