New build fine tuning

[maxkracht]

The build looks nice. Fine tuning is always the hard part.

The "fat" channel is very dark and quiet, so I suspect I'll need to revisit how I wired my input jacks to switch that cathode bypass cap in and out (might end up using maxkracht's "cloacal" input jack trick using a stereo jack. I suspect some weirdness is happening with signal getting bled to ground there.

I don't know if I would trust the footswitch function of my multipurpose input on anything remotely high gain. I have only used that in champs and other little SE amps.

but I do remember getting 185V at the cathode of the cathode follower, and I know 12AX7s are rated for a 200V maximum between the heaters and cathode. So while there probably is still some wiggle room in terms of preamp voltage, I'm thinking that's as close to maximum as I'd want to get

Elevate your heater and you don't have to worry about it and you get a bonus of potentially lower hum.

The headroom issue does remain, though. It produces more drive than most non-master Marshalls I've tried or owned. My concern is less that there's more overall drive available and more that I wish it would be more responsive to guitar volume and picking dynamic changes.

I doubt the slightly larger cathode bypass cap is what is causing this. Do you have a schematic?

[cdemike]

The build looks nice. Fine tuning is always the hard part.

Thanks, maxkracht!

I don't know if I would trust the footswitch function of my multipurpose input on anything remotely high gain. I have only used that in champs and other little SE amps.

I haven't been using the input jacks in the fullest sense of "cloacal," but rather borrowed the idea to differentiate between the two inputs by switching half of V1's cathode bypass capacitor. Still really itching to try the full version with a stereo jack, but I'll need to find the right build, I think.

I doubt the slightly larger cathode bypass cap is what is causing this. Do you have a schematic?

Up-to-date schematic attached! (Edit: caught minor discrepancy between the schematic and the built version, which used a 33k NFB resistor in series with the NFB control rather than the listed 27K). I agree about the bypass cap, and the way I currently like running it should be electrically equivalent to a lead-spec amp with low filtering: running into the bright input for 680nF bypass on each side of V1, 2.2nf coupling cap from V1 via the "F.A.C." control, 4.7nF bright cap, 680nF bypass on V2 rather than 1.38uF, maximum NFB (33k), 32uF screens filtering, and 50uF mains filtering. I'm really liking the transient response and how easily it gets the honky, rubbery live Jimmy Page sound, so I think I'm close with where things are, but really would like to have it be more responsive to my guitar's volume control. I have noticed that it behaves closer to how I'd like it to when I have the tone stack switched to the 66k/250pF setting, so that might be a clue as to what's happening.

Edit part 2: Found a mistake in the schematic in the tonestack switch. I confirmed that it's correct in the amp, but the capacitors in the schematic should be labeled 500pF, and the order on the top "fork" should be reversed so that there is one 500pF cap engaged in the switch position shown in the schematic and both 500pF caps in series for an effective 250pF when in the opposite position, corresponding with 66k total slope resistance.

[bepone]

like for the anode mixer with separate cathodes in the first stage!

[maxkracht]

Ha, looked at the schematic before the message and was going to ask about the treble caps. If you like the bigger slope resistor, you could easily try even bigger to see if that changes things. I think I would play with the volume control and or grid leak values before the CF. Easy to slap a resistor in parallel with one or both and see how it changes the feel. Could also try a split plate load switch like in a Komet, if you need more switches...

[cdemike]

like for the anode mixer with separate cathodes in the first stage!

Thanks! So far I'm really liking it. I'm pleasantly surprised with how there's less noise with the paralleled cathodes by how switching one side's cathode bypass cap makes a noticeable but subtle difference. It makes a big difference in terms of the transient response.

Ha, looked at the schematic before the message and was going to ask about the treble caps.

Yeah, sorry about that. I'll draw a version with less eraser marks and that doesn't have silly mistakes once I reach a version I'm happier with. I'm also hoping to record some sound clips once it reaches that point. I'll post them here once that happens.

If you like the bigger slope resistor, you could easily try even bigger to see if that changes things. I think I would play with the volume control and or grid leak values before the CF. Easy to slap a resistor in parallel with one or both and see how it changes the feel.

Liking the idea of messing with the grid leaks! I do the amp work outside since our cat thinks she's a goat and tastes whatever is on the ground, so that's our solution to keep her from snacking on stray solder... Anyway, it's less than 20 degrees out this weekend, so I'm liking being able to clip in alternate values there while I wait for the weather to improve. After I read your post about the slope resistor I did some fooling around with the tone stack simulator and decided to see how the touch responsiveness changes with different middle control settings. I am finding that it is considerably more responsive with lower middle settings, so I think most of the undesirable drive is coming from the phase inverter. Hitting V2 less hard coming out the volume control also might help resolve that, especially since it's unsurprisingly more responsive at lower volume settings or if I run a boost pedal in front at less-than unity gain. I'm considering bumping the filtering at the phase inverter up from 30uF or 80uF to see if that helps, though it seems like that would be pretty subtle?

Could also try a split plate load switch like in a Komet, if you need more switches...

Haha I definitely got a little carried away with the switches on this build... I haven't tried split plate load resistors. I've seen versions with parallel coupling caps with the switching happening there -- seems like that'd be the better approach in terms of noise and in terms of safety if I wired it to a push/pull vs. a full-sized switch. Hopefully the grid leak value tweaks will get me where I need to be before I get carried away with even more switches. I'm also finding that Fischerman's point about the all-access style build is ringing truer and truer as time goes on, and that I'm not really using the coupling cap control, V2 cathode bypass cap switch, or NFB control. So I'll accept your roast on the switches

[cdemike]

I think I got this to a place where it's behaving more like I wanted it to, so I thought it I'd pass on what I ended up doing and say thanks to everyone who helped me figure out how to get this working the way it does now.

As a starting point, I wanted to see if reducing the voltage divider after V1 would help improve touch sensitivity per maxkracht's suggestion. That helped, but it was pretty subtle when I had it set to the overall amount of drive I wanted with my guitar's volume all the way up. There was also a change in the sound that I didn't love when the resistance to ground went up that is hard to describe. It was also subtle, but the transient didn't feel as punchy and the sound overall became somewhat harsh. That makes sense since the bright cap basically makes the following resistor to ground function as a high-pass filter, and I noticed a similar effect on the 4-hole Marshall's I've played in the past when using the bright channel un-jumpered and playing with the normal channel's volume control (that was really the inspiration for that section of the circuit since I tend to run the normal channel's volume in those circumstances all the way up or close to it). So while I think that was a step in the right direction, I returned it back to the original setup since I couldn't get past the brightness.

The impact of the slope resistor on the responsiveness got me thinking that the phase inverter might be the prime culprit for the lack of responsiveness I was looking for, since the output fed into the phase inverter seems to make a significant difference. I should have recognized it earlier since the 12BZ7 test moved me significantly closer to where I wanted to be, but here we are. Still fun to roll the tubes there and see how that impacted things. I did some digging around previous threads about getting more headroom out of the phase inverter and noticed that my phase inverter voltage was not low but also not particularly high. There's a ton of really valuable information here, and the people in the Dumble section helping each other dial their amps in provided great insight. So I decided to go ahead and mess with voltages in the B+ string. I started with getting a full set of voltage measurements:
OT CT node: 436V
Screens node: 434V
PI node: 355V
V2 node: 306V
V1 node: 289V

The limiting factor would be the cathode voltage on the cathode follower, which remained at 185V (good to know my wall voltage is relatively stable!). I started with 5.2K replacing the 10K right after the choke and ended up with 390V at the node feeding the PI's plate resistors. That's in line with the high end of what I've seen in Marshall circuits, though I know some Fender designs ran them even higher. With that setup (rectifier > choke > 5.2k > 10k > 10k) I was getting exactly 200V at the cathode follower's cathode resistor, which, as I understand it, is over maximum since the 6.3VAC would put the heater-cathode voltage at a maximum of ~203.5V without signal. So my options were to elevate the heaters, switch to DC heaters, or lower the B+. I tried 8.2k in lieu of 5.2k, which raised voltage at the PI node to about 375V which didn't seem to be appreciably higher than 355V. So next I raised the 10k resistor between the PI and V2 nodes to 22k, which got me similar voltages to where I started with 315V at V2, 304V at V1 and 190V at the cathode follower's cathode. The amp is surprisingly more responsive now, and I can get actual clean sounds of the amp with my guitar's volume at 4 and moderate-to-light picking pressure playing chords. Those differences carried over when I swapped the 12BZ7 back to a 12AX7, and I think I liked the slight increase in compression I got with the 12AX7 though they honestly sounded pretty similar. I'm not aware of new production 12BZ7s, so I'm trying not to put a ton of hours on the 3 I have. With the current setup I can pick as hard as I want playing single notes, and now I have a much wider range of usable clean range on my guitar's volume. I should mention that I did consider changing the taper of the guitar's volume pot and did play with pickup heights which made a difference too. I opted against changing the taper of the pot, since my other amps clean up well with this guitar. In any event, the amp sounds great -- exactly as intended, really scratching the itch to have an amp that is kind of does the Jimmy Page's MSG sound, but also very versatile and useful for the other playing I do (mostly country). I spilled a glass of water in my laptop, but when it's repaired I'll record and post clips. Proud of this one!

[cdemike]

Took me long enough, but I finally made clips. My apologies for the sound quality; these were recorded on a cheapo Amazon microphone, but I'm planning to record some better clips with some nicer microphones soon-ish.

Since the last post, I kept tinkering since I can't leave a good thing alone. I biased the second gain stage cooler (cathode resistor is 2K now) to increase headroom and get more clipping on the same side of the waveform which the cathode follower clips based on the Merlin approach. The switch controlling that gain stage's cathode bypass capacitor value now now just takes a 680nF bypass capacitor in and out of the circuit by shorting a 220K resistor between the capacitor and ground. Leaving it un-bypassed dramatically increases headroom, but it feels a little too stiff for my taste. It did significantly increase headroom and improved its responsiveness to changes in guitar volume or picking attack. This stage might get tweaked further in the future. Separately, I was able to confirm that the red plating KT88 was indeed a bad tube. I didn't buy them new, so I'm hesitant to call it EH's fault; seems more like me learning my lesson about being a cheapo and buying used tubes in mysterious condition. There was an as-new set of Gold Lion KT88s that were about the cost of a pair of new JJs on Reverb, so I thought I'd give them a try. They sound great and were in the amp when these were recorded.

Mostly thinking out loud, but I'm flirting with replacing the NFB control with a master volume since I basically always leave it at maximum NFB/minimum resistance. The amp also basically always lives with 2.2nF effective capacitance coupling the first stage to the volume pot via the FAC control, so that makes 2 seldom-used controls (as predicted earlier in the thread regarding all-access style builds). Given there being 2 redundant control spots, I'm considering adding a Trem-O-Nator via an auxiliary board either reorganizing V1 to free up a spare triode or use a MOSFET. I'm hesitant to use a V1 triode for the LFO for a number of reasons related to layout, so it'd either be a MOSFET or I'd get rid of the parallel triodes on the input stage and reassign the triodes' functions borrowing from the recent thread about DC cathode followers being in different tubes. The latter would likely mean using half of V2 as the LFO. Using a tube LFO would be the bigger project and I'm hesitant to tear apart an amp I generally really like. Either way, any future changes besides possible tweaks to the second gain stage's cathode resistor are probably decently far down the road; I'm taking time to enjoy the amp as much as possible as it is now.

Clip 1: https://youtu.be/Vj1Dfv_Aihk
Clip 2: https://youtu.be/BujpYDWQ9Sc

[cdemike]

Decided to add the trem -- will report back how that goes. I did a few things with the schematic that I haven't seen before, so I'll be excited to know if it'll work.

I also redid the filtering scheme to see if I could get closer to the vocal quality I hear in Jimmy Page's MSG sound. I decided to try even lower filtering, but I'm still not able to quite get the kind of transient response that you can hear at 8:24 when he bends the B played on the G string: https://youtu.be/MoHojt4iWoU?si=rFEeCeOO_N_GW6Yp&t=502

To me it almost sounds like a wah, and I've gotten closer to this by progressively lowering filtering. However, I'm finding I'm hitting a plateau with the current filtering level (32uF reservoir, 32uF screens, 64uF PI, 30uF V2, 16uF V1). I had 16uF on V2 before, but that didn't make a ton of difference in this regard, so I'm wondering if anyone has any pointers in terms of getting this kind of transient response.