never say never - 300B amps...

[dsavitsk]

Probably more than that. 9V winding will be more like 6 to 8W per heater, depending on the diodes used and the resistance in the winding.

 

[andrew_whitham]

OK, but whats reasonable? what are the DC heated 300b guys doing with their amps, any suggestions? For the sake of reliability lets work on 8-10w dissipation from the TO220 package - how to loose that?

Any advice very welcome. I think this is the last stumbling block before I can get something that would work. (how well is a different matter and the reason for the experiment)

Cheers

Andy

 

[Sheldon]

You don't want to trap that heat inside the chassis. So that means that you have to transfer it to the outside air. I don't know what your chassis looks like, so it's hard to make a definitive recommendation. If your chassis is made of aluminum or some other metal that conducts heat well, you could mount the chip directly to the underside of the chassis (with electric insulator pad, of course). If the chassis is steel, it would be better to mount the chip to a reasonably stout piece of aluminum, say 3mm thick and 3x3 cm. Then mount that flat against the chassiss with thermal grease. Or use a similar sized piece of aluminum angle. With a steel chassis you need a decent area of a good thermal conductor to transfer the heat to the chassis.

Or, you could bring the chip to the outside surface of the amp and mount it on a finned heatsink. I'd go with something around 3x3x5 cm. A piece of aluminum angle would work ok here too.

Sheldon

 

[andrew_whitham]

Or, you could bring the chip to the outside surface of the amp and mount it on a finned heatsink. I'd go with something around 3x3x5 cm. A piece of aluminum angle would work ok here too.

Sheldon

Something might be wrong then, I used a finned heatsink for testing, of about the size you mention and it was too hot to touch. Could it be oscillating? I can't see anything on the (20mhz)scope and in theory I'm dissipating 6 or 7 watts with my 9v transformer. (right?)

The case will almost certainly be alloy heatsink at the moment then although I really did fancy trying one of those wooden ones (a'la bottlehead)

oh well I do have more heatinks to try...

Andy

 

[Sheldon]

Something might be wrong then, I used a finned heatsink for testing, of about the size you mention and it was too hot to touch. Could it be oscillating? I can't see anything on the (20mhz)scope and in theory I'm dissipating 6 or 7 watts with my 9v transformer. (right?)

The case will almost certainly be alloy heatsink at the moment then although I really did fancy trying one of those wooden ones (a'la bottlehead)

oh well I do have more heatinks to try...

Andy

Let's define terms a little better. When I quoted the size, it was for a piece of aluminum (or copper) that was to be attached to a metal chassis with thermal compound. Therefore, the "heatsink" is actually the entire chassis. The mounting piece is just to allow adequate transfer to the larger chassis. You need a pretty large piece of material to dissipate 6 or 7 watts. For a flat piece of aluminum 3mm thick, and standing vertical in free air, I'd guess something on the order of 50 square cm should be fine. You are eventually transferring the heat to the air, so it is a function of size and airflow. Heat can be transferred by radiation and convection. You can use a smaller radiator if you have air moving past the part. If you are in a vacuum, a much larger radiator is required.

Sheldon

 

[andrew_whitham]

Let's define terms a little better. When I quoted the size, it was for a piece of aluminum (or copper) that was to be attached to a metal chassis with thermal compound. Therefore, the "heatsink" is actually the entire chassis. The mounting piece is just to allow adequate transfer to the larger chassis. You need a pretty large piece of material to dissipate 6 or 7 watts. For a flat piece of aluminum 3mm thick, and standing vertical in free air, I'd guess something on the order of 50 square cm should be fine. You are eventually transferring the heat to the air, so it is a function of size and airflow. Heat can be transferred by radiation and convection. You can use a smaller radiator if you have air moving past the part. If you are in a vacuum, a much larger radiator is required.

Sheldon

oops I completely misunderstood, sorry. The "problem" then is stemming from the nature of the prototype. where the 'sink is in free air. By the sounds of it, I'm not getting anything too odd which is a relief. Although I will try this with a bigger heatsink to be sure I'm tempted to consign this to the 'working' pile and move on...

And I can say then that the case will (probably) be alloy

Fortunately the Tubecad software has arrived too. So many hours can now be spent trying to get a nice input stage

Cheers

Andy

 

[zarniwoop]

well, I hear people saying that the heatsink is too hot to touch. That doesn't say a lot, really. Those regulators can operate at 125C, usually, so if your's is like 80C than its no problem.

I'm running lm317s on 6b4g filaments with finned heatsinks, inside chassis, no connection to chassis. They get very hot but they haven't died yet, so I'm satisfied.

 

[Sheldon]

They can run pretty hot with no problems, but that's not the entire story. It depends on what else you have under the hood. If it's only resistors and wire, no problem. If you have electrolytic caps, their operating lifetime is very much affected by temperature. In that case, better to transfer the heat to the top plate, and allow free air flow from the bottom, to keep any sensitive underneath components cooler.

One other thing to keep in mind is that hotter temperatures will dry out some thermal compounds eventually. Doesn't mean you can't operate hot, but check things occasionally. Whenever possible, I like to keep heat generators (except for tube envelopes and some power resistors) cool enough to touch for a couple of seconds. I figure it can't hurt. The heat's gotta get out eventually, why let it build up?

Sheldon

 

[andrew_whitham]

Got it!, bigger heatsink required is all. Still within the realms of 'sensible' ended up with about 6cm x 6cm with a few fins. This is using the LM338T with its 3v dropout and associated warmth too.

So having played around with the Tubecad stuff I'm convinced enough to try the high current game. All other things being equal theres a useful reduction on 2HD

so I'm now in need of a +/- 60v signal from the 5v pk - pk supplied by my cd player

here comes the kicker though. I want to leave room for a bit of feedback . can anyone suggest a linear tube that'll suffice? I'm looking I suppose for an ECC88 (6dj8) with more gain. maybe around 50? (12AT7 is not linear enough)

Andy