Triple build wire vs. insulation tape? 4

[MagicSmoker]

Is it possible for an SMPS transformer to make it through hipot/safety agency approval testing using triple build wire on all of the primary and secondary windings rather than the usual tape in between each winding? I could really cut down leakage inductance by multifilar winding the transformer with triple build wire rather than separating the two sides with tape.

Now, what if the transformer is to be used in an isolated DC-DC converter for electric vehicles, where the traction pack might be 300-400V? And which UL standard might best apply in this situation, anyway? Surely not 583, as this isn't a forklift?!

Thanks for any pearls of wisdom that may come my way; less thanks, but still an appreciation of any rotten tomatoes.

 

[MagicSmoker]

A followup of sorts, despite the underwhelming response to what I thought was a pretty good question...

The UL document covering my application hasn't been finalized - when it does come out it will be UL 458A.

The document that comes closest to being relevant is UL 2202, but even it is centered around line-operated equipment (ie - chargers) and at least in that case it doesn't seem acceptable to substitute triple/quad build wire for the usual wrap of tape between primary and secondary.

At any rate, I couldn't waste any more time on this crap so I wound a prototype transformer myself in the usual fashion (except I didn't bother to interleave the windings to reduce proximity effect) and the performance was fine as-is. Measuring the leakage inductance got me the correct values for the RC damper across the primary the first go and even though it would be nice to tighten everything up by ditching the layer tape, it obviously isn't critical. More a curiosity at this point...

 

[MagicSmoker]

Thanks to whomever gave me a star... I guess at least one other person is curious about this, too.

One last ditch effort to coax a response from the eng-tips crowd...

The transformer in question is for a 12V to 24Vx2 dc/dc converter. The secondaries have to float w/r/t each other at up to, say, 400V but otherwise, the transformer strictly deals with "SELV" voltages. Nothing is connected to the AC line so UL 5085-x doesn't seem to apply.

Thus far my approach to this is to buy the standards I *think* might apply from

http://www.e-standard.org

("seems legit" being the operative meme, here...) and then reading through them until my head turns to mush from all the bureaucratic legalese, only to find out I need to buy some other standards.

Suddenly I understand why compliance/safety/testing consultants make such good money...

 

[electricpete]

I'm not very familiar with the type of equipment you're describing (small transformers). I'm somewhat familiar with large transformers, and most familiar with motors.

You're trying to get rid of the major insulation (between windings) by beefing up the minor insulation (between turns). Since the major insulation withstand requirement is so much more than the turn requirement, I'd think that would result in bigger windings.. which might might counteract the effect of less interwinding-spacing and make you end end up with higher leakage reactance (the opposite of what you were trying to do).

Also, I am under the general impression that film (enamel) insulation on conductors has generally good withstand strength but can be non-uniform such that there are tiny pinholes spread at various places along the conductor... not a problem for turn insulation as long as the pinholes don't line up. Might be more of a problem for the interwinding insulation, since the voltage is higher which can increase the distance that pinholes need to be separated, especially if contaminated. There may be difference in mechanical characteristics of film vs mica-paper tape as well.

Just some thoughts fwiw (since no-one is chiming in), I may be way off base.

=====================================
(2B)+(2B)' ?

 

[electricpete]

Here's what I was picturing when I suggested you may have increase in leakage reactance: core form transformer with lv winding forming a cylinder nearest the core, hv winding forming a cylinder outside the lv. The farther out the hv winding is spread from the hv winding, the more flux (from outer filaments) flows in the the volume occupied by the hv winding and is not linked by the lv winding... increases hv winding leakage reactance.

You mentioned multi-filar winding...I'm not very familiar with that. Certainly may change the conclusion.

=====================================
(2B)+(2B)' ?

 

[MagicSmoker]

Thanks for replying, electricpete.

I get what you are saying about possibly increasing leakage L from substituting wire with thicker insulation for insulation tape between windings. This is why if you are going to use triple/quad build wire you have to wind all of the windings "multifilar" (ie - twist all of the wires together for each winding then wind all of the windings at once). This construction method is commonly used for toroidal gate drive and small dc/dc converter transformers as it results in the lowest possible leakage L, but it is not entirely clear to me it is acceptable for my particular application. Then again, it does not seem to be entirely clear to my preferred custom transformer vendor, either (hence this thread!)

At any rate, I went ahead and bought a couple pounds of #28 quad-build wire from MWS as they had some in stock from a run they did a few years ago. The insulation is solderable polyurethane-nylon - my favorite kind for prototyping, just dip the end in a solder pot to strip it - and is NEMA 83C spec, so supposedly capable of withstanding 5000V. I'll wind up a transformer with it and see if it performs any better in the prototype and if so, I'll hipot it at twice the voltage I think I have to withstand (which is 1.5kV) and see if it holds up.

 

[electricpete]

NEMA MW-1000 calls out requirements for magnet wire. 83C is a designation defined within MW-1000.

MW 83-C - POLYURETHANE OVERCOATED WITH POLYAMIDE FILM INSULATED ROUND COPPER MAGNET WIRE FOR SOLDERABLE APPLICATIONS, THERMAL CLASS 180

INSULATING MATERIAL The conductor shall be coated with a dual film. The underlying insulation is
based on polyurethane resins (1.4.2). The superimposed coating is based on polyamide resins (1.4.2).

The performance characteristics are defined, including “HVDC continuity test” of section 3.9. That is an interesting test…. how you would actually test the insulation of a whole spool of magnet wire?. They do it a very logical way: The conductor is mechanically fed through grounded brushes while conductor is energized at dc. When any current flows (indicating a tiny defect in the insulation), it is recorded as one defect. Does a defect mean failure? Heck no. For @28 quad build magnet wire, NEMA MW1000 specifies the maximum number of defects per 100 feet is 2,000 (!). That is a little bit disconcerting at first. But again in the context of typical application on turn-to-turn insulation, even with this large number of defects the chance of two defects lining up next to each other is miniscule. As voltage increases if contamination is present, the chance gets larger. Clearly this is not suitable for ground insulation… I’m sure you have some other insulation to protect the conductors from contacting grounded core (it is obvious from mechanical considerations if not electrical considerations). My only point is that magnet wire insulation is fundamentally different than tape insulation in the allowance of finite number of defects for length. May or may not be an issue for your application…. You know better than me.


=====================================
(2B)+(2B)' ?

 

[electricpete]

Whoops, sorry. 2000 was the test voltage for that configuration.

Maximum Fault Count per 100 feet is 25 for single-build, 10 for double-build, and 0 for triple-build or quad-build. That should make you feel better about quad build.

=====================================
(2B)+(2B)' ?

 

[electricpete]

Yet another correction. The previous post gave allowable fault count for aluminum magnet wire.

Here is for copper:
Maximum Fault Count per 100 feet:
single-build: 15
double-build: 5
tripe-build: 3
quad-build: 3

So the triple and quad build copper are not as good as the triple and quad build aluminum for some reason. Your configuration copper quad build would have 3 faults allowed per 100 ft.

=====================================
(2B)+(2B)' ?

 

[MagicSmoker]

*argh* No wonder no one knows the answer to whether it is okay to mulifilar wind all of the windings together in one shot, or if they need to be separated with tape!!!

Even though it is extremely unlikely that two pinhole faults would be next to each other in the separate windings in a multifilar wound transformer, I don't think I want to take any chances with botching the agency testing over this. A bigger snubber resistor and/or a higher Vds rated MOSFET to deal with the higher leakage L from using inter-layer insulating tape seems a lot more palatable.

Thanks for digging up the information on the magnet wire testing - I hadn't really thought of that before and even though it doesn't specifically answer my original question it gave me enough support to decide not to go with multifilar windings.