RCA 86A sidechain transformer?

[bobschwenkler]

Two bad transformers at the same time! I just replaced the interstage with a Sowter replacement and it looks like the SC transformer is also bad. I read that most any 1:1CT will work?

I still don't have much an understanding of tube rectifiers, anyone care to explain how this rectifier/transformer combo is working here?

 

[CJ]

current flows from hot to cold, but not from cold to hot, thus rectification,

transformers provide the ac signal that gets rectified,

 

[plumsolly]

The Collins 26-C uses a 20K:80K in that spot, with the same rectifier/triode tube. I emailed that manual to you at your website email.

Best,

Ben

 

[bobschwenkler]

Haven't gotten very far on this yet... My transformer knowledge is pretty lacking, do I need to be specific about its specs with the B+ running through one side of it? I remember reading that only gapped transformers should have DC put through them. Or something like that?

Or should I just find any 1:1CT that'll handle the AC signal level and DC voltage?

 

[CJ]

yes, dc xfmr needs a gap,

you have a schemo of the 86a?

 

[plumsolly]

Attached the schematic. I believe you also need to consider the DCR of the primary as that, along with 152.9ohm resistor, provide the load R for the triode section.

Best,

Ben

 

[bobschwenkler]

Anyone have any ideas where I should start looking or what production (or out of production) model might be a suitable replacement?

Hope I'm not asking too much, I appreciate the help.

 

[CJ]

ok, it says T-4  K-901075-502  RCA Stock Number 43580

DCR PRI:2070  ohms
DCR SEC:2070 ohms

this tells you right away that is bi-filar wound, and has a ratio of 1:1.

so now you try for the bobbin size and wind fine wire till you get 2070 ohms,

there is a splice on the sec that must be brought out for the CT,

from emmr, king of the boat anchors we have:

It's a 1:1, one side center tapped, DCR on each side of 2070 ohms.  Any quality 1:1 interstage with a CT on one side should work as a replacement."

maybe another Sowter innerstage would work, checking for DC on the Ba6a,

ok, interesting side chain, not the typical dual rect/smoothing circuit, they have a triode/rect tube combo that feeds itself,

the choke from the power supply feeds the weird 152.8 ohm or what ever it is resistor, then the resistor feeds the xfmr, which feeds the cathode biased tube, so the pri needs to pass dc, i bet the Ba6a Sowter innerstage xfmr would have problems with that, no wonder the T4 round can is just as big as the rest of the iron,

also, there is a slight bit of dc on the sec, it comes from the zero adjust pot, it feeds the input trans CT also, they are tied together, but the 3.3 meg resistor keeps the current way down so not to worry there, it is mainly the pri we have to fix, a kraft paper gap on a 625 EI ,

47 K resistors across the xfmr, so 100 K total Z at our standard center frequency1000 hz = how many turns?
2piFL=6.28*1000*L=100K
6,280L=100,000
simplify since this is a freq dominated formula anyway,
6L=100
L=100/6 , 10/6 is a repeating decimal of 1.66666 so ad a decimal place and we have

17 henries.

seems low,but remember, we have to use silicon steel for the DC.
and we have a gap, which lowers the henries even more, so from Mag Met catalog, we have

L(henries)=0.3326^-8KN^2uAC

K is stacking, with a gapped core it will be lower that the typical value of 0.9 so we just leave it for now.

simplify 0.33^-8*Turns Squared*Perm-AC = Henries for 625 EI.

we know Henries, it is 17, so

17H=0.33^-8N^2uAC, where uAC equals permeability of the material,

what is perm for silicon steel? (permalloy would saturate with the dc)

about 400 to 500, use 500,

17H=0.33^-8*N^2*500

we are after turns so we can see if it resembles a 2070 ohm dcr transformer,

17/500=0.33^-8*N^2
0.034/0.33^-8=N^2
10,000,000=N^2
N=3200 turns.

that is for an un-gapped transformer

so we could probably double that to 7500 turns which with small wire, could add up to 2070 ohms.

you can find the mean length turn for 625 EI square stack bobbin,
then figure feet of copper and find the wire size,

7500 turns seems high to wind bi-fi, the hi-freq loss would be horrendous to to Leackage C from the 2 wires being wound together, API does it but only uses 287 turns,

so now me thinks this is a dual coil job, which would help with the high dc voltage,
so now we re calculate using a UI lam, or maybe even a C core, :-\







the reason your xfmr is bad is because of all the dcr and plate current, makes it get hot, and those things were on 24/7.