Xits X10 = liverpool half power?

[Colossal]

OK, here is a draft schematic based on everyone's analysis and observations. I have made every effort to get it right but there could be some mistakes. Anywhere where a part value is unknown, there is a ? above that part.

If anyone sees things that need correcting, please post them here and I'll update the drawing until we're all convinced that nothing more can be done with the information and photos we have and we'll call it done.

<u>Items that need sorting out:</u>
1) Power supply dropping resistors <b>(still outstanding)</b>
2) Grounding scheme <b>(resolved)</b>
3) Screen resistors on the power tubes <b>(pretty much resolved)</b>
4) Cathode bias R and bypass C on the power tubes <b>(pretty much resolved)</b>
5) Size of dropping resistors in front of rectifier diodes <b>(still outstanding)</b>
6) Size of sag resistor just after rectifier <b>(resolved)</b>
7) Is the distribution of B+ correct as drawn (i.e. B+1, B+2, B+3, etc) <b>(resolved)</b>
8 ) Is the pair of 220k/220k grid leaks grounded? <b>(pretty much resolved?)</b>
9) Anything else?

<u>[EDIT: REVISED DRAFT 7/22/10]</u>
1) Updated screen resistors to 100R.
2) Changed V1 coupling cap from 0.022uF to 0.01uF.
3) Moved sag resistor from before standby switch to between B+1 and B+2 caps, redefined as "choke" resistor. Assigned value as 1k 10W based on similarities in X10 to Trainwreck Express and Liverpool designs.
5) Added power tube cathode bias values of 100R/250uF based on similar Vox designs and apparent reduction in power tube drive from small value grid leaks (approximately 110k/100k with MV). Assumed that 220k/220k pair are grounded.
6) Comments added in red explain rationale behind assigment of guessed values.

<b><u>[EDIT: REVISED DRAFT2/20/2011]</b></u>
Attachment removed. Please see pg. 6 for revised schematic

[Colossal]

So im sorry for keeping on about the MV, Martin i got your reply when i posted the question on metros site. Cheers dude.
With the 220K's grounded, we end up with them in parallel to the pots and saftey R's in the MV.
Now if a suggested value of 250k with 2M2 were used, we end up with about 110K total resistance. Sooo, is that value ok as a grid leak? 220k or 470k is what we see on some other 18 watt builds.
What happens with the lower resistance here?
Even using higher value pots a saftey R's, we will never get close to the 220k values, with 1M pots and 10M R's would still only get to 175k or so.
Just need to know what effect that has on operation.
Cheers doods!

Azatplayer,
I have been thinking about this long and hard, trying to think about what the builder might have been thinking.

If the 220k/220k grid leaks are NOT grounded so that their combined, fixed value (440k) is intended to work in parallel with the MV, why not just put a single 440k resistor in that position, parallel to the cut cap? This was mentioned before but the choice could have been aesthetic; just to give the PI section of the board that "classic" look. But considering the meticulous layout, everything is carefully planned and attention is paid to economy and practicality so my guess is that the pair <i>is</i> grounded.

There is an AX84 design for a 2xEL84 amplifier which uses 100k/100k grid leaks. This would certainly limit how hard the power section is being driven so if the MV actually is 250kA || 2M2 then the combined grid leaks would be fairly low at 110k as you mentioned. That is certainly reasonable. I think we need consider how much gain is being developed in the preamp and PI to see what might make sense for the power section so that it doesn't end up a fizzy mess. We don't know the voltages either.

The transformers are custom wound Mercurys. Hard to make any inference about them. The classic 18W Marshall sound depends on 8k OT primary, anything less and it's not quite the same. This amp has some Voxy and Liverpool DNA as previously mentioned.

[martin manning]

So im sorry for keeping on about the MV, Martin i got your reply when i posted the question on metros site. Cheers dude.
With the 220K's grounded, we end up with them in parallel to the pots and saftey R's in the MV.
Now if a suggested value of 250k with 2M2 were used, we end up with about 110K total resistance. Sooo, is that value ok as a grid leak? 220k or 470k is what we see on some other 18 watt builds.
What happens with the lower resistance here?
Even using higher value pots a saftey R's, we will never get close to the 220k values, with 1M pots and 10M R's would still only get to 175k or so.
Just need to know what effect that has on operation.
Cheers doods!

When the MV is turned down, the grid leak resistance seen by the power tubes approaches zero... that's not a problem. As seen from the other side, the 111k equivalent resistance will load down the PI, but maybe that sounds good? The PI coupling caps have to be chosen accordingly, and 0.047u looks like a pretty good bet, I get a roll-off of 23 Hz with that combination, figuring in the output impedance of the PI.

[Darkbluemurder]

A total loading of 110k on the EL84 pair would definitely not be out of line given the fact that EL84 are very input sensitive. A design advantage would be that the distortion would primarily be created in the PI and be well controllable by the PPIMV.

[Darkbluemurder]

Colossal,

Thank you very much for your draft and effort, it is really appreciated.

I agree to the signal path of your draft. Since it is a Vox-derived design I guess that there is a 100R/250uf combination on the power tube cathodes and 100R/5W screen resistors. Given the fact that the grid drive to the power tubes has been reduced that would be a safer environment for the EL84 compared to amps of similar heritage.

On the power supply it seems strange that there is neither a choke nor a resistor between B+1 and B+2. The amp does not have a choke for sure.
Could it be that the "sag resistor" is really the resistor between B+1 and B+2?

Cheers Stephan

[M Fowler]

What about that switch on the back panel between the fuse holder and speaker jack?

[Darkbluemurder]

What about that switch on the back panel between the fuse holder and speaker jack?

Impedance selector?

[M Fowler]

I had first thought of that but can't following the wiring from the pictures we have.

[rooster]

I will add that on V1, the cathode caps are probably switchable. It seems likely that either the .1uf cap or the 22uf cap would be used - and not at the same time. Any indication/reference to a switch here?

Also the screen resistors have to be 100 ohms. It doesn't make sense to reinvent the wheel that is working so well.

As to the PPIMV, I will say that the video tells me it is working very well. But I am having a hard time with the 220K resistors. Maybe here the wheel has been reinvented, or possibly they simply remain from the original layout that didn't use a MV, and are no longer in the circuit. Dunno. The LarMar doesn't need these resistors and I think this (the LarMar) is what I am hearing in the video clip, having installed about 8 of them in various amps now, including Express's, Rockets, and reisue DRs. YMMV.

Lastly, at the PI, do you guys really think that is a .01 to ground? That's pretty light I think but, well if you are seeing this on the Sozo cap then it must be. For that matter, are you guys sure about that .022 exit cap off V1A? This seems a little fat considering the amp and the sound bites. Remember that the AC30 uses a 500pf cap here.

But all this aside, this is an interesting study, and, as the original poster stated, the gtr player on the clip is a great demo guy playing what sounds like a great amp.

[Colossal]

Colossal,

Thank you very much for your draft and effort, it is really appreciated.

Sure, glad it is of use!

I agree to the signal path of your draft. Since it is a Vox-derived design I guess that there is a 100R/250uf combination on the power tube cathodes and 100R/5W screen resistors. Given the fact that the grid drive to the power tubes has been reduced that would be a safer environment for the EL84 compared to amps of similar heritage.

Excellent rationale, I will make the change to the draft adding your comments at that location.

On the power supply it seems strange that there is neither a choke nor a resistor between B+1 and B+2. The amp does not have a choke for sure. Could it be that the "sag resistor" is really the resistor between B+1 and B+2?

I was wondering about that too and it does not make sense that there is no resistor between B+1 and B+2. I never thought of it, but that is a *very* good guess that the big black sag resistor could actually be bridging the two 30uF caps (and it makes technical sense). My guess would be 1k or 2k2 10W since we can see screen resistors on output tube socket pins. Any guesses on the values for the screen resistors?

Well done Stephan

[fishy]

This is all very cool stuff.
The 22uf and the .1uf in parallel on V1 is intriguing. Could that be a similar application as used in power supplies providing a better path for high frquency signals? Maybe has a tonal benefit. Haven't tried anything like that so I don't know. It does seem a little different and that's the only reason I can think of unless it is switchable.

Anyone have another reason why it would be that way and not as Rooster suggested?

[Colossal]

I will add that on V1, the cathode caps are probably switchable. It seems likely that either the .1uf cap or the 22uf cap would be used - and not at the same time. Any indication/reference to a switch here?

There is no other switch on the amp than the bright switch. BUT! It's hard to make out a silver mica cap on the switch from the gut shot, but I'm pretty sure it's there (and guessing about 100pF). About that 0.1 bypass cap: you may be very well right and it could be part of the "Sparkle" (bright) switch as it is grounded separately than the cathode R and C for V1. So maybe when you switch in the 100pF bright cap, it is also grounding that 0.1uF cap. You can clearly see a black ground lead that is separate from the ground for V1's cathode R/C. Paralleling a film cap (if that is what is being down) with the 22uF should add a perceived addition of clarity and help punch up that 22uF. So very cool Rooster, perhaps that 0.1uF is part of the Sparkle switch. Hard to tell though from the photo. Maybe others can weigh in on that but that's an interesting idea.

Also the screen resistors have to be 100 ohms. It doesn't make sense to reinvent the wheel that is working so well.

My guess would also be 100-220R.

As to the PPIMV, I will say that the video tells me it is working very well. But I am having a hard time with the 220K resistors. Maybe here the wheel has been reinvented, or possibly they simply remain from the original layout that didn't use a MV, and are no longer in the circuit. Dunno. The LarMar doesn't need these resistors and I think this (the LarMar) is what I am hearing in the video clip, having installed about 8 of them in various amps now, including Express's, Rockets, and reisue DRs. YMMV.

I'm really thinking that 220k/220k pair is grounded. Otherwise there is 440k fixed in parallel with the LarMar MV (not that that's bad or anything...). Why not just use a fixed resistor if they are not grounded? Just speculating though...

Lastly, at the PI, do you guys really think that is a .01 to ground? That's pretty light I think but, well if you are seeing this on the Sozo cap then it must be.

Definitely. There is a clear shot of it in one of the other photos. 103k Sozo.

For that matter, are you guys sure about that .022 exit cap off V1A? This seems a little fat considering the amp and the sound bites. Remember that the AC30 uses a 500pf cap here.

No, not sure about that, just a guess. And, in looking at it right now, I'm inclined to believe that's 0.01uF too. Good call about it being too fat. That stage is fully bypassed so might get a bit boomy with 0.022uF. I think you are right...should be 0.01uF.

But all this aside, this is an interesting study, and, as the original poster stated, the gtr player on the clip is a great demo guy playing what sounds like a great amp.

Yeah, Greg V is a hot player. If he owns all the great amps he's demoing...well, he's got some killer tones!

[Darkbluemurder]

Paralleling a film cap (if that is what is being down) with the 22uF should add a perceived addition of clarity and help punch up that 22uF. So very cool Rooster, perhaps that 0.1uF is part of the Sparkle switch. Hard to tell though from the photo. Maybe others can weigh in on that but that's an interesting idea.

For that matter, are you guys sure about that .022 exit cap off V1A? This seems a little fat considering the amp and the sound bites. Remember that the AC30 uses a 500pf cap here.

No, not sure about that, just a guess. And, in looking at it right now, I'm inclined to believe that's 0.01uF too. Good call about it being too fat. That stage is fully bypassed so might get a bit boomy with 0.022uF. I think you are right...should be 0.01uF.

Ad 1 - Paralleling the e-cap with a film cap should add some clarity in the highs. I did that on my Club Deluxe as well and left it in even though it did not make that much of a difference to me. I doubt that it is part of the sparkle switch.
Ad 2 - first coupling cap: it is definitely the same size as the PI input and grounding caps - so it must be a 0.01uf.

I think we're getting very close - even if we don't have it 100% we will have a recipe for a very good sounding amp!

Cheers Stephan

[Darkbluemurder]

My guess would be 1k or 2k2 10W since we can see screen resistors on output tube socket pins.

If I look at what's been done in other amps, especially the Trainwreck Liverpool, 1k/10W would be logical so we should take that until somebody proves different.

Cheers Stephan

[Colossal]

This is all very cool stuff.
The 22uf and the .1uf in parallel on V1 is intriguing. Could that be a similar application as used in power supplies providing a better path for high frquency signals? Maybe has a tonal benefit. Haven't tried anything like that so I don't know. It does seem a little different and that's the only reason I can think of unless it is switchable.

Fishy, the concept of paralleling film caps with electrolytics to improve their response or shunt high frequency hash to ground can get you kicked in the nuts on some forums. I personally have done it and I can't say for sure if it helps in power supplies, but I will say that the amps I've done it too, they were very quiet with no PS buzz or hum. Very quiet. However some will argue vehemently that to parallel a film cap at say 0.1 to 1% of the electrolytic's value will invite parasitic ringing and inductance. It's hotly debated in hi-fi circles and I still have yet to see a definitive answer. That said, I have seen amps with all film power supplies (using Solen fast caps..Carr Amplifiers does this) and the guys on AX84 have done this and the consensus is that film caps (usually polypropylene in the large uF sizes needed for PS caps) provide faster response, clearer and punchier low end, and the general perception of more width. I have researched this exhaustively and this is what I've read. Some dudes will get all sour on you and say no, no film caps are no better than an electrolytic. Capacitance is capacitance, but one must take the dielectric into account as well. On precision alone, a film cap usually has much tighter tolerance. A film cap will typically have a precision of +/-5% of the value and a comparable electrolytic (we're talking about high uF values) can be -20%/+80%. I for one don't mind the huge size nor the higher cost of the giant film caps. I plan to use an all-film PS in some upcoming builds to hear for myself. I personally have noted better, tighter low end response when bypassing with film caps on preamp cathodes. I've used WIMA polyester caps (22uF) and notice the difference. And I don't feel this is the result of bias from "expecting" to hear better low end. I have no physical evidence to support this though, only what my ears tell me. Also, Jackie Treehorn just built a bitchin' Rocket with all film caps in the power supply. We're waiting for clips on that, but I'll bet it sounds killer.

Anyone have another reason why it would be that way and not as Rooster suggested?

I just responded to Rooster on this, but my guess would be it is either grounded separately or might be tied into the Sparkle (bright) switch which seems plausible.