Re: RM503 ongoing repair saga T620 UNWOUND.



Everything I have seen is consistent with what you have said here.

But for the gap size.

I went back and re-measured because I did quite a bit of cleanup after the
first set of measurements.

The 503 core that I have the two sides are not the same.
One half all the legs, left/middle/right are equal length.

The other half the middle leg is .013" to .015" shorter. Which is less then
the .018 reported from my earlier not as clean test, most likely from epoxy
on the two left/right legs that have the other epoxy that held the .020
plastic spacer in the outer two legs, the center was fairly clean from the

That would match your measurements, if you mean each core half middle leg is
ground 0.007" lower and the total gap with two matching sides in there
actual position would have a total center gap of .014"

When I did some tests on the core with the saturation tester you can push
the two together and really change the curve if they are not very tightly


-----Original Message-----
From: TekScopes@... [mailto:TekScopes@...]
Sent: Thursday, March 26, 2015 11:03 AM
To: TekScopes@...
Subject: Re: [TekScopes] RM503 ongoing repair saga T620 UNWOUND.

There is something that needs to be said to all future
transformer hackers regarding epoxy:

First, in the failed brown epoxy, it is soft always, but
when heated to 130C, it becomes like chewing gum. It lacks
much cohesive strength. It is easily scraped, and picked
out of many things.

Second, not all epoxy is failed brown epoxy. The epoxy
tek used to glue the two outer core legs on the 647 core
is really good epoxy... unlike the brown stuff they potted
the coil in. 130C won't even faze it. You have to heat
the good epoxies to 150, or even 160C to soften them to
the chewing gum stage... I have seen some that needed
even higher temperature, but fortunately not on tek

If the core halves are bound with fiberglass tape, the
epoxy is going to be the bad sort, and 130C will do. If
there is nothing binding the edges, it is good epoxy, and
150-160C will be necessary.

A nice way to separate the halves is to make 4 wedges out
of fiberglass PCB material, that taper from 1/2 inch to
5/8 inch over 4 inches, or so. Wedge them in pairs into
the slots on either side of the winding... be firm, but
not aggressive. Pop the transformer into an oven set to
130C, or 150C, and walk away for 15 minutes. When you
come back, there will be a pile of core pieces, and
coils and wedges. I hope you put something under the
transformer to catch any stray parts.

After you have separated the halves, without breaking
them, you need to take a knife blade and scrape the epoxy
from the joining surfaces. The knife won't harm the
ferrite, if you take it easy... and there is no reason to
work very hard... if it gets difficult, pop it back in
the oven for a few minutes.

The gap can be easily measured on a clean core by
assembling the core EE direction, aligned like my printed
text... legs to back. And using a feeler gage to measure
the gap. In every core I have seen, from the late 60's
forward, it will be 0.007 inch, which is 0.18mm.

[Old waxy 545 transformers can be ground on no legs, with
tape gap spacers, or ground on one core half with the
full gap, or ground on both core halves, each with 1/2
the gap.]

If your gap isn't rock solid stable, none of your measurements
will be stable. That is because the act of the gap not
shrinking as the magnetic forces increase is what stores the
energy in a gapped inductor.

Fortunately, the only really critical thing in these
transformers is the HV insulation techniques you use in
making them.

In the early days of experimenting, I used to wrap a cobby
old piece of electrical tape around the outside of the core
to hold the halves together while I was testing the winding,
and sometimes as the core heated, the tape's adhesive would
let go, and the core would flop open, and yet it still worked.

-Chuck Harris

'John Snyder' Kochcal@... [TekScopes] wrote:

Any numbers would help me because the 547 winding spec is available.

And by the way when I say .020" gap I should be saying .020" additional

For the 503 core I have not measured the center gap directly with the two
halves together(which I should)

But one half of the core has a center leg of 0.612" and the outside leg is
0.631 on one side and 0.628 on the other side. As clean of epoxy as I

The other half the leg is .630 on all legs from top to bottom.
So a ".020 gap core" as I have been saying would have three gaps of about
.020 right side, .38 center and .020 left side.
I measured the no side gap core configuration(which would be about a .018
center gap) with the same 16 turn 26 gage coil test coil at a saturation
about 9.5A and 110.8 uh (no resistance comp math)

And on the GR1657 at 1KHz Series .1080 mH with a Q of 4.4
And at 1KHz parallel .1135 mH with a Q of 4.4

With a DC resistance Rdc of 0.1496 ohms.

And I guess that is the measurement I would like to get some other
for which I think your numbers could be turn adjusted for.

My GR1657 is repaired and not calibrated, clearly reads in the right range
based on a bunch of inductors, but I have no real calibration "standard"


Join to automatically receive all group messages.