Destructive Analysis of a DR35 Coil Failure

Custom Search

since 03/8/2014

Recently one of the list members suggested that I look at the GM DR35 coil because it has spade lug connectors instead of the more difficult to use sockets of the DR37T. So I ordered a couple from www.rockauto.com.

Last night I wired one up, put it in the oil, set the amperage setting for a modest 3 amps @100 volts AC in (about 90kV out) and loaded it with my "corona dummy load". The purpose was to run it for awhile to make sure the insulation could withstand the constant over-voltage and to see how durable it was.

About 20 minutes in, the output suddenly stopped and the driver went to a high pitched rep rate. The primary waveform on the scope looked OK, only attenuated. I held a ground wire near the corona load and got maybe 1/4 of spark. It was becoming apparent that there was a turn-to-turn fault in the primary winding.

I removed the coil from the oil. It was cool but with the power off, it quickly got too hot to touch from heat soaking out from the interior.

The first step in figuring out what happened was to re-measured the good DR35 and the DR37T and see what the electrical differences were. I got the following results

Value	DR35	DR37T
Primary inductance	7.27 mH	5.95 mH
Secondary inductance	33.6 H	28.6 H
Primary/Secondary Ratio	68:1	69:1

Based on the similarities of the AC characteristics, I started thinking one of two things:

The coil was wound with cheap aluminum wire or
The actual windings were very similar in size.

I had intended to set up a Kelvin bridge and measure the respective coil resistances but instead this morning I decided to dissect the coil and do a destructive analysis of both its features and its failure mode. What I found is very interesting.

Here are the two coils side by side. The DR35 is much larger than the DR37T and it has a higher high voltage tower. Looked promising.

This is the DR35 waveform operating with the corona dummy load. Significantly cleaner than the DR37T. One thing I did notice was that the flyback pulse height did not vary much with input amps variations. It was acting more like transformer than an induction coil.

Here is an overview shot of the coil after I took the Porta-Band to it. The interesting parts are labeled. Click on the thumbnail for a high resolution shot. A burnt odor emitted from the saw as it cut into the primary.

The important thing to notice in this picture is how thick the epoxy is and how long the thermal path is.

What this looks like to me is that they took basically the same winding structure as is in the DR37T and simply molded more epoxy around it to make it look larger. Typical Chinese trick.

Not evident in this photo but there is very little space between the epoxy assembly and the core so very little oil can flow through the inner space to cool from the inside. The DR37T has considerably more space.

Here is a detail shot of the secondary. Notice how the secondary is wound on a sectional bobbin. That is designed to limit the voltage stress on any one group of windings. Notice the very thick insulation (thermal and electrical) between the primary and secondary. And finally, the large amount of space between the edge of the sectional bobbin and the outer surface. A lot of useless filler epoxy.

The next step in this evolution is to order a DR37T from www.rockauto.com and a DR35 from NAPA and then X-ray them to see how they're constructed.

The RockAuto coil is made in China while the NAPA coil is made in Poland. That may be significant, as I've been impressed with products made in Poland.

My fear is that we're just going to have to bite the bullet and pay for NAPA quality.

-----

---