Resistors for dummy load?

[tubetvr]

There seems to be no info on capacitance, but inductance figures

2uH is very good for a wirewound resistor but will still have a effect on very high frequencies.

Using multiple non wirewound resistors it is possible to make something that is much better but the problem is of course how to measure the performance. When I built my dummy load about 10 years ago, I used Resista metal oxide resistors and measured the performance using a network analyser at my work, the dummy load behaves like an almost perfect 8 ohm load up to at least 10 MHz.

I have now built a similar load using resistors I found here in Tokyo, but I have no way of refering the exact performance but I think the performance is similar.

Regards Hans

 

[Christer]

Using multiple non wirewound resistors it is possible to make something that is much better but the problem is of course how to measure the performance. When I built my dummy load about 10 years ago, I used Resista metal oxide resistors and measured the performance using a network analyser at my work, the dummy load behaves like an almost perfect 8 ohm load up to at least 10 MHz.

I certainly see your point about the usefullness of a low-reactance
load. I considered buying the NHS resistors, but decided to go
for a large number of cheap surplus wire-wounds instead. That
saved me money both on the resistors and on not needing
heatsinks. Maybe, it was a good decision after all, since
I didn't waste money on
something that's better but maybe not good enough.
However, then immediately arises the question about the
inductance of metal-oxide resistors. I don't think any metal oxide
resistors are easily available locally and I haven't been able to
dig up any info at all about those metal-film that are easily
available here.
Considering that these are often spiral-cut AFAIK it seems not
obvious they should have a low inductance, although they seem
to usually exhibit a low capacitance. Do you have any info on
typical values for L and C for "standard" metal film resistors and for
metal-oxide ones?

I have now built a similar load using resistors I found here in Tokyo, but I have no way of refering the exact performance but I think the performance is similar.

That is always a problem, how to assess the quality of lab
equipment unless you have even better lab equipment to measure
it with? Of coure, if using a precision resistor with know or
guaranteed low C and L in series, with the dummy load it should
be possible to measure
the phase angle and impedance at various frequencies.


Edit: I temporarily forgot the Caddock film resistors which are a
bit expensive and needs heatsinking, but they do spec inductance
in the datasheets: 20nH for the 100W type and 10nH for the
lower rated ones.

 

[wintermute]

Hmmm lots of activity while I was asleep thanks for all the suggestions guys..... the 30 jaycar 5W wirewound resistors (60 for two channels) is indeed a cheap option (about $11) but I don't know how I'd mount them, 30 in parallel would certainly make the inductance negligible as well.

I've had another look at the Arcol range, and one of the things I like about them is that they say the 75W ones can handle 45W with no heatsink, and the 100W ones can handle 50W with no heatsink....... I suspect they are pretty robust.

Tony.

 

[Christer]

I've had another look at the Arcol range, and one of the things I like about them is that they say the 75W ones can handle 45W with no heatsink, and the 100W ones can handle 50W with no heatsink....... I suspect they are pretty robust.

Well, if you look at them they already look like heatsinks.

Add a cheap computer fan or something, and you can probably
go quite a bit higher.

 

[tubetvr]

Do you have any info on typical values for L and C for "standard" metal film resistors and for metal oxide ones?

Stored somewhere in Sweden I have extensive data on Resista metal oxide and metal film resistors I would imagine that similar size resistors will have similar values of capacitance and inductance, but I can not remember any exact values.

One old rule of tumb I learned is that a 1/4 W standard resistor have a capacitance of ~2pF almost regardless of type, (carbon composite, carbon film or metal film) the capacitance is dependant mostly on the size.

The capacitance is also almost independant on resistor value but as expected the inductance is dependant on how the resistance film is spiralised.

I did some quick calculations on the expected inductance of the 5W metal oxide resistors I bought here in Tokyo. The resistors are 8mm in diameter and 25mm long and have totally 6 turns of resistance film, (this is for 39 ohm resistors). The calculated inductance for one resistor is then ~0.08uH. As I parallell connect 10 resistors and then connect 2 of these packages in series the total inductance is 2*(0.08/10) = 0.016uH. This low inductance is for the resistors themselves and eventual inductance of connecting wires need to be added to this value but it should not be a problem to keep the total inductance low.

BTW, the reason why I prefer metal oxide instead of metal film is because the metal oxide types can withstand higher power.

Regards Hans

 

[tcpip]

Auto headlamps

Anyone tried ordinary incandescent lamps meant for auto use, rated for 12V? Unlike mains-rated lamps, these will run at proper rating, when driven by an amp or its transformer (I like to load-test my power transformers to see load regulation), which means they will glow and get hot, which means they'll give you the right rated impedance. (There's a huge difference in impedance of incandescent lamps when cold and hot. All ratings are for when they are hot.)

So, if you want to pump in 36V rails into this load, put three lamps in series. They are inexpensive, need no heat-sinking, and are widely available. The only headache is some ingenuity needed to mount them in a stable way, in case you want a reusable permanent setup. And each lamp is rated for 50W quite often. Make sure they are ordinary incandescent lamps, not some new hi-tech gas-filled alternative... those may have different inductance characteristics. However, they'll all behave well at rated currents. The only doubt I have is how they'll behave with an AC supply, since they get DC in cars.

I was also wondering about using the El Cheapo 5W wirewound resistors in parallel, all fixed using Araldite to a heatsink. Shouldn't they be good enough, unless you're specifically looking for high-frequency measurements?

Tarun

 

[tubetvr]

Anyone tried ordinary incandescent lamps meant for auto use, rated for 12V?

I have an old book describing simple transmitters and receivers built with valves, the circuits are very simple with one valve transmitters and regenerative receivers. In that book it is described to use incandescent lamps as dummy loads when testing transmitters as they are apparently relatively free of inductance and capacitance. One problem I see is that the resistance of a incandescent lamp is changing with applied voltage which is not a good thing when measuring audio amplifiers.

BTW, incandescent lamps can be used as simple current regulators as current trough the lamp is more or less constant, (if current would go up the resistance goes up which counteracts the increased current)

Regards Hans

 

[peranders]

You mean a heatsink from an RC!!!? How big is that? Or maybe
it wasn't a heatsink for the thyristors?

Not for thyristors, something else, dont know what. The weight was a couple of kg's. Not that huge but very handy.

 

[wintermute]

Hi Tarun,

That's not a bad Idea, I might go to the local auto shop and see what I can get, non halogen globes should be pretty cheap too.

I suppose what all this really comes down to is whether or not the idea is to do accurate measurements of freq response, power output etc, or simply a way of loading the amp to check for things like oscillation, PS ripple, clipping behaviour etc. For me (at least at the moment) it's more the later.

Tony.

 

[AuroraB]

Something doesn't stick here.....

First- there is massive discussions on the nonlinearities of power resistors, inductance and capacitance included....

Then the suggestion is to use light bulbs,--- a highly nonlinear resistor in response to temperature. The light bulbs may have very little reactive elements, but releives heat badly. Thus a filament temp of 200 deg. is hardly noticable as a light glow, but changes the resistance almost 100%.Remember resistance is directly proportional to delta-T ( in Kelvin). You would have to use a lot of lamps to avoid the heating effect.

Good old fashioned carbon composition resistors, or metal comp, is probably the cheapest way to get something reliable. Immersion in oil is really not a problem either. Use a paint can with a good lid,- just avoid topling it!

Or- buy the expensive bits..................

 

[tcpip]

That's not a bad Idea, I might go to the local auto shop and see what I can get, non halogen globes should be pretty cheap too.

Precisely my thoughts. While you're at it, also look for holders or mounts of some sort, so that you can set all them bulbs up in some way.

I suppose what all this really comes down to is whether or not the idea is to do accurate measurements of freq response, power output etc, or simply a way of loading the amp to check for things like oscillation, PS ripple, clipping behaviour etc. For me (at least at the moment) it's more the later.

Precisely. I too wanted to build dummy loads basically to test (i) whether my power amp PSU (specially the transfi) heats up after an hour or two of full-load use, (ii) whether my power amp behaves, (iii) how hot my power amp heatsink gets. I think that this sort of test is a must if you want to trust the reliability of the poweramp for many years to come. In such situations, I don't want to vary the signal frequency much (it'll either be 50-100Hz or something "standard" like 1KHz), and I don't want to take any but the grossest of measurements like output current, rail voltage, temperature, etc. So a bit of inductance or capacitance doesn't bother me, as long as the power amp doesn't get into oscillation.

For precise bandwidth measurements, I might not even want to take things to full load, and if I do, I'll think about pure-resistive loads then. Presumably, I'll be rich enough by then to buy a couple of hundred linear resistances with very low tempco and low L and C.

Tarun

 

[tcpip]

Something doesn't stick here.....
First- there is massive discussions on the nonlinearities of power resistors, inductance and capacitance included....
Then the suggestion is to use light bulbs,--- a highly nonlinear resistor in response to temperature.

I too was confused.

However, even light bulbs can be treated as quite stable impedances, if you choose bulbs which will be driven to "full load" by the amp. Hence my choice of auto bulbs, which are designed to glow brightly and reach a stable thermodynamic condition with just 12V. A few such bulbs in series can be good test loads for amps at full power and low frequency, provided you choose the number of bulbs in series to match the voltage output by your amp at full load.

Trying to use incandescent lamps, and then fussing with them to ensure they remain cold, is too unstable an approach, I feel. You must drive them so that they glow at full brightness.

One assumption here is that you'll be pumping a steady signal of whatever frequency but at unchanging amplitude into your amp. Varying the output power rapidly will allow the bulbs to cool and re-heat, thus sharply changing their impedances. It all depends on what kind of tests you want to do.

And to find out the impedance of the bulb, it's of no use measuring it with an Ohmmeter when cold. The best approach, it seems to me, is to take the rated wattage and voltage of the lamp, and calculate --- not measure --- the impedance, and then realise that this impedance is valid only if the rated voltage is applied on the lamp from a low-impedance high-current source, i.e. a power amp (not an Ohmmeter).

This auto-headlamp approach was suggested to me by The Great Audio Guru Angshu.

Tarun

 

[AuroraB]

I agree in principle to using lamps as loads, - this has been done. However- during my army days in the early 70s, we used som special bulbs as RF loads, but I cannot recall what it was......

Personally I think there is only two easy choices- wire wounds or paralled compositions. Otherwise- no way around the expensive stuff...

Actually, -some years ago, I needed to test a 24V / 30 A PSU in a hurry,...
I soldered a handful of "Dales" at hand to two wires and hung it from my desk.....after a while there was some strange thumps...looking closer, I found the resistors on the floor.....

Heatsinks are absolutely recommended....!!

 

[wintermute]

I soldered a handful of "Dales" at hand to two wires and hung it from my desk.....after a while there was some strange thumps...looking closer, I found the resistors on the floor.....

Heatsinks are absolutely recommended....!!

Originally I asked the question about inductance, simply because I didn't know if there was some nasty thing that it could do, which would make my amp look like it had problems when it didn't.

I've wondered a few times whether some of the nasty things I've seen were in fact due to the cheap jaycar 5W resistors I currently have (only two in parallel so 10W, although good for 100W for about 20 seconds before they start smelling!)

Seems like the only real concern is whether or not you want absolute perfection in the measurements, which at this stage I don't. It's certainly been an interesting discussion though!!!!

Tony.

 

[JOE DIRT®]

this is the resistor I was referring to in my previous post