1W JTM build

[TriodeLuvr]

The only issue I see is the 1K cathode resistor at the outputs. If the 9.4V reading is correct, and you have about 220 VDC between the plates and cathodes, you're operating the tube cold, at less than half its dissipation rating. Looking at the plate curves below, you're operating near the red marker dot located on the -10V grid line. The tube should be operating near the upper marker, which would be about -7V. Again, assuming 220 VDC plate to cathode (the vertical red line in the chart), the -7V grid line intersects the 220V line at about 10 mA (each section). The cathode resistor for a single section then would be equal to 7 ÷ 0.010 (E÷I), or 700 ohms. For a resistor common to both sections, use half the value, or 350 ohms. The nearest common value (moving in a direction that won't over-dissipate the tube) is 390 ohms. The resistor will dissipate 0.14W, so use a 1W for longevity.

Jack

[Bieworm]

I could do 390R with a parallel 3k3.. that's 348 ohms. I only use at least 1W , generally 2W mf.

[TriodeLuvr]

Use the 390 ohm by itself first. Then measure the DC voltage across the resistor and from one of the anodes to ground. Those two measurements will allow you to calculate the anode to cathode voltage and current flow. I would use 5W as a good upper limit (the sections together are rated for 5.5W).

Jack

[Bieworm]

Use the 390 ohm by itself first. Then measure the DC voltage across the resistor and from one of the anodes to ground. Those two measurements will allow you to calculate the anode to cathode voltage and current flow. I would use 5W as a good upper limit (the sections together are rated for 5.5W).

Jack

Hey Jack, would the lower power tube operation be related to the subnotes? Because the goal is a 1 to 2 watt amp, not 5. The guy I'm building it for is out for that low watt full throttle amping. He's already got a Cornford Harlequin , that's 5 to 6 watt. And a little too loud for that purpose. OTOH, I want to supply him a great sounding little tone monster...it already looks awesome!!!

IMG-20210620-WA0004.jpeg

[TriodeLuvr]

Anode dissipation and output power are two different things. Running the tubes at 80-90% of the dissipation rating will provide output of about 1W to 1.2W.

Jack

[Bieworm]

Did the 390R cathode resistor, removed the parallel Rk of the PI back to 56k. Swapped the 33uf for 47uf on the preamp filter cap.
Pretty good I think..
Listen:
https://www.dropbox.com/s/7hhl7iabpo0q3 ... 1.m4a?dl=0

20210621_205738.jpg

Should I be doing something about V1a voltages?

[TriodeLuvr]

Sounds good to me! Are you using the same amplifier settings that produced those ugly subharmonics before?

I'm surprised the DC voltages at V1a and V1b aren't more similar. The anode and cathode resistors are the same, right? Maybe you should confirm their actual values with an ohmmeter. This is a strange situation, because V1a shows more anode voltage than V1b, indicating less current flow. However, V1a also shows more voltage on its cathode, and that would mean more current flow than V1b. It only makes sense if the resistors aren't the same values.

Your voltage chart doesn't say if the +222V on the outputs is measured from the anode to cathode or from the anode to ground. I'll assume it's anode to ground, which means we have to subtract the voltage on the cathode resistor to know the voltage across the tube itself. So, the tube is operating with 215V at the anode, -6.8V bias, and it's flowing 17.4 mA. With these numbers, the tube is dissipating 3.7W total. That's all close to the tube chart, so the tube is operating as predicted. You could probably reduce the value of the cathode resistor a little more, but it might not be worth the effort. The original Marshall design that I've seen online uses a 470 ohm cathode resistor. You're already down to 390 ohm, and it seems you have the sound you want.

EDIT: Don't know how much B+ is applied to the outputs in the Marshall design. That effects the allowable value for the cathode resistor. Brimar and GE show identical values for "Typical A1" operation of the 12AU7 with 250V on the anode. Dissipation with their values is 2.63W per section, very close to the maximum of 2.75W. If your amp can hold the 215V on the anode, you should achieve the same dissipation (2.63W) with 12mA/section and 5.9V on the cathodes. That would mean a resistor value of 246 ohms. So, our previous estimate of 350 ohms is actually conservative, and that's because we calculated for 2.5W each section, not 2.63W. The bottom line is that you have quite a bit of room to move concerning the cathode resistor value, so adjusting it to suit the sound might be a good approach. Just remember that the higher the dissipation, the shorter the life of the tubes.

Jack

[Bieworm]

I will have to check those plate resistors on V1 indeed.
The 222V is , like you assumed, anode to ground.
I tried a 2k2 parallel with the 390R, but is wasn't an upgrade, so I removed it.
I think it sounds pretty good now.. a little on the edge of the subtones.. but not dissonant..rather in a harmonic way.
I'll settle for now..and let it sink into my head

Thanks Jack! You gave me some good guidance.

[TriodeLuvr]

Cool. I added an edit to my earlier post.

Jack

[Bieworm]

In that context my amp now is biased at 34% dissipation. Isn't it better practice to have at least 60 to 70%?
Is it then possible there is crossover distortion at act, causing the weird subtones?
Maybe this is a good idea: put a 220 Ohm resistor in series wit a 220 Ohm 4W pot and adjust the bias by ear?
https://www.gotron.be/componenten/passi ... o-lin.html

BTW, the 100k anode resistors on V1 are identical and read 99k. I think I should do some tube rolling to rule out the difference doesn't start there

[JMPGuitars]

It's definitely an improvement, but there's still lacking articulation. Between about 1:30 to 2:30, and 3:30 to the end sound the best. Any time there's a lot of distortion, notes are getting lost.

I'd go with the hotter bias / slightly higher B+ and see what that sounds like. Try adding the .047µF film cap in parallel to the preamp filter cap and see if it affects the unwanted stuff.

Thanks,
Josh

[TriodeLuvr]

In that context my amp now is biased at 34% dissipation. Isn't it better practice to have at least 60 to 70%?
Is it then possible there is crossover distortion at act, causing the weird subtones?
Maybe this is a good idea: put a 220 Ohm resistor in series wit a 220 Ohm 4W pot and adjust the bias by ear?
https://www.gotron.be/componenten/passi ... o-lin.html

BTW, the 100k anode resistors on V1 are identical and read 99k. I think I should do some tube rolling to rule out the difference doesn't start there

You’re running the tube at 3.7W, which is 67% of the maximum allowable dissipation (5.5W). If you want to be sure it’s class A1, and that there’s minimal chance for crossover distortion when it’s driven hard, change the cathode resistor to 240 ohms. That should put the operating point almost exactly where Brimar and GE data sheets suggest, even if B+ sags a little when you draw the additional current.

I don’t think it’s necessary to add adjustable bias. That’s up to you, of course, but IMO, it’s best simply to have the outputs at their best electrical operating point. If you want to use the output stage to make tone changes, an adjustable B+ would be a better option. Honestly though, I don’t think any of this is appropriate in a 1W amplifier.

Be sure to measure the cathode resistors at V1, not just the 100Ks. It might be a good idea to re-solder the connections at the sockets and the various components, too. Something seems wrong, but I don’t think it’s the tube itself.

Jack

[Bieworm]

It's definitely an improvement, but there's still lacking articulation. Between about 1:30 to 2:30, and 3:30 to the end sound the best. Any time there's a lot of distortion, notes are getting lost.

I'd go with the hotter bias / slightly higher B+ and see what that sounds like. Try adding the .047µF film cap in parallel to the preamp filter cap and see if it affects the unwanted stuff.

Thanks,
Josh

The hardest thing is the low wattage.. it so contra intuitive on many areas. The fact that the volume knob is maxed out when doing tweaks and such.. how often do we do that with an 18 watt? The 18 watt behaves different too when dimed, and I forget to compare the situations for both amps in state of maxed out. Very confusing...

FWIW, there already is a .1uf parallel to the 47uf filter cap. Do I need to add another .047?

[TriodeLuvr]

More about V1…. A simulation with a “perfect” tube and your power supply voltages indicates each section should have +138V on the anode and +1.16V on the cathode. However, bear in mind that tubes are much more variable than most of the solid state devices we’ve gown accustomed to, so all your measurements are reasonable in this respect. What’s not reasonable are the relative currents in the anode and cathode resistors.

I’ll spare you the math, but the bottom line is this: Based on your voltage measurements, V1a has 0.71 mA flowing in the anode resistor and 0.88 mA in the cathode resistor. V1b has 0.78 mA flowing in the anode resistor and 0.68 mA in the cathode resistor. This is a simple series circuit at DC consisting of the anode resistor, the tube, and the cathode resistor. This means the current flowing in all three components must be identical. It doesn’t matter if the tube looks like 50K ohms or 250K ohms, the currents must be equal within each section. Here are a few ways readings like this can occur:

1. The amplifier hadn’t warmed up enough before making the measurements; current/voltage was still changing.
2. Line voltage was changing while making the measurements.
3. The anode resistors or the cathode resistors are unequal.
4. Cathode bypass caps are leaking.
5. The 1M grid resistors are allowing contact potential bias to form on the grids (negative voltage). Unlikely, but can be measured with a high impedance meter.
6. The tube is gassy, allowing current to flow out the grids. Similar to above, unlikely, but can be measured.
7. The input impedance of your meter is loading down the anode measurements. Also unlikely, because the measurement difference is reversed between the two sections. Cathode current is higher than the anode in one; lower in the other.
8. High resistance solder joints are affecting voltage drops and current.
9. Some combination of the above.

I’ll add anything else if I think of it.

Jack

[geoff 1965]

what about the filtering between V1a and V1b anodes Jack,could that account for the difference?