|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
68.148.189.116
In Reply to: RE: WECo 91A/B--Long, but hopefully informative posted by gilmorneau on November 22, 2014 at 12:06:13
Thank you for taking time time to write such a detailed response, it actually is a big help, and it settles some points too.
1. Input tubes will be metal 6J7. 6J7 has lower distortion than 6SJ7, and it avoids the need for a weird tube socket. Metal version because of shielding.
2. The tone shaping network will require some playing with to get right.
3. I will use PSUD-II to model various power supply configurations, but I plan on staying with multiple series connected electrolytics, like western electric did... The challenge will be specifying an appropriate choke.
4. Output tx puts me at a loss. You seem to have tried them all, basically I would say I will buy it on spec ... Best compromise between distortion at 20 hz, -3db high frequency limit, and price. Not everyone has this info though so it may take some gambling to get it right.
5 I plan to start with as close as possible circuit to original and then optimize it, so wiide bandwidth is absolutely necessary from output tx
Follow Ups:
First, a couple corrections/edits to my post above.
I initially came up with a sensitivity of 0.25V but on re-thinking it, it's probably even more sensitive than that--around 0.05V input for full output. Still a 'back of a napkin' casual calculation, but I think it's closer to correct. Obviously, some sort of attenuation will be required for home use.
Also, I said the 'field supply' had a resistance of around 1000 ohms. What I meant, of course, was that the field winding on the speaker had about 1000 ohms (used as the choke for the amp's power supply filter).
Regarding the series connected capacitors on the original, WECo did this in order to reliably get the necessary voltage rating out of the electrolytic capacitors available at the time. There's no need to do it, and no advantage, if you can source capacitors with sufficient voltage ratings. OTOH, if you do decide to use series capacitors, it's essential to parallel the voltage dividing resistors across them to ensure that each cap gets equal voltage.
Regarding output transformers: As noted below, the original WECo 171A trans were wound on (what is now) an archaic type of lamination, namely a CIC type, which I understand is no longer available. There are those who salvage the cores from other WECo transformers and chokes in order to rewind them to something like 171A specification. There is (or was) at least one vendor in the USA who resells 171A on cores rewound in Japan. Note that the original 171A and its clones have very limited primary inductance, on the order of 10Hy. If you gap them for more DC in the primary, you get less inductance. If you gap them for more inductance, you get less DC in the primary before saturation. Nature is cruel.
The Peerless/Magnequest TFA-204 is the closest thing available new AFAIK. It is a good little transformer, wound on conventional EI lams, but requires adjustment to the original circuit values in order to perform optimally. The original 91 amps pulled 80mA through the transformer, the TFA 204 isn't happy with any more than 60mA.
I am unfamiliar with the Amtrans option mentioned by GSH, though I'm not surprised they're not mentioned in a 1935 catalog if they're really meant as model 91 replacement transformers--the model 91 wasn't introduced until 1936. Does anyone have a 1936 or later Amtrans catalog with showing the transformers GSH mentions? Please post.
I have to agree that any of the three 'original' types just outlined will be of limited utility for full-range, hi-fi use. Maybe adequate for reproducing an impression of 1930's sound (if you had the right speakers), but get a bigger, wider bandwidth transformer for highest satisfaction. I like the Partridge best, but they have become difficult to source.
Several people have mentioned a reference to 6J7 having lower distortion than a 6SJ7. Radiotron Designer's Handbook (4th ed.) page 510. It has probably been reprinted since, but there it is.
In his 'YAWE91' thread, Thorsten advocates for using POSITIVE feedback to boost the high frequencies a bit. WECo used frequency selective NEGATIVE
feedback to achieve the same end. Thorsten's is a viable (though tricky) approach, but is not authentic to the original model 91 ethos, which seems to be the OP's intent.
Almost hate to indulge them by mentioning it, but as I predicted in my point 5 above, the thread-jackers have apparently arrived. It would be funny if it weren't so tedious. In what sense is a suggestion to try a DC 2A3 amp helpful to someone explicitly inquiring about an authentic WECo model 91 clone? Sure, a DC 2A3 can make for a satisfying little amplifier, no doubt. But it's not the ONLY way, and it's not even the best. People who listen to other topologies are not ignorant--they may know things the DC 2A3 crowd doesn't. There's more than one way, for sure. But it needs to be its own discussion thread.
Any ideas what these are?
http://www.ebay.com/itm/300B-SE-tube-output-transformer-one-pair-Similar-Western-Electric-171A-for-91A-/121490405993?pt=Vintage_Electronics_R2&hash=item1c49647a69
They sure do LOOK the part.
No direct experience with them, but the description is not very detailed, and there are no photos of the transformer core/windings. Might be original WECo laminations in there, but you'd think he'd be showing them off if so. I expect a dissection would show a fairly small, EI lam transformer. If anyone's bought them and looked inside, do share what you've found.There's a company in Japan that I know rewinds to genuine WECo cores as used in the 171A. Their version of the 171A sells for US$1995/pr, last I checked. TubesUSA can get them for you.
Edits: 11/24/14
I'm surprised you haven't heard of these. They were regarded higher than the Thordarson True Fidelity series, the only other pre WWII transformers to state hi-fi as other than 100-10Khz. Paul Mundt, the aforementioned hustler believed some Amertrans OPT's were installed on NEW WE91's, especially if it was destined for a small theater system and NOT the monitor. Who knows, but they are good! In the 1935 catalog, they suggest their parallel feed reactors in combination with their existing selection for SE use.
I'm familiar with AmerTran, I just wasn't aware that they offered a series feed single ended transformer suitable for use in a WECo 91 circuit. Wouldn't happen to know the reference number, would you?
no, but I think I also remember seeing them mentioned in those large Altec service binders, Graybar, etc... I think those were called an upgrade.
Maybe I'm dreaming. And yes they seem to have been series fed, not the 1935 versions. I actually saw them on 91 amps. Wires came out the top, flanges on the bottom.
Another awesome post! Thank you again. The RDH4 was my bed time reading for a couple years in the 1990s so that makes a lot of sense that is the source - thanks for the reference!
I am now very curious as to how the Shindo labs 91a manages excessive gain, since they seem to be one of the only "clones" to maintain two gain stages plus negative feedback.
I understand very well that there is no need for the series connected electrolytic capacitors, my idea is just to build it as close to original as possible, with the exception of the output trabsformer - this project will take a bit of financial outlay and i don't intend to spend a pile of money on an amplifier which can't achieve wide bandwidth performance.
What model partridge do you have specifically ? Difficult to source is not impossible to source, and that is quite a name to recon with.
The Shindo amp has much less gain than a 'stock' 91-B because it isn't the model 91 circuit, really. Tube compliment on the Shindo is one 310, one 311, one 300B and (I think) one 5R4. There's a big difference between a 310 and a 311. The 310 is a high gain, small signal pentode. The 311 is a power pentode with much less gain (but capable of more power) than the 310. I have no more insight into how Shindo utilizes it than anyone else, but I suspect it's used for its power, to provide a driver with more 'gumption' than just a 310. It might be pentode wired, might be triode wired, or it might be wired as a cathode follower, direct coupled (or transformer coupled) to the output tube. That would really give the amp some 'drive' (and much less gain than a 'stock' 91-B).
The Partridge transformers I have carry part number TK7441. The Partridge literature I have lists the TK4519 and TK4663/2 as having the same specifications.
To the best of my knowledge (gilmorneau, correct or expand if you have something!) the original output transformer was built on a core shape that has not existed for decades. Not having seen one, or the winding design, or heard one, I won't even guess about performance!
The Peerless TFA-204 was apparently designed as a replacement unit for the Model 91. I have experience with the Magnequest version of this design (Mike has the design archives) which, while a wonderful transformer in its own right, cannot handle the Model 91's 70mA of DC plate current without moving into saturation territory on loud bass signals. I would guess that the movies being shown in the forties and early fifties did not normally have much bass below 50-100Hz, so this would not have been a problem. But for these reasons I would suggest getting a larger and more capable output transformer for audiophile use.
I thought the 6SJ7 was an octal tube?
Or are you speaking of the WE 310?
I read an article a long time ago, maybe in audio magazine, maybe radio&tv news, to be honest I forget the origin. The author tested all the small signal pentodes, and the conclusion was that the 6J7 has better performance than the 6SJ7, 6AU6, etc.
What I meant about sockets, is I would prefer to use an octal socket instead of the old fashioned 6 pin socket the 6C6 and the 310A use.
The external grid cap is most likely the biggest contributor to the lower distortion.
In general, I've found that tubes with separate grid/plate caps sound better.
"The external grid cap is most likely the biggest contributor to the lower distortion.
In general, I've found that tubes with separate grid/plate caps sound better."
Is the lower distortion caused by being able to run higher voltage, lower current on the tube, because of the plate cap?
GEC made a KT88 with a grid cap that is much sought after. I would have thought China/Russia would have reproduced by now. George Anderson at Tubelab has a favorite tube that he says is a 300B with a plate cap. Can't remember the name.
Big speakers and little amps blew my mind!
It's because it is not the same tube internally, it's not related to the grid cap.
Are you quite certain of this?
You may be correct - but it seems to me that it was too easy to simply take a 6J7 and move the grid connection to the base.
Lots of tubes are different when you wouldn't expect them to be:
i.e. 6N7 is different than the 6SN7
the 6Q7 has a lower mu than the 6SQ7
The 6J7 has a lower transconductance than the 6SJ7, and better linearity
Probably the 6SJ7 is better optimized for RF use.
I hear you - but I'm still unconvinced that the biggest difference is not (all but) eliminating the grid to plate capacitance.
The data sheets show the same .005pF control grid to plate.I must not be understanding this thread.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 11/23/14
"It's because it is not the same tube internally, it's not related to the grid cap."
Cool. Still it seems that many tubes with grid caps have good reps for good sound. Wonder why that is?
Jamie
Big speakers and little amps blew my mind!
MHF,
"I read an article a long time ago, maybe in audio magazine, maybe radio&tv news, to be honest I forget the origin. The author tested all the small signal pentodes, and the conclusion was that the 6J7 has better performance than the 6SJ7, 6AU6, etc."
I've read the same thing too somewhere. Are yours British made?
Jamie
Big speakers and little amps blew my mind!
Yes, a friend in the UK sent them too me as a gift years ago. I just put them back in the amp... Vesus the RCA 6J7 I have to say the difference if any is sure not obvious. I thought the RCAs sounded better before but now I am not so sure.
Yeah, maybe stick with 6J7 as the safer option.There are good versions of the UX6 sockets available, but availability of the 6C6 tubes is starting to get patchy. I have over ten pairs remaining, but two pairs have broken (cap gripped the loose grid connections a little too tightly). Probably my fault though, oops.
Cheers.
"In the beginner's mind there are many possibilities, in the expert's mind there are few." Shunryo Suzuki
Edits: 11/22/14 11/22/14
I would prefer an octal tube as long as it's within a couple % of the sound I heard with the WE 310A (actually what I heard used a Soviet version of the 310A)
Post a Followup:
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: