Believe it or not, a DSL20 schem exists in the cybers. Questions about power scaling.

[Whatizitman]

EDIT: pdf added.

From what I gather, DSL20 does not use triode mode on the EL34s to power scale. Just low (20w), and really low (10w), plate voltages. There are a few mosfets, but it doesn’t look they are tied to the B+ like with a VVR, either. The PT apparently has several taps. You can see on the left of the schem the two separate rectifiers that switch to the output B+, screens, and PI.

A marshallforum thread a while back the OP measured pretty low voltages on the EL34 plates, in full and low modes.

https://www.marshallforum.com/threads/50w-ot-in-a-20w-marshall-dsl.114100/page-3

Some googling brought up some posts here and there of amp builders that have made low voltage EL34 builds, so I guess it’s possible. So my question is about OT primary impedance. Doesn’t that change with voltage? Does it matter enough to compensate somewhere? I can’t decipher anything in the circuit to indicate that.

As a sidenote, is it possible the OT has an extra tap? 4ohm would be really, really nice. Not a good pic. But it’s the green wire.

 

[The Ballzz]

Yes, most pages of the schematics are available, except the one with the transformers and rest of the power supply! I have all the other pages. The DSL40 and DSL100 (latest models) are the same story! If you ever come across that page, I'd be indebted to you for sharing it with me. Please be aware that Marshall seems to have someone (or several someones) scouring the innernest looking for leaks of their proprietary documents and often come down as hard as they can on whatever websites allow them to be posted. The old adage "Big Brother is watching you!" is much more realistic than most want to believe!

And as far as the power reduction not actually being a Pentode/triode arrangement, the sound and dynamics degradation in the lower power modes on the DSL20, 40 & 100 is hauntingly similar to that configuration! A little better, but not by much. Now the system they call "Power Stem Technology" in the ORIGIN 20 & 50 is absolutely stellar, by comparison! On the down side, their are other issues with that series that do not endear them to me.

FWIW, I currently own a DSL20CR, a DSLR 40CR and have owned an ORIGIN 20! The DSL20 is a rockin' little amp, with the both the GAIN and VOLUME on the CLASSIC GAIN channel fully cranked, the ULTRA GAIN channel set to match that volume and tamed with a great attenuator, to required volume levels! The DSL40 is going up for sale!

Just My Experience,
Gene

 

[Dacious]

I had a Mini Silver Jubilee 20 watt and remember reading a tech detailing that the low power switch drops about 1/3rd of the B+ voltage.

It used cathode bias on the EL34s at around 300 volts B+ for full power and 200 volts B+ for 5 watt IIRC.

To me the tone didn't change much but the bottom end got a lot looser.

Fender used lower B+ with fixed bias with the Fender 75 which was 15 watts on low power and the 100/25 watt Evil Twin of the late 80s/90s.

 

[The Ballzz]

I had a Mini Silver Jubilee 20 watt and remember reading a tech detailing that the low power switch drops about 1/3rd of the B+ voltage.

It used cathode bias on the EL34s at around 300 volts B+ for full power and 200 volts B+ for 5 watt IIRC.

To me the tone didn't change much but the bottom end got a lot looser.

Fender used lower B+ with fixed bias with the Fender 75 which was 15 watts on low power and the 100/25 watt Evil Twin of the late 80s/90s.

Thank You Sir, for confirming that suspicion. While I've not been familiar with how well that worked on the Fender models mentioned, I have used enough amps with the Hall VVR and London Power Scaling to discern that dropping the B+ also requires some sort of adjustment of how hard the power tubes are driven to keep the sound more consistent. There's been a lot of discussion as to whether it is best to use a pre PI master volume, post PI master volume or include the PI in the voltage drop. On the Marshalls, not being variable seems to make it easier to put any one of those changes on a switch to keep the balance correct. Still begs the question of what works best, from a tonal and dynamics response perspective? It would be interesting to understand which method Marshall uses on their various multi power level systems. Just for reference and future possible design!

Always Curious!
Gene

 

[Jon Snell]

Marshall Amplification, VOX and many more manufacturers have copyright on their schematics because they pay their staff wages from manufacturing novel amplifiers.
These current service manuals are only available to authorised service departments who sign a guarantee that they will not disclose the service information. If they do, they will no longer have the status as authorised service department.
It is not difficult to work out what goes where as you look at them, a transformer is a transformer after all and has a limited amount of windings. The clever bit is how they are wound alongside the balance of copper to iron.
I am unable to share diagrammes, sorry. As pointed out most are available on the interweb but not all complete.
As for triode connected EL34s, they are normally switched if required.

 

[Whatizitman]

Well, I don't want to put anyone out of work. But if something makes it to the interwebz, AFAIC it's fair game. I'm speaking from a technical standpoint. Not philosophical. I have no dog in that fight.

In any case, someone somewhere can trace their amp. Having a schem available just speeds up the inevitability of people at some point figuring out how to repair and mod their own PCB amps. That's all I'm gonna say on that.

So I should probably clarify. My main question is about impedance load on an amp with power scaling, WRT to the OT and output impedance. I think that applies to any amp with power scaling. Not just modern Marshalls. But I'm having zero luck finding consistent info on this. It's clear to me now that there are plenty of examples of low voltage EL34 amps in the wild. What is not clear to me is how power scaling, when utilized in some form or another, impacts the OT beyond subjective sound preferences.

How does this circuit (or any other similar design) compensate for impedance load changes on the OT? Or does it even need to? I understand the PT and OT are designed for the circuit. But it still has two power levels. Both are low voltage. One is just lower than the other. So regardless of how low the voltage is, the plate load impedance must be different between the two settings. Yet the speaker load doesn't change.

Why do I want to know? Well.... let's just say I might have future plans for my DSL20CR. It's mine, and I've had it long enough now to no longer feel the need to protect it from my tinkering brain and hands. There. I said it.

Before I potentially change and/or ruin my amp for good, I want to better understand the hows and whys of the design choice of PT, OT, and circuit design to utilize EL34 at very low powers, with what I consider decent results. It's a good sounding amp. But good sounding never stopped anything from being butchered eventually.

 

[Whatizitman]

OK maybe I think I got it(?). Since a transformer does not have a "set" load impedance or optimum range, but instead reflects the load on the secondary, am I right in ascertaining that the output tube performance is really what we're concerned with?

If the tube can handle a plate load impedance reflective of low voltage and/or current, all's good?

EDIT: Speaking of EL34s.... Here's a very relevant thread.

https://el34world.com/Forum/index.php?topic=15583.0

EDIT EDIT: OK there it is. Lower voltage AND lower current (because the cathode resistance doesn't change), means load impedance doesn't change. Typical for my undiagnosed ADHD brain, I have to wrestle through all the noise before I can get down to the fundamentals. Makes sense now.

I think we're done here(?).

 

[The Ballzz]

Hey @Whatizitman
Thanks for the link to that discussion! Some very interesting takes on the topic. I'm VERY interested in how Marshall achieved their multi wattage settings on the Origin 20 & 50, As well as how they're doing it on the DSLs. We need that missing schematic page (or a close inspection, from a reverse engineering perspective) to confirm whether either or both of the schemes are using separate taps and/or windings of the tranformer, or dropping the voltage in a different manner.

I've contemplated doing a version of VVR/power scaling, but instead of having it variable, have it on switches for different levels. I think once you remove the variable potentiometer from the system, the sometimes problematic MOS FET may no longer be needed. You see, I don't really care about wasting a tubes capability I'd happily use a quad of KT88s, neutered down to 7 or 8 watts, if it sounded great! I realize, that's a kinda drastic example, but I hope you get my point. It's all about the tone, response and the feel, baby!
Just Ponderin'
Gene

 

[Dacious]

The Mesa Dynapower system used their own version.

The 5 watts setting is one tube. Some of the bigger amps like the early Mark amps it cut in two out of four tubes. There was even one series which had EL34s and 6L6s in the same amp.

The 5/15/25 watt amps click extra capacitance in on the power section which combined with big iron in the transformers gives them the perceived jump in power.

Having had the Express 5:25, Transatlantic 15 and 30 and Mark V:25 it's effective. You notice on the 25 watt setting or 40 watt setting on the TA30 there's a lot more bass, and more headroom on the high power setting.

So if you want breakup at a lower power setting especially on the Vox and Tweed settings it's there, then if you meant more bass and clean grunt in the Bassman and Plexi settings it's there.

 

[Whatizitman]

Hey @Whatizitman
Thanks for the link to that discussion! Some very interesting takes on the topic. I'm VERY interested in how Marshall achieved their multi wattage settings on the Origin 20 & 50, As well as how they're doing it on the DSLs. We need that missing schematic page (or a close inspection, from a reverse engineering perspective) to confirm whether either or both of the schemes are using separate taps and/or windings of the tranformer, or dropping the voltage in a different manner.

I've contemplated doing a version of VVR/power scaling, but instead of having it variable, have it on switches for different levels. I think once you remove the variable potentiometer from the system, the sometimes problematic MOS FET may no longer be needed. You see, I don't really care about wasting a tubes capability I'd happily use a quad of KT88s, neutered down to 7 or 8 watts, if it sounded great! I realize, that's a kinda drastic example, but I hope you get my point. It's all about the tone, response and the feel, baby!
Just Ponderin'
Gene

My very untrained and unreliable eyes seem to confirm (at least to me) that the DSL20 uses two different taps from the PT for voltage 'scaling'. No mosfet or anything variable, as in what I understand the first V in VVR to be. Just a simple switch between the power taps. I also read the schem to indicate that voltage to the preamp is constant, and only the output section gets the switched power. I'm no expert at schems, though, so I could be reading that wrong. In any case, I can't think of any good reason to scale the preamp voltages on a master volume amp. A major question I had was how the output plate impedance load was affected by the voltage change, and in turn what impact that had on the OT, speaker impedance, etc... After going through the Hoffman/EL34 site thread, it occurred to me that load impedance would stay relatively constant since the cathode doesn't "vary", or at least self-adjusts in the case of DSL20, as it's cathode biased. V x I both are able to change. Voila, no need to change speaker loads.

Until someone far more knowledgeable than me comes along and corrects me, this is my story I'm sticking to.

My other main question was about how low voltage EL34 circuit can even work. But apparently it does, and is not uncommon. I'm good with that.

I'll post a full chassis pic. There's so many terminal wire connectors noted on the schem, that it's hard (at least for me) to see how the sections fit together without looking at the physical connections and numbers on the PCB. Y'all can ascertain for yerselfies.

 

[DADGAD]

Thank you for posting the schematic. I have a copy of the DSL1 schematic (found on the Marshall forum) and except for the output power, looks very similar in gain stages and voicing to the DSL20. I Have an early DSL5C that needs the EQ on the green channel cleaned up. It has a problem with too much bass and too much high end. But the red channel sounds great.

 

[Whatizitman]

Thank you for posting the schematic. I have a copy of the DSL1 schematic (found on the Marshall forum) and except for the output power, looks very similar in gain stages and voicing to the DSL20. I Have an early DSL5C that needs the EQ on the green channel cleaned up. It has a problem with too much bass and too much high end. But the red channel sounds great.

High gain channels by design typically have low frequency roll off to keep muddiness down from cascaded gain stages. Typical marshall 2 channel amps have shared EQ, so it makes things a little tough to balance out lows and highs between the channels. It's a common complaint, and one that I share. I rather use pedals for boosts and such on the same channel, than switch channels during a song.

 

[The Ballzz]

@DADGAD
I also have an earlier DSL5C and did some fairly simple mods that worked well, but approached it from the opposite perspective of what you mentioned. You may want to do some of the testing I'm going to suggest.

> First, the CLASSIC green channel is only too bass and top end heavy if you switch to it from a well dialed in ULTRA channel. Conversely, if you crank all the way up, the GAIN and VOLUME of the CLASSIC channel, you can dial in a very nice, full range, almost "Plexi-ish" on the edge of breakup sound, with lots of complex tonal character, just not quite enough "oomph!" Although if you are a fan of boost/overdrive/distortion thingies, you could mostly live on that channel, as @Whatizitman alluded to above!
>Next, if you switch to the ULTRA red channel, with all the same EQ settings, you'll likely find it lacking a bit of bottom and maybe high end. Now if you dial this channel in for best possible sound, and switch back to the CLASSIC channel, you'll experience the exact anomalies you've described.
> Then, looking at the schematic tells me that if we modify the CLASSIC channel, at V1A, it will also change how it drives the ULTRA channel, as the V1A portion of the CLASSIC channel becomes part of the signal chain of the ULTRA channel. Note here that the actual EQ/tone stack is not dedicated to one channel or the other, but is shared by both, so that is not the part to modify, as it comes long after (in the circuit) the channels have been switched. Again, changing the EQ will affect both channels. This led me to suspect the best place to modify is the ULTRA channel at V1B, to make it sound the way we want when the EQ is set to optimize the CLASSIC channel. This channel is also a bit easier and straight forward to modify and actually where the real disparity between the two channels lies!
> And Then, I chose to eliminate some of the components surrounding the grid of V1B, and change the values at it's cathode, all to more closely mimic the treatment of V1B of a JCM800 2203/2204. I also changed a couple components between the plate and switch RL1B, to be closer to 2203/2204 values.
> So Now, I did these few mods all above the board (to avoid removing/replacing the board multiple times just for testing), by clipping component leads and "scabbing" on components, with the "threat" that once I found the best recipe, I'd remove the board once and make the mods more cosmetically pleasing and permanent! I've not yet found it necessary to follow through with that threat, but have been using the amp like this with the mods in place for several years and am fairly pleased with the results! It's been long enough since doing the mods that I don't currently have the specifics to share, as hand drawn "scratch" schematics have been misplaced and time has clouded the memory, but I could likely figure out and find them if you were truly interested!
>So Finally, its likely high time that I finally follow through with that threat, given that I see no real need to ever want to put it back to the original configuration! I just haven't quite decided if it might be advantageous to still tweak a couple component values! There's also the fact that I've determined that while the amp is certainly loud enough for most of my applications, once cranked to get the power amp into the tonal picture, I'm not a fan of the dual triode, push/pull power section, as that's just what it sounds like, a dual pentode amp that has been neutered by switching to triode mode, with all the ensuing lack of dynamics, character, etc! It's still a nice little amp, with a properly functioning effects loop to sit beside my desk to use when my real amp(s) are either at the rehearsal space or in the band truck! I still need an attenuator to use it cranked up for bedroom type levels!​

I will likely finalize and document the mods and be selling it off, to allow someone else the pleasure of being able to brag about having the best sounding DSL5C they've ever plugged into!
Just Sharin'
Gene

 

[DADGAD]

High gain channels by design typically have low frequency roll off to keep muddiness down from cascaded gain stages. Typical marshall 2 channel amps have shared EQ, so it makes things a little tough to balance out lows and highs between the channels. It's a common complaint, and one that I share. I rather use pedals for boosts and such on the same channel, than switch channels during a song.

@DADGAD
I also have an earlier DSL5C and did some fairly simple mods that worked well, but approached it from the opposite perspective of what you mentioned. You may want to do some of the testing I'm going to suggest.

> First, the CLASSIC green channel is only too bass and top end heavy if you switch to it from a well dialed in ULTRA channel. Conversely, if you crank all the way up, the GAIN and VOLUME of the CLASSIC channel, you can dial in a very nice, full range, almost "Plexi-ish" on the edge of breakup sound, with lots of complex tonal character, just not quite enough "oomph!" Although if you are a fan of boost/overdrive/distortion thingies, you could mostly live on that channel, as @Whatizitman alluded to above!
>Next, if you switch to the ULTRA red channel, with all the same EQ settings, you'll likely find it lacking a bit of bottom and maybe high end. Now if you dial this channel in for best possible sound, and switch back to the CLASSIC channel, you'll experience the exact anomalies you've described.
> Then, looking at the schematic tells me that if we modify the CLASSIC channel, at V1A, it will also change how it drives the ULTRA channel, as the V1A portion of the CLASSIC channel becomes part of the signal chain of the ULTRA channel. Note here that the actual EQ/tone stack is not dedicated to one channel or the other, but is shared by both, so that is not the part to modify, as it comes long after (in the circuit) the channels have been switched. Again, changing the EQ will affect both channels. This led me to suspect the best place to modify is the ULTRA channel at V1B, to make it sound the way we want when the EQ is set to optimize the CLASSIC channel. This channel is also a bit easier and straight forward to modify and actually where the real disparity between the two channels lies!
> And Then, I chose to eliminate some of the components surrounding the grid of V1B, and change the values at it's cathode, all to more closely mimic the treatment of V1B of a JCM800 2203/2204. I also changed a couple components between the plate and switch RL1B, to be closer to 2203/2204 values.
> So Now, I did these few mods all above the board (to avoid removing/replacing the board multiple times just for testing), by clipping component leads and "scabbing" on components, with the "threat" that once I found the best recipe, I'd remove the board once and make the mods more cosmetically pleasing and permanent! I've not yet found it necessary to follow through with that threat, but have been using the amp like this with the mods in place for several years and am fairly pleased with the results! It's been long enough since doing the mods that I don't currently have the specifics to share, as hand drawn "scratch" schematics have been misplaced and time has clouded the memory, but I could likely figure out and find them if you were truly interested!
>So Finally, its likely high time that I finally follow through with that threat, given that I see no real need to ever want to put it back to the original configuration! I just haven't quite decided if it might be advantageous to still tweak a couple component values! There's also the fact that I've determined that while the amp is certainly loud enough for most of my applications, once cranked to get the power amp into the tonal picture, I'm not a fan of the dual triode, push/pull power section, as that's just what it sounds like, a dual pentode amp that has been neutered by switching to triode mode, with all the ensuing lack of dynamics, character, etc! It's still a nice little amp, with a properly functioning effects loop to sit beside my desk to use when my real amp(s) are either at the rehearsal space or in the band truck! I still need an attenuator to use it cranked up for bedroom type levels!​

I will likely finalize and document the mods and be selling it off, to allow someone else the pleasure of being able to brag about having the best sounding DSL5C they've ever plugged into!
Just Sharin'
Gene

My point about the DSL "green channel" can be seen on the attached schematic, in the area that is circled. The clean/green channel has EQ emphasis, before the gain pot. That emphasis is bypassed for the red chanel. In my amp, I believe the treble bleed (C30) makes the clean channel too strident and the rest is likely emphasizing too much bass. The red channel in my DSL sounds great.

 

[Whatizitman]

I agree that the green channels sound more plexi to my ears than the red channels, in part because of the more full low end. Try a clean boost into an unmodified green channel, and Mr. Hyde will come out.

The only "mod" I've done so far is to change V2 to 5751. It helps tame the red channel a bit, and takes a little of the ice pick out of the green. Red channel sounds like a modern Marshall. But front loaded green (a boost for extra gain stage) makes the green sound vintage big and crunchy. But it turns the red channel to mush. So boost stays on the green.

But..... Confession time. The reason I started this thread was to get answers to some questions I have before I decide to gut my DSL20 and handwire in an 800 pre section. Until someone tells me otherwise, I'm now going on the assumption that the DSL20 PT and OT will work fine, maintaining the EL34s. Ideally I would like to make without too much chassis altering so it's all reversible. Obviously not plug and play. But able to get back to stock as much as possible. Chassis will be have to be drilled for mounted pre tube sockets, and room made for discrete jacks, etc..., for starters. It will be one channel, and initially use only the full voltage setting. Keep it simple. I'm checking out @robrob's micro 800 board.... I'm starting to think it's possible without having to get too crazy with workarounds.

I should add that I have no good reason to do this. Only that I have the amp, it's starting to gather dust, and it will be far, far cheaper than a full kit or scratch build. And when the call to tinker comes, you have to answer.

 

[The Ballzz]

Anything without an fx loop limits the useful range of time based effects... it ends up with the amp needing to be a pedal platform, which defeats the tone arguments.

Yeah Ive never been an FX loop user because I dont use that many time based effects.

My point about the DSL "green channel" can be seen on the attached schematic, in the area that is circled. The clean/green channel has EQ emphasis, before the gain pot. That emphasis is bypassed for the red chanel. In my amp, I believe the treble bleed (C30) makes the clean channel too strident and the rest is likely emphasizing too much bass. The red channel in my DSL sounds great. View attachment 833515

Well sir,
It will likely help if you work from a DSL5C schematic, instead of the DSL15! I'll PM it to you in pdf

I also apologize for mentioning GAIN and VOLUME on that channel, as I also have a DSL20CR & DSL40CR and was just a bit confused! And I guess its a matter of taste as to whether you want to voice the ULTRA to be more like the CLASSIC or vice/versa?
Just Sayin'
Gene

 

[The Ballzz]

OOPS, Duplicate Post!
Sorry!
Gene

 

[The Ballzz]

@DADGAD
For some reason, I seem unable to start a conversation/PM with you. Maybe you have that function turned off or blocked? I'm not comfortable posting the document publicly!
Let Me Know?
Gene

 

[DADGAD]

@DADGAD
For some reason, I seem unable to start a conversation/PM with you. Maybe you have that function turned off or blocked? I'm not comfortable posting the document publicly!
Let Me Know?
Gene

@The Ballzz Thank you, Gene! I had a conversations box unchecked. Try it now.

 

[Dansz]

My point about the DSL "green channel" can be seen on the attached schematic, in the area that is circled. The clean/green channel has EQ emphasis, before the gain pot. That emphasis is bypassed for the red chanel. In my amp, I believe the treble bleed (C30) makes the clean channel too strident and the rest is likely emphasizing too much bass. The red channel in my DSL sounds great. View attachment 833515

I also think the stock red channel in my DSL5C sounds great, and the stock green classic channel is too bright and boomy. I took the approach of modifying the premphasis circuit of the green channel to make it sound more like a fender tweed bassman and balance well with the red channel. Here's what I did...