Thoughts on Jimmy Page’s Super Bass Marshall?

[Roe]

that amp seem like a standard superlead but with parallel .01u caps instead of .022u (something that can be seen on a few amps from this period).

Personally, I remember trying the GE6550As and found them to sound too hard and sterile for guitar. I sold them to an organist and bass player, where they work better.

I'm inclinded to believe the superbass was lightly modded, first with 6550s and a bright cap. Possibly, the 12AU7 driver came later. not sure

[Reeltarded]

I wonder if it’s possible to build a low wattage version and still have the sound?

The amp would weigh a ton and be deafening loud. I’m not a fan of putting KT-88’s in a Marshall using the stock transformer set as they have to be biased very cold and they really don’t sound better than EL-34’s in my opinion.

The problem is making that volume the instrument hears. So, take 9lbs of decent mahogany and stand closer than one meter. Let me know where to find an extra 6db. My 100w anecdotes are many and the pickups are scraping the cloth on the cabinets.

Majors act like Jimmy Page's sound, but don't sound like it.

Physics is a bitch.

You can make the sound, just not the feel.

[Reeltarded]

that amp seem like a standard superlead but with parallel .01u caps instead of .022u (something that can be seen on a few amps from this period).

Personally, I remember trying the GE6550As and found them to sound too hard and sterile for guitar. I sold them to an organist and bass player, where they work better.

I'm inclinded to believe the superbass was lightly modded, first with 6550s and a bright cap. Possibly, the 12AU7 driver came later. not sure

The PI has to be the answer to most of the question. That is where they make or don't make all that snotty attitude. Otherwise it does sound like an SL with mild gain dialed into the curve of a large bypass, like you described. Also, no 34s.

[Roe]

I wonder if it’s possible to build a low wattage version and still have the sound?

The amp would weigh a ton and be deafening loud. I’m not a fan of putting KT-88’s in a Marshall using the stock transformer set as they have to be biased very cold and they really don’t sound better than EL-34’s in my opinion.

The problem is making that volume the instrument hears. So, take 9lbs of decent mahogany and stand closer than one meter. Let me know where to find an extra 6db. My 100w anecdotes are many and the pickups are scraping the cloth on the cabinets.

Majors act like Jimmy Page's sound, but don't sound like it.

Physics is a bitch.

You can make the sound, just not the feel.

I happen to have some insane LP replicas - old top of the line Tokais. they do help. but the bridge pickups sound a little to fat. so sometimes a SG with T-tops sound better here

[Reeltarded]

And even with much more distortion I am using PAFs, but honestly, $99 Duncan SH1 is made to do this thing really well.

Buy only necks or only bridge and love the imbalance. Drop the neck mounted one a little lower. Nobody made positional pickups before the insanity started chasing bullshit.

[Roe]

Try these amp settings:
Presence-10
Bass-9
Mid-3
Treb-5
Bright vol-6

[Reeltarded]

I have a small piece of information. If you take your screw coils out and test both coils..

Unwind the screw coil until it is -5% value of the slug coil.

Don't tell anyone I ever said this.

[Roe]

I have a small piece of information. If you take your screw coils out and test both coils..

Unwind the screw coil until it is -5% value of the slug coil.

Don't tell anyone I ever said this.

Is this for the bridge or neck pickup?

[Reeltarded]

Both. Reel off wire from the screw coil to -5% of slug coil on both pickups.

Imbalance opens them up a way you probably like.

[cdemike]

Been doing some experimenting with my amp and have found a few things:
-I think y'all are right about the bass-spec tone stack after some really close back and forth listening between playing the amp and listening to the Song Remains the Same album. Especially on things like Celebration Day, the sound is nasal but in a different way than how the lead-spec tone stack sounds, but I'm able to dial in something with the right kind of nasal with the bass-spec tone stack.

-Miles alluded to this in the thread that informed the design choices behind my amp, but high B+ seems to be a contributor as well. So I raised the B+ using Martin's suggestion from the B+ lifter thread in the technical section placing dual secondaries from a small toroidal transformer in series with the existing PT. I thought the amp sounded close with the 434V B+ it had before, but it how has 464V at the OT CT, and it sounds closer. I previously compensated for the difference in the preamp B+ with some dropping resistor value changes, but bringing them closer to stock also sounded better following the B+ increase.

-Another piece of the puzzle Miles previously alluded to was the relaxed preamp filtering, and I've gradually reduced the amp down 32uF to 16uF and 20uF at the first and second nodes, respectively. This helped introduce more of the rubbery/chewy transient response I hear on that album.

-Changes making V1 equivalent to bass-spec (fully bypassed 1k5 Rk, 22nF coupling cap) sounded further away from the goal posts to my ears, significantly reducing midrange and taking away honk/quack. Same story with the 680nF cap bypassing the second gain stage's cathode. However, my amp does not have a bright cap(only the 470k+500pF/470K peaking network), and I suspect the 4700pF bright cap would compensate. Checking graphs on some frequency response curves, the corner frequencies seem very similar to me, though the effect from two 680nF caps bypassing the first two stages sounded way too midrange forward in a previous version of the amp that had the same cathode bypass arrangement and a bright cap.

Ck.PNG

bright cap.PNG

-Removing the second gain stage's cathode bypass cap also sounded further to my ears, especially in terms of the "texture" of the overdrive with a severe reduction in the chewy/rubbery transient response.

My takeaways from these observations is that Page's amp may indeed be closer to bass-spec with the 100nF coupling caps to out of the PI and bass-spec tone stack. I do think that there needs to be some kind of highs and midrange boost, but I'm not certain if that would come most convincingly from the the bright cap on the one hand or the combination of the smaller coupling cap and the 680nF bypass caps on the first two gain stages. The difference in the amp's response without the second gain stage bypassed makes me think it's the latter.

B+ also seems to be an important piece, and based on the very wide range of B+ readings documented in late 1968 to early 1969 1959/1992’s, it seems to me that erring higher than lower would get closer. But the squishy filtering also seems important.

Haven't had a chance to experiment with 12AU7 in the phase inverter yet, but still plan to test it after clearing some other projects off the docket.

[Mark]

Interesting observations, keep us posted, I would like to see how it all turns out.

[cdemike]

I've really gone off the rails with this Page Marshall project spending way more time and effort than I thought I ever would playing the same 15 songs over and over, but I think I have this amp the closest it's been so far. A lot of the time and effort went into chasing down things that ultimately ended up being dead ends. For example, I kept revisiting the lead-spec tone stack since it hit the phase inverter harder, which I think a key to the sound – with the LTPI over driving with the right amount of filtering at the node, I was able to get the most vocal transient response and maximize the honk and snarl I hear in the Song Remains the Same. Ultimately, though, I wasn’t able to dial out a harsh upper mid spike around 2khz without losing those same qualities when using the lead-spec tone stack.

I mentioned that changes clean up the phase inverter I lost a lot of snarl and honk and one of the paths I tried was to center bias a 12BZ7 in that position since I don’t have a 12AU7 on hand and cleaning up the PI even with the other tube seemed like it’d be adequate to point in the right direction. So a bunch of load lines and resistor substitutions later, I actually found that with the raised PI and preamp B+ my amp has that the 12BZ7s I have on hand biased near center with the stock Marshall 100k/82k/470R setup: with the PI node getting 390V, I was getting a bias voltage of ~2V, depending on the wall voltage at that particular time. As an aside, one plus with using the 12BZ7 to experiment is that I also had a reasonable sense of how much change there was in headroom given the similarities between it and a 12AX7. In any event, changing the PI to get more headroom took the amp a step further away from the qualities I’m hearing from mid-70s LZ as described previously, so I decided not to try the 12AU7. Experimenting with the phase inverter did help me realize how important it is to dial in the filtering at the PI node for this particular sound, and I’m pretty sure 100uF, at least with my power supply, is right with 64uF being too unclear and with transients hitting too softly.

This would be a good point for me to apologize for flip-flopping on my findings, since the next places I looked were at the first two gain stages and the tone stack and further experimentation has gone against some of what I reported in my last post before this one. One thing I tried was to see whether the earlier suggestion that the 4n7 bright cap would have a similar effect as the cathode bypass caps in terms of overall high-pass filtering as described in that post, and I did find that, while the 680nF bypass caps sounded similar in terms of frequency response, that leaving the second gain stage’s cathode un-bypassed sounded truer to “the sound,” especially in terms of compression and transient response. Correspondingly, changing the first gain stage to a fully-bypassed setup with a 22nF coupling capacitor compensated for the lack of low end resulting from the addition of the 4n7 bright cap. I don’t like the idea of replacing cathode capacitors, so I took the opportunity to A/B test pairing a 1k5 cathode resistor with 20uF, 50uF and 100uF bypass capacitors and then to compare those against a red LED. Consistent with what Merlin and others have written in the past on using forward biased diodes to provide bias voltage, I wasn’t able to hear much difference in this specific application. So with things moving in the right direction, I re-consulted the articles that have come out about the Sund Dragon Super Dragon, it occurred to me that one place Tony Frank might have look to early on while trying to increase Page’s Marshall’s headroom would be the second gain stage’s cathode, especially since doing so not only impacts the headroom of that specific stage but also the bias point of the cathode follower. I found that I was able to get closer to the sound by increasing the second gain stage’s cathode resistor since that helped restore some of the compression and “chewy” transient response that got reduced with the other changes above. Increasing the value past the factory values of 820R/1K sounded progressively closer, and in keeping with the general goals outlined in those articles (I.e., maximizing dynamic range and increasing headroom), I settled on 1K5 to aim for center bias. This setup was pretty squishy, and that change significantly reduced the upper treble response. Adding a 50nF bypass cap brought that back while keeping the responsiveness and headroom.

I additionally did a deep-dive on whether there could be any difference stemming from the use of ceramic tone stack capacitors in Marshalls around the same serial number range as Page's amp (again, relying on the assumption that Page's amp may have had the same substitution as Sakurai's amp inferring from the proximity in serial numbers) and learned there's significant voltage-dependent non-linearity in class 2 ceramic capacitors. I highly doubt Marshall was using class 1 capacitors given that the 10nF disc caps were an ersatz substitution in the first place, and that many other changes were made to circuits building a pattern of Marshall selecting components based on cost savings. With those capacitors blocking nearly 200V and likely not rated more than 500V, it seems likely to me that there would be an audible effect on the amp by switching the tone stack caps to class 2 ceramics. This graph is from a datasheet for a Y5P class 2 ceramic cap and demonstrates the relationship pretty clearly:

y5p bias voltage vs capacitance.PNG

Switching to 4x 10nF Y5P class 2 caps paralleled in lieu of the 2x 22nF polyester caps did indeed make a very significant difference in the sound. That difference also tracked with the projected lower in-circuit capacitance, with significantly more midrange emphasis and making the amp come across as darker. I’m really not trying to be the “mojo components” guy here, but ceramic capacitors’ nonlinearity is well documented in manufacturer literature and was historically associated with the proliferation of mica caps in HF equipment, so I feel there’s a pretty strong case to be made that in this specific instance there’s clearly a significant difference between capacitors nominally rated at the same capacitance.

I’ll try to get some clips recorded this weekend to share more tangible findings, but I’m thinking this is pretty close.

[Roe]

I thought the amp sounded close with the 434V B+ it had before, but it how has 464V at the OT CT, and it sounds closer.

Around 500-520v should work here. a 100w amp is essential to this tone and response.

Anyway, how high are your preamp voltages?

[cdemike]

a 100w amp is essential to this tone and response.

I’m not certain I understand why a 100w amp specifically would be needed for this, unless it’s purely a volume thing – that would make sense if the question were sustain and decay, so it doesn’t seem like that would be specifically the cause. Put another way, shouldn’t a 50w amp behave electrically similarly enough to a 100w amp if the power supply were scaled appropriately for similar power supply sag, if the OT were constructed for similar saturation despite lower power input, and if speaker power handling were proportionally reduced? Obviously there are real-world implementation issues with sourcing or constructing an OT to saturate the same way with the lower power handling. With the breadth of PTs available and with workarounds to fine tune the power supply response, however, I’m not sure this would be specifically something that couldn’t be addressed in a 50w build. With the Super Bass specifically, scaling NFB shouldn’t be too difficult since I understand most from that period connected the NFB tap to the speaker jack. Since Page’s amp has 2x speaker outputs and there seems to be consensus that he was using Celestion speakers during LZ, a 100w amp would need to be running into two cabinets, and thus the NFB tap would connect to the 8 ohm tap. Since output power scales by sqrt(2), the signal magnitude would be roughly 0.707x weaker in a 50w amp, but luckily moving the NFB tap to the 16 ohm tap would increase the NFB signal by sqrt(2). So there wouldn’t need to be any adjustment to the NFB resistor and thus retain an identical frequency response in the presence control network as in a 100w amp while providing the same magnitude of NFB signal.
Is your concern primarily the OT or is there something else I’m not considering?

Around 500-520v should work here.

With regard to the B+, I've actually been aiming for 500-520 since I first ordered my PT. When I installed that auxiliary PT I was also aiming for about a 50V bump based on it being 34V and pretty similar in terms of current rating to the primary PT. However, it’s clear with the auxiliary transformer actually installed that I’d need a stouter B+ booster. Since another work-around solution could begin to approach the cost of a new PT, I've been considering the PT swap to address the root of the issue, but I'm trying to do as much reading as I can before pulling the trigger since I've missed the mark twice now for B+ projections with this amp. Furthermore, I recognize that power supply sag is specifically a critical part of the sound, so in addition to finding a PT that supplies adequate voltage, it also would need to sag the same amount despite the lower current demand from the smaller output section. It seems like there’s not sufficient evidence to say specifically what PT was in Page’s amp during LZ with any degree of certainty, but it is known that the amp currently has a 1203-80, which seems like a reasonable performance benchmark given that it’s also roughly period-correct for what would have been originally in the amp. From what I have researched, I understand 1203-80 should sag about 90V under maximum signal (https://forum.metropoulos.net/viewtopic.php?t=41489), so I think a PT that sags a similar amount with the scaled reduction in power section current demand would be best.

Anyway, how high are your preamp voltages?

B+ voltages and circuit details are as follows:
Reservoir node (32uF, 464V) > Choke > Screens node (32uF, 462V) > 8K2 > PI node (100uF, 390V) > 22k > Preamp node 2 (20uF, 306V) > 22k > Preamp node 1 (16uF, 274V)
I raised the value of the dropping resistor between the PI and second preamp node from the "stock" 10k since I was getting over the 200V heater-cathode voltage maximum at the CF's cathode as the preceding gain stage's cathode resistance crept up over testing. In the most recent round of revisions focused on the second gain stage's cathode, the only voltage I was tracking closely was the heater-to-cathode potential since that was the only "hard limit," and tuning from there would consequently mean raising or lowering only the dropper between the two preamp nodes assuming the CF was kept as close maximum as would be possible with standard value dropping resistors between the PI and second preamp node.
The reasoning behind the 22k dropping resistor between the preamp nodes is that this amp only has a bright channel since it was built from the outset aiming for this sound. I've only ever seen pictures of Page plugged into the bright channel and with half the triodes drawing current from the first preamp node, doubling the dropper seemed appropriate to compensate, which I think bears out in terms of the similarities in voltages at both nodes when referenced against contemporary Marshall schematics and voltage references.

Regarding changes to the preamp power supply , it seems that there's something at least resembling a consensus "hunch" that Frank likely didn't stray too much out of the Marshall ballpark when modifying Page's amp; between the discussion of the Major's operating points for the 12AU7 driver and Frank's relationship to Unicord, that seemed to be at least a reasonable starting place. So I dug through some Marshall schematics to see what the highest preamp voltages I could find were to establish some voltage benchmarks. The commonly-circulated Unicord Marshall voltage chart had the highest overall voltages with 530V at the OT CT, but the preamp voltages were not meaningfully different than a JTM100 Black Flag schematic that lists the B+ rail voltages. I did note that Black Flag had B+ measurements at the reservoir and screens nodes very similar to my amp, but about 70V shy of the highest readings on the Unicord document at the reservoir node. I thought that was reassuring in context of the similar 8K2 resistor between the PI and screens nodes between the Black Flag and my amp.

marshallvoltagechart.gif

JTM100 schematic (black flag).gif

I did some recordings today of the amp in its most recent iteration:
https://soundcloud.com/mike-913267560/s ... immy-plexi

I’m thinking it might be helpful to try the suggestions coming out of this thread and to collect findings here. It’s not my goal to hijack this thread, though, so I’m happy to make a different thread if this is becoming too focused on my amp specifically.


Edit: forgot to mention that I experimented with shorting the dropping resistor between the preamp nodes (basically the same B+ rail topology as a JTM45). I don’t think it sounded closer — too glassy and hard. So while high preamp voltages are a piece of the puzzle, it seems like 290-300v may be the upper limit for the input gain stage node voltage when aiming for this specific sound.

[Roe]

a 100w amp is essential to this tone and response.

I’m not certain I understand why a 100w amp specifically would be needed for this, unless it’s purely a volume thing – that would make sense if the question were sustain and decay, so it doesn’t seem like that would be specifically the cause. Put another way, shouldn’t a 50w amp behave electrically similarly enough to a 100w amp if the power supply were scaled appropriately for similar power supply sag, if the OT were constructed for similar saturation despite lower power input, and if speaker power handling were proportionally reduced? Obviously there are real-world implementation issues with sourcing or constructing an OT to saturate the same way with the lower power handling. With the breadth of PTs available and with workarounds to fine tune the power supply response, however, I’m not sure this would be specifically something that couldn’t be addressed in a 50w build. With the Super Bass specifically, scaling NFB shouldn’t be too difficult since I understand most from that period connected the NFB tap to the speaker jack. Since Page’s amp has 2x speaker outputs and there seems to be consensus that he was using Celestion speakers during LZ, a 100w amp would need to be running into two cabinets, and thus the NFB tap would connect to the 8 ohm tap. Since output power scales by sqrt(2), the signal magnitude would be roughly 0.707x weaker in a 50w amp, but luckily moving the NFB tap to the 16 ohm tap would increase the NFB signal by sqrt(2). So there wouldn’t need to be any adjustment to the NFB resistor and thus retain an identical frequency response in the presence control network as in a 100w amp while providing the same magnitude of NFB signal.
Is your concern primarily the OT or is there something else I’m not considering?

The 100w amps have different transformers, power amps and higher voltages, which affect tone and response/dynamics. It is much easier to get close to the page tone with a 100w amp than a 50w one really. just try to compare a few amps. Tweaking a 50w would help but it will not be the same thing really. and it is not mainly about volume