I'm close to fire-up. I could actually get it going and do the biasing here at work (I don't have NEAR enough DMM's to get the job done at home, plus they have the neat little clippy probes), since all the boards are done. I'm waiting for my Alps pot to come in to finish the front panel.
And as I said before some final cosmetic stuff as well. Should look totally killer. I received my straw hat LEDs last night for my back-lit, chrome polished aluminum DIY knobs. Just have to figure out how to glue veneer to aluminum, and I think I have the epoxy that will do the trick.
And as I said before some final cosmetic stuff as well. Should look totally killer. I received my straw hat LEDs last night for my back-lit, chrome polished aluminum DIY knobs. Just have to figure out how to glue veneer to aluminum, and I think I have the epoxy that will do the trick.
I listened and I'm not worried. I'm a mechanical engineer who has dealt with thermal issues before. Do I know anything about amps? Hardly. I guess we'll see.
How thick do you think the anodizing is?
Would it be easy to remove?
Conrad heatsinks polish the mounting surface for the very reason you state.
It's usually spec'd at 0.005" or less build-up. Two ways to remove it: grind or milling machine. Anodize will chew up end mills too. Best way is to plunge a couple thousands to break through, and then traverse. People have been building with the 4 and 5U enclosures without any problems, so I don't think you should bother with the anodize. Best case, you'd be able to bias a little bit higher if the heat is tamed a bit more.
0.005'' is 0.000127 meter.
30 W/m·K divided by 0.000127 makes 236,220 W/K for 1 square meter.
For a TO247 (21 x 26 mm), that makes ~129 W/K
aka an additional thermal resistance below 0.01 C/W
Hard anodise is more likely to be around 1thou (1mil, or 0.0000394m), ordinary/soft anodize is likely to be <1thou.
The dye in (in rather than on) the anodise may increase the Thermal Resistance, but still bloody brilliant.
Alox thermal spacers are usually 40thou (0.001m) to 160thou thick (0.004m) and some prefer these to using 1 or 2thou mica for best thermal conductivity.
The dye in (in rather than on) the anodise may increase the Thermal Resistance, but still bloody brilliant.
Alox thermal spacers are usually 40thou (0.001m) to 160thou thick (0.004m) and some prefer these to using 1 or 2thou mica for best thermal conductivity.
W/m^2 degK is area, not thickness. Can you imagine dissipating 236,000W in one square meter? Me neither.
heat transfer (W), q, is (1/((1/h)+(t/k)))*A*deltaT
h=heat xfer coefficient (generally estimated)
t=thickness
k=thermal conductivity (in W/m^2 degK)
A=area
anodize conductivity: 30W/m^2 degK
free air convection on vertical plane, delta T of 30 degC (303 degK): coefficient is 5 W/m^2 degK
Area 1m^2 (imagine a flat sheet, 1m^2 in area)
(1/((1/5)+(.000127/30)))*1*303 = 1515 Watts
Last edited:
W/m·K for thermal conductivity is derived with dimension analysis from Fourier's law. (for 1-dimensional heat flow)
Q = - Lambda · A · dT/dx
Q is heat flow (W)
A is area (m2)
dT = temperature difference (K, negative for positive flow)
dx is thickness
W = Lambda · m2 · K/m
=> Thermal conductivity = (W · m) /( m2 · K) aka W/m·K
(Thermo 101, 1st year Delft university, gawd knows how long ago. Which can only imply I'm so insane, I should have become a physics teacher for stupid private school girls with perky breasts)
A 2N5684 has a Pd of 300W, it can easily dissipate 150W.
A TO-3 package has a surface area of ~700 mm2
150W / 700mm2 = 215,000 W/m2
Yes, I can. (thanks for the reminder)
I think I just spit out my coffee! Ah yes, my mistake. I was mixing up the units of thermal conductivity and the heat transfer coefficient.
For the P1 and P2 trimmer pots on the gain stage, do I turn them completely CCW to start the biasing procedure? I did some measurements and the pot doesn't go from 0 to 500. It just sorta hangs around the 250 area. I went full CCW AND CW and it only changes a few ohms. Bought Bourne 3296 series, style W. I'm stumped.