Peavey 5150 Output Transformer

[CJ]

more trouble inside the heavier sec, cameras do not focus well on copper wire for some reason,

 

[CJ]

first crossover ever seen on wire this big,  bi-fi middle wind,

looks like #16, from a strictly eyeball point of view,

 

[CJ]

inner primary,

here is the trouble spot right here, maybe a nick in the wire during winding or a small defect in the wire coating, probably a shorted turn by the looks of the burned paper,

rewind with some Essex triple build and we will be able to use 10 stomp boxes in front of the 5150 without arcing out,

 

[CJ]

layer under the shorted turn got scorched a bit, but is is just carbon from the paper, the wire was not damaged,

but we do have more crossovers, this time on the primary,

 

[CJ]

done with the unwind, first layer is 32 turns which bumps us from 8 to 16 ohms,
can we estimate the total sec turns?

well if we know that turns and Z are a quadratic function,

then 8 to 16 ohms = Z ratio of 2

turns will be the root, so 1.414

we have 32 turns increasing our Z by a factor of 2

so 1.414 times our 8 ohm tap turns will give us the extra turns needed to make 16 ohms,

algebra can now be implemented,

1.414 x Y = Y + 32  where Y is our 8 ohm turns,  solve for Y,

1.414 Y - Y = 32

0.414 Y = 32

Y = 32  / 0.414

Y = 77.2947,  not partial turns allowed so we expect our 8 ohm tap to have 77 turns

so total sec turns are 77 + 32 = 109

cross check:  109/1.414=77

what will be our 4 ohm tap turns?  77/1.414=54.5, so it s/b either 54 or 55 T.

we already have the turns count, lets check the theory against the actual and put it on the print in a few minutes,

here is Layer 1>

 

[CJ]

core tube shows no sign of collapse, this is a good thing,

 

[CJ]

here is the basic diagram, reminds me of the Peerless  16431 in some ways,

would sound killer in a home stereo if stacked with good iron,

here we don't even have .018" but .025" barn roof steel,

was wondering about the low Henries, maybe they wanted that classic old school sound,

or maybe they wanted it on the cheap,

 

[CJ]

here are the winding instructions with every layer shown,

we have some reverse winds but they are only one layer so all you do is swap the start and finish and exchange the lead colors,

this is a very well engineered transformer, with a few character defects from the manufacturing process

better DCR balance could have been achieved by alternating pri sections, but this miught have upset the great response,

 

[CJ]

ok we found a bobbin that fits a 2 inch stack of 138 EI, had to cut the center divider out,

changing winding structure-

16 ohm wind eliminated as no one uses 16 ohm cabinets
this will provide more room for insl which means a simple coil due to less leakage C,

we recycled the heaver gauge sec wire and used new primary wire,

increased turns from 1348 to 1514 fro more pri inductance(bass)

 

[CJ]

getting ready to test, stacking 25 mil I bars is a treat,

lams did not bend as much as 14 mil stuff does when you do a tear down,

old lead wires used, 16 ohm org left out, more room to work with,

 

[CJ]

increase in inductance shown below, remember that L increases as turns squared,

so adding 163 turns to a 1348 wind gives us a 35 percent increase in Henries, something this core steel  could certainly use,

structure uses heavy 16 ga wire first, this is so it will not cut into pri winds,

this also lowers DCR as the MLT is smaller, 8 ohm wind put right on top of this,

so wind structure goes from 5-4 to pri sec,

what will this do to freq response?

well since we used pre bent sec wire which does not lay as flat, and added extra pri-sec insl (more room by leaving out the quad wind 16 ohm section) we now get 10 Hz to 120 K Hz, wtf, over?

so much for rolling off the hi end, and pri DCR is more balanced, 61 ohms and 66 ohms,

so we get a more reliable transformer with better response and less sec DCR simply by removing the 16 ohm section,

going to bake and dip in some varnish and plug it in, hopefully no more hi-freq osc and volume drop out,

 

[CJ]

here is the rewind print,

with the nylon bobbin we can run out our margins, so we squeezed 17 turns of the big bi fi wire per layer, if we use 51 turns instead of 54, we go from 4 big layers to 3, more room for insl, less C ,

combine reduced sec turns with increased pri turns means we go from 2200 plate load to 3300, increased tube life due to less load, a welcome treat for modern tubes that sell for 17 bucks each,
and flux goes from 18 KG to 16 KG which might clean up that nasty 6 stage preamp a bit,

our core perm came down a bit, this means we have more leakage inductance, due to the simplified winding structure, we went from 4 pri sections to 1, and from 5 sec sections to 2,  and the pri is further away from the center leg which is now occupied by the speaker wind, but we might have more sec inductance so it all evens out,

we have a slight rez peak at 170 K Hz, this is due to the changed coil structure, a small trade off considering the ease of winding, 3 sections instead of 9, and we save two pri spices,

 

[trobbins]

Did you confirm along the way that there was a primary winding break - via resistance test?  Or was the likely issue from arcover between layers or between turns where the insulation was damaged?

When you reduced the margins because of the new former, did the creepage from primary to secondary reduce, or could you keep that the same with margin tape or wrap around insulation?

Did they noticeably pad the insulation layering to reduce interlayer capacitance?  I think it was standard not to use impreg, so as to minimise capacitances.

 

[CJ]

that primary wire damage was like when you break your leg but the bone does not stick out like Joe Theisman, a fracture, but not a compound fracture, thus the wire breaking apart during unwind, could have been arcing to the previous layer, when the overdrive channel was used, the volume would drop with the knob set past 5, so maybe that is when the arcing, if any occurred, we will find out when we put the transformer back in the circuit weather we have solved the problem,

we used a nylon bobbin, so leaky margins are not a concern,,

they had a tough time getting all that wire on this coil, so there was never more than 1 layer of clear poly combined with 1 layer of kraft.

 

[trobbins]

CJ, this is like a detective novel - sorry to bug you with queries:

In the region of the disintegrated kraft paper (trouble.jpg), do you recall seeing any damage to the clear poly or the winding layered on top?  Is the winding layer shown in 5162.jpg the same layer as showing through the brunt kraft in trouble.jpg, or the layer underneath?

With respect to creepage distances, I think what you are describing is that the original windings would have had a creepage distance between primary and secondary windings that is larger than what you have now wound (because you reduced the margin distance on each layer).  But the new winding will have a very large creepage distance between primary and core, whereas creepage could have been quite short in the original (depending on how well that clear poly was positioned).

 

[CJ]

yes when that wire gets pushed outwards, it gets dangerously close to the core, which is grounded, without the nylon bobbin, it can arc over to the core,

there was no poly insl between layers, only paper, Nomex would be good in a coil that  has to handle all the artifacts generated by the crazy gain in this amp,

we strap each pri layer down with a layer of yellow tape in case the varnish does not get to every spot on the coil, and we add a 1/4 ich layer of thicker poly over that at each end of the coil to keep the ends of the layers from getting together,

you can get away from using varnish on line level stuff, but not power stuff,

Dennis Hoyer who winds Mac and Marantz coils tells of a story where McIntosh wanted him to wind transformers with no varnish so as to get better hi end, he told them they were crazy, but they insisted, he said OK, but would not be liable, well, they came back about a year later with snapped wires, a chunk of copper singing a song in the key of 10 K Hz will not last long,

 

[trobbins]

Hi CJ,  just trying to clarify for my own edification 

yes when that wire gets pushed outwards, it gets dangerously close to the core, which is grounded, without the nylon bobbin, it can arc over to the core,

The original appeared to have clear poly sheet between the edges of layers of the winding assembly, and the core.  Perhaps that poly sheet was not adequately acting as a solid insulation barrier, with some creepage to core occurring at the edges of the poly sheet?

we strap each pri layer down with a layer of yellow tape in case the varnish does not get to every spot on the coil, and we add a 1/4 ich layer of thicker poly over that at each end of the coil to keep the ends of the layers from getting together,

I think you are describing a layer of 'margin tape' - where the margin tape is somewhat similar in thickness to the winding wire, and buts up on one side with the end winding of a layer, and the other side of the margin tape is the outer edge of the total winding assembly.  Each layer of turns would than have margin tape applied on each side of the turns.