• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

Kofi Annan in: "Tube Amp for Multi-Way Speakers"

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hey! It's been a while since I've posted on the forums, but I'm looking to get back into the DIYAudio world now that my kid is 12 and has adopted a policy of aggressive indifference toward me! Hooray for vindictive apathy!

Anyway, a friend of mine wants to build a tube power amp that can push some reasonably efficient (maybe 93dB?) speakers. He's still working on the speaker selections (he's interested in some classic Altec Lansing designs-- think Voice of the Theater, but updated) and wants me to put the amp together.

Given that he'll probably need to push some wattage to achieve the desired SPL , I was thinking about either a push-pull EL34 or KT88 design, but it's been a while and wanted to get some feedback from the community.

I have looked through the EL84 Baby Huey thread and guessing that this design would work well, but wanted to see if anyone could recommend something that they built or feel would be a good fit for some multi-way speakers.

I'd be glad to go outside the EL84 / KT88 topologies as well. NO GREAT IDEA WILL BE REFUSED!

Thanks in advance for any advice you can offer.

As always,
Kofi
 
Apply Paul Joppa's 102 dB. rule to a 93 dB. sensitivity and you get 8 WPC. The 30 or so WPC you get from PP 5881 equivalent Russian 6Π3C-E (6p3s-e) tubes will be plenty, even with losses in the crossover.

Depending on O/P "iron" quality, either Mullard or Williamson style small signal circuitry will be just fine.
 
I mean if he is interested in altec/western electric stuff, why not build a western electric amplifier? A 124 or a 142 would be fine. Heck if you want to shoot for the moon, you could build a 143.

If you aren't stuck on building a PP amplifier, a single ended 300b amplifier pairs very nicely with vintage horn speakers.
 
Hey! Thanks for all the replies!

Apply Paul Joppa's 102 dB. rule to a 93 dB. sensitivity and you get 8 WPC. The 30 or so WPC you get from PP 5881 equivalent Russian 6Π3C-E (6p3s-e) tubes will be plenty, even with losses in the crossover.

Aha! Thanks for the insight, Eli. Looks like the 5881 is a 6L6 / 6L6G only more rugged (increased dissipation ratings). I'm assuming that either these, EL84s of KT88s will push about the same power out, so I should be able consider any designs using these output tubes?

Depending on O/P "iron" quality, either Mullard or Williamson style small signal circuitry will be just fine.

I am guessing you mean I should use something like an EF86 for the first-stage voltage gain and the ECC88 as the splitter, correct?

I mean if he is interested in altec/western electric stuff, why not build a western electric amplifier? A 124 or a 142 would be fine. Heck if you want to shoot for the moon, you could build a 143.

If you aren't stuck on building a PP amplifier, a single ended 300b amplifier pairs very nicely with vintage horn speakers.

Actually, he has no idea what an Altec Lansing / WE amp / speaker is. He just thought the looked cool. Also, I would like to see if I can find something with some depth of documentation online as he wants it done in about 4 - 6 weeks, which will limit my build -> blow it up -> wallow in self pity -> find the strength to go on -> rebuild amp cycle.

Hey welcome back!

Great to be back! Thanks, man!

I'm currently leaning toward the Claus Byrith EL84 Push Pull Amp. I remember reading this paper years ago and thinking it would be a fun build, but there may be better designs out there. It's basically a modified Mullard 5-20.

Any comments on this design?

And thanks again!

Kofi!
 
EL84s are good for approx. 15 WPC. KT88s are good for 60+ WPC.

The best "vintage" version of Mullard style circuitry I'm aware of is the Harman/Kardon Citation V. Notice the high gm theme in the small signal types. 12BY7s are becoming scarce. So, a 6922 cascode is provided as a substitute high gm/low CMiller voltage amplifier.

Possible enhancements to the Cit. 5 are a CCS in the LTP tail, instead of a resistor, and employing the ECC99 (yet higher gm) as the LTP.
 

Attachments

  • Citation-V-full-final.gif
    Citation-V-full-final.gif
    273.3 KB · Views: 990
  • Cascode 6922 Mullard Front End.jpeg
    Cascode 6922 Mullard Front End.jpeg
    636.5 KB · Views: 970
Thanks, Eli!

I actually misstated the Claus Byrith amp-- it's actually an EL34 push pull, not an EL84. That should be 35W output according to the design docs (I think there's a good amount of NFB in the amp driving down the power). 35W should be fine, right?

I'll review the amp design you posted and will likely post some questions about it soon.

Thanks!
Kofi
 
.... now that my kid is 12 and has adopted a policy of aggressive indifference toward me!

It looks like I'm not the only one with a 12yo at home, lol :rofl:

@Eli
How about another high gm pentode like the EF184 as replacement for the 12BY7?
I don't need the gain as my version is too sensitive as it is with trioded EF86. I'm not keen on heavy gNFB but might give it a try.
 
this is a thread screaming for Pete Millett's Engineer Amp. Link to Pete's page, or search for the many threads here.

Fast build time, scalable up to around 200W if George and other extremeists are to be believed, proven board and circuit, and Pete's usual documentation and happy free consulting if your questions are halfway intellegent or better.

Thanks for the idea, aardvarkash10. Unfortunately, I'm concerned that the fixed bias will be an issue and that the aesthetics of having a visible / hot / touch-able / ouch-able heat sink won't work for my friend. He will likely never calibrate the bias when it drifts out of whack and the optics would be a no-no. Note that this will be going into a restaurant he owns and I'm trying to limit his legal liability.

Sorry!

It looks like I'm not the only one with a 12yo at home, lol :rofl:

Wow! There's two of us? Let's get another and start a support group!

@Eli
How about another high gm pentode like the EF184 as replacement for the 12BY7?
I don't need the gain as my version is too sensitive as it is with trioded EF86. I'm not keen on heavy gNFB but might give it a try.

The EF184/6EJ7 looks interesting as a voltage amp. The 6922 cascode has (IMO) the big advantage of using an in production type.

Readily-available tubes are a indeed requirement for this project, I'm afraid.

Eli-- I am working up a version of the schematic for the Citation amp you recommended using the 6922 front end. I'll post my version shortly to see if what I drew up is what you were talking about (spoiler alert: it probably won't be as I tend to misunderstand things-- just Google my posts and you'll see what I mean...).

Also, how about the PP-1C from Allen Wright? Would that be a decent candidate?

Eli,

Whipped up this version of the Citation V using the 6922 approach you drew up. Is this what you were thinking?

Not sure if this is exactly correct so feel free to critique.

Kofi
 

Attachments

  • Citation V with 6922.PNG
    Citation V with 6922.PNG
    170.2 KB · Views: 522
What is shown as R7 & R8 (2X 500 Kohm parts) is, in fact, a 1 MOhm trim potentiometer. The "set & forget" pot. is adjusted to yield 3 V. of 6922 self bias.

PSRR in cascodes is very poor. At a minimum, well filtered B+ is needed and regulation is (IMO) indicated.

The GNFB loop was omitted.

I'm concerned that the fixed bias will be an issue and that the aesthetics of having a visible / hot / touch-able / ouch-able heat sink won't work for my friend. He will likely never calibrate the bias when it drifts out of whack and the optics would be a no-no. Note that this will be going into a restaurant he owns and I'm trying to limit his legal liability.

The H/K Cit. 5 is very much a "fixed" bias amp. So, let's use its "idiot resistant" Mullard style small signal circuitry, but "clone" a self biased O/P tube setup. The Heath W5M provides that template.

Edcor's CXPP100-6.6K O/P trafo will provide the magnetic headroom needed to ensure that the GNFB induced deep bass error correction signal does not saturate the core.

When legal liability rears its FUGLY head, a substantial tube cage is a must. Keep those mischief making little fingers out of places they definitely don't belong.
 

Attachments

  • Heath W5M Schematic.gif
    Heath W5M Schematic.gif
    84.9 KB · Views: 522
  • Heath W5M Voltage Chart.JPG
    Heath W5M Voltage Chart.JPG
    56.3 KB · Views: 504
Also, how about the PP-1C from Allen Wright? Would that be a decent candidate?

I have built the PP-1C and then turned it into the PP-2C, its big brother. Clean and clear amps, Allens sonic signature, but not super powerful as they are biased into Class A. Also contain silicon bits and pieces which some are allergic to.

Whatever you build, please document it in a big thread including all the dumb things you did :) - your previous threads were priceless and very informative. Unless, heaven forbid, you don't do dumb things anymore?
 
What is shown as R7 & R8 (2X 500 Kohm parts) is, in fact, a 1 MOhm trim potentiometer. The "set & forget" pot. is adjusted to yield 3 V. of 6922 self bias.

Right-- I represented this as two 500K resistors as I wasn't sure how to implement a trim pot in LTSpice. Too dumb / lazy to figure that out.

PSRR in cascodes is very poor. At a minimum, well filtered B+ is needed and regulation is (IMO) indicated.

Here comes the ignorance:

If I wanted to regulate the cascode supply with some transistors, would that do the trick? Maybe with the 10M45S? Or if I just added a decent-sized choke to the supply would that do?


The GNFB loop was omitted.

I am definitely going to need a parent to help with the GNFB implementation...


The H/K Cit. 5 is very much a "fixed" bias amp. So, let's use its "idiot resistant" Mullard style small signal circuitry, but "clone" a self biased O/P tube setup. The Heath W5M provides that template.

Hmmm... so, I updated the schematic in LTSpice. Does the attached look correct? It appears this would run the 6L6GCs at around 98mA. I assumed a 330R cathode resistor would be OK here, but maybe lower would be better?

Edcor's CXPP100-6.6K O/P trafo will provide the magnetic headroom needed to ensure that the GNFB induced deep bass error correction signal does not saturate the core.

Eight weeks of lead time. Sheesh. Better get the order in now.

When legal liability rears its FUGLY head, a substantial tube cage is a must. Keep those mischief making little fingers out of places they definitely don't belong.

Speaking of which, anyone know how I can ensure that my friend / his restaurant bear the legal liability of someone gets hurt? Maybe I need a lawyer to draft something?

I have built the PP-1C and then turned it into the PP-2C, its big brother. Clean and clear amps, Allens sonic signature, but not super powerful as they are biased into Class A. Also contain silicon bits and pieces which some are allergic to.

I really would like to build this one, but I'd be concerned that 8W wouldn't do it for a multi-way speaker and a bunch of people in a restaurant. Maybe I could adapt it?

Whatever you build, please document it in a big thread including all the dumb things you did :) - your previous threads were priceless and very informative.

Will do, as always!

Unless, heaven forbid, you don't do dumb things anymore?

Ha ha ha ha ha ha ha haaa haww HAWW HAAWWAAAWHAAAAGGHHH *COUGH* *COUGH* *COUGH* uhhh.. uhh. uhhhhhh......

No. Never happens anymore. Never. Sometimes. Mostly.

Kofi

6L6GC Amp Self Bias.PNG
 
Administrator
Joined 2004
Paid Member
Make sure the thing is in a properly grounded, fully enclosed metal box, properly fused, that there are no accessible points with HV on them, and the tubes are not accessible without removing a cover first. Make sure that the end user understands that it must be properly ventilated and that regular maintenance is required. (For which you must be paid.. :) )

Note I am not a lawyer, but on this front news might not be so good - without an LLC I doubt you can be legally excused from personal liability in the event that misuse or a malfunction causes property damage or personal injury. The end user may indemnify you against such loss but that will not be binding on the insurance company insuring his restaurant. The risk is pretty small but I do recommend you talk to someone qualified to advise you in this regard since you raised the issue.

In some jurisdictions there may be further complications around requirements for UL/CSA/cUL certification for anything used in a public venue. (LA comes immediately to mind.)
 
Thanks, Kevin.

I am definitely concerned about this issue. I am confident that I can make the amplifier safe for proper use and handling, but I can't guarantee someone won't drop it while it's plugged in, spill a mai tai in it and slurp it off a tube with their tongue, attempt to hold up their end of a double dare by sticking their fingers in a socket, etc.

I already have an LLC (for a management consulting business), but I am thinking product liability insurance would be a good thing to have. I've reached out to a few lawyer friends but no one seems too keen on responding as of yet.

Also, tube cages-- I don't think Atlanta has a Tube Cage Dept location here. Where does one find these?

Kofi
 
Administrator
Joined 2004
Paid Member
Hammond makes nice chassis and matching cages. I recommend aluminum as the steel chassis are hard to work with. (Kills punches, etc.) You can get them from Antique Electronic Supply or through their distributor business.

Long ago I thought about making a little FM stereo solid state/tube hybrid receiver for commercial use, but the business tanked long before I ever got there.

Yes, LLC should help. You probably need a lawyer familiar with product liability law I would suspect. Maybe you can make him an amp in return for legal advice! :D
 
I would go for a parallel push-pull with EL84's. Pentode with 10% cathode feedback for 30W power easy. Possibly more.

A reasonably priced transformer for this would be the Toroidy 4K:
TTG-CFB4000PP - Tube output CFB transformer [4kOhm] Cathode Feedback Push-pull - Shop Toroidy.pl

The front-end only needs to be capable of swinging more volts (i.e. 10% of the output voltage). But it's not a big deal really as the input impedance of the output stage with CFB is more benign.

N.B.
When reading the specs, where it says 10% Ra for cathode fbk windings in reality is 10% of total primary turns. So it wrongly specified but the transformer works quite well.
 
Last edited:
Hammond makes nice chassis and matching cages. I recommend aluminum as the steel chassis are hard to work with. (Kills punches, etc.) You can get them from Antique Electronic Supply or through their distributor business.

Long ago I thought about making a little FM stereo solid state/tube hybrid receiver for commercial use, but the business tanked long before I ever got there.

Yes, LLC should help. You probably need a lawyer familiar with product liability law I would suspect. Maybe you can make him an amp in return for legal advice! :D

Thanks again, Kevin. I just found a lawyer that will help me with this issue. I'll keep you posted.

I would go for a parallel push-pull with EL84's. Pentode with 10% cathode feedback for 30W power easy. Possibly more.

A reasonably priced transformer for this would be the Toroidy 4K:
TTG-CFB4000PP - Tube output CFB transformer [4kOhm] Cathode Feedback Push-pull - Shop Toroidy.pl

The front-end only needs to be capable of swinging more volts (i.e. 10% of the output voltage). But it's not a big deal really as the input impedance of the output stage with CFB is more benign.

N.B.
When reading the specs, where it says 10% Ra for cathode fbk windings in reality is 10% of total primary turns. So it wrongly specified but the transformer works quite well.

Thanks for the reply, 45!

I will look into that once I have the EL34 fully modeled. I believe that there are some complications related to the balance of the EL84s in the parallel pair, correct? I would love to make an amp that had all of the benefits of an EL84 push pull only MORE POWER!

Any designs you would recommend?

Also-- @Eli: I have been noodling around in LT Spice with the Citation V design and input scheme you posted and wanted to see about using the the EL34 (I happen to have a quad of those already).

Would you (and everyone) mind taking a look at the design I attached and tell me how horribly awful I am at this? Also, would I need to have a regulated supply if I CLCRC the snot out it (see PSU below the audio circuit)?

Kofi
 

Attachments

  • EL34 Amp Self Bias with PSU.PNG
    EL34 Amp Self Bias with PSU.PNG
    201.9 KB · Views: 339
Because of current hogging, forget PPP EL84/6BQ5s in a self biased amp.

Different O/P "iron" for EL34s than 6L6 family tubes. Edcor's CXPP60-4K should be OK, but make certain to roll infrasonic noise off at the amp's I/P. A single RC high pass pole whose F3 is in the 16 to 18 Hz. range gets the job done.

There's little to be gained by reinventing the wheel. Use Mullard's original 5-20 as the starting point, for a self biased PP EL34 O/P stage.

BTW, Mullard was trying to sell tubes that they manufactured, when the 5-20 design was published. The topology is fine, but the high RP/low gm 12AX7/ECC83 is a poor choice for the LTP.
 

Attachments

  • Mullard 5-20 Schematic.jpg
    Mullard 5-20 Schematic.jpg
    380.3 KB · Views: 327
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.