Peerless S-217-D Pultec EQP1a OPT Redesign

[jackies]

Nice ringing! Can't you just damp it enough with even more serious azz-zobel?
What exactly is "poly tape", like, electrical tape?
I used some capton tape, is that overkill?

 

[CJ]

3M #56 or similar,

3M #27 cloth for taping down lead wires,

 

[CJ]

3M #27

 

[CJ]

3M #69 is a cut above #27,  takes more heat,

Kapton has it's place, like tesla coils and DIY tazers for that neighbor who lets his dog go on your front lawn, 

 

[jackies]

CJ thanks for your kind reply!
I will order some and try, whenever I attempt to wind something next time!

 

[AusTex64]

What's the latest on this project? Thanks!

 

[CJ]

abck on this project tonight,  going to rewind thye coil with  screens to try and reduce phase shift on the fb wind,

even 30 degrees is enough to cause trouble as there are other reactive elements in the circuit (caps)

so the fb wind has to follow the pri wind, this means we need to limit capacitance between sections which is the point of the screens,

 

[CJ]

decided to try it without the screens, they are a pain, xtra hookup,

we have sec-pri-sec now, FB wound with pri, bifi style, maybe a mistake,

sec is now linking up better, more Henries,

turns - 300T-1340T-CT-1340T-300T  60 +60 fb

36 H sec, 650 H pri

flat to 60K, Bessel roll off at 400K

will measure shift on fb today

 

[CJ]

still oscillating, 180 phase reverse from fb wind at 120 kHz with this latest wind,

so we add screens and go back to pri-fb-sec coil struture,

that way we only need two screens, with sec-pri-fb-pri-sec we need 4 screens, not enuff room for the extra tape and wire screens,

but wait, why not make one last ditch effort to avoid screens by using reverse winding?

lot's of permutations, rev wind pri, rev wind pri and sec, rev wind pri and fb, rev wind sec and fb, rev wind pri , sec and fb, rev wind fb,  rev wind sec,

6 ways to win, start with the easiest first, rev wind fb,

going to keep the sec-pri-fb-pri-sec stucture as sq wave ringing was dramatically reduced over the pri-fb-sec structure,

 

[CJ]

we have a new coil wound,

putting the feedback wind first, then the secondary, then the primary, then the other half of the secondary which is reverse wound for more inductance and balanced capacitance,

this puts the fb wind away from the primary, if we use the start of the first secondary as our ground lead, we will have an earthy wind of big #32 which will be a nice juicy screen between the pri and fb in order to reduce C which should improve phase shift.

if this does not work, we will use fb-sec-pri next, adding even more space between the critical fb wind and the primary.

 

[CJ]

this version sounds pretty good, we have 723 H pri ind at 20 Hz and 5 v-rms
we laced the core lap 3 to provide a little extra headroom,
sec ind is 34.5, so 723/34.5=20.9
root of 20.9 = 4.57 which matches our turns ratio down there at 20 Hz, so we have balanced linkage of the pri and sec winds,

still had some osc problems, but this could be due to the amp layout, noticed the lack of the 20 puff cap across the 12ax plates, when we added some C the osc went away, will test on the scope tomorrow,


here is the schemo so you can see the cap, chassis wiring might be enuff to create this 20 pf depending on your layout Pultec must also have had osc problems thus the cap,

 

[CJ]

alright this geometry looks good, good to 100k with no pri-fb phase shift,

here is the 100k dual trace shot.

 

[CJ]

sq wave looks pretty good too,

non linear regression, don't know why,  different from y(t)= A*e^-t * cos (wt + theta)        (what?  )

 

[CJ]

fb wind finally flips at about 300K, but is off smaller amplitude which is good, most phase flips occur at a rez peak many times the orig voltage,

what does this mean? the amp should be stable as the 300K positive fb is a lot lower in amplitude,
and the higher freq at phase flip means we can filter this out with a smaller cap which will mean more fidelity,
also noticed i had no zobel on the tail end of the HS-29 (goes across 12ax grids), so adding that, and the 20 pf on the plates should cure any osc problems,

we used fold back winding on the fb to minimize C, this means wind one layer of 71 turns then coming back across the layer to start the new layer,

this OPT also has the exact same turns as the S-217-D, done by not using interlayer insl on the 3150 turn pri which means more winding room, freq still flat to 100K,

 

[CJ]

here is a winding print, no DCR,  left it at work, will check later,

Henries go up with level,

 

[jackies]

Great work CJ, I didn't quite understand, did you try this last design with fb winding first in the actual circuit yet?

 

[CJ]

yes, it sounds great in the circuit, had to put a cap across the 12ax7 plates ut that may just e the circuit layout, so tonight we will lay out the circuit differently to see if we can get rid of the cap and still be stable,

 

[CJ]

we are going to put this thing inside a UTC mu shield and fill it with wax,

those nested cans found in the 100X are not really mu metal, but some blend of metals that are cheaper but work well enough, there are three of them nested inside each other on the stock input, we will just use one,

we taped the core up with glass tape so the wax does not bond to the lams and inhibit magneto-restrictive forces,

 

[CJ]

we tape the bottom seam of the can to keep the poly-crystalline wax from running out,

that is the crock pot next door, looks like a DuKane 2A3 can or something,

 

[CJ]

all done with the pour, took about 5 laps, this stuff sets up like concrete, you can still find it, searching micro-crystalline wax would probably get you more hits,