Sprague KoolOhm

Nobody has responded to my post, so I'm emphasizing: Sprague's famous Koolohm molded power resistors were popular for decades, and were famed for performance other resistors couldn't equal.

I've been measuring the self-inductance of all kinds of wire-wound power resistors. Series inductance is the parameter that limits the usefulness of modern wire-wound resistors in many applications, from high- current pulsing to H-bridge monitoring to RF to vacuum-tube transmitter stages (oops)...

So far I've failed to find parts to equal these old beauties, within a factor of 10 or more in most cases, and so I wonder what's become of them. I'm asking what we're all supposed to be using now. Consider: the appealing thing about them was they were standard, readily available parts, not special-order Aryton-Perry windings with a high price and 1k minimum orders, etc.

I've got a few answers, but I really want to hear from ya'll out there first. Somebody knows what went down, and I'm eager to hear about it.

They're not in the same power league, but Caddock makes some thick-film power resistors that are flat up into the 5+ GHz range. The smaller ones, like 950 ohms stuck in an SMA female, make nice 8 GHz scope probes.

The MS series goes up to 22 watts.

John

Nice trick.

Thanks, John! Yes I know about some modern candidates resistors that are either low power, or rather expensive. I'm wondering about modern low-inductance Jellybean parts.

Both ohmite and sei/stackpole still offer NI variant wirewound resistors through distribution. Vishay have them too (NS NH from Dale).

Some types in the Digikey catalog Ohmite WNseries to 2W axial Vishay NH series to 250W plate-mounting.

I recon there must be better non-inductive alternatives..

Is it ayrton-perry or aryton-perry? Seems to be some confusion there. Always thought it was the first. BOTH are used in the Ohmite catalog.

RL

How about some numbers? What inductances are you measuring and at what frequency? Which resistors gave which numbers?

Big question, but I'll try to give some short answers. Most power resistors are wirewound types. Exceptions are wire-strip resistors with values below 0.3 ohms, carbon-composition resistors, some metal oxide types, and ceramic-composition resistors (rare). I eliminate common film types, etc, from this list, because they cannot handle high pulse-power or surge levels.

Unless they use a special winding technique, such as Aryton-Perry, wirewound resistors have high inductances compared to composition types, and typically become inductive above 5-15MHz (1 to 5W parts with values above 20 ohms) or above 0.5 to 5MHz (for values from 1 to 20 ohms). For example, standard 3-ohm wirewound resistors I measured became inductive as low as 430kHz (1.17uH for an Ohmite 10-watt part) to 2MHz (275nH for an IRC 3-watt AS-2 part).

A 5-watt Sprague KoolOhm part measured 48nH, or about 10MHz, which is 5 to 20x better than those others. A 2-watt 3 ohm carbon is 2-3x better yet, about 20MHz. A straight wire, bent and fixed in the same position as the 2W carbon, had nearly the same inductance, 14nH.

Well, I am finding some low-inductance parts, even in my own inventory. For example, DigiKey stocks Yageo's low- cost RSF metal-oxide series, but only in the 1 and 2-W sizes (Yageo makes them from 1/2W to 5W). In contrast, the KoolOhm resistors went to 10 and 20W or thereabouts. The cool thing about KoolOhm was that they were low-cost standard high-power parts, yet you got low-inductance Aryton-Perry windings, or the equivalent.

I'm sure somebody knows...

There is an interesting history of Zandman and Vishay ( named after his hometown in Lithuania) at

which covers a little of the demise of Sprague. Sprague was just getting into semiconductors at the time the founder R.C. Sprague retired in 1962 and his son took over. It was a disaster from that point on with the company losing hundreds of millions attempting to develop a semiconductor business. Sprague ownership then passed to General Cable and from there to Penn Central Railroad of all things. Penn Central eliminated everything except for the capacitor product lines. That's all that was left when Vishay moved in, looking to expand into the capacitor market, possibly unimpressed with KoolOhm design or market share, although that particular product may have been liquidated prior to Vishay's run to take over the entire world of component manufacture.

Yes, I have some of the WN series, sadly DigiKey carries only a few values, and the most interesting ones are out of stock. In the 1-watt and below region, there are lots of interesting surface-mount resistors. It's the higher power levels and 0.4 to 100-ohm region that's difficult, expensive, and hard to find in stock. KoolOhm, R.I.P.

$100 is too expensive (Newark) for a Jellybean resistor!

But I like RCD's non-inductive HPD series, in TO-220 and TO-247 cases, $6.50, etc., at Newark. Say, why does a 50-watt TO-220 power resistor cost more than a highly- sophisticated 50W power MOSFET? Sometimes it makes sense to rig a way to replace the power resistor with a MOSFET.

Let's keep looking.

Aryton-Perry, I think, caps because they were two blokes.

Le Wed, 07 Nov 2007 04:27:17 -0800, Winfield a écrit:

Vishay, NS010 10W, 50 ohm 1% is stocked at mouser. $3

Lots of other (unfortunatly unstocked) values.

--
Thanks,
Fred.

That sounds like a very useful part. Do we have the part number right, I can't get it to come up.

Nevermind, you said Mouser! I looked up DigiKey.

--
http://www.caddock.com/Online_catalog/power/Power.html

That's about all that's in an Agilent 54006A probe, $2500.

John

[...]

Do they do speaker cables too?

--

John Devereux

I think it's been quite a while since you could order non-inductive Sprague resistors off the shelf. NI windings were always special order.

Don't these wound things begin to look capacitive - Ayrton-Perry having starts and finishes wound adjacent to one another on a fairly large body?

It might be cheaper and more versatile to construct your own non-inductive power pattern using penny smd parts on FR4

I had samples once of porcelain-on-steel thick film resistors that you might look at.

You might arrange to get some kind of special terms, if you're ordering for an educational facility - good targeted advertising.

RL

Only the caps are left of Sprague. To me it all looked like a major corporate shriveling-up process. Maybe one of those 3rd generation downfalls.

I used to buy Dale for low-L resistors, they come in a heatsink mountable enclosures:

Look for N, NI parts that are low-L, maybe they are low enough for what you need to do.

--
Regards, Joerg

http://www.analogconsultants.com/

What do you mean, they "become" inductive? All my wire wound resistors are inductive from the lowest frequency, achieving a (very low Q) series resonance at some higher frequency, where they are capacitive for a while.

I don't understand "...measured 48 nH, or about 10 MHz...". Do you mean "...measured 48 nH, AT about 10 MHz..."?

Why don't you sacrifice (sob!!) one of your Koolohms and see how it's built?

Again, I don't understand. What is happening at 20 MHz?

Farnell stock as standard, the planar, non inductive, thick film types, BPC(Bi tech) and (Tyco) MCP. These are quite cheap and have proved surprisingly rugged.

Now, now, you know exactly what I mean. Low-value power resistors are well characterized as a resistance in series with an inductance. They "become" primarily inductive above f = R / 2pi L, where the inductive reactance exceeds the resistance.

No. A good LCR meter will show the same value for series inductance over a wide range of measuring frequencies (to be sure of the measurements, I do a quick check to verify that I'm measuring with a high enough frequency to get a sizable phase shift from the inductance, e.g. 5MHz or more if needed). --- Sorry, here's a missing datum, the 48nH measurement was on a 3-ohm 5-watt KoolOhm resistor, so f = 3 / 2pi*48nH = 13MHz. Oops, where'd 10MH come from? That was from an earlier 66nH measurement, except after re-zeroing the meter I got 48nH. There are just too many numbers floating around. :-)

Yep!

L = R / 2pi f = 23nH, or 2x better than 48nH (another measurement yielded 66nH, hence the 2x - 3x estimate).

The carbon is close to a piece of similarly-shaped wire.

The breakpoint frequency. It's convenient to say the part acts primarily as a resistor until a 3dB point, afterwhich it acts primarily an an inductor. Plot Z vs. frequency: a straight line from DC to a breakpoint, then rising with f. You know the drill, I assume.