Ideas for amp rebuild... - diyAudio
Go Back   Home > Forums > Amplifiers > Tubes / Valves

Tubes / Valves All about our sweet vacuum tubes :) Threads about Musical Instrument Amps of all kinds should be in the Instruments & Amps forum

Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Reply
 
Thread Tools Search this Thread
Old 22nd May 2004, 10:28 AM   #1
hifiZen is offline hifiZen  Canada
diyAudio Member
 
hifiZen's Avatar
 
Join Date: May 2001
Location: Mountain View, CA
Question Ideas for amp rebuild...

I knew this day would come eventually... my beloved tube amp which I built back in '96 is really sounding tired now, and seems to break down entirely too often. So, I think it's time to breathe some new life into this old friend. Unfortunately, though I've learned a lot about electronics since 1996 (I got my EE degree in the intervening years), it's been a very long time since I've looked at tube circuits. So, I'd like to get some fresh ideas and insight I can apply to the rebuild...


Here's what I'm starting with:

- fairly standard Mullard push-pull type circuit
- input tube is a 12AT7, 1/2 per channel
- phase splitters are 12AU7, 1 per channel
- output tube is E34L in ultralinear or triode connection (switchable), fixed bias, 2 per channel
- 6600 ohm plate-plate output trannies, 60W (Hammond 1650P)
- Power trannie is from the Assemblage ST40... 760VCT@350mA w/ 65V tap for bias supply and 6.3V filament winding
- solid state rectifiers and zener / mosfet regulators


I don't want to waste what I've already got, so that means keeping the existing transformers, and probably the output tubes as well, though I'm not stuck on the E34L. I'd like to stay as close to 50W as possible... I was thinking either UL class AB1 or perhaps triode class AB2, but I don't know much about running the EL34/6CA7 in triode class AB2... Perhaps I should use a different tube entirely? Maybe even a real triode like the sovtek 6B4-G? I'm sure there are lots of possibiities I've overlooked.

From my recent reading, I'm intrigued with the possibility of adding an interstage trannie and running the input tube (say, an ECC99 or 5687) with a solid state CCS, Pimm-style. That would give me a nice linear input and takes care of phase splitting, but kinda rules out global feedback. Of course, with class AB I think I'd definitely want to have some feedback around the output - perhaps something like in Norman Koren's "Local Hero" amplifier. Comments?

Anyway, I haven't made up my mind on any of this yet, and I have a feeling I've got a lot more catching up to do with all the new tube designs out there, so I'm open to any suggestions - different circuit types / phase splitters, alternate output tubes / class / connections, etc...

Also, anyone with relevant LTSPICE designs they want to throw my way are more than welcome. I'm going to try and simulate a few options before I build.

Thanks in advance, folks... I'm hoping this will result in a major improvement, and really looking forward to it! I'm also really happy to be back in vacuum-land. It seems so much has happened since I last paid a visit.
__________________
- Chad.
  Reply With Quote
Old 22nd May 2004, 12:49 PM   #2
SY is offline SY  United States
diyAudio Moderator
 
SY's Avatar
 
Join Date: Oct 2002
Location: Chicagoland
Blog Entries: 1
You've got a good, solid amp to start. My own inclination would be to evolve it, not do anything too radical. Were it MY amp, I would (in order of importance):

1. Upgrade the regulation. If the zener/mosfet regs you mention are simple source followers, I'd be looking at something a bit better. Of all the regulator schemes I've tried, Joe Curcio's has been the best and most reliable.

2. Better output tubes. There are some "super" NOS 6BG6s floating around at very attractive prices. They appear to be identical with the classic 7027s, including the 30W plate dissipation rating. And they work GREAT with 6600 ohm transformers.

3. Input stage. You can do very well with the Williamson topology- there's a reason it's a classic- but you'll do even better if you use better tubes. In the octal world, there's 6SL7s and 6SN7s of various flavors.


PS- re your avatar, "Is this a trick question?"
__________________
You might be screaming "No, no, no" and all they hear is "Who wants cake?" Let me tell you something: They all do. They all want cake.- Wilford Brimley
  Reply With Quote
Old 22nd May 2004, 01:07 PM   #3
diyAudio Senior Member
 
fdegrove's Avatar
 
Join Date: Aug 2002
Location: Belgium
Hi,

Quote:
2. Better output tubes. There are some "super" NOS 6BG6s floating around at very attractive prices. They appear to be identical with the classic 7027s, including the 30W plate dissipation rating. And they work GREAT with 6600 ohm transformers.
Great tube indeed, just keep in mind they need a topcap for the plate supply.

Cheers,
__________________
Frank
  Reply With Quote
Old 22nd May 2004, 08:22 PM   #4
diyAudio Member
 
Join Date: Apr 2004
Location: Monroe Township, NJ
On the 6BG6 front, adapters can be either fabricated or purchased that allow both the "super" 6BG6 and the 6L6GC to work in the amp. The current production SED (Svetlana) 6L6GC has a good reputation. BTW, SND Tube Sales (www.vacuumtubes.com) has stock of the 6BG6 and the adapters.

A lesson taught to me by Jim McShane, Mr. H/K Cit. 2, is that if the B+ is regulated, then the bias supply should be regulated too.

If you CCS load the voltage gain block, you don't need global NFB, as CCS loaded triodes are VERY linear. An ECC99 wired as a true Schmidt (differential) phase splitter would give you a very convenient means of applying NFB. Instead of grounding the non-inverting triode's grid, you use it as the point to which NFB is applied. Put a CCS in the cathode circuit of the splitter. A N Channel enhancement mode MOSFET is ideal for the CCS job. A "final" touch is to connect the "arms" of a 100 Ohm pot. to the 2 cathode resistors. The wiper of the pot. gets connected to the CCS.
__________________
Eli D.
  Reply With Quote
Old 23rd May 2004, 12:21 PM   #5
hifiZen is offline hifiZen  Canada
diyAudio Member
 
hifiZen's Avatar
 
Join Date: May 2001
Location: Mountain View, CA
Excellent! These are just the kind of suggestions I'm looking for. The 6BG6 looks like a very good choice indeed. Thanks SY, and thanks Eli for pointing me to the source. I know very little about the 6L6 types... guess I've got some reading to do. What do you recommend for these: ultralinear, pentode, or triode mode (I know this is sometimes a matter of personal taste)?

After looking at Lundahl prices, I've decided to shelve the interstage transformer idea. heh. A Schmidt phase splitter could be just the ticket though - very similar to the LTP I've got now, and should perform very well with a CCS and a negative rail. I also like the possibility of bringing local feedback from the output to this point. I think that's at the top of my list of options right now.

The voltage regulators are indeed just simple source followers. The Curcio regulators look awfully good, if a little complex for a tube amp. I'm wondering if simple gas tube shunt regs would do the trick if the circuit isn't too demanding. I like the look of those things.

Gee, I really like where this project is headed...

Quote:
Originally posted by SY
"Is this a trick question?"
Hehe. I see you recognize which Calvin strip the picture is from. I think it has to be one of my absolute favorites, along with the snowman series. One of the oldest pictures of me captures my first experience with tools... holding a hammer and pounding away at a board with some nails in it. I must have been been barely able to walk at that age, but the experience seems to have set the tone for the remainder of my life.
__________________
- Chad.
  Reply With Quote
Old 23rd May 2004, 01:33 PM   #6
SY is offline SY  United States
diyAudio Moderator
 
SY's Avatar
 
Join Date: Oct 2002
Location: Chicagoland
Blog Entries: 1
Quote:
What do you recommend for these: ultralinear, pentode, or triode mode (I know this is sometimes a matter of personal taste)?
If one mode were clearly better than the others, all amps would be made that way. Think of your amp as a test bed so that you can find out for yourself what the tradeoffs really mean in the real world.
__________________
You might be screaming "No, no, no" and all they hear is "Who wants cake?" Let me tell you something: They all do. They all want cake.- Wilford Brimley
  Reply With Quote
Old 25th May 2004, 09:50 PM   #7
diyAudio Member
 
Majestic's Avatar
 
Join Date: Mar 2004
Location: Melbourne
Default Mullard 5-20

hmm..., I have a pair of Mullard-circuit based 20watt monoblocks on the bench right now too! So I'll be watching this thread with interest.


I think my circuit is known as the Mullard 5-20. It uses a 12Ax7 driver/phase splitter feeding a pair of P-P EL34's. Interestingly, it uses a 6BR7 as the input pentode instead of the more common EF86.

Morgan Jones in his analysis of this type of circuit (in Valve Amplifiers , pp. 222-227 of the first edition) seems to feel that it warrants an auto-bias circuit for the output tubes. I'm not sure I want to go that far yet (...nor have I had any expereience with how this mod might sound)

The only other suggested mods for this circuit that I've seen around are from Uncle Ned, who offers some sensible comments about the Dynaco modified Mullard circuit.





(P.S. IMHO the Time-Travel Calvin and Hobes series just cant be beat! :-)
{There's about 30 strips in the whole series}
__________________
check out my avitar and go easy on me ok? :-)
  Reply With Quote
Old 27th May 2004, 11:11 AM   #8
hifiZen is offline hifiZen  Canada
diyAudio Member
 
hifiZen's Avatar
 
Join Date: May 2001
Location: Mountain View, CA
Default Re: Mullard 5-20

Quote:
Originally posted by SY


If one mode were clearly better than the others, all amps would be made that way. Think of your amp as a test bed so that you can find out for yourself what the tradeoffs really mean in the real world.
Fair enough. I suppose what I really meant to ask was what mode you would choose, given a 6600ohm OPT... I've already pretty much been through this with the EL34, where my personal preference is for ultralinear mode over pentode. Triode mode sounds very nice, though the top end is a little rolled off, and I have wondered if this was simply due to increased miller capacitance seen by the driver tubes. I'll find out soon, as I'm almost finished my SPICE model. At least in theory, the lower plate resistance in triode mode should improve the OPT bandwidth, yes?

In any case, I will probably retain the mode switch, so I can 'taste-test' and run different modes to suit the situation.


Quote:
Originally posted by Majestic
hmm..., I have a pair of Mullard-circuit based 20watt monoblocks on the bench right now too! So I'll be watching this thread with interest.
Thanks for the links.

The Mullard is a pretty common circuit. I found a schematic somewhere of a version which had an all-triode front end... the input triode was a 6AQ8/ECC85, which I replaced with the 12AT7. With the my current tubes, my circuit has much less open-loop gain than your version with pentode + 12AX7 LTP must have. In fact, I can even run my version with feedback disconnected, and it sounds just fine. However, it really sings with a moderate level of feedback. The bass really firms up a lot, and it just sounds 'cleaner'. Too much feedback, and it loses that special tone, and starts to sound hard... too much like a transistor amp. I suspect many of the stock Mullard circuits could benefit from a reduction in open-loop gain, such as provided by replacing the input pentode with a triode... and in your case, maybe converting from an 'AX7 phase splitter to a medium-mu tube as well.

Speaking of feedback, I read an interesting article in tube cad journal about a variable feedback scheme - basically employing a shelving filter to increase the open loop gain at low frequencies... this seems appealing, and I might try it out, depending on how it looks in SPICE.

Auto-bias is worth investigating, at least from the standpoint of getting the most out of the OPTs, or maybe if you don't fancy making periodic adjustments to your amplifier. I haven't gotten far enough with my re-design to start thinking about these sort of tweaks too much. Still, I'd be interested in seeing what other auto-bias circuits people have built, and what their sonic effects have been. I'm sure it's a topic which has bounced around here before. Another 'tweak' in the back of my mind is changing to regulated DC filaments. Unfortunately, my 6.3V winding doesn't lend itself well to rectification and regulation at these current levels. It'll take some good Schottky diodes and very low dropout regulators to get that working... probably a choke as well. the 6BG6 turns out to be an asset here, drawing only 0.9A compared to the EL34's 1.5A. I think that'll be last on the list, and I'll just try to leave space for this stuff inside the chassis.

If all that weren't enough, I'm still churning over various voltage regulation schemes in my head... whether to go with a stiff feedback regulator scheme such as the Curcio type suggested by SY, or to go with something simpler like a gas tube shunt. ... And there's the question of whether or not to regulate the output B+ (along with the negative rail - thanks Eli for pointing that out). Mainly, I want to avoid that hard 'transistor' sound, and am apprehensive that overdone regulation could be a contributing factor, particularly where the AC currents don't cancel or balance in-circuit. Anyway, a lot more reading and thinking lies ahead.

Quote:
(P.S. IMHO the Time-Travel Calvin and Hobes series just cant be beat! :-)
{There's about 30 strips in the whole series}
Ah, yes. The time travel series. Good stuff. And then there's always Spaceman Spiff. I never got tired of reading those ones. In fact I really wish there had been more. Hmm... I'm suddenly drawn toward my bookshelf...
__________________
- Chad.
  Reply With Quote
Old 27th May 2004, 11:31 AM   #9
SY is offline SY  United States
diyAudio Moderator
 
SY's Avatar
 
Join Date: Oct 2002
Location: Chicagoland
Blog Entries: 1
Well, depending on how they're run in a given setup, the source impedance seen by the OPT's primary can be lower in UL than in triode. Indeed, you surmise correctly, triode outputs can suffer from Miller effect, though that's minimized with low mu output tubes and/or (better yet) low source Z from the driver.

Which output topology I'd choose depends on the transformer, the tubes, the available voltages, desired feedback schemes, and what load I want to drive. It's not as clear-cut as some would like, where "Oh, yah, triodes are best!" or "Ultralinear isn't musical!" sorts of shibboleths overrule engineering analysis. That's why I think it's worthwhile for you to build the amp with the capabilities of changing the OP stage configuration- it becomes a useful test bed for you to get a more intuitive feel for the trade-offs involved in choosing output stage topologies.

Hmmm, 811s in a Circlotron might work well into 6600 ohms... (just kidding)

Autobias is very, very difficult in anything other than a class A amp.
__________________
You might be screaming "No, no, no" and all they hear is "Who wants cake?" Let me tell you something: They all do. They all want cake.- Wilford Brimley
  Reply With Quote
Old 27th May 2004, 02:48 PM   #10
diyAudio Member
 
Majestic's Avatar
 
Join Date: Mar 2004
Location: Melbourne
Default Re: Re: Mullard 5-20 style mods

Quote:
Originally posted by hifiZen

...In fact, I can even run my version with feedback disconnected, and it sounds just fine. However, it really sings with a moderate level of feedback...
Ahh yes, that was one thing that I was looking forward to doing was decreasing the amount of feedback that my own amps currently have. Anecdotally, Iíve heard that for those of us who generally donít like feedback; that a humble 5dB or so (exact figure will vary upon individual circuit) is a good compromise between musicality vs. Ďaccuracyí. Not wishing to invalidate SY's sensible comments, I naturally note that YMMV.

The variable-frequency feedback scheme you describe sounds very interesting to me. I think I've actually passed over it myself in my own 'net travels, but Iíll try and look it up again.

I'm just getting started in my own understanding of SPICE modeling. I really need a beginner's book on the subject, preferably with an 'audio' slant to it. Still, I'll keep plodding along...
Norman Koren has some good stuff about tube-amp SPICE-modelling if you're yet to run across it yourself.


Biasing
With the subject of fixed bias vs. cathode bias, many claim that despite fixed's engineering superiority, that cathode bias actually sounds better. There is a legitamate reason for this that purportedly lies in the nature of the amp's overload/overshoot characteristics; with cathode bias having a much speedier recovery. I can't find a link to the article right now, but I'll repost shortly if I do.

The fact that so much power gets waisted kinda still "irks" me, however. Which is why I'd like to hear some reliable opinions from other DIY-erís whoíve had experience with the sound of auto-biasing circuits . In the absence of any, it looks like I might just have to try and bite the bullet myself! (...when I can afford the parts )


To regulate or not
Building a good sounding one does appear to be no-trivial task. My own experience is not enough to be of help, but Thorsten Loesch claims that his own tube projects "...started singing when (he) stopped using active regulation and instead built meaty enough passive Supplies..."
Iím sure good sounding ones can still be made. Joe Curcio's is probably one, and by all accounts Allen Wright's is another. Just to name two.
It sounds one of these might be the way to go if you already have regulation present.


Circlotrons
Kidding aside, I've always been attracted to the symmetry of a good circlotron design schematic. It just seems like a sensible 'fit' with PP (PPP) or balanced topologies. However, as with most things audio, none of this is any guarantee of one sounding good.... (or more properly rephrased: sounding better than a well-designed conventional topology). I'm still yet to hear one myself.




(PS You're right. Spaceman Spiff is also pretty hard to beat - damn all his stuff was just so good)
__________________
check out my avitar and go easy on me ok? :-)
  Reply With Quote

Reply


Hide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
KEF - rebuild into OB ? poynton Multi-Way 1 16th August 2007 08:09 AM
RE XXX...Rebuild? DKAudio Subwoofers 0 16th January 2006 03:01 PM
MC-30 rebuild, any ideas? markp Tubes / Valves 18 18th November 2004 03:36 PM
rebuild offroadbum Multi-Way 2 25th February 2004 04:11 AM
ideas for holiday rebuild of APT 1 power amp? mlloyd1 Solid State 6 29th November 2001 01:28 PM


New To Site? Need Help?

All times are GMT. The time now is 01:34 PM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright ©1999-2014 diyAudio

Content Relevant URLs by vBSEO 3.3.2