I'm working on my class A amp design and I had an idea to use aluminium oxide washers for the output stage transistors. These washers have impressive thermal resistance <0.3 deg/W....
The question that I have is as follows. Has anyone used them and are the results positive? Also what is the appropriate method of using them? Do they provide a good contact on their own or should one still use a very thin layer of standard thermal paste to quarantee a good contact.
I have not seen these washers in real life. I'm just thinking of ordering them...
Thanks for any feedback
The question that I have is as follows. Has anyone used them and are the results positive? Also what is the appropriate method of using them? Do they provide a good contact on their own or should one still use a very thin layer of standard thermal paste to quarantee a good contact.
I have not seen these washers in real life. I'm just thinking of ordering them...
Thanks for any feedback
There probably is a difference
What you need to use for "optimal dielectric conditions" is Beryllium Oxide washers (ref radio frequency stuff). Illegal in Europe, but I think you can get them from Canada in TO3 form factor. TO247 washers are unavailable. More on this under the X-100 thread in solid state.
These should be better because they help reducing the apparrent capacitance of the device.
Aluminum Oxide washers are just alu washers which have an oxide layer on both sides and are probably no better than Silpad for dielectric properties.
AluO and BeO units are prone to performance reduction given the need for thermal grease.
Petter
What you need to use for "optimal dielectric conditions" is Beryllium Oxide washers (ref radio frequency stuff). Illegal in Europe, but I think you can get them from Canada in TO3 form factor. TO247 washers are unavailable. More on this under the X-100 thread in solid state.
These should be better because they help reducing the apparrent capacitance of the device.
Aluminum Oxide washers are just alu washers which have an oxide layer on both sides and are probably no better than Silpad for dielectric properties.
AluO and BeO units are prone to performance reduction given the need for thermal grease.
Petter
hifi said:
Ergo, I used SIL-PAD K10 (or SIL-PAD 2000) in my Zen (I obtained samples of each). They are the same thermally, one is a higher (MIL?) spec. They are used dry. Being a born sceptic, I confirmed the 0.2C/W by my own measurements - good stuff! Bolt them down reasonably tight, but don't go mad - you'll distort the flange on the TO3 and compromise the thermal joint. These washers 'squish' down a little and are therefore also non-reusable. They are non toxic.
Good luck.
Pete
Thanks to everyone for your replies.
The reason I consider the aluminium oxide washers is that these are the best solution offered by ELFA, a part supplier in Sweden... As I live in Estonia (near Sweden and Finland) it is very hard to find suppliers for such small orders with reasonable shipping costs.... If someone knows an online part supplier (especially in Scandinavia or Europe) that accepts international orders I would be interested.
Ergo
PS. to /micke
In our country the beryllium problem most likely wouldn't scare people as there are so many hazardous materials, building materials, etc around as a legacy from Soviet Union times that most don't think of it much as one would most likely be exposed to it somewhere anyway...
The reason I consider the aluminium oxide washers is that these are the best solution offered by ELFA, a part supplier in Sweden... As I live in Estonia (near Sweden and Finland) it is very hard to find suppliers for such small orders with reasonable shipping costs.... If someone knows an online part supplier (especially in Scandinavia or Europe) that accepts international orders I would be interested.
Ergo
PS. to /micke
In our country the beryllium problem most likely wouldn't scare people as there are so many hazardous materials, building materials, etc around as a legacy from Soviet Union times that most don't think of it much as one would most likely be exposed to it somewhere anyway...
http://www.farnell.com
they are a lot better than elfa. quicker orders and cheaper price they are uk based...its cheaper and the shipping goes faster (also the shipping cost is less) for me to order fron farnell than elfa altough i live in sweden!
/micke
they are a lot better than elfa. quicker orders and cheaper price they are uk based...its cheaper and the shipping goes faster (also the shipping cost is less) for me to order fron farnell than elfa altough i live in sweden!
/micke
If the Aluminium Oxide washers are the one I'm thinking of, white 2-3mm thick ceramic, I have heard them called 91 day wonders, 1 day after the warranty end the washer fails. Of course they heve most likley improved in the past few years.
As to the sound, I believe YBA used a three layer insulator in their amps, the middle layer was copper foil, supposedly helped with capacitance to the heatsink/chassis?
I found a similar scheme in an old mainframe PSU, the power transistors were mounted via 2 layer ceramic washers, when I seperated the washers I found one of the interal faces was metalised nd included a grounding wire.
aavid/thermalloy make a pad called A-Dux
<a href="http://www.aavidthermalloy.com/products/options/pads.shtml">A-Dux</a> without Adhesive</b>
Thermal Impedance 0.16°C/in<sup>2</sup>/W
Swedish Distributors for aavid/thermalloy are listed <a href="http://www.aavidthermalloy.com/bin/catdbctry.pl?Country=Sweden&GoCountry.x=12&GoCountry.y=16"><b>HERE</b></a>
theirs a sample request form on their site also.
<center><font size="+1" face="fantasy">Regards
James</font></center>
As to the sound, I believe YBA used a three layer insulator in their amps, the middle layer was copper foil, supposedly helped with capacitance to the heatsink/chassis?
I found a similar scheme in an old mainframe PSU, the power transistors were mounted via 2 layer ceramic washers, when I seperated the washers I found one of the interal faces was metalised nd included a grounding wire.
aavid/thermalloy make a pad called A-Dux
<a href="http://www.aavidthermalloy.com/products/options/pads.shtml">A-Dux</a> without Adhesive</b>
Thermal Impedance 0.16°C/in<sup>2</sup>/W
Swedish Distributors for aavid/thermalloy are listed <a href="http://www.aavidthermalloy.com/bin/catdbctry.pl?Country=Sweden&GoCountry.x=12&GoCountry.y=16"><b>HERE</b></a>
theirs a sample request form on their site also.
<center><font size="+1" face="fantasy">Regards
James</font></center>
Toxicity of BeO
BeO is not a problem until you start grinding it up to get dust -- it is perfectly safe until that point. The effects when inhaled (ground up dust) is similar to that of Astbestos (but possibly a little stronger). It is remarkable that two materials with such opposite thermal performance have similar negative effects on the human body.
I used to work with radical equipment which had a BeO window. BeO really is a remarkable material -- thermal and nuclear "aperture" are probably closest to "nothing" you can get while the dielectric constant is nice and low.
If you get BeO washers, get the thickest units to maximize electrical performance. The conductivity of the washer is so high that you are unlikely to notice the difference.
Petter
BeO is not a problem until you start grinding it up to get dust -- it is perfectly safe until that point. The effects when inhaled (ground up dust) is similar to that of Astbestos (but possibly a little stronger). It is remarkable that two materials with such opposite thermal performance have similar negative effects on the human body.
I used to work with radical equipment which had a BeO window. BeO really is a remarkable material -- thermal and nuclear "aperture" are probably closest to "nothing" you can get while the dielectric constant is nice and low.
If you get BeO washers, get the thickest units to maximize electrical performance. The conductivity of the washer is so high that you are unlikely to notice the difference.
Petter
grounding middle layer
tvi,
If they grounded the middle layer, this would remove any positive (capacitance) effects. Thus, this grounding scheme was probably something else (such as bootstrap which would be VERY effective). There is however the alternative that they did not know what they were doing -- something which seems quite prevalent in the audio industry in general (no comment versus this manufacturer in particular).
Now, if we want to borrow from YBA, we could use SilPad's onto machined copper blocks which are bigger than the transistor you would loose very little thermal performance to heatsinks if you use say a large piece of SilPad. Now connect the copper block to a suitable bootstrap point in your circuit ...
If you want the best, do what the RF people do, go with BeO. In Estonia it MIGHT be possible to get "new old stock" or pillage old russian equipment -- that is if they used transistors and not tubes in such positions.
The next step would be to go with tubes which have integrated heatsink, thermal washers, low capacitance etc. ...
Petter
tvi,
If they grounded the middle layer, this would remove any positive (capacitance) effects. Thus, this grounding scheme was probably something else (such as bootstrap which would be VERY effective). There is however the alternative that they did not know what they were doing -- something which seems quite prevalent in the audio industry in general (no comment versus this manufacturer in particular).
Now, if we want to borrow from YBA, we could use SilPad's onto machined copper blocks which are bigger than the transistor you would loose very little thermal performance to heatsinks if you use say a large piece of SilPad. Now connect the copper block to a suitable bootstrap point in your circuit ...
If you want the best, do what the RF people do, go with BeO. In Estonia it MIGHT be possible to get "new old stock" or pillage old russian equipment -- that is if they used transistors and not tubes in such positions.
The next step would be to go with tubes which have integrated heatsink, thermal washers, low capacitance etc. ...
Petter
I found the article about the Mica/Copper/Mica insulators, it was in the September 1987 issue of Hi-Fi News and Record Reviews.
Its mentioned in and interview with the designer of YBA,
<i>"Also, I use Copper mica between devices and ther radiator (heatsink) to break the capacitance that will always exist with mica. This non-linear capacitance affects the signal, giving what we call 'the transistor sound' because the case is acting as the collector. I have measured this at 10kHz, it is not exactly an attenuation but more a sort of noise about 60dB down caused by the limited quantity of electrons moving through the device."</i>
In the next issue Paul Miller reviewed the YBA 1 Power amp and noted:
<i>"the devics have a mica/copper/mica sandwich between them and the heatsink, with a series wire-link from all points of the copper to earth"</i>
Maybe he is taking it to a quiet earth point not the possibly noisier chasis earth?
The amp review always stuck in my mind because the amp uses an unusual inductively coupled feedback loop via a winding around the 10uF polyprop input cap, this winding is inductively coupled to the output. This could also explain the Mica/Cu/Mica, is it possible he needed them to keep this feedback scheme stable?
I wonder if YBA uses something like a <a href="http://www.bergquistcompany.com/thermal/pdf_2000/SilPadDG_sheild.pdf">SilPadDG sheild</a> theses days?
<center><font size="+1" color="#000000" face="times new roman">Regards
James</font></center>
Its mentioned in and interview with the designer of YBA,
<i>"Also, I use Copper mica between devices and ther radiator (heatsink) to break the capacitance that will always exist with mica. This non-linear capacitance affects the signal, giving what we call 'the transistor sound' because the case is acting as the collector. I have measured this at 10kHz, it is not exactly an attenuation but more a sort of noise about 60dB down caused by the limited quantity of electrons moving through the device."</i>
In the next issue Paul Miller reviewed the YBA 1 Power amp and noted:
<i>"the devics have a mica/copper/mica sandwich between them and the heatsink, with a series wire-link from all points of the copper to earth"</i>
Maybe he is taking it to a quiet earth point not the possibly noisier chasis earth?
The amp review always stuck in my mind because the amp uses an unusual inductively coupled feedback loop via a winding around the 10uF polyprop input cap, this winding is inductively coupled to the output. This could also explain the Mica/Cu/Mica, is it possible he needed them to keep this feedback scheme stable?
I wonder if YBA uses something like a <a href="http://www.bergquistcompany.com/thermal/pdf_2000/SilPadDG_sheild.pdf">SilPadDG sheild</a> theses days?
<center><font size="+1" color="#000000" face="times new roman">Regards
James</font></center>
James
Mica exhibit very good insulating properties, this is known since 100 years and over. I never have measured such non-linearities. A mistake, I presume. The most important thing is to use the exact quantity of silicon grease to cancel the air layer between the mica and the metal (which must be perfectly flat machined).
Regards, P.Lacombe
Mica exhibit very good insulating properties, this is known since 100 years and over. I never have measured such non-linearities. A mistake, I presume. The most important thing is to use the exact quantity of silicon grease to cancel the air layer between the mica and the metal (which must be perfectly flat machined).
Regards, P.Lacombe
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.