1W JTM build

[Bieworm]

Well going to take that leap into the 1W territory. A dude wants me to build it. Going to build it according to the schematic MATEC drew, but with a TMB tone stack.
Details will follow...soon

[Bieworm]

ok here goes...

took the schematic from Matec and added a TMB tone stack. Also want to add a MV, so I drew 3 scenarios for that. The amp has a cathodyne PI so there are some options.

Option 1: is the MV right behind the tone stack. The disadvantage is that it's before the last stage of amplification. I feel it's a better option after that stage.

Option 2: is the MV right before the PI. I think I have to eliminate the NFB on the schematic to do that? I read it's the better option for a cathodyne PI for it will give me control over the infamous cathodyne double frequency blocking distortion? Is there a way to maintain the NFB loop in this configuration?

Option 3: is the post PI MV, trainwreck type 3. I can maintain the NFB loop if it's in that spot. But then again, will I call for the blocking distortion after all?

Isn't that NFB loop cleaning things up too much. Note that it IS on the schematic, so.. it must have it's advantages?

schematic MV options.png

[Bieworm]

Ok, I think it will be option 1. But I'm wondering if it wouldn't be better to add the TMB as a CF. Isn't there a chance the TMB will pull too much signal, since it's normally a single tone knob? Or am I worrying over nothing?

[Bieworm]

Heater wiring done.

20210606_101617.jpg

[TriodeLuvr]

Looks like a fun project! You might consider adding a voltage divider from B+ to elevate the heaters. The cathode bias method seems like wasted effort when the output stage consists of a couple 12AU7s with only a few volts available. Effective designs for this technique use self-biased power pentodes that can provide +20V or +30V from the cathodes.

Jack

[Bieworm]

Looks like a fun project! You might consider adding a voltage divider from B+ to elevate the heaters. The cathode bias method seems like wasted effort when the output stage consists of a couple 12AU7s with only a few volts available. Effective designs for this technique use self-biased power pentodes that can provide +20V or +30V from the cathodes.

Jack

Thx. I haven't thought about it that way. Good tip.

[TriodeLuvr]

Ok, I think it will be option 1. But I'm wondering if it wouldn't be better to add the TMB as a CF. Isn't there a chance the TMB will pull too much signal, since it's normally a single tone knob? Or am I worrying over nothing?

The 12AX7 is a real weakling when it comes to driving anything from the anode. I totally trust Marshall's decision to use a CF for this.

Jack

[Bieworm]

I understand. But I'm going to try it first from the anode. There is a 3rd gain stage after the tone stack, so maybe that's sufficient enough. I'm worried I will get some weirdness in that cathodyne PI if I push it too hard , because that's a notorious problem for that type of PI.
I can easily change the circuit for the CF configuration if necessary.

[TriodeLuvr]

I understand. But I'm going to try it first from the anode. There is a 3rd gain stage after the tone stack, so maybe that's sufficient enough. I'm worried I will get some weirdness in that cathodyne PI if I push it too hard , because that's a notorious problem for that type of PI.
I can easily change the circuit for the CF configuration if necessary.

It's not about the gain; that can be shifted up or down as needed. The issue is that the higher impedance of the common cathode driver will interact with the tone stack. I suspect it affects the contour and probably also reduces the range of available control. Maybe I'll run this through the simulator later to see how significant it is. You can bet Marshall didn't waste that cathode follower stage in the original design.

Jack

[TriodeLuvr]

I simulated the 1W tone section shown in schematic #1 and compared it to the original stack used by Marshall in the 2104. The result with all controls centered is plotted below. There are two primary differences.

First, the 1W tone section has about half the voltage gain of the Marshall. This is due to loading of the 12AX7 anode by the stack. Second, the difference between the mid trough and the low frequency peak in the Marshall design is about 7 dB. The difference in the 1W stack is about 13 dB. I would question whether that's usable in a 1W amp, but of course, it can be flattened by turning down the bass control.

If you're interested in knowing more about the differences and characteristics of these two stacks, Duncan Monroe's Tone Stack Calculator is the tool to use. You can model both stacks easily, and also define the drive impedance and load. That lets you tell the software whether the stack is being driven by a CF or an anode. I would recommend doing this before making a final decision on component values for the stack.



Jack

[JMPGuitars]

Speaker characteristics will play a big part in this too though. And keep in mind, some of those Marshall tone stacks can be a bit ice-picky. Though I would love to hear an A/B comparison in the same circuit of both options.

[Bieworm]

I simulated the 1W tone section shown in schematic #1 and compared it to the original stack used by Marshall in the 2104. The result with all controls centered is plotted below. There are two primary differences.

First, the 1W tone section has about half the voltage gain of the Marshall. This is due to loading of the 12AX7 anode by the stack. Second, the difference between the mid trough and the low frequency peak in the Marshall design is about 7 dB. The difference in the 1W stack is about 13 dB. I would question whether that's usable in a 1W amp, but of course, it can be flattened by turning down the bass control.

If you're interested in knowing more about the differences and characteristics of these two stacks, Duncan Monroe's Tone Stack Calculator is the tool to use. You can model both stacks easily, and also define the drive impedance and load. That lets you tell the software whether the stack is being driven by a CF or an anode. I would recommend doing this before making a final decision on component values for the stack.



Jack

To what specific Marshall are you comparing th 1W? A 2204 or plexi?

[TriodeLuvr]

It's the Marshall 2104 MV. I think the 2204 is the same tone stack, not sure about the Plexi.

[TriodeLuvr]

Speaker characteristics will play a big part in this too though. And keep in mind, some of those Marshall tone stacks can be a bit ice-picky. Though I would love to hear an A/B comparison in the same circuit of both options.

The ice pick might not be the stack itself. The sims I did today indicate the Marshall stack at mid-position actually has less treble boost than the 1W under discussion. FWIW, in the process of building my 36W, I constructed the entire 2104 MV circuit in a simulator. There's a HUGE amount of high frequency boost designed into this circuit external to the tone stack. In fact, based on the response plots the simulator produced, I decided to omit the 470 pF peaking cap located just ahead of the preamp gain control. I read later that this is a common modification by Marshall owners. Even without it, this amp is extremely crisp and very detailed and forward in the upper ranges. It's not harsh or strident, but I wouldn't want more.

Jack

[Bieworm]

Thanx Jack!
I'm going to build it per schematic first and see from there. It seems the JTM145 sports the same TS like I drew it out earlier. Except the 1st and 2nd lug on the mid pot is linked on the JTM145.
I was thinking about putting in smaller caps and larger resistors on the TS so I can parallel those to tweak the values.
That CF is always an option in case of necessity. But this is a 1W amp, so there's always gonna be OD anyway..

[Bieworm]

Had a bummer earlier this week. The faceplate I ordered at Modulus Amplification arrived 2 weeks after ordering. Sadly it is a mirror image of what I expected to get. Because I already drilled all the holes and added the power supply there were only 2 options:
1. Order a new , correct one
2. Reverse the amp. Attach the faceplate to what would have been the backside, meaning the pots would be positioned above the heater wiring.

I have seen Geoff's work and he did this on several amps before, so I figured I'd be all right doing the same. After all there are only 4 wires coming from the pots to the board, one that even travels to the grid of V2a. This wire will now only be like 6cm long, where it would have been 20cm shielded wire crossing the entre amp. Don't think the build will suffer from this configuration.

[TriodeLuvr]

Sorry to hear about that. It's hard enough building an amp without something like that happening. Hope it works out OK.

Jack

[Bieworm]

It's alive!!! You were right Jack, about the excessive low end. I got it under control by disconnecting the cathode bypass cap of V1a and also leaving out the extra 820R 220uf network.
Sounds fat and creamy!! Awesome !!!
It's tweaking time though... i'd like to have more mids. Going to swap the .022uf cap + 25k pot for a .01uf and 50k pot, and go the jalapeno route... should I also change the 56k slopen resistor?
And how the F*** can I calculate the bias with the 12AU7? Plate voltage is 260V and Cathode measured 8.6V

[JMPGuitars]

It's alive!!! You were right Jack, about the excessive low end. I got it under control by disconnecting the cathode bypass cap of V1a and also leaving out the extra 820R 220uf network.
Sounds fat and creamy!! Awesome !!!
It's tweaking time though... i'd like to have more mids. Going to swap the .022uf cap + 25k pot for a .01uf and 50k pot, and go the jalapeno route... should I also change the 56k slopen resistor?
And how the F*** can I calculate the bias with the 12AU7? Plate voltage is 260V and Cathode measured 8.6V

Rob's calculator covers a 12AU7: https://robrobinette.com/Tube_Bias_Calculator.htm

According to that, you're at 4.6W, and 78% dissipation assuming I entered everything correctly.

[Bieworm]

I tweaked like I said. Put the tremolo TMB tone stack and pots in it. Added zener diodes to lower B+. All voltages are within 10%, so that's fine.
Here's the measurement data:

20210612_122833.jpg

20210612_123514.jpg

20210612_123647.jpg