|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
71.212.196.9
I got a wild hair to build "old-timey" monoblock amplifiers from the 1950s, in part because I have a lot of the parts on hand and also because the aesthetic appealed to me. I've had very good luck in my amateur builds of the past couple years. Looking at this schematic I wonder if I could get some advice from the gents here. I have 4 questions:
1. Are there any modern but simple changes that would improve this 1950s design? This is a Grommes 216BA. I won't change the 12AU7s, nor add CCS. My PS will be:
400-0-400>5U4 or 5R4>20uf film 500V>6H choke>40uf 500V>dropping R>20uf>etc. 430VDC B+2. What does that connection w/ 220K R from 6BG6 plates to the cathode of the driver 12AU7 accomplish?
3. Since I will be using OPTs with 8ohm speaker tap only, how much would I adjust the feedback resistance? I have modern speakers. My OPTs are 5.5Kohm, (w/ no UL taps) though this design calls for 7Kohm. (If I can find 6.6 - 7Kohm OPTs I'll use 'em)
4. Should I use the triode connection shown or supply G2 from a separate ~400V DC source off the B+ line? I am using tube data from old sources, tdsl.duncanamps.
I plan to provide the bias of -40 to 45V from small separate transformer, not the arrangement shown.
Thanks for any experienced advice.
Pete
Edits: 05/22/17 05/23/17Follow Ups:
with it, allowing you to upgrade the PSU's storage (capacitance).
This will be a good thing to do.
It will lower noise (ripple) and will give the amp more slam.
Warmest
Tim Bailey
Skeptical Measurer & Audio Scrounger
Nah, I've got 3 VERY good (more modern) homebuilt amplifiers with diode rectification, Cree Schottky. This is purposely an old-school build, will have a different look and sound. I have slam if I need it.
I can drive my fancy newer Mercedes but still appreciate a classic old MGA. :)
Edits: 05/29/17
You've received a lot of good advice and suggestions so far. I'll throw my $.02 into the ring at this point...
1. "Are there any modern but simple changes that would improve this 1950s design?"
I suggest replacing the original output tubes with 6L6GC or 6P3S-E. That will eliminate the high voltage wiring above the chassis. Replace the 12AU7s with 'SN7 octals. I would use a solid state power supply for better regulation. Also, the filter design needs to be qualified for output impedance over frequency. Some older designs perform poorly in this regard. I never buy parts or cut the chassis before analyzing a proposed PS filter with SPICE. Several people here can do that for you if you want.
2. "What does that connection w/ 220K R from 6BG6 plates to the cathode of the driver 12AU7 accomplish?"
It's a nested feedback loop. I would eliminate it and use one loop only. Including the power transformer in the loop does no harm.
3. "Since I will be using OPTs with 8ohm speaker tap only, how much would I adjust the feedback resistance? I have modern speakers. My OPTs are 5.5Kohm, (w/ no UL taps) though this design calls for 7Kohm. (If I can find 6.6 - 7Kohm OPTs I'll use 'em)"
That's one of the most difficult tasks you'll face in terms of optimization. I generally apply about 12dB NFB to pentode amps, then work it from there using square waves, distortion measurements and various loads (including capacitors). OPT primary impedance isn't critical for pentode use; your 5.5K transformers will be fine.
4. "Should I use the triode connection shown or supply G2 from a separate ~400V DC source off the B+ line? I am using tube data from old sources, tdsl.duncanamps."
I would not make it switchable like the original design unless you plan to use this as a guitar amplifier. For hi-fi use, it's not possible to optimize for both modes in the same amplifier. Triode operation can bring sonic benefits, but only if the support components are very high quality and your speakers are efficient enough that you don't need the additional power provided in pentode mode. Also, the OPTs need to be greater than 5.5K for best performance in triode mode. Regarding screen voltage, the most common method is to use a dropping resistor and good quality bypass. Regulation can be beneficial, but I wouldn't include it in a first-time build.
I'm sure you realize by now that there are many ways to skin this cat. There's nothing particularly outstanding about the Grommes design, and it shouldn't be considered an immutable standard for your work going forward.
--------------------------
Buy Chinese. Bury freedom.
"It's a nested feedback loop. I would eliminate it and use one loop only. Including the power transformer in the loop does no harm."
Two things; one, Global loop NFB is the worst for creating stability issues, and IME it sounds the worst. This NFB is also balanced, and reduces the output impedance directly which helps with bandwidth in and of itself. It also linearizes/reduces distortion faster than SE on a per dB application.
Second, why would you put the power trans in the loop? I suspect you meant OPT...LOL
cheers,
Douglas
Friend, I would not hurt thee for the world...but thou art standing where I am about to shoot.
Placing the power trans in the loop allows you to modulate the AC mains. Your neighbor can enjoy tube sound too!
--------------------------
Buy Chinese. Bury freedom.
Thanks for your advice on this, in particular about triode operation. I have built several amps inc 6L6GB/C and mention them below, but certainly don't expect anyone to dig thru all that. :) The plate caps are sort of the whole point of this build. :)
I once made a preamp around a beautiful cast alloy Vol knob from a 1948 EH Scott amp.Seems like another way to treat screen voltage is with a 15H 40mA choke, instead of resistor. Unfortunately, R is high in those chokes.
I could do screen regulation; I have room for OBx type tube regulator, since I'm doing this old-school. They will handle the current but not sure about 250-300V.
Researching screen regulation on this august site shows lots of fights and opinions. Consensus is rare. But I guess that keeps it interesting. :)
Thx
Edits: 05/25/17 05/25/17
I don't think a screen choke is useful for push-pull. The simple RC approach has proven to work well. If you want to regulate, a series pair of 0A2 or 0D3 will get you to +300V.
OK, I understand the aesthetics you're after. If I were going down that road, I'd consider the 1625/807. A quad of those has a nostalgic look that's hard to beat.
--------------------------
Buy Chinese. Bury freedom.
That's good advice - i will keep screen connections simple. Probably can convert to Triode later to see what that's like. I don't think I need a ton of power. My only triode amp (my first build) was w/ EL34s.
Funny you should mention 1625 and 807s - I have good quads of each of those. BUT you have to commit to the oddball 5 and 7 pin sockets; can't plug in alternatives.
I believe those numbers are maximums. That's a lot of juice, will try to stay in the moderate zone.
Datasheets indicate that this tube type is rated for 20W maximum anode dissipation. TDSL shows that as well in the abbreviated specifications. However, TDSL shows AB1 operating conditions with 450V at 116mA idle current. That's over 26W per tube. Operating at full power, or 210 mA, pushes that to 47W.
TDSL shows these same operating conditions for the 6L6 and 6L6GC. In fact, the latter is the only one in this series that can be run at those power levels.
--------------------------
Buy Chinese. Bury freedom.
Yeah, I knew that current draw was crazy high, but I'll have adjustments to each tube, at least that's the plan. I will measure mA at each tube.
OK, sounds like you've got it under control. Have fun with the build!
--------------------------
Buy Chinese. Bury freedom.
Sorry, I had forgotten that the 6BG6G is the ST bottle. It's also one of the nicer looking tubes. Guess there's not much reason to use the less common sockets, although the 1625/807 do have a higher dissipation rating. Speaking of which, watch out for the limit if you're building a supply over +400V. The 20W rating will be easily exceeded at higher voltages.
--------------------------
Buy Chinese. Bury freedom.
I trick I have used when lacking a bias tap is to use a reversed filament transformer.
Take a 12vct unit, 100ma is more than anough, and hook the centertap and one of the secondary legs to your filament buss. This will give you 60vac to play with for the bias supply.
Errr...connecting half a 12.6V to 6.3 VAC will deliver full primary voltage( as in 120V ). If you hook the whole 12.6 winding up to 6.3, you will get 60VAC out on what used to be its primary...:)
cheers,
Douglas
Friend, I would not hurt thee for the world...but thou art standing where I am about to shoot.
Thanks. I would have to find a very small xformer and have 2 that match, in this case. Most of my 12V formers are for filaments, 2A+.
Question.
How would other parameters change if I ran this tube in pentode mode, since i don't have UL taps. Tube specs say under 400V on the screen. Sourced from dropping B+ ?
Thought I'd do triode mode but looking at options.
Thanks
It's a beam power tetrode, not a true 3 grid pentode, but that's essentially semantics.
Max. open loop linearity in full pentode mode is obtained by regulating g2 B+ at a fraction of anode B+. Nasty IM distortion gets held down.
Eli D.
Thanks, I know it's a tetrode but the data sheets call it pentode mode. I see the people are tired of this thread.
Time to start drilling holes in aluminum.
Thx
before you go cutting holes, you should figure out what your g2 voltage target is...:) or leave room to do it later.
cheers,
Douglas
Friend, I would not hurt thee for the world...but thou art standing where I am about to shoot.
The supply needs to be well reg'd, but it is not 'a fraction of B+' that is important, it is exactly what the combination of B+, idle point, load mixes with the effects of g2 voltage. The data OP's are for max power, probably best if you don't know why you're doing it to reduce the g2 voltages by at least 10%, maybe 15%. Depends on who wrote the data sheet.
cheers,
Douglas
Friend, I would not hurt thee for the world...but thou art standing where I am about to shoot.
Probably the best AB1 operating condition for 6BG6G pentodes is 750 V 30 mA plate, 350 V regulated screen. A lot of power with very low open loop distortion. The only caveat is 10-12K output transformer. If you have super-6BG6GA tubes, those can be taken even further to 800-850V, 400V, and 40 mA, with even better results.No 6L6 version is capable of pulling such stunts.
Edits: 05/24/17
It is a variation on the Williamson theme. Compare it to the self biased Heath W5M.
The suggestion to use 6CG7s instead of the less linear 12AU7 is quite sensible. Remember, Williamson used 6SN7s and the 6CG7 is electrically "equivalent" to the 6SN7.
20 muF. in the 1st filter cap. position is good, if the rectifier is a 5R4, but more is OK, if a 5U4 is used.
5R4 data sheet
5U4 data sheet
Eli D.
I noticed when I had gone over 4uf for the first cap,the smaller choke would buzz.
"For every complex problem there is an answer that is clear, simple, and wrong" H. L. Mencken
That's a pretty wild set up. I already built my KT66 amp, a HK Citation V clone, with Sherwood S8000 OPTs. Sounds very good, punchy and dynamic w/ new GL KT66s. I don't run them hard, 42mA at each tube.
What output trafo did ou use in the KT66 and did you use GNFB?
"For every complex problem there is an answer that is clear, simple, and wrong" H. L. Mencken
Hi Mike, the Citation V clone amp uses OPTs from a Sherwood S8000, at 7Kohm. Not huge transformers but very sweet. I built the amp for 6L6GB and 6P3S-E and it was great but decided to try KT66, and though it is run mildly (385V plates, about 42mA on each tube, adjustable) it has real punch and great sound. Susan Tedeschi and her Blues band - Wow!I used the original specs for FB; probably not correct for the new modified set up but it works: 10K resistor and 100pf cap. Without FB the amp barely functioned, sounded terrible.
I have other amps that you can remove fb and and I hear no difference, (but would obviously measure differently).Pete
Edits: 05/24/17
Thanks for that - actually I was studying my old (1960 original! - with voltages) Heathkit W5M manual that made me consider variations, such as higher filter choke value and the 5R4 rect. But then I'm deviating from the Grommes design a lot, taking chances w/ things I can't properly measure.
What about plugging 12FQ7 into 12AU7 sockets?
Pete
"What about plugging 12FQ7 into 12AU7 sockets?"
That will work, if and only if the 'U7s are wired for "12" V. heater power. OTOH, the 12BH7 is 100% pin compatible with the 12AU7 and, as long as the doubled heater draw is OK, drops right in. While (perhaps) not quite as linear as the triode found in 'SN7s, 'FQ7s ..., the 12BH7 is much more linear than the 12AU7. IMO, the only audio jobs the 'U7 triode is suited to are "concertina" phase splitter and cathode follower. Inside the bottle, local, NFB saves the day, in those applications.
Eli D.
Thanks for that info, I have lots of 12BH7, and the other tubes. But not so sure about that much increased heater draw.
Pete
12FQ7s are getting quite scarce. It would be kind of you to let the guys who need them have first crack at them.
Besides, the heater winding in the amp is a 6.3 volt winding - so a very minor rewiring is all that's needed to plug in 6CG7/6FQ7s.
12FQ7 isn't a direct plug in for 12AU7 anyway. It can't run on 6V, no connection to pin 9.
Pete
You dont have to change the circuitry for the 12AU7 but I would use adapters and run 6CG7s because the difference in distortion and bandwidth is noticeable and measurable.Steve is right in that you want to beef the bias supply. You can use a 20uf film cap there but a lytic there is fine.
Why do you want to cut the charging cap to 20uf off the 5U4? I would use a 50uf film cap there and keep the existing choke because if you go to a 6Hy,that can kill dynamics and roll off the high end.What is the factory value of the choke? If it is 6Hy,then you should be ok because it was probably calculated.The 220k resistor from the 6BG6 plate is a loop feedback resistor for the driver.One more thing,if the output trafos in the amp are small,you may have to use some global NFB to keep the amp stable.
"For every complex problem there is an answer that is clear, simple, and wrong" H. L. Mencken
is only going to reduce stability margins. Any amount greater than Zero is going to make the amp *LESS* stable.
The balanced NFB applied to the driver cathodes does a lot less to wreck stability margins on a per dB applied than a global loop.
cheers,
Douglas
Friend, I would not hurt thee for the world...but thou art standing where I am about to shoot.
"The balanced NFB applied to the driver cathodes does a lot less to wreck stability margins on a per dB applied than a global loop.....[because there are less reactive components in the loop path]."
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Thanks very much Mike. (we go back about 10 years, I reckon). This Williamson circuit is reminiscent of my old Heathkit W5Ms (now sold), which I meticulously restored. It used 5R4s which I understand require smaller 1st cap. Also I am running out of room underneath this monoblock chassis. I plan to use 5R4 to drop B+ a bit (but have MANY diff 5V rectifier tubes). I'll do 30f film.Honestly did not know increasing choke would "kill dynamics and roll off the high end". I use 10H - 20H in preamps, but that's a different animal. Schematic for this calls for 2.6H. Anyone want to trade some good 6H and 9H chokes for a pair of 2-3H chokes? :)
Speaking of the Hk W5Ms, I switched the 12AU7s in those to 6CG7s - there was no benefit whatsoever - I may have even lost something. What if I plugged in 12FQ7s (which I have) or even 12BH7 (have dozens) to the signal tube positions? There isn't room for 6SN7s.
Pic shows my chassis layout test. (A choke sitting where the OPT will go.)
Thank
Petein Denver
Edits: 05/23/17 05/23/17
i would like to see the explanation for HOW a power supply choke can kill dynamics and roll off the top end. Anybody? The power supply should be as rock stable as possible. Anything that affects that stability is going to affect sound quality.
What I was saying was,if the choke is totally miscalculated for circuit,it can change the impedance of the power supply and how it relates to the load that it is feeding.This can cause loss of dynamics and or freq response.It happens mostly at lower frequencies like 120hz.
I figured this out years ago when I took the choke out of a Citation 2 and put it in the Golden tube SE-40 and we are talking two amps that draw close to the same current. When I put the choke from the Deuce into the SE-40,everything got muffled and dynamics just washed out. The Choke in the SE-40 was a bit larger,but the choke in the Deuce was like .7hy but being both amps had plenty of capacitance of either side of the choke,I didn't think it would be a big deal.
As I have said,I didn't mean to imply that it wouldn't work however,just throwing a high inductance choke that looks impressive into a power-supply can cause issues that one might not expect.That's why I asked Pete if it was the one already designed for that circuit.
I'm one of these people that believe in measuring the basic fundamentals when you make any electrical changes.
"For every complex problem there is an answer that is clear, simple, and wrong" H. L. Mencken
In preamps,current is much lower of course.A 6hy could work just fine if you don't saturate the core by drawing excessive current and Heathkit was able to execute that circuit quite well.
There are a couple things at play here.One,Heath is using an opt trafo that is almost double the impedance of the typical 6L6/KT66 circuit.Most KT66/6L6 PP circuit work in the range of 6.6k.The 16309 trafo is like 11.9k and the 16448 is like 10.9k.The Heath choke is 150ma so that is adequate for that circuit.The circuit is also UL and all these things play a role in that wide bandwidth,low distortion circuit.
"For every complex problem there is an answer that is clear, simple, and wrong" H. L. Mencken
NT
With 400VAC to rectify, and a DH rectifier, root2 times your AC voltage is a good, absolute minimum capacitor rating. It will be over 500V...:)
The plate to cathode connection is NFB...:) I'd bump the current in their circuit( lower value resistor) up and drop the global NFB connection.
cheers,
Douglas
Friend, I would not hurt thee for the world...but thou art standing where I am about to shoot.
Doug
He may not be able to totally dropping the NFB because the transformers have small iron content a lot of time.He could add a 100 resistor in series with the 1k and reinsert the loop between them and add a bypass lytic. That would lower it quite a bit but being the amp also has that 220k to the driver,that tells me they probably did a lot of tweaking to make the amp stable with the existing iron.I'm speculating but it makes sense.
"For every complex problem there is an answer that is clear, simple, and wrong" H. L. Mencken
That is, of course, the problem (of several) using old schematics as a guide. One can't fully know the old iron used, and what compromises were made to work around the iron's specs/shortcomings. A picture of the amp shows BIG OPTs. Mine are smallish, from Hammond organ amp. I used them in a previous build so they sound full range.Tube data shows 6BG6G in class AB1 - 450V anode, 400V grid2, -37V grid1 bias, and 5.6Kohm impedance on OPT. So that was sorta was I was going for.
Thanks for info,
Pete
Edits: 05/23/17
For a given amount of FB, the longer the path the less stable. The RCA SP10 and 20 used a similar topology, and had no global loop.
Aside from the idea of building with 12AU7 because that is what is on the original, bean countered schematic...well, good luck wid that. Or, build, learn, and then decide how to apply what you learned for further improvement...LOL
cheers,
Douglas
Friend, I would not hurt thee for the world...but thou art standing where I am about to shoot.
I agree but look at the iron in the SP20.It is humongous and the amp is low power.All I'm saying is, Grommes used that output trafo with GFB and loop FB to the driver and much of that was probably to compensate for the output trafo. I havven't seen the OPT trafo in the amp but if he runs it at full power and bandwidth,he may run into problems.
"For every complex problem there is an answer that is clear, simple, and wrong" H. L. Mencken
Edits: 05/24/17
This is what I'd do. You may end up liking the 12AU7.
I may. A surprising number of high-end preamp/amp manufacturers still use 12AU7 tubes. They aren't doing it because they're beancounters.
Thanks for "input" :)
In a preamp they are fine but in an amp where the driver is much more demanding,there I would use a 6CG7.
"For every complex problem there is an answer that is clear, simple, and wrong" H. L. Mencken
I'd put more filtering on the bias supply. And an adjustable pot. Maybe, change R27 to 22K-ohm and add a 10k linear pot? But, the values may change when you tap the bias supply off a new transformer. Looks like you'd need around -45VDC.I'd sub a 100mfd/100VDC cap (C3) on the bias supply.
Edits: 05/22/17
Yes, that's good advice. I will make a small perfboard circuit w/ small Tamura trans and diode bride (not half-wave as original). I don't have a lot of real estate for another pot. 1/2W pot would be ok? Thanks
Pete
1/2-watt pot will be fine. I'd put that pot into the bias supply. You are prolly going to using old-stock tubes. These tend to range all over the place, when it comes to idle bias dissipation settings.
Good luck on the project! I love old-school stuff. Tell us how you like those 6BG6B tubes.
8^)
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: