Power, not energy. Power is the rate of flow of energy. Energy is measured in joules, 1 watt is 1 joule per second. Heat is a kind of energy and thermal dissipation is just a fancy way to say heat flow.
Talking of physical units, heatsinks are rated by thermal resistance which has usings K/W (or deg C/W). This is the reciprocal of thermal conductance which has units W/K.
The power rating of the resistor is the highest dissipation it can support without damaging itself *** if its cooled aggressively ***. Give it an inadequate heatsink and it'll burn out well before that datasheet rating.
You'd never run a 100W resistor continuously at 100W anyway, just as you'd never draw 5A continuously from a 5A power supply - this is simply pushing things too hard and leaving no slack at all.
I'd suggest not pushing more than 75W through the "100W" resistor, and using a heatsink thats less than 1K/W thermal resistance. Ideally 0.5K/W (a large heatsink).
Note that heatsinks designed for forced-air cooling can be much more efficient at removing heat, an old CPU fan assembly might be up to the job and more compact than a passive heatsink.
2x4=8? 2R5x3=7R5? close enough?
What's the issue? Why did you think I specked 300W resistors and said they would handle 1200W for 40$? 4x2R=8R... 1200W handling, too.
Last edited:
You'd never run a car motor at redline all day, but you do with a boat motor... I always wondered about that... I figure a car motor SHOULD do the same, if you wanted it to... Chrysler will melt and die though - I killed a rented Sebring in 2 weeks.
Because the rating changed somewhere.
It certainly SHOULD handle 100W continuous -- That's what it says on the label. I get derating... But this seems more than that.
I had a 50W 1R resistor TO-220 that needed to dissipate 4W in normal operation (lesson learned, TO-220 is only good for 15W or so) package as part of an RC filter. It exploded on the inrush even though it was bolted to a ~0.5W/C sink. Now I use a giant 40W WW part (integrated heatsink like a tube?)
No. Apparently, you need LN2 to reach that spec just like any modern part with a screw hole...
Last edited:
Mark: For your choices of 0.75K/W or 0.5K/W, those would heat up by 56.25K and 45K, respectively. Did you pick those values because ~50K seemed reasonable?
Something I’ve worried about is a heat sink that can’t effectively cool the component. Is there a factor for the amount of heat a heat sink can dissipate? Like, if I got a 0.5K/W heat sink, is there any implication that it will be able to absorb 100W? Could it absorb 1000W and just get really really hot?
Something I’ve worried about is a heat sink that can’t effectively cool the component. Is there a factor for the amount of heat a heat sink can dissipate? Like, if I got a 0.5K/W heat sink, is there any implication that it will be able to absorb 100W? Could it absorb 1000W and just get really really hot?
Would a 100W lightbulb melt with a 1kW power? Never mind the filament, what about the glass?
For short peaks, yes. for anything more than tens of milliseconds, probably not. It would explode. Ask me how I know? (12V 50W bulb into 120V because it was for TV production and they used the same connector (and yes, I know that's square law, but the idea is the same)).
For short peaks, yes. for anything more than tens of milliseconds, probably not. It would explode. Ask me how I know? (12V 50W bulb into 120V because it was for TV production and they used the same connector (and yes, I know that's square law, but the idea is the same)).
Last edited:
Also, aren't you using this as a testing rig? You don't plan to run it full out for hours, do you?
The only thing with a 100W rating in this equation is the resistor. The heat sink is only rated as “0.5K/W”. That tells me how quickly it will heat up, but nothing about how much heat it can absorb. How do I know what size heat sink to apply to a given part based on that? Could I put a 2K/W heat sink on a 100W resistor if I was ok with it reaching 200C at maximum operating load?
But yes, practically I will only be loading up this rig for short periods, and I can certainly provide active cooling. I’m mostly trying to understand the calculations that go into heat sink selection so I can do this on my own in the future.
This LINK
is not this LINK
For a visual of my previous example... The TO-220 is about the same size as the mounting hole and tab on the 40W...
The patch of compound on the heatsink is all that remains of the popped resistor. The giant traditional resistor is seen in the upper right...
It's a hand wired modified De-lite amp using tubes to drive depletion mode MOSFETS as followers that are loaded with 300W light bulbs BTW...
The TO-220 parts are shown between the TO-247 parts...
As an aside, the SMPS and boost on the left have been cooked to 120°C a few times and don't seem to care...
is not this LINK
For a visual of my previous example... The TO-220 is about the same size as the mounting hole and tab on the 40W...
The patch of compound on the heatsink is all that remains of the popped resistor. The giant traditional resistor is seen in the upper right...
It's a hand wired modified De-lite amp using tubes to drive depletion mode MOSFETS as followers that are loaded with 300W light bulbs BTW...
The TO-220 parts are shown between the TO-247 parts...
As an aside, the SMPS and boost on the left have been cooked to 120°C a few times and don't seem to care...
Attachments
Last edited:
A heatsink does more than just absorb. It moves the heat from the hot device into the metal of the 'sink, thence to the air.
The math is easy - the heatsink you mention is rated at 0.5K/W, so it will increase in heat by 1/2 a degree Celsius for every watt you put into it.
You want to dissipate 400W
400 x 0.5 = 200
So your heatsink will heat up by 200deg C, or 392F.
AND your room (and therefore the heatsink) will start around 20C (68F) making the heatsink rise to about 430F, which is more than an oven used for baking.
For your application you are going to need a big resistor (expensive) or an array of smaller resistors and a big heatsink. (also expensive, but probably a bit easier)
And a fan, as moving even a little air makes a huge difference.
How hot are you willing to let the 'sink get? I'd suggest 60-70 C as a maximum, as that's getting close to will burn you on contact, and hurts when touched. More than that and you will get immediately burned. The resistors will always be very hot.
The back of the napkin says you'll need a 12 inch piece of this 8.000" Wide SERRATED FIN Extruded Aluminum Heatsink - HeatsinkUSA and a fan.
The math is easy - the heatsink you mention is rated at 0.5K/W, so it will increase in heat by 1/2 a degree Celsius for every watt you put into it.
You want to dissipate 400W
400 x 0.5 = 200
So your heatsink will heat up by 200deg C, or 392F.
AND your room (and therefore the heatsink) will start around 20C (68F) making the heatsink rise to about 430F, which is more than an oven used for baking.
For your application you are going to need a big resistor (expensive) or an array of smaller resistors and a big heatsink. (also expensive, but probably a bit easier)
And a fan, as moving even a little air makes a huge difference.
How hot are you willing to let the 'sink get? I'd suggest 60-70 C as a maximum, as that's getting close to will burn you on contact, and hurts when touched. More than that and you will get immediately burned. The resistors will always be very hot.
The back of the napkin says you'll need a 12 inch piece of this 8.000" Wide SERRATED FIN Extruded Aluminum Heatsink - HeatsinkUSA and a fan.
Last edited:
Ok, so, 6L6, in your example with the 400* hearsink, the attached resistor could be just fine, it’s the practical considerations of not wanting 400* aluminum on my desk that dictate necessary cooling?
400W is A LOT of energy. You have to put it somewhere, and there is no getting around the physics of that.
The resistor(s) will only be fine because of the heatsink.
The resistor(s) will only be fine because of the heatsink.
Yes it is. That's why I suggested big resistors that can handle peaks and not require sinking.
Nevermind the cost difference...
If you plan on making this a continuous install, you could use a fan?
Nevermind the cost difference...
If you plan on making this a continuous install, you could use a fan?
During my PA days, I routinely dump 100W resistors in a bucket of water and test 200+W amps without any problem.
Overread that🙄.
The point I wanted to make is that metal shell resistors can dissipate themselves quite a bit.
As an example: with 2W on them they barely get warm.
The resistors from post 1 will get rather hot with 2W on them. They are not as robust but probably as expensive being metal film.
Still 4-12pcs. (the more the better) metall shell resistors plus heatsink are expensive and you might want to look at the ones from post 14.
You can look at types made in china but you´ll never know how well these are made and will have to go for 12pcs. at least.
50W 100W 0.5R 1R 2R 4R 6R 8R 10R 20R Aluminium Power Metal Shell Fall Draht Widerstand 0,01 ~ 100K 0,5 1 2 4 6 8 10 20 100 ohm|Resistors| - AliExpress
Forced convection makes heatsinks much more effective so you will not need a 0.2K/W heatsink (or so) in the end.
See some examples here:
Dummy Load
Amplifier Dummy Load for THD measurements / non-inductive?
Audio dummy load
This was another suggestion in one of those threads but there´s no guarantee the resistance will be a good fit:
https://www.amazon.com/-/de/dp/B08912X4YN/ref=cm_cr_arp_d_product_top?ie=UTF8
Last edited:
There is an important between those black 'ribbed' wound power resistors and those gold ones with the mounting holes.
The ribbed wound ones can handle much higher temperatures than the gold ones.
Even if both can handle the same power, say 200W, the ribbed ones can handle that in free air because they can easily heat up to the point where the solder starts to melt, with no ill effect.
The gold ones have to be heatsinked because they can not handle so high a temperature.
I guess that's why they have mounting holes and a flat back ;-)
Jan
The ribbed wound ones can handle much higher temperatures than the gold ones.
Even if both can handle the same power, say 200W, the ribbed ones can handle that in free air because they can easily heat up to the point where the solder starts to melt, with no ill effect.
The gold ones have to be heatsinked because they can not handle so high a temperature.
I guess that's why they have mounting holes and a flat back ;-)
Jan
I think so too and would choose the ribbed ones + fan without the use of a heatsink. That said the datasheet for the gold ones from Vishay (and others) suggest operating temperatures of 200°+; Vishay actually state 250°C.
Needless to say we don´t want to get anywhere near as even leadfree-solder will melt at that point. Also the inner construction of those gold ones will probably suffer at those temp.; plus you will need to derate them heavily and will have to use even more pieces for your 400W load:
Wirewound Resistors – Construction, Pulse Load and Types – Passive Components Blog
Under "R 2.2.6 Adjustable" you will find another example of the "ribbed ones" that are adjustable.
Here´s a list of pretty much all varieties on page:
High Power Resistors - Adjustable Power Resistors | Ohmite Mfg Co
Just put your resistors in a bucket of water for your test.
Or find an old heating element of some sorts.
Unless you’re planning on going into the amplifier output testing business…
Why spend $100 to sell a $400 - $500 item?
Or find an old heating element of some sorts.
Unless you’re planning on going into the amplifier output testing business…
Why spend $100 to sell a $400 - $500 item?
Good tips on the ribbed guys. So I could setup those up, no heat sink, put a fan on them, and be in pretty good shape, huh? Sounds like a plan…
How does the “resistor in water” plan work? I always heard water was bad with electronics. Do I have to keep it to like, one resistor per jar, so voltage can’t short through the water to other resistors?
Phase: I’ve been warning to build a load box for a while to test various amps I want to build. This particular case called for much higher power demands than I’d thought, but I figured “hey, why not, better too much than too little”. So I’m here asking questions to determine feasibility and practicality.
How does the “resistor in water” plan work? I always heard water was bad with electronics. Do I have to keep it to like, one resistor per jar, so voltage can’t short through the water to other resistors?
Phase: I’ve been warning to build a load box for a while to test various amps I want to build. This particular case called for much higher power demands than I’d thought, but I figured “hey, why not, better too much than too little”. So I’m here asking questions to determine feasibility and practicality.
Yeah, if you find a good deal on those, go for it.
I´d do it similar to this guy in post 21:
Audio dummy load
Get 2-3 pcs. if you can and a 12cm fan over them or similar and you´re set.
Having performed humidity bias tests in my previous job I can´t recommend high heat plus high humidity/drowning them in water for anything electronic.
Those ribbed ones will look old in no time and contacts will wear at the very least.
Have a look at post 21 above. They look 50years old with just lots of heat, oxygen and some humidity from the air.
Here's another option that would cost about the same as a big heatsink and chip resistors --> https://www.mouser.ca/ProductDetail/Ohmite/PFE5K8R50E