RCA Ba6a Innerstage XFMR

[CJ]

directdriver you are a Saint! Thank you!

i will fix the print.
most likely the topology goes top 1/2 coil a in series with bottom 1/2 coil b , and bottom 1/2 coil a to top 1/2 of coil b to get balanced linkage, as has been seen in other transformers,
your sec dcr adds up to 1/4 the dcr i deduced so 1:1 seems like a solid assumption, along with Sowter and Doug.

when you saw terminals 10 and 12 with two wires, do you mean tiny magnet wires from the coil or cloth leads leading from the solder tags to the terminal plate?

I have taken apart a CG-133 and unfortunately the primary wire size is too small to handle the plate current,

 

[emrr]

I have taken apart a CG-133 and unfortunately the primary wire size is too small to handle the plate current,

I dunno, consider the 22K plate resistors coming off the transformer, it's fairly current limited. The Gates Dual Limiter has a pretty small IT in comparison, and it's direct feed. Of course I haven't compared your CG-133 print to something like the A-18 or A-19 used in the UA limiters, I think you did one of those? We certainly tested some pretty small transformers and they all did fine, though we didn't run them long term where cooking would turn into a problem.

 

[CJ]

stock primary for BA6a is #36 AWG good for 25 circular mils,

#44 AWG is good for 4 circular mils,

if the 36 is getting hot, i would be nervous about the #44, however, that model is fairly cheap next to some of the other vintage stuff for sale,

here is a revised print, taking several liberties with the specs until a real one comes along which should not be anytime soon,

 

[directdriver]

"when you saw terminals 10 and 12 with two wires, do you mean tiny magnet wires from the coil or cloth leads leading from the solder tags to the terminal plate?"

I can see cloth wires coming out of a square can inside. That can appears to be nickel color or maybe it's a mu metal can. It is covered with black tar or whatever potting material.

 

[CJ]

Nice! that verifies that there is no coil jumping going on, just two coils wired in parallel.

that xfmr could be wired 1:2 and called a different part number, i will check the catalog for more clues,

 

[directdriver]

CJ: "that xfmr could be wired 1:2 and called a different part number, i will check the catalog for more clues"

I suspected the same since seeing it's an enclosure inside an enclosure that maybe the inner can is just an off the shelf model rewired for the job. Who knows?!

 

[CJ]

i am rewinding the original with random wind instead of layered paper, hopefully the capacitance will be ok, the shields ought to help,

the #44 i had came in at 0.0015" so i am using #42, if the sec DCR comes out like yours i will adjust the print accordingly.

it would be nice to use the thicker wire as breakage and wire tension is a lot less of a concern,

hopefully this rewind will work out and i can send this thing back to Nielsk for a test drive,

fresh modern wire out to make it last for a long time,

 

[emrr]

I suspected the same since seeing it's an enclosure inside an enclosure that maybe the inner can is just an off the shelf model rewired for the job. Who knows?!

Standard multi-can shield. The 86 limiter is 2:1 step-down to the output stage, I've never seen a UTC version of that one but it may have used this coil with a different part # as a replacement part. Other replacement part candidates would be the 76 console PGM amp, and the separate rack mount versions of same.

 

[vinyvamos]

Amazing work CJ! When you state that the lams are butt-stacked with no gap, to me this is still a "gapped" core as they are not interleaved, no? I have recently done a tear down of one of these so I will check all my data against yours and report back. The core in mine was bright and shiny and upon substituting in 50% nickel UI39 lams I got the same inductance. 80% nickel was too much. My core did indeed have paper in there to create a gap and based on my calculations a plain steel core could not get to 130H in this setup (EDIT: even without paper in gap). I don't have steel lams in this size though so cannot test it yet.

Confirmed from here also that it is 1:1 with the secondaries in series-parallel configuration. DCR here was 2400R pri and 2500R sec. I rewound on a nickel based core with paper in gap got all electrical readings to match within about 5%, including all parasitics and the primary self-resonance frequency.

 

[vinyvamos]

Also, I'm pretty sure that I did a DMG spot test on the original lams which came out positive, meaning there is nickel present. Gonna check my messy notes!

 

[vinyvamos]

Im waiting to get the serial number of the BA6 that my defect trafo came from (Subterranean studios CA) and I believe that your defect unit CJ is from a BA6B. I also did inductance measurements without paper in the gap and then the inductance was much higher than the 130H. Could it be that they started off with nickel cores and paper in gap, then to cut costs they resorted to Steel core and no paper in gap?

I vaguely remember a post from you CJ a long time ago echoing something from Brian Sowter stating that the original UTC trafo was nickel core and gapped. Am I trippin or wha? Very unusual I know, but that is what I found in my teardown. Can 50% nickel lams still rust like yours have with enough moisture??

 

[CJ]

I just finished rewinding the coils , I am scotch brighting the lams to get the rust off so I can get the whole stack back in there so I can check Henrys tomorrow,

Yes it is still a gapped core when butt stacked, there being a natural air gap from space between the lams from imperfections, stamping burrs, etc

There is a list of values for natural air gaps for various cores in Transformers for Electronic Circuits by Grossner.

 

[vinyvamos]

Look forward to hearing how the readings go CJ. Yes exactly, still an air gap due to imperfections in the metal edges. That's how I do the plate choke for my OP6 preamps. But my question is why did they gap it? To avoid plummeting LF response upon imbalance of the 6SK7 tubes under GR?? The Sowter 9830 is not gapped. That was the number mentioned in the quote some posts ago, and I also remember reading that Sowter changed the spec early on for this trafo. They at least changed the nickel to steel ratio

 

[CJ]

I think Sowter might have used a different take a part I did on the cg 133 which has a part nickel core,

Dang, I meant to say I rewound the ba6a coil and will do the 86a next,

 

[vinyvamos]

The ones that I have wound so far I did without glassine paper between wire layers and it at least wound fine on my machine and all electrical specs were spot on including parasitics. Prolonged testing will show whether the wires are too free and loose in there, and fatigue will break them eventually (I hope not!). UI39 laminations happen to be nearly exactly the same size as the originals, albeit all the butt joint will be at one end of the core. This will require a different clamp than what UTC had used, one that goes longitudinally. The only imperial lamination that I found to match is the 101L and I simply could not source that without getting it custom-stamped for me ($$$$$$).

If people here are interested in interstage trafos wound to the UTC/RCA spec then please let me know. It would be good to get an idea of demand before I order more laminations. My main intent with all the research is to build BA6 limiters as a product, just like my OP6-based pre. Still need to tear down an output trafo, any dead ones sitting around??

Oh and on a side-note, the BA6 input trafo uses the exact same laminations, and it is a 1:8.6 ratio, not 1:4 as the Sowter one is . I will also start to wind those in the next months.

 

[MidnightArrakis]

This guy has been the subject of a lot of talk over the years by people who want to diy the infamous BA6a compressor, or fix a dead one.

Rare, expensive and prone to failure.

Pictures tomorrow,

[Pictures tomorrow] -- Here's a photo of a current-day "clone" chassis today:




>> "ROTATABLE" Photo of the ASA-6A Vari-Mu Compressor:
https://audio-scape.sirv.com/ASA-6A/ASA-6A_14.jpg?scale.option=fill&scale.width=2700




/

 

[vinyvamos]

They have done a remarkable job on the aesthetics! Love how they got the font correct on the dials, attention to detail. The audio transformers are apparently made by Altran. All power iron is Hammond. Separate filament transformer, probably to get the filament DC for the 6sk7 tubes as the stock Hammond will only have 6V3, 5V and HT. I just don't know how they sell it for $3k and still make some pocket money. Amazing!

 

[drtechno]

Standard multi-can shield. The 86 limiter is 2:1 step-down to the output stage, I've never seen a UTC version of that one but it may have used this coil with a different part # as a replacement part. Other replacement part candidates would be the 76 console PGM amp, and the separate rack mount versions of same.

A lot of these were cloned just using radiola III transformers for almost everything.

 

[CJ]

been checking the rewound transformer,

inductance was 300 Henrys for both pri and sec, and held steady as frequency went up, due to the stacking,

i take back my statement on the core being steel after the readings with the full amount of turns, and the comments upstream,
probably 50/50, inductance would go way up with a lap 1 core,

i used #37 and #42, got it to fit but left out the screens,

frequency response was not the best, starts dragging current at 20 KC,

an adjustment to the coil would be to try #38 and #43 to cut down on capacitance (smaller wire, smaller capacitor plates) and put more space between sectors, use the screens this, time,

this should bring it up to 40 kc,

with the high turn count, i doubt that the original had stellar freq response, maybe they wanted to limit frequency and level,

would be nice to know response of a working innerstage transformer, a 27 K resistor on the sec helped dampen the response,

dcr with #37 pri wire was about 500 each coil so 1 K series,
dcr for #42 sec wire was 3 K each coil so 1.5 K parallel

dcr on sec went up to 3.2 K per coil once installed on the core due to digital ohm meter not dealing well with the core,

 

[CJ]

dismal sq wave response, never seen cross over distortion in a transformer,

i am a failure,