1968 Plexi component values?

[Fischerman]

Thank you. Looking at the materials you shared, I realize that the component values ​​I had in mind for the SLP circuit were exactly those of the '68 model.
Since I built it referencing the Mark Huss Plexi 6V6 and the 1959/1987x reissue schematics, it seems I effectively built it to '68 specifications from the start.
I had experimented with changing a few component values ​​after looking at various versions of the circuit found online, but I ultimately ended up reverting to the '68 or reissue specifications.

Below is the schematic for the amp I am currently building. I am repurposing a single-channel Deluxe Reverb clone that I built about 15 years ago.
I’ve incorporated switches to toggle between Bassman, JTM, and SLP modes.

KakaoTalk_20261007_155553151.jpg


After checking the links, I have a clear understanding of the overall evolution of the circuit. Thank you for sharing such valuable information.

I'm no switching expert, far from it, but I'm pretty sure at least some of those switches will pop or thump pretty good when switching them. I'm almost certain the NFB switch is going to have a decent thump. My best success with NFB switching is to start with the 'high resistance' value, then switch resistors in parallel with the high value to achieve the lower resistance settings. Or way easier is a pot in series with a minimum NFB resistance value but that sort of depends on location/layout.

Edit: and for the cathode cap/cathode split switching, you can put a resistor of say 15k or 27k between the lugs of the switch to quiet the pops. It only needs to be ~10x the cathode resistor value.

[psychepool]

I'm no switching expert, far from it, but I'm pretty sure at least some of those switches will pop or thump pretty good when switching them. I'm almost certain the NFB switch is going to have a decent thump. My best success with NFB switching is to start with the 'high resistance' value, then switch resistors in parallel with the high value to achieve the lower resistance settings. Or way easier is a pot in series with a minimum NFB resistance value but that sort of depends on location/layout.

Edit: and for the cathode cap/cathode split switching, you can put a resistor of say 15k or 27k between the lugs of the switch to quiet the pops. It only needs to be ~10x the cathode resistor value.

Yes, as you mentioned, a loud pop noise occurs with most of the switches. However, since I don't intend to toggle the footswitch while actually playing—and mostly use the amp at home with an attenuator—it doesn't bother me much. Regarding the NFB switch, for instance: while a parallel wiring setup for a 3-position switch would place the highest resistance value at the center position, I preferred having the resistance increase sequentially from one side to the other, so I used an On-On-On switch instead. Naturally, this results in some pop noise, but I find this configuration much more to my liking in terms of operation.

[cdemike]

A significant part of what I associate with the circa 1968 100w amps is the very relaxed filtering, especially by comparison to the amps from 1969 and later (I'm thinking specifically of later amps with serial numbers with the prefix SL/A; these amps generally reflect what's shown on the Unicord schematic). While most 1968-dated 100w heads had 50uF reservoir values (same as the the later 100w amps), screens filtering in early 1968 was often 16uF (2x 32uF capacitors in series) compared to 50uF (2x 100uF in series) for the screens node in the SL/A amps. As is the case with a lot of Marshalls in the late 60s, though, there are exceptions, with some 1968-dated amps having been documented as having 50uf at the screens node (2x 100uF in series) like the SL/A-labeled amps. Preamp filtering was more relaxed in most amps from early 1968 with the V1 and V2 nodes each having either 32uF or 20uF versus the later amps having 50uF as shown in the Unicord schematic. As you refine your amp, you may want to experiment with different filtering values if you're interested in building something inspired by the 1968-dated amps. Roe has an outstanding summary of the evolution of the early 100w amps: https://roef.folk.ntnu.no/JTM100.htm

With the exception of the V1 cathode resistor switch, it's possible to reconfigure all of the switches on that schematic to be pop-free and usable while the amp is running. The negative feedback resistor in particular can just be a pot wired as a varistor in series with the minimum value NFB resistor (in your schematic a 100K pot in series with a 22K resistor would give you all the same options as the switch). The tone stack can be switched by reconfiguring the slope resistors and treble capacitors in series and placing a voltage reference resistor across the second treble capacitor (schematic below). Regarding the mixer resistors switch, I think most of the amps with 270K mixer resistors had 560pF peaking capacitors rather than the 470pF found on amps with 470K mixer resistors. The V2 cathode bypass capacitor switch can be made noise-free if the switch shorted a large-value resistor referencing one side of the capacitor to ground rather than connecting the same side of the same capacitor to ground as it is shown currently. I don't think there'd be a way of changing the V1 cathode configuration with the amp on without some noise, though using an optocoupler or a BJT as a switch could be made to work without a pop, especially if you used an RC network to slow the switching timing.

On the negative feedback piece, you may want to consider that your output section will be considerably less powerful than a 100w amp, so your NFB signal will also be considerably smaller at the phase inverter. So to get the same magnitude NFB signal, you'll need to reduce the NFB resistor's value to account for the wattage difference. More on scaling NFB here: https://ampgarage.com/forum/viewtopic.php?t=31905

I'm guessing the reverb setup is drawn in error on the schematic since the "return" triode's plate being directly coupled to ground via the level pot (it seems like the amp seems to be working without issue now), but I wanted to point it out in case I misunderstood and the reverb isn't currently set up on the amp.

Popless tone stack switch.PNG

Edited for grammar and clarity.

[psychepool]

A significant part of what I associate with the circa 1968 100w amps is the very relaxed filtering, especially by comparison to the amps from 1969 and later (I'm thinking specifically of amps with serial numbers with the prefix "SL/A"). While most 1968-dated 100w heads had 50uF reservoir values (same as the the later 100w amps), screens filtering in early 1968 was often 16uF (2x 32uF capacitors in series). By comparison, the compared to 50uF for the screens node in the SL/A amps. As is the case with a lot of Marshalls in the late 60s, though, there are exceptions, with some 1968-dated having been documented as having 50uf at the screens node (2x 100uF in series) like the SL/A-labeled amps. Preamp filtering was more relaxed in most amps from early 1968 with the V1 and V2 nodes each having either 32uF or 20uF versus the later amps having 50uF as shown in the Unicord schematic. As you refine your amp, you may want to experiment with different filtering values if you're interested in building something inspired by the 1968-dated amps. Roe has an outstanding summary of the evolution of the early 100w amps: https://roef.folk.ntnu.no/JTM100.htm

With the exception of the V1 cathode resistor switch, it's possible to reconfigure all of the switches on that schematic to be pop-free and usable while the amp is running. The negative feedback resistor in particular can just be a pot wired as a varistor in series with the minimum value NFB resistor (in your schematic a 100K pot in series with a 22K resistor would give you all the same options as the switch). The tone stack can be switched by reconfiguring the slope resistors and treble capacitors in series and placing a voltage reference resistor across the second treble capacitor (schematic below). Regarding the mixer resistors switch, I think most of the amps with 270K mixer resistors had 560pF peaking capacitors rather than the 470pF found on amps with 470K mixer resistors. The V2 cathode bypass capacitor switch can be made noise-free if the switch shorted a large-value resistor referencing one side of the capacitor to ground rather than connecting the same side of the same capacitor to ground as it is shown currently. I don't think there'd be a way of changing the V1 cathode configuration with the amp on without some noise, though using an optocoupler or a BJT as a switch could be made to work without a pop, especially if you used an RC network to slow the switching timing.

On the negative feedback piece, you may want to consider that your output section will be considerably less powerful than a 100w amp, so your NFB signal will also be considerably smaller. So to get the same magnitude NFB signal, you'll need to reduce the NFB resistor's value to account for the wattage difference. More on scaling NFB here: https://ampgarage.com/forum/viewtopic.php?t=31905

I'm guessing the reverb setup is drawn in error on the schematic with the "return" triode's plate being directly coupled to ground via the level pot since the amp seems to be working without issue now, but I wanted to point it out in case I misunderstood and the reverb isn't currently set up on the amp.

Popless tone stack switch.PNG

Thank you for the great suggestion. While most of the switches currently produce "pop" noise—and it is by no means a low-level noise—it hardly bothers me in my actual usage environment. Because the chassis is so compact, my priorities were, first, "Does it perform the desired function?" and second, "Is it easy to build?" The measure you suggested involves adding components that might seem trivial in a typical scenario, but in my specific situation, it would be quite a burden.
Please refer to my previous posts for details on the chassis:

https://ampgarage.com/forum/viewtopic.php?f=6&t=38953
https://ampgarage.com/forum/viewtopic.php?f=6&t=39047

Replacing the NFB resistor with a potentiometer is also good idea. I built a Bogner Ecstasy clone a while back where I added various power amp controls—inspired by the Friedman BE-50 Deluxe—including an NFB knob. I find it very useful and wanted to apply it here (along with a "Depth" mod), but there is simply no room left on the control plate.

I’m currently testing a few additional switches for the SLP mode, and I’m considering fixing the NFB at a suitable value and removing the switch entirely—again, due to the lack of space. I’d prefer not to have to do that, though.

I’ll be sure to read through the NFB-related material you linked. To be honest, I have no theoretical knowledge in this area; the values ​​I’m using now are based on information I got from Gemini and ChatGPT. I’m not sure if the answers were technically correct, but I’m quite satisfied with the results. That said, I do think a slightly lower resistance value might be better for the JTM45. I’d like it to be a bit smoother. I’ll reconsider the appropriate resistance value based on that information.

You were right about the reverb circuit. There is a 10nF capacitor between the return stage plate and the pot, which was missing from the schematic. I built it according to the schematic in the link below (with the Dwell fixed at 1M), and it’s working perfectly.

https://el34world.com/Forum/index.php?a ... tach=44427

[psychepool]

I am currently deliberating on adding a few final switches and adjusting some component values—specifically regarding the SLP mode.
There is no single "correct" answer here, so I am not looking for a definitive solution; rather, I’d like to hear your personal preferences.
I have already begun some testing, but I am curious to know what you think.

1. SLP Bright Channel Cathode Resistor Value
In my mind, the 'default' value for the SLP bright channel cathode resistor is 2.7kΩ. However, I’ve noticed that 820Ω resistors were used in early split-cathode SLP models and some reissue circuits.
I am currently working on a setup that allows for easy testing by switching a 1.2kΩ resistor in parallel with the 2.7kΩ one. I’m curious: would you choose the 2.7kΩ or the 820Ω?

2. SLP Bright Channel Coupling Capacitor
Currently, in SLP mode, a 2.2nF capacitor connects to pin 3 of the volume pot via a switch.
I’d like to add a mode—selectable via a switch—that either places a 22nF in parallel or otherwise switches to a 22nF.
I’ve already tested this configuration; the sound becomes much "beefier," and I think it would be a useful feature, but I’m weighing the options because space for additional switches is limited. I know some vintage Plexi amps (specifically split-cathode models) used 22nF coupling capacitors—what has been your experience with that sound?

3. SLP Bright Channel Bright Cap
I tested four different values—none, 100pF, 470pF, and 4.7n—and ultimately decided to settle on 4.7n as the fixed value for SLP mode.
However, as you know, applying a 4.7n cap causes the bright channel to overdrive even when the volume pot is set to just "1."
If you turn the volume down significantly, the sound becomes quite piercing.
So, I’m considering adding a mode without a bright cap for the SLP setting. I found that the 470pF cap didn't make the sound darker compared to the 4.7n; instead, it narrowed the frequency range of the signal, resulting in a more "ripping" or harsh, fizzy tone. I’m thinking of using either 100pF or no cap at all, and I actually preferred the "no cap" option; it felt much rawer and grittier. The Suhr SL Series MK1 models also use a "no cap" spec.
Which values ​​do you prefer?

4. V2 Cathode Bypass
Currently, I have a 680n cap paired with an 820R cathode resistor, toggled via an "on-on" switch.
However, I’ve noticed that some circuits use surprisingly high values.
For instance, the Mark Huss Plexi 6V6 (2008) circuit uses a 4.7uF cap, and the Suhr SL Series MK2 apparently uses a staggering 250uF value.
Since it’s a simple matter of swapping the "on-on" switch for an "on-off-on" switch, I’d like to add another option.
I’d love to hear from anyone who has experimented with values ​​other than 680n.

5. Power amp coupling capacitor value
As you can see in my schematic, the coupling capacitor currently connecting the PI output to the power tube grid is a 100nF.
I chose this value partly because I was repurposing parts from a Deluxe Reverb clone I already owned, and also to accommodate the Bassman/JTM modes included in the design.
However, while I can't predict how my preferences might change down the line, I currently use the SLP mode most of the time.
I also feel that 100nF might be unnecessarily large—offering no real benefit—even for the Bassman/JTM modes.
So, I am considering swapping it for a 22nF capacitor.
If you were building a circuit that incorporates both JTM45 and SLP modes, which value would you choose: 100nF or 22nF?
Is there any significant advantage to using a 100nF capacitor in a Bassman/JTM circuit?

That’s all. Please share your experiences or preferences.

[Fischerman]

Marshall's are my wheelhouse. In my experience, the optimal combination varies from amp to amp. Some amps sound great with an 820R on the split cathode Bright channel while other get too loose/boomy in the lows. It also of course depends on your target tone. Is it at the dirty end of the spectrum like EVH or the clean end of the cranked Marshall spectrum like Jimi Hendrix.

That said, here's my preferences. I tend to prefer 100W, that makes a difference in some of these values.
1. If it's split I tend to like 2k7.
2. I prefer 2n2. But I do run my EVH Marshall with Bass on 10.
3. Big ass 4n7/5n OR no cap. 500pF might be the worst option ever. IMO. EDIT: I have 47pF on my Jimi head.
4. I prefer larger values than 680n. I tend to use 25uF (never 680n anymore), 25uF sounds like a boost. It's just 'more of everything'. I also REALLY like no cap here. My Jimi head has no cap, my EVH head has a switchable 330uF. 330uF doesn't sound much; if any, different than 25uF. I usually leave it OFF tho.
5. I like 100n for the older circuits (for the Jimi Hendrix tones) and 22n for post '68 SLP circuits (for the EVH tones).

I think you'll find it's more a combination of things that get the best clean-ish tones and then a different combination for the best dirty tones. Those early amps are supposed to be quite bassy, that's part of the sound. The NFB plays a big part in how clean or dirty the amp wants to be. A lot of NFB (27k, 16ohm tap) for the early amps and for EVH levels of distortion, the NFB needs to come way down (100k, 4ohm tap) and bass tamed a bit (2k7 and 2n2 pre, 22n output).

[Roe]

Marshall's are my wheelhouse. In my experience, the optimal combination varies from amp to amp. Some amps sound great with an 820R on the split cathode Bright channel while other get too loose/boomy in the lows. It also of course depends on your target tone. Is it at the dirty end of the spectrum like EVH or the clean end of the cranked Marshall spectrum like Jimi Hendrix.

That said, here's my preferences. I tend to prefer 100W, that makes a difference in some of these values.
1. If it's split I tend to like 2k7.
2. I prefer 2n2. But I do run my EVH Marshall with Bass on 10.
3. Big ass 4n7/5n OR no cap. 500pF might be the worst option ever. IMO. EDIT: I have 47pF on my Jimi head.
4. I prefer larger values than 680n. I tend to use 25uF (never 680n anymore), 25uF sounds like a boost. It's just 'more of everything'. I also REALLY like no cap here. My Jimi head has no cap, my EVH head has a switchable 330uF. 330uF doesn't sound much; if any, different than 25uF. I usually leave it OFF tho.
5. I like 100n for the older circuits (for the Jimi Hendrix tones) and 22n for post '68 SLP circuits (for the EVH tones).

I think you'll find it's more a combination of things that get the best clean-ish tones and then a different combination for the best dirty tones. Those early amps are supposed to be quite bassy, that's part of the sound. The NFB plays a big part in how clean or dirty the amp wants to be. A lot of NFB (27k, 16ohm tap) for the early amps and for EVH levels of distortion, the NFB needs to come way down (100k, 4ohm tap) and bass tamed a bit (2k7 and 2n2 pre, 22n output).

1 1k
2 2n2
3 5nf
4 680n switchable
5 18nf

[psychepool]

With the exception of #5, I’ve installed switches and tested for all the other points.

01.jpg

02.jpg

Although some of these serve overlapping functions, each offers a slightly different nuance, which makes me want to keep them all.
I could satisfy this desire by installing a few push-pull pots, but then I have to ask myself: is it really worth making the wiring that messy?
Since I haven’t even tested these for 30 minutes yet, I think I’ll need another week or two of testing before making a final decision.

To be fair, a proper assessment might require swapping the power amp coupling capacitor to 22n, but based on my experience so far, here are my thoughts:

1. If I had to choose one, I’d go with the 2.7K option. But the 820R setting is also great—it adds a punchy low-mid presence and, naturally, increases the gain. Above all, it provides a warm, smooth overall character. The 2.7K setting, by contrast, has a more "metallic" quality, which aligns more closely with my mental image of an SLP-style amp.

2. Again, if I had to pick, I’d choose the 2.2n. Actually, if I were using the bright channel on its own, I’d prefer the 22n. My DIY Bogner Ecstasy build also has a toggle to switch between 4.7n and 22n, and I almost always leave it on 22n. However, since the SLP also has a normal channel, I think the 2.2n is the better choice here. It also feels somewhat redundant with item #1, though the nuances are distinct.

3. If I had to choose only one, it would definitely be the 4.7n—no question about it. The way its signature mid-high frequencies push forward with such boldness is exactly the Marshall reference sound I’ve always had in mind. That said, the no bright cap option offers a huge advantage; by removing the bright cap, you can finally achieve a clean tone even on the SLP’s bright channel. With the 4.7n, you have to keep the volume pot turned down extremely low to get anything resembling a clean tone, but that results in an unpleasant, grating sound—and if you try to compensate by turning up the normal channel, it immediately kicks into crunch. (Though, to be fair, that might just be because I only use one input.)
If I had to pick just one option from everything I’m currently testing—one that I’d consider an absolute must-have—this would definitely be it.

4. I’m also quite fond of this feature. The none, 680n, and 250uF settings each serve a distinct purpose.
The none mode offers the most composed and transparent sound. The 680n setting boosts the mid-highs, giving the amp a much more aggressive character, while the 250uF setting adds thickness across the entire frequency spectrum. The 680n setting alone is certainly useful, but since I also like the none mode, I’ll definitely be installing a switch. And since I’m adding a switch anyway, there’s no reason not to include the 250uF mode as well.

I’ll continue testing and replace the 100n coupling cap to 22n. If this were an amp used exclusively as an SLP, I would have built it with a 22n capacitor from the start without a second thought; in fact, I suspect I might actually prefer the 22n value even when using the JTM/Bassman mode.

Marshall's are my wheelhouse. In my experience, the optimal combination varies from amp to amp. Some amps sound great with an 820R on the split cathode Bright channel while other get too loose/boomy in the lows. It also of course depends on your target tone. Is it at the dirty end of the spectrum like EVH or the clean end of the cranked Marshall spectrum like Jimi Hendrix.

That said, here's my preferences. I tend to prefer 100W, that makes a difference in some of these values.
1. If it's split I tend to like 2k7.
2. I prefer 2n2. But I do run my EVH Marshall with Bass on 10.
3. Big ass 4n7/5n OR no cap. 500pF might be the worst option ever. IMO. EDIT: I have 47pF on my Jimi head.
4. I prefer larger values than 680n. I tend to use 25uF (never 680n anymore), 25uF sounds like a boost. It's just 'more of everything'. I also REALLY like no cap here. My Jimi head has no cap, my EVH head has a switchable 330uF. 330uF doesn't sound much; if any, different than 25uF. I usually leave it OFF tho.
5. I like 100n for the older circuits (for the Jimi Hendrix tones) and 22n for post '68 SLP circuits (for the EVH tones).

I think you'll find it's more a combination of things that get the best clean-ish tones and then a different combination for the best dirty tones. Those early amps are supposed to be quite bassy, that's part of the sound. The NFB plays a big part in how clean or dirty the amp wants to be. A lot of NFB (27k, 16ohm tap) for the early amps and for EVH levels of distortion, the NFB needs to come way down (100k, 4ohm tap) and bass tamed a bit (2k7 and 2n2 pre, 22n output).

As you mentioned, having a clear goal would make choosing easier; since I lack a specific objective, the decision-making process has been a bit difficult.
I primarily use guitars equipped with humbuckers, and I’ll likely use the "SLP mode" mostly for hard rock or metal—though I also quite enjoy that distinct, crisp, mid-high-focused crunch.
Most of your observations align with my own experience. I especially agree that the 500pF value is the worst; it lets only the high frequencies through, resulting in a thin, piercing tone that’s painful to the ears—it sounded bad regardless of whether I was using low or high gain.
Regarding point #4, I personally preferred the 680n over the 250uF (though "none" is also a good option). Since they serve such distinct purposes, I’ll probably end up keeping all three as choices.

1 1k
2 2n2
3 5nf
4 680n switchable
5 18nf

It seems our perspectives are quite similar. Thanks for suggesting the 1K option for point #1—I’m definitely curious to test that out. I’ve found that the 820R tends to sound overly bloated with high gain (though it does sound great when a boost is engaged).
As for the 18n value in point #5, that’s a new one for me. I’ve been using Mallory 150 in this amp, but I don’t recall ever seeing an 18n version in that line. Is there a noticeable difference compared to a 22n?

[psychepool]

I also have a question I’d like to ask.
The 680n cathode bypass capacitor for V2 was already in use previously.
Currently, it’s wired to a switch alongside a 250uF capacitor; back then, it was mounted on the circuit board and connected to the switch.

Typically, engaging a 680n capacitor should boost the mid-high frequencies, increase gain, and also raise the noise floor—much like what happens in the current setup. However, when it was mounted on the circuit board, it didn’t really boost the mid-highs; instead, it made the overall sound feel "thicker" and significantly reduced the noise.
I certainly prefer the current result, but I’m curious: why did the noise decrease back then? The high-frequency range also felt a bit more subdued.
In that previous configuration, one side of the 680n was connected to the circuit board’s ground, while the other side was connected to an otherwise unused point, which was then wired to the switch along with pin 3 of V2.

[Roe]

With the exception of #5, I’ve installed switches and tested for all the other points.

01.jpg
02.jpg

Although some of these serve overlapping functions, each offers a slightly different nuance, which makes me want to keep them all.
I could satisfy this desire by installing a few push-pull pots, but then I have to ask myself: is it really worth making the wiring that messy?
Since I haven’t even tested these for 30 minutes yet, I think I’ll need another week or two of testing before making a final decision.

To be fair, a proper assessment might require swapping the power amp coupling capacitor to 22n, but based on my experience so far, here are my thoughts:

1. If I had to choose one, I’d go with the 2.7K option. But the 820R setting is also great—it adds a punchy low-mid presence and, naturally, increases the gain. Above all, it provides a warm, smooth overall character. The 2.7K setting, by contrast, has a more "metallic" quality, which aligns more closely with my mental image of an SLP-style amp.

2. Again, if I had to pick, I’d choose the 2.2n. Actually, if I were using the bright channel on its own, I’d prefer the 22n. My DIY Bogner Ecstasy build also has a toggle to switch between 4.7n and 22n, and I almost always leave it on 22n. However, since the SLP also has a normal channel, I think the 2.2n is the better choice here. It also feels somewhat redundant with item #1, though the nuances are distinct.

3. If I had to choose only one, it would definitely be the 4.7n—no question about it. The way its signature mid-high frequencies push forward with such boldness is exactly the Marshall reference sound I’ve always had in mind. That said, the no bright cap option offers a huge advantage; by removing the bright cap, you can finally achieve a clean tone even on the SLP’s bright channel. With the 4.7n, you have to keep the volume pot turned down extremely low to get anything resembling a clean tone, but that results in an unpleasant, grating sound—and if you try to compensate by turning up the normal channel, it immediately kicks into crunch. (Though, to be fair, that might just be because I only use one input.)
If I had to pick just one option from everything I’m currently testing—one that I’d consider an absolute must-have—this would definitely be it.

4. I’m also quite fond of this feature. The none, 680n, and 250uF settings each serve a distinct purpose.
The none mode offers the most composed and transparent sound. The 680n setting boosts the mid-highs, giving the amp a much more aggressive character, while the 250uF setting adds thickness across the entire frequency spectrum. The 680n setting alone is certainly useful, but since I also like the none mode, I’ll definitely be installing a switch. And since I’m adding a switch anyway, there’s no reason not to include the 250uF mode as well.

I’ll continue testing and replace the 100n coupling cap to 22n. If this were an amp used exclusively as an SLP, I would have built it with a 22n capacitor from the start without a second thought; in fact, I suspect I might actually prefer the 22n value even when using the JTM/Bassman mode.

Marshall's are my wheelhouse. In my experience, the optimal combination varies from amp to amp. Some amps sound great with an 820R on the split cathode Bright channel while other get too loose/boomy in the lows. It also of course depends on your target tone. Is it at the dirty end of the spectrum like EVH or the clean end of the cranked Marshall spectrum like Jimi Hendrix.

That said, here's my preferences. I tend to prefer 100W, that makes a difference in some of these values.
1. If it's split I tend to like 2k7.
2. I prefer 2n2. But I do run my EVH Marshall with Bass on 10.
3. Big ass 4n7/5n OR no cap. 500pF might be the worst option ever. IMO. EDIT: I have 47pF on my Jimi head.
4. I prefer larger values than 680n. I tend to use 25uF (never 680n anymore), 25uF sounds like a boost. It's just 'more of everything'. I also REALLY like no cap here. My Jimi head has no cap, my EVH head has a switchable 330uF. 330uF doesn't sound much; if any, different than 25uF. I usually leave it OFF tho.
5. I like 100n for the older circuits (for the Jimi Hendrix tones) and 22n for post '68 SLP circuits (for the EVH tones).

I think you'll find it's more a combination of things that get the best clean-ish tones and then a different combination for the best dirty tones. Those early amps are supposed to be quite bassy, that's part of the sound. The NFB plays a big part in how clean or dirty the amp wants to be. A lot of NFB (27k, 16ohm tap) for the early amps and for EVH levels of distortion, the NFB needs to come way down (100k, 4ohm tap) and bass tamed a bit (2k7 and 2n2 pre, 22n output).

As you mentioned, having a clear goal would make choosing easier; since I lack a specific objective, the decision-making process has been a bit difficult.
I primarily use guitars equipped with humbuckers, and I’ll likely use the "SLP mode" mostly for hard rock or metal—though I also quite enjoy that distinct, crisp, mid-high-focused crunch.
Most of your observations align with my own experience. I especially agree that the 500pF value is the worst; it lets only the high frequencies through, resulting in a thin, piercing tone that’s painful to the ears—it sounded bad regardless of whether I was using low or high gain.
Regarding point #4, I personally preferred the 680n over the 250uF (though "none" is also a good option). Since they serve such distinct purposes, I’ll probably end up keeping all three as choices.

1 1k
2 2n2
3 5nf
4 680n switchable
5 18nf

It seems our perspectives are quite similar. Thanks for suggesting the 1K option for point #1—I’m definitely curious to test that out. I’ve found that the 820R tends to sound overly bloated with high gain (though it does sound great when a boost is engaged).
As for the 18n value in point #5, that’s a new one for me. I’ve been using Mallory 150 in this amp, but I don’t recall ever seeing an 18n version in that line. Is there a noticeable difference compared to a 22n?

820R resistors may read as high as ca 1k. if you really crank an amp 15nf may differ noticably from 22nf, mostly from a quicker response with less blocking distortion

[Fischerman]

Regarding the 18n, Valvestorm has been selling 18n NOS Mustards for a long time. Also 27n and 39n but sometimes they aren't available. Then they come back sometimes. 22n NOS Mustards have gotten quite expensive, these oddball values are relatively cheap. Also, the little red 47pF dogbones sound really great as bright caps, I'll often put one on the Marshall Normal channel and almost always in a BF Fender.

I'm not saying Mustards sound better (or that they don't), I'll just say I like using them. In my EVH Marshall I used 18n at the PI input and 27n for the PI outputs. Since so much of the Marshall crunch happens in the PI, the 18n keeps it a little bit more tight. I also used 18n for the Normal channel coupler but I never use that channel on the EVH. In my Jimi head, I used 18n for Normal channel, PI input, and tone stack with 36k slope/510pF treble. But that amp is 560v so I used 600v rated caps for the PI outs. The EVH is always variac'd down to 90vac so I don't worry about using 400v Mustards there.

[Roe]

Regarding the 18n, Valvestorm has been selling 18n NOS Mustards for a long time. Also 27n and 39n but sometimes they aren't available. Then they come back sometimes. 22n NOS Mustards have gotten quite expensive, these oddball values are relatively cheap. Also, the little red 47pF dogbones sound really great as bright caps, I'll often put one on the Marshall Normal channel and almost always in a BF Fender.

I'm not saying Mustards sound better (or that they don't), I'll just say I like using them. In my EVH Marshall I used 18n at the PI input and 27n for the PI outputs. Since so much of the Marshall crunch happens in the PI, the 18n keeps it a little bit more tight. I also used 18n for the Normal channel coupler but I never use that channel on the EVH. In my Jimi head, I used 18n for Normal channel, PI input, and tone stack with 36k slope/510pF treble. But that amp is 560v so I used 600v rated caps for the PI outs. The EVH is always variac'd down to 90vac so I don't worry about using 400v Mustards there.

for the PI input, I prefer 1nf on bassy amps with nfb and 2n2 otherwise

on the normal ch. 3-5nf is fine.

[psychepool]

820R resistors may read as high as ca 1k. if you really crank an amp 15nf may differ noticably from 22nf, mostly from a quicker response with less blocking distortion

To be honest, I do like the 820R mode—it adds a pleasant warmth compared to the 2.7K mode—but I’ve been a bit conflicted because, when combined with other settings, it can sometimes feel like the sound gets overly "bloated." I’ll try testing some higher values ​​as well; thanks for the recommendation.
As for the power amp coupling capacitor, I have no intention of going below 22n for now—especially since I also have the Bassman/JTM modes available. I’ve even considered testing a 47n cap, though I suspect I’ll likely end up sticking with 22n.

Regarding the 18n, Valvestorm has been selling 18n NOS Mustards for a long time. Also 27n and 39n but sometimes they aren't available. Then they come back sometimes. 22n NOS Mustards have gotten quite expensive, these oddball values are relatively cheap. Also, the little red 47pF dogbones sound really great as bright caps, I'll often put one on the Marshall Normal channel and almost always in a BF Fender.

I'm not saying Mustards sound better (or that they don't), I'll just say I like using them. In my EVH Marshall I used 18n at the PI input and 27n for the PI outputs. Since so much of the Marshall crunch happens in the PI, the 18n keeps it a little bit more tight. I also used 18n for the Normal channel coupler but I never use that channel on the EVH. In my Jimi head, I used 18n for Normal channel, PI input, and tone stack with 36k slope/510pF treble. But that amp is 560v so I used 600v rated caps for the PI outs. The EVH is always variac'd down to 90vac so I don't worry about using 400v Mustards there.

The core concept behind this build is "recycling." As you can see from the way the amp is constructed, using expensive or rare components would be completely unnecessary.
Since this isn't a circuit designed with careful selection and precise configuration—but rather one intended for free-form experimentation and constant switching—it’s really just a fun toy.
Seeing the builds others have created with such carefully curated components makes me want to try putting together a project like that myself someday.

for the PI input, I prefer 1nf on bassy amps with nfb and 2n2 otherwise

on the normal ch. 3-5nf is fine.

You’re using some remarkably low values ​​there. I’ve never seen such low values ​​used between the phase inverter and the power amp before. Doesn’t that end up altering the preamp’s tone a bit too much?

[pdf64]

for the PI input, I prefer 1nf on bassy amps with nfb and 2n2 otherwise

on the normal ch. 3-5nf is fine.

You’re using some remarkably low values ​​there. I’ve never seen such low values ​​used between the phase inverter and the power amp before. Doesn’t that end up altering the preamp’s tone a bit too much?

I think @Roe means the coupling cap at the input of the LTP phase splitter, between treble control wiper and LTP grid, normally 22nF in 5F6A / Marshalls.
60s Fenders typically used 1nF (intended for open back cabs) or 500pF (sealed cabs) there.

[Roe]

820R resistors may read as high as ca 1k. if you really crank an amp 15nf may differ noticably from 22nf, mostly from a quicker response with less blocking distortion

To be honest, I do like the 820R mode—it adds a pleasant warmth compared to the 2.7K mode—but I’ve been a bit conflicted because, when combined with other settings, it can sometimes feel like the sound gets overly "bloated." I’ll try testing some higher values ​​as well; thanks for the recommendation.
As for the power amp coupling capacitor, I have no intention of going below 22n for now—especially since I also have the Bassman/JTM modes available. I’ve even considered testing a 47n cap, though I suspect I’ll likely end up sticking with 22n.

Regarding the 18n, Valvestorm has been selling 18n NOS Mustards for a long time. Also 27n and 39n but sometimes they aren't available. Then they come back sometimes. 22n NOS Mustards have gotten quite expensive, these oddball values are relatively cheap. Also, the little red 47pF dogbones sound really great as bright caps, I'll often put one on the Marshall Normal channel and almost always in a BF Fender.

I'm not saying Mustards sound better (or that they don't), I'll just say I like using them. In my EVH Marshall I used 18n at the PI input and 27n for the PI outputs. Since so much of the Marshall crunch happens in the PI, the 18n keeps it a little bit more tight. I also used 18n for the Normal channel coupler but I never use that channel on the EVH. In my Jimi head, I used 18n for Normal channel, PI input, and tone stack with 36k slope/510pF treble. But that amp is 560v so I used 600v rated caps for the PI outs. The EVH is always variac'd down to 90vac so I don't worry about using 400v Mustards there.

The core concept behind this build is "recycling." As you can see from the way the amp is constructed, using expensive or rare components would be completely unnecessary.
Since this isn't a circuit designed with careful selection and precise configuration—but rather one intended for free-form experimentation and constant switching—it’s really just a fun toy.
Seeing the builds others have created with such carefully curated components makes me want to try putting together a project like that myself someday.

for the PI input, I prefer 1nf on bassy amps with nfb and 2n2 otherwise

on the normal ch. 3-5nf is fine.

You’re using some remarkably low values ​​there. I’ve never seen such low values ​​used between the phase inverter and the power amp before. Doesn’t that end up altering the preamp’s tone a bit too much?

Famously, Fender used 500pf and 1nf. Both work fine. without NFB, you need ca. 1n5 for full frequency response for guitar. With NFB. less than 1n5 is needed, but the NFB loop may collapse at full distortion. Still, smaller caps work fine, by tighting up bass, giving a quicker dynamic, tactile response, and by minimizing blocking distortion.

I also like 15-18nf on the PI coupling caps instead of 22-100nf. The response is quicker and tighter, with far less blocking distortion. And there is enough bass for guitar. just give it a try