Long discussion.. ok let's cut it short.
How do they all compare sonically??
I am quite surprised that nobody has claimed yet that it "gives depth to the soundspace" , that it causes "sonic bliss" or that it cures bad breath, although now that I think of it, the last claim *can* be verified.
Modern ones are single use.
The mid 1850's ones were washable and reusable.
Of course as thick as a party balloon.
Back to thermal insulators, after a few dozen thermal cycles under high pressure, the rubbery material tends to take a fixed shape, matching surface imperfections.
In theory you can pull and replace them in the *exact* same position, which is sort of iffy.
For DIY use, I have had good luck with using automotive Anti-Seize compound on Transistors and CPU's and the thermal transfer is quite good. It is cheap, only $5-7 for a large tube, so if you waste some, it is no big deal. Also, the transistors "un-stick" themselves easy if you have to remove them (great for car amps)
AN adhesive I use, with also good results is COPPER RTV sealant, not the regular kind. On projects I want to be semi-permanent, that stuff is awesome.
AN adhesive I use, with also good results is COPPER RTV sealant, not the regular kind. On projects I want to be semi-permanent, that stuff is awesome.
Where to buy 247 mica insulators
Hello all,
Very interesting thread and I have learned much reading your experiences and preferences. Thank you! I am an old school tech that would prefer to do things the traditional way with mica and goop. Unfortunately my supply of 247 style mica insulators is dwindling. Can anyone here point me to a reasonable source to replenish my supply?
Thanks!
Hello all,
Very interesting thread and I have learned much reading your experiences and preferences. Thank you! I am an old school tech that would prefer to do things the traditional way with mica and goop. Unfortunately my supply of 247 style mica insulators is dwindling. Can anyone here point me to a reasonable source to replenish my supply?
Thanks!
Hi Mike
TO-247MICA - NTE ELECTRONICS MICA FOR TO-247 CASE | Newark element14
or quick one with free shipping
To 247 Transistor Mica Insulator Insulation Sheet x 50 | eBay
Anyway verify all dimensions if is ok for your transistors
i know is normalised standard but some time
they have differents throughole diameters.
Kind regards
Hi Mike
TO-247MICA - NTE ELECTRONICS MICA FOR TO-247 CASE | Newark element14
or quick one with free shipping
To 247 Transistor Mica Insulator Insulation Sheet x 50 | eBay
Anyway verify all dimensions if is ok for your transistors
i know is normalised standard but some time
they have differents throughole diameters.
Kind regards
Soundhappy and labjr
Thank you for pointing me to these sources. Thanks to your help I have 100 pieces on order.
Fantastic with all the new micas is diy potential for many amplifiers
Each one can be divised on two layers with cuter its easy to do so
100=200 pcs
Just to work out the math -
The devices I'm using only come in TO-220, unfortunately.
Using a microscope, a razor, cussing, and about a 50% yield, I was able to get a handful of mica insulators split down to around 1.7mil (43um).
I have coming some aluminum oxide (aka "alumina", "Al2O3") insulators, 600um thick. 10pcs to 220 Ceramic Insulator Plates High Temperature Thermal Conductivity | eBay
According to
the thermal conductivity of mica is 0.7 W/(m*K).
The thermal conductivity of alumina is 24W/(m*K).
To compare, divide the thermal conductivities by the thicknesses:
mica: 16279
alumina: 40000 ....alumina wins by better than 2x!
So, I'll go with the alumina insulators.
(BTW, most TO-220 alumina insulators I've run into give a thickness of 1000um, so using those would make the race closer, but still alumina wins. Hope the ones I'm getting are as thin as they claim to be, and are actually alumina - they only say "ceramic"... also hope that they don't crack when I tighten down the devices!).
The devices I'm using only come in TO-220, unfortunately.
Using a microscope, a razor, cussing, and about a 50% yield, I was able to get a handful of mica insulators split down to around 1.7mil (43um).
I have coming some aluminum oxide (aka "alumina", "Al2O3") insulators, 600um thick. 10pcs to 220 Ceramic Insulator Plates High Temperature Thermal Conductivity | eBay
According to
the thermal conductivity of mica is 0.7 W/(m*K).
The thermal conductivity of alumina is 24W/(m*K).
To compare, divide the thermal conductivities by the thicknesses:
mica: 16279
alumina: 40000 ....alumina wins by better than 2x!
So, I'll go with the alumina insulators.
(BTW, most TO-220 alumina insulators I've run into give a thickness of 1000um, so using those would make the race closer, but still alumina wins. Hope the ones I'm getting are as thin as they claim to be, and are actually alumina - they only say "ceramic"... also hope that they don't crack when I tighten down the devices!).
I have successfully mounted TO-247 type cases directly to black-anodized heatsinks, with the standard goop, of course. That should be about of thermally conductive as it can get, short of somehow growing a diamond film onto the heatsink.
Not a procedure I'd usually recommend, however. The anodizing is extremely thin and easily broken by any debris trapped between the surfaces, and the edges of any holes have to be beveled with no rough edges projecting up. Definitely for experimental purposes only.
Not a procedure I'd usually recommend, however. The anodizing is extremely thin and easily broken by any debris trapped between the surfaces, and the edges of any holes have to be beveled with no rough edges projecting up. Definitely for experimental purposes only.
Last edited:
The ebay link is in my post, but no hard data there. I did find some Chinese made alumina insulators of the same dimensions on alibaba, so I'm reasonably confident that the ones coming are alumina too (likely by the same manufacturer).
You do use goop, like with mica (I have some alumina ones from Mouser, thicker though). They are very hard and rigid, moreso than mica. They DO NOT bend!
I didn't check Kerafol, not having a source for diy quantities of it. If you have thickness and thermal conductivity data for it, divide it out and post the result. Unless it works without goop, then the comparison wouldn't work.
You do use goop, like with mica (I have some alumina ones from Mouser, thicker though). They are very hard and rigid, moreso than mica. They DO NOT bend!
I didn't check Kerafol, not having a source for diy quantities of it. If you have thickness and thermal conductivity data for it, divide it out and post the result. Unless it works without goop, then the comparison wouldn't work.
The insulators arrived, they measure 63um on my old micrometer. I wrote the vendor, he says they are indeed alumina. I already have a couple fets mounted on the heatsink with them, and set up a constant current circuit to set power consumption for a test. But i have to wait till a thermocouple meter and probe arrive bfore I have any new data to report. Maybe tomorrow..
A data point
I got the thermocouple meter in today and did some measurements. The test setup was:
I measure temperature a 3 points across the top, 3 across the bottom, and just above where the FETs were mounted after the temperatures stabilized. And on the transistor tabs.
Across the heatsink, the average of temperatures was 60.47C. (Other than at the FET mounting points, all temps were within 1 degree of 57C, so sources are well distributed). Based on the overall average, the heatsink is calculated to be 0.74C/W (1.35 W/C).
At the mounting points, average of the 2 temps was 62.8C, so effective heatsink calculated from that would be 0.79C/W (1.27W/C).
And, finally, the insulators: 5.2 W/C (0.19 C/W).
So, yeah, those insulators are pretty good!
...and 50W is about all I'd want to run on one of these heatsinks.
Theta jc of. the part is 0.7K/W, so the junction would be at 80.3C. Max rated is 175C, so no issue there, but heatsink much hotter might cause burns and bake nearby electrolytics..
I got the thermocouple meter in today and did some measurements. The test setup was:
- One of the hatsinks from this smallish power amp enclosure:
DIY new 2412B Full Aluminum Enclosure/mini AMP case/power amplifier box/chassis-in Consumer Electronics on Aliexpress.com | Alibaba Group. Dimensions are: 246mm long, 110mm high, 25mm deep, 35 vertical fins each 20mm long and 2.2mm thick - Two TO-220 MOSFETs mounted using the 0.6mm ceramic alumina insulators and goop, each centered vertically and about 1/3 of the way from each end
- Each device dissipating 25W (50W total), no fans, heatsink fins open
- 23.5C ambient
I measure temperature a 3 points across the top, 3 across the bottom, and just above where the FETs were mounted after the temperatures stabilized. And on the transistor tabs.
Across the heatsink, the average of temperatures was 60.47C. (Other than at the FET mounting points, all temps were within 1 degree of 57C, so sources are well distributed). Based on the overall average, the heatsink is calculated to be 0.74C/W (1.35 W/C).
At the mounting points, average of the 2 temps was 62.8C, so effective heatsink calculated from that would be 0.79C/W (1.27W/C).
And, finally, the insulators: 5.2 W/C (0.19 C/W).
So, yeah, those insulators are pretty good!
...and 50W is about all I'd want to run on one of these heatsinks.
Theta jc of. the part is 0.7K/W, so the junction would be at 80.3C. Max rated is 175C, so no issue there, but heatsink much hotter might cause burns and bake nearby electrolytics..
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.