Marshall JCM 900 OPT 100 Watt Dagnall C3070

[CJ]

used tape and a poly wrapper, now we are down to scraping enamel off with a razor blade so we don't waste solder heating the leads, and we don't have to sniff that deadly wire insl. either, if it were true enamel like the old days then we would skip the razor and sniff all night,  ;D

5 turns outer wind , needed to get to the other side of the coil to the orig term site,

 

[CJ]

getting those quad leads sorted and tinned, still have to re-lam and do the bake and dip,

 

[CJ]

best part of the job, mindless lam stacking, let's see, 1 lam this way, 1 lam that way, and maybe some I bars for desert, 

 

[CJ]

stacking done, time to test this thing,

 

[CJ]

we ended up with 1000 turn pri instead of 810, kept the sec turns the same at 80,
so new ratio is 1000/80 = 12.5:1 for 16 ohm speaker load, used to be 10.5:1,

so Z ratio is 12.5^2 = 156.25,  so for a 16 ohm speak, the EL34 quad set will see

156.25 x 16 = 2500 P-P.  used to be 1764 P-P, seems low so the new ratio should add a little tube life,

inductance went up nicely even with a few lams missing due to varnish build up,

here is the chart>

 

[CJ]

frequency response was a nice surprise, do not know if the quad winds made it better as original xfmr was shorted so we could not test it,

we used 265 T on the first pri and 235 T on the outer to kind of balance DCR, which came out to about 21 ohms ea, measures 16 ohms now that the the coil is on the core, digital meters do weird things when measuring transformer ohms,

varnish will probably reduce the flat part down to 60K as it will add capacitance when it gets between the turns and layers,

should sound pretty good, we will install it back into the amp and give it the torture test,

 

[PRR]

> used to be 1764 P-P, seems low

A lot of EL34 data suggests a condition with 3.4K load per pair. Seems brutal but HAS worked for many amps over many years. Two pair 1.7K. Someone just copied book-values.

3.4K/pair runs the loadline through the fat part of the knee. 5K/pair will cut above the knee. Power and distortion may change. In general it may be a more "6L6 like" sound (since 6L6 is traditionally not loaded so hard).

That's if your load is 8.000 Ohms. With real speakers wandering 7 Ohms to 50 Ohms over the audio band, I don't think nominal 3.4K/5K is any big difference.

> varnish will probably reduce the flat part ... as it will add capacitance

That's the thing about LOW impedance.... capacitance matters less. What do you think, 1,000pFd? 1,000pFd and 5K Ohms is 33KHz, two octaves above any useful gitar sound. 1,000pFd is a wild guess, it may be much less.

I don't think capacitance governed design. I think the designer had some HIGHLY optimized dimensions for Williamson OTs (a Williamson can be mighty fussy), and just copied them into this lower-spec job.

 

[CJ]

good thing we wound this for 2500 ohms, just noticed that this amp originally came with 5881's, has  6L6GC  JJ's in there, wonder why they did not use a different OPT,

2 grand for a 6L6GC Marshall, new owners must be accountants,

RCA manual says a pair of 6L6 tubes want 5000 ohms P-P so 1000 turns should keep them happy,

your right, varnish won't make any audio difference, and i noticed that when the neg feedback resistor is hooked up, the high end drops off a bit, they had a 47 pf across the feedback resistor but after playing with different values, we just left it out, old Marshall's did not have this cap anyway,

plugged it in and set all the controls to 11, tried to break the OPT for 1/2 hour but no luck, it sounds great so it's back to the Peavey Roadmaster  to see if we can fix that, .

 

[reverbus]

You made my day CJ
I have  Marshall 9200 with one dead  C3070 OPT.
Someone did rewind it, but obviously without correct data.
So we'll do it right as you did.

Thank you very much to post your data and procedure.