MusicRadar

RothwellAudio · 11-07-2009, 12:46 PM

Quote:

Originally Posted by ecc83

I think everyone will know intuitively that a 400x400mm aluminium plate 10mm thick with a big R bolted to the middle of it will not reach the same temperature, ever, as one 400x400mm but 0.5mm thick for the same continuous power input. Were this not the case we could make smashing heatsinks out of beer can material!

Of course materials "absorb" heat, that is why they use synthetic diamonds for laser windows and molten sodium in valve stems.

Ok, the whole thing is complicated by convection and such matters but "heat" is a basic, low grade form of energy and it causes a different temperature rise in different materials.

It is one of the triumphs of (relatively) modern science that we can now solve the conundrum of why things do not get to an infinitely low or high temperature.

Dave.

How does this answer the question of specific heat as opposed to continuous heat?

ecc83 · 11-07-2009, 01:09 PM

"How does this answer the question of specific heat as opposed to continuous heat?"

Not sure how that was exactly phrased but I will have a go!

If we can, for a moment lose the complications of convection, an object will reach a certain temperature when the heat input balances the heat output by radiation. The ally plates will have very similar surface areas and so radiate about the same amount but we know the big one will never get to the same temperature as the thinnner one. This is because it takes a specific amount of heat energy to raise a specific number of aluminium atoms thru' X dgrsC. As the temperature of a body rises so too does the energy of the radiation. First infra red then visible red and so on, each "jump" in energy level taking more energy than the next (where is a physicist when you need one!).

Heatsinks of course use all three methods of cooling (except in space!) we want a nice chunk of metal and fins and if it can be black anodized so much the better. Ally is good, copper would be better, M steel would be cheaper but not nearly so good!

Do not BTW confuse heat capacity with thermal conductivity, water has a massive heat capacity but is a hiss poor conductor.

Dave.

martinw · 11-07-2009, 04:55 PM

Quote:

Originally Posted by ecc83

Of course materials "absorb" heat, that is why they use synthetic diamonds for laser windows and molten sodium in valve stems.

Dave.

Sorry Dave but you're wrong there. The sodium in valve stems (which is only molten when the engine is running obviously) is used to transfer heat out of the valve, not retain it.
Of course materials do absorb heat, but that's not their function.
As the valve reciprocates, the sodium is thrown continuously from end to end, picking up heat from the head, and then dissipating it to the stem which is immersed in oil and thence out to atmosphere via the oil cooler.

The key word in the case of the original post is "dissipate". All the heat passed through the resistor in question is dissipated to the air around it, once the initial temparture increase is complete and the system has become stable, and barring the small amount retained by each component (which will vary).

Because the mass and temperature of the lower or higher wattage resistors are inversely proportional, all else being approximately equal, the change in temperature of the air will be the same, or near enough.

Reliability of the actual component however is a different issue!

ecc83 · 11-07-2009, 06:11 PM

"is used to transfer heat out of the valve, not retain it."

So why not use aluminium, lead, tin,

And do you have another solution to the thin/thick plate example?

Why make heatsinks out of expensive aluminium?

Dave.

ecc83 · 11-07-2009, 06:24 PM

Ooops, they might not melt but mercury would and tho nasty is a bloody sight less so than sodium.
Dave.

martinw · 11-07-2009, 09:42 PM

I assume sodium is light, cheap, liquid at the appropriate temperature, and otherwise suits the application.

Surface area is what matters with heatsinks, so you're right that thin plates would be fine, but its impractical, the heatsinks need to be strong enough. I've seen some pretty thin clip-on heatsinks.
Look at aircooled motorbike engines as an example; as bikes develop, the weight (mass) of the castings goes down, but fin area increases to cope with power outputs that typically doubled, along with the waste heat. More surface area, less mass.

Aluminium is cheap: $1800 per tonne, versus $6500 for copper, $15,000 for tin and $17,000 for nickel. Steel has too high a thermal resistance.

RothwellAudio · 11-08-2009, 05:52 PM

Quote:

Originally Posted by ecc83

I think everyone will know intuitively that a 400x400mm aluminium plate 10mm thick with a big R bolted to the middle of it will not reach the same temperature, ever, as one 400x400mm but 0.5mm thick for the same continuous power input. Were this not the case we could make smashing heatsinks out of beer can material!

Of course materials "absorb" heat, that is why they use synthetic diamonds for laser windows and molten sodium in valve stems.

Ok, the whole thing is complicated by convection and such matters but "heat" is a basic, low grade form of energy and it causes a different temperature rise in different materials.

It is one of the triumphs of (relatively) modern science that we can now solve the conundrum of why things do not get to an infinitely low or high temperature.

Dave.

Ok, I've given this some thought and here is my analysis.
To take your example of two heatsinks 400x400mm, one 1mm thick, the other 10mm thick; if you were to put a specific (and small) amount of heat into both by switching on the resistor for say 10 seconds, the thicker heatsink would certainly remain cooler than the thinner one. However, if the resistor is switched on continuously there will come a point when the system reaches equilibrium. If the resistor is generating 1 watt, equilibrium will be reached when the heatsink is radiating/convecting/conducting 1 watt, ie the same power is going out as is coming in. Will the temperature of the two heatsinks be different? No, I don't think so, because the thick heatsink has no more ability to radiate/convect/conduct than the thin one. In fact, the thick heatsink would be very inefficient because despite having ten times the mass it has (almost) no greater surface area and will be no better at losing heat. The difference between the two heatsinks will be that the thicker heatsink will take longer to reach equilibrium and take longer to cool down after the power is switched off. That will be the effect of its greater thermal capacity.

To get back to the question of electrolytics being cooked by cathode resistors, would a small resistor at a higher temperature do more damage than a big resistor at a lower temperature? Well, consider an analogy. Consider heating a room with a big radiator at a temperature of, say, 60 degrees. Would that heat the room better than a candle off a birthday cake? The candle might be at a temperature of 1600 degrees but it can't put out the same amount of heat as the big radiator, so despite being at a lower temperature the radiator will heat the room more than the candle.
Similarly, a small cathode resistor a higher temperature won't heat an adjacent capacitor any more than a big resistor at a lower temperature. If both resistors are generating 1 watt, the heating effect will be same, despite the difference in temperature.
The best way to keep the capacitor cool is to move it away from the resistor. Even just 10mm or so will be a big help. Also, if the cap and resistor are mounted on a tag board, maybe drilling a few ventilation holes on the board to improve air flow will help, too.

ecc83 · 11-10-2009, 01:21 PM

Ok,
I have rigged two 150R's in series. One 6W MF other 1WMF centered approx 80mm apart.

The supply is a stabilized 24.07Vdc. these are the findings in free air...

After 5 mins or so Tamb =20C T6W =55C, T1W= 86C ( at 11.38am)

At 11.54 am

Tamb=22C T6W=55C T1W=85/86C*

Now which one do you want 5mm from an electrolytic?

*It was quite difficult to measure the smaller R temp' I kept burning my pinkie! I would like to have araldited a temp probe to each R but Times Wing'ed Chariot.....

Dave.

martinw · 11-10-2009, 01:34 PM

Your experiment is incorrect.
The surface temperature of the r's isn't the issue. We'd all accept that one gets hotter.

The question is how much of that heat is radiated or convected into an adjacent component.

Repeat the experiment, but mounting a typical electrolytic cap (47uF?) a set distance away from each resistor, i.e. with centrelines the same in both cases; the distance surface-surface being slightly more in the case of the smaller resistor, only fair right? (so you 5mm comment is slightly out).

The measure the surface temperature of the caps, on the side away from the heat source of course.

All good fun, but of course you're right in that I always use 5-6W or more cathode resistors, and mount them as far away from anything else as I possibly can.

dogface · 11-10-2009, 01:51 PM

Quote:

Originally Posted by ecc83

Ok,
I have rigged two 150R's in series. One 6W MF other 1WMF centered approx 80mm apart.

The supply is a stabilized 24.07Vdc. these are the findings in free air...

After 5 mins or so Tamb =20C T6W =55C, T1W= 86C ( at 11.38am)

At 11.54 am

Tamb=22C T6W=55C T1W=85/86C*

Now which one do you want 5mm from an electrolytic?

*It was quite difficult to measure the smaller R temp' I kept burning my pinkie! I would like to have araldited a temp probe to each R but Times Wing'ed Chariot.....

Dave.

No one ever doubted that a smaller resistor would get hotter, did they? What was in doubt was all that stuff about heat and thermal capacity and heatsinks keeping things cool by soaking up heat rather than dissipating it and ignoring "complications of convection".
As was pointed out, a candle is a hell of a lot hotter than a radiator, but who wants to rip out their central heating and replace it with a candle?

11-07-2009, 09:42 PM	#36
martinw Member Join Date: Nov 2008 Posts: 60	I assume sodium is light, cheap, liquid at the appropriate temperature, and otherwise suits the application. Surface area is what matters with heatsinks, so you're right that thin plates would be fine, but its impractical, the heatsinks need to be strong enough. I've seen some pretty thin clip-on heatsinks. Look at aircooled motorbike engines as an example; as bikes develop, the weight (mass) of the castings goes down, but fin area increases to cope with power outputs that typically doubled, along with the waste heat. More surface area, less mass. Aluminium is cheap: $1800 per tonne, versus $6500 for copper, $15,000 for tin and $17,000 for nickel. Steel has too high a thermal resistance. __________________ www.mjwamps.com

11-10-2009, 01:34 PM	#39
martinw Member Join Date: Nov 2008 Posts: 60	Your experiment is incorrect. The surface temperature of the r's isn't the issue. We'd all accept that one gets hotter. The question is how much of that heat is radiated or convected into an adjacent component. Repeat the experiment, but mounting a typical electrolytic cap (47uF?) a set distance away from each resistor, i.e. with centrelines the same in both cases; the distance surface-surface being slightly more in the case of the smaller resistor, only fair right? (so you 5mm comment is slightly out). The measure the surface temperature of the caps, on the side away from the heat source of course. All good fun, but of course you're right in that I always use 5-6W or more cathode resistors, and mount them as far away from anything else as I possibly can. __________________ www.mjwamps.com

Thread Tools
Show Printable Version Email this Page
Display Modes
Linear Mode Switch to Hybrid Mode Switch to Threaded Mode

MusicRadar

Buy here

From the makers of

11-07-2009, 01:09 PM	#32
ecc83 Senior Member Join Date: Sep 2006 Location: northampton uk Posts: 2,614	"How does this answer the question of specific heat as opposed to continuous heat?" Not sure how that was exactly phrased but I will have a go! If we can, for a moment lose the complications of convection, an object will reach a certain temperature when the heat input balances the heat output by radiation. The ally plates will have very similar surface areas and so radiate about the same amount but we know the big one will never get to the same temperature as the thinnner one. This is because it takes a specific amount of heat energy to raise a specific number of aluminium atoms thru' X dgrsC. As the temperature of a body rises so too does the energy of the radiation. First infra red then visible red and so on, each "jump" in energy level taking more energy than the next (where is a physicist when you need one!). Heatsinks of course use all three methods of cooling (except in space!) we want a nice chunk of metal and fins and if it can be black anodized so much the better. Ally is good, copper would be better, M steel would be cheaper but not nearly so good! Do not BTW confuse heat capacity with thermal conductivity, water has a massive heat capacity but is a hiss poor conductor. Dave.

11-07-2009, 06:11 PM	#34
ecc83 Senior Member Join Date: Sep 2006 Location: northampton uk Posts: 2,614	"is used to transfer heat out of the valve, not retain it." So why not use aluminium, lead, tin, And do you have another solution to the thin/thick plate example? Why make heatsinks out of expensive aluminium? Dave.

11-07-2009, 06:24 PM	#35
ecc83 Senior Member Join Date: Sep 2006 Location: northampton uk Posts: 2,614	Ooops, they might not melt but mercury would and tho nasty is a bloody sight less so than sodium. Dave.

11-10-2009, 01:21 PM	#38
ecc83 Senior Member Join Date: Sep 2006 Location: northampton uk Posts: 2,614	Ok, I have rigged two 150R's in series. One 6W MF other 1WMF centered approx 80mm apart. The supply is a stabilized 24.07Vdc. these are the findings in free air... After 5 mins or so Tamb =20C T6W =55C, T1W= 86C ( at 11.38am) At 11.54 am Tamb=22C T6W=55C T1W=85/86C* Now which one do you want 5mm from an electrolytic? *It was quite difficult to measure the smaller R temp' I kept burning my pinkie! I would like to have araldited a temp probe to each R but Times Wing'ed Chariot..... Dave.