Inductance of wirewound power resistors

[dick_freebird]

Working on a test jig to take a 400V GaN power device
to the cyclotron for single event survivability and spurious
drain current testing. This voltage swing needs to be fully
exercised, but my 'scope probes say 300V, yadda yadda.

Between a resistor divider and a pulse transformer, the
divider looks cleaner if a bit slower. Transformer has the
nice property of being a non-dissipative load but what
we specifically want, may be longer lead time than
remains to the trip.

"On" power level says I need 150W resistors and some
values plus high pulsed duty will put me in wirewound.
Thick film values are very limited at this power level.
Looking specifically at TE Connectivity (looks like used
to be Ohmite?) FSC150 range.

My interest in waveform fidelity wants board C and
resistor L minimized. TE Connectivity datasheet says
"low inductance" but no quantitative info found in
the 4 pages of Google hits from "HSC150" "inductance"
"resistor". Only thing I saw was one thesis where the
MOSFET load was ringing pretty badly, and wirewound
resistors were used.

Couldn't find anything quantitative on TE Connectivity
(was Tyco Electronics) web site.

What I'm looking for, is any kind of reasonable rough
inductance value for a 1Kohm 150W like this guy (one
of the other divider values):

https://www.digikey.com/en/products/detail/te-connectivity-passive-product/HSC1003K9J/5587209

Anybody ever measure, or see a number for this
kind of wirewound?

 

[D.A.(Tony)Stewart]

These are bifilar wound so L cancels out except for the length between terminals so based on wire aspect ratio assumptions I expect 6 nH/cm

 

[aliarifat794]

For high-power dissipation, consider using multiple lower-wattage resistors in parallel. This can reduce the inductance and improve thermal management.

 

[D.A.(Tony)Stewart]

10 of 25W 10k in parallel with daisy chain ESL and ambient rise? Not much gain I think. But if you twist minimal equal lengths of AWG 24 then you can reduce ESL in connection. Then consider ESL/R is in picoseconds!

 

[KlausST]

Anybody ever measure, or see a number for this
kind of wirewound?

I can´t give values for the mentioned resistor.

But years ago I measured a 5W resitor in ceramics ... those 20 x 5 x 5 mm THM ones with leads on both ends.
If I remember right it was a 47 Ohms one ... but at 1MHz they were far from 47 Ohms.

Klaus

 

[D.A.(Tony)Stewart]

With 1k and 100ns step response, HOW would one define low ESL? by HF BW?
ESL= Tau*R = 1e-7 * 1e3 = 1e-4 H = 100 uH

YOu measure it with a tuning cap. Winding methods and materials for low R and high R vary a lot.

Wirewound Resistor | Resistor Materials | Resistor Guide

What is a Wirewound Resistor? A wirewound resistor is an electrical passive component that limits current. The resistive element is an insulated metallic wire that is wound around a core of…

eepower.com

 

[FvM]

Arcol is specifying detailed inductance values. The have low inductance and standard types.

 

[D.A.(Tony)Stewart]

Good find Frank. .My estimate of 100 uH wasn't far off but these are the std R's not the low ESL

--- Updated Friday at 7:23 PM ---

Low ESL concrete resistors are custom.

A better choice is the ceramic power resistors

Tubular Ceramic Resistors​

Non-inductive, high voltage, high power tubular resistors are available in a wide range of standard sizes, ceramic materials terminations and mounting hardware. Bulk construction advantageously produces an inherently non-inductive resistor; and it allows energy and power to be uniformly distributed through the entire ceramic resistor body – there is no film or wire to fail.

 

[FvM]

Rereading post #1, I want to contradict the statement about thick film resistors. You get planar thick film resistors with kV voltage and multi kW pulse power rating, e.g. from Vishay or TE.

 

[Easy peasy]

put some zeners across the Gan ( lots of ) to limit the peak spike - this allows a wider choice of resistors and wiring L.

 

[D.A.(Tony)Stewart]

or choose the non- ESL ceramic resisitor

 

[mtwieg]

I'd expect chassis-mount film resistors to have far lower inductance than equivalent wirewound inductors (including "non-inductive" wirewound). But actual data on inductance is difficult to come across.
Vishay makes many series of film resistors, and uaully provides a max inductance spec. Like <100nH for LPSxxx and RCH series, or <50nH for RPSxxx series)
Riedon PF2700 series specifies 14nH (suspiciously low, actually)
Ohmite doesn't seem to provide any inductance spec

Lack of available resistances is annoying, expect to combine several resistors in series/parallel if you need a particular resistance. Might be better off using many DPAK or TO220 packages.

I have some non-inductive wirewound resistors at work (either TE or ohmite), maybe I can try getting ESL measurements on them. What sort of resistances are you interested in?

put some zeners across the Gan ( lots of ) to limit the peak spike - this allows a wider choice of resistors and wiring L.

Perfectly fine for protecting the FETs, but the OP mentioned "waveform fidelity" so ESL is still a problem.

 

[Easy peasy]

Perfectly fine for protecting the FETs, but the OP mentioned "waveform fidelity" so ESL is still a problem.

Is this an issue if you are switching GaN fets ? the Vds is either high or low.

 

[dick_freebird]

Is this an issue if you are switching GaN fets ? the Vds is either high or low.

Looking for spurious drain pulses in a realistic(-ish)
switching setup. The paper I saw showed ringing tails
persisting for a large portion of the flat-top. Hard to
think how I'd trigger and count off that mess, other than
full scale restitch events (which are possible, we think,
with the integrated gate drive). Still we want to count
and quantify amplitude of lesser, intra-final-stage
spurious or definitively declare the absence (with
data).

We do have "DC" jigs but that is not the predominant
application and everybody will ask if I don't show the
as-used test case.

 

[Easy peasy]

Spurious drain pulses are current - where the Vds goes down - so zener clamps do not interfere with this.

the usual way is looking at a ( often common ) shunt resistor - to pick up spurious drain pulse events - i.e. those not linked to intended gate pulses.