How to find out a resistor's power rating?

[John_doe]

Greetings to everyone, my first post here.
I'll try to be as concise as possible as I do have the tendency to drone on.
I've had a need to stress test a 24V PSU at 100W. To this end I ordered a few 10 Ohm 100W resistors from a Chinese seller on eBay at about $0.94 USD apiece which I thought was a great price seeing how, more than a decade ago, I bought a couple of 50W resistors at a reputable local electronic component store for what I only remember to be an obscene amount of money.
When the new resistors arrived, they were only slightly bigger than my 50W resistors and I decided to test their power rating.
My two 50W resistors are 5.1 Ohm ARCOL HS50 and look exactly like this 3.3 Ohm resistor (not my photo). The 100W resistors look like this (sellers photo).
Here's what I did: I've connected 1 of each resistor (50W and 100W) to a different PSU as to give them exactly 25W each (roughly 11.4v and 16V respectively) and measured their temperature.
Both resistors' temperature were rising at about the same rate until at about 2-3 minutes they reached about 250°C(482°F) and stayed there (if my cheap GM320 Thermometer is to be believed).

But after my test I wondered... Being that I know even less about thermodynamics than I know about electronics, can I really infer anything from my results?

To sum up, my questions are:
1. Can the power rating of the new resistors be deduced from my test?
2. If not, IS there an easy way to find out the power rating of an unknown resistor?

Any ideas will be welcome! Thanks in advance.

[Zero999]

That type of resistor is designed to be screwed to a heat sink. Did you use one?

There are two possibilities:

1) You bought 50W resistors, marked as 100W.
2) The 100W resistors can withstand higher temperatures, than the 50W ones.

There's only one test which can be done and it will be destructive. Increase the power applied to the resistor, until either its value drifts out of its tolerance range (measure the voltage and current across the resistor, during the test an apply Ohm's law ) or it starts to deteriorate physically, i.e. emit smoke, discolour or explode.

[Ian.M]

As you don't know the maximum rated temperature of the resistors and failed to heatsink them adequately your test was useless.  I know you failed to heatsink them as the ARCOL HS50 ones shouldn't have any hotspots exceeding 200 deg C surface temperature if properly heatsinked. (Mounting surface <50 deg C).

It appears you have already overheated and may have damaged the ARCOL  ones as their datasheet says derate to zero power at 200 deg C.  http://www.arcolresistors.com/wp-content/uploads/2014/03/HS-Datasheet.pdf

If you want to compare them properly you'll need a very good heatsink e.g. a water cooled copper block, with a temperature probe in an insulating sleeve in a smal hole drilled from the back avoiding the water passages to sense the temperature of the resistor mounting surface. 

[Damianos]

The results of your experiment are according to theory (simplified)!
By applying the same power they:
- increased their temperature in the same rate, because they have about the same mass and (as they seem, probably due to the similar materials...) the same thermal capacity.
- reached the same final temperature, because they have about the same external surface area, they are colored similarly and in the same environment, so they dissipated the same amount of energy.
But this test is not relevant with what you wonder!

The power rating of a device (resistor, transistor, transformer ...) it is an indication of its' relative durability when it operates in some ranges, defined by its' specifications: voltage, current, temperature...
So it is needed to study the data-sheets to decide which model is capable for your application, before the purchase. By the way, have in mind that a resistor, except of the resistance and power rating, has also other characteristics, like insulation and inductance.

About testing the resistors:
You can evaluate some characteristics of them, like measuring the resistance and variation of it, the inductance, the insulation, the recovering after a temperature increase or decrease ...
The maximum power of the device can only be a destructive test with prerequisites as already described by others earlier.

[John_doe]

I did not know these resistors MUST be heatsinked before use to get anywhere near their rated power. I assumed (maybe foolishly) that their power rating is specified at 25°C ambient temperature without a heatsink and that a heatsink can be applied in an enclosure or to increase the rating. But now looking at the datasheet for the HS50 that Ian linked, I see that they specified a "standard heatsink" of 535 square cm and I don't understand what they mean by that. Furthermore, I fail to see how this method of rating is accurate to any extent... Unless there is actually a "standard heatsink", wouldn't the rating change from the quality of the heatsink (shape, copper or aluminum, fins, etc.)? Can the manufacturer specify that in order to attain the rated power, the resistor must be submerged in liquid nitrogen? And in this case can they still be advertised as 100W (or any other rating)? Is the lesson to be learned here is that if you intend to buy unbranded power resistors you must make sure they come with a datasheet?

Well, I guess that's more of a consumer issue than an electronics (electric?) issue. My apologies about the rant, I'll get back on topic now.

There's only one test which can be done and it will be destructive. Increase the power applied to the resistor, until either its value drifts out of its tolerance range (measure the voltage and current across the resistor, during the test an apply Ohm's law ) or it starts to deteriorate physically, i.e. emit smoke, discolour or explode.

I can do that. But first I must understand the heatsink requirements.

If you want to compare them properly you'll need a very good heatsink e.g. a water cooled copper block, with a temperature probe in an insulating sleeve in a smal hole drilled from the back avoiding the water passages to sense the temperature of the resistor mounting surface. 

That seems overly complex... And much more time consuming than I can afford.

The maximum power of the device can only be a destructive test with prerequisites as already described by others earlier.

Yes, I now understand and am willing to do that.

So my question now is: Seeing as the resistors from the Chinese eBay seller did not come with a datasheet, and assuming I only want to ballpark the power rating of these resistors, what heatsink will suffice? Can it be passively cooled? A picture of such heatsink will help.

Again, thanks to everyone who replied.

[Ian.M]

The point of using a good active cooled heatsink for testing would be to maintain as near as possible a constant temperature at the thermal interface (base surface), so your measurements can be taken in the steady state.

You'd check for hotspots on the rest of the resistor surface using either a calibrated thermal camera, or a temperature indicating crayon chosen for the maximum permissible temperature rise once you know what heatsink temperature you can maintain.   For testing on the cheap you could probably get away with a heavy finned Aluminum heatsink suspended with the fins immersed in an ice bath (to maintain a constant temperature), though it would actually be better mounted fins vertical, with screws and sealent to the inside of a hole a bit smaller than its base area, near the bottom of the side of a container so convection would bring in fresh chilled water as the warmed water rose.   Mill a small slot  (or drill a row of depth controlled holes and dremel it out by hand if you don't have a mill), for the sensor under the resistor footprint and put the sensor on a sliver of expanded polystyrene to insulate it and keep it pressed against the resistor, with a smear of heatsink compound.  The slot should be as small as possible so its an insignificant fraction of the resistor base area.  A bare wire tip type K thermocouple is probably the best option for an off-the-shelf sensor as its the smallest cheap and readily available one.  Otherwise, the smallest possible three leaded SMD temperature sensor, soldered directly to Kynar wire wrap wire should work well enough.  One of the key issues is mimimizing the temperature differential along the leads so Kapton or foil tape them to the heatsink surface for an inch or so.

The HS50 is rated for a maximum surface temperature of 200 deg C and the Temp. Rise & Power Dissipation graph on page 1 of the datasheet indicates that a 150 deg C surface temperature rise is expected at full power.  Therefore its  heatsink must be big enough or actively cooled well enough to keep it below 50 deg C for full power.  Above that, up to 200 deg C de-rate linearly.   The table on page 2 has a column: 'Typical surface rise HS mounted' which I *ASS*U*ME* is in deg C/W, (to correspond to the graph on the previous page).  The standard heatsink specifications read like they are for a simple plate of that thickness and area, almost certainly aluminum, as it doesn't call for anything special.

You cant 'ballpark' the heatsink without testing as you have no idea of the expected surface temperature rise per watt so cant calculate the maximum heatsink temperature.  A high end fan cooled CPU heatsink may work if it has enough flat base area for the resistor and you can drill and tap it for the mounting holes, but especially for the 100W resistor, that may not be good enough to reach its rated power if the ambient temperature is significantly above 25 deg C.

[John_doe]

*ASS*U*ME*

Yes, I am aware of the age-old adage, but, to err is human, and so is to make (unfounded) assumptions.

Anyways, it seems like I have my work cut out for me. I DO want to preform the test you described but due to time contraints it won't happen anytime soon, and again, thanks for your help.

I have watched several people on Youtube test the same power supply I wanted to test and noticed they use different resistors that looks like this:

No cooling of any sort was being used in the videos I watched.
I'd like your opinion: If I purchased some of these resistors rated for 100W, again, unbranded and supplied without a manual, can a test be made without cooling at < 25°C to find out if they are really 100W ?

[Brumby]

No cooling of any sort was being used in the videos I watched.

For this style of resistor - there most certainly was some cooling being used.  It is air.  These resistors have the resistive element very close to the surface, allowing the surrounding air to be able to take heat away.  They would have been in open air which allows for this to take place quite effectively, especially when the resistor has an increased surface area.

[John_doe]

Apologies for not being clear on this point, yes, my use for these resistors will be in open air inside an air-conditioned house but without any "active" cooling (I.e. fan).
Under these conditions can I expect to get close to the rated power?

[John_doe]

I have another somewhat related question: I see these 100W 100Ohm Potentiometers such as this one:
https://www.amazon.com/Ceramic-Wirewound-Potentiometer-Resistor-Rheostat/dp/B00K82VPH2

And I'd like to know if they are 100W at any resistance value or only at 100Ohm.
That is, if I set the dial at 5 Ohm, do I have the equivalent of a 5 Ohm 100W resistor or a 5 Ohm 5W resistor ((100/100)*5)?

[Ian.M]

The green ceramic bodied resistors running at full power can get hot enough to melt solder on their connecting tags.  You need to crimp and bolt the terminals and use wire with high temperature non-flammable insulation for the last few inches to the resistor.   If you want to get the full rated power they'll need to be mounted horizontally to a vertical metal surface to get unrestricted airflow (driven by convection) over them (unless their datasheet specifies forced airflow).  If they start glowing when observed in a dark room you are over-running them.   Fan cooling is *strongly* recommended.

Pot power ratings are always for the whole track.   100W  100R means the track can carry 1A without overheating.  Turn it down to 5R and its still only good for the same 1A, so only for 5W.

[rstofer]

24V isn't very high, why not just toss the resistors in a bucket of water?  I would prefer distilled water which is readily available (used in steam irons).

That assumes, of course, that the leads are no more than 24V to ground.  I don't know which PS you are testing so the safety of this is up to you.

But it's how I would do it.  I remember a Heathkit RF dummy load that was noting more than a resistor arrangement in a 5 gallon can of oil.

https://www.eham.net/reviews/detail/13123

[HB9EVI]

I do not comment the resistor issue; but did you maybe think about an electronic load to test the PSU? That's kind of state of the art to do so; much easier in handling and controlling the parameters like current flow than with resistors.
A simple voltage controlled current sink with a bunch of heavy duty FETs and low resistance shunts; it's easy and scalable.