Call Search
     

New to Ham Radio?
My Profile

Community
Articles
Forums
News
Reviews
Friends Remembered
Strays
Survey Question

Operating
Contesting
DX Cluster Spots
Propagation

Resources
Calendar
Classifieds
Ham Exams
Ham Links
List Archives
News Articles
Product Reviews
QSL Managers

Site Info
eHam Help (FAQ)
Support the site
The eHam Team
Advertising Info
Vision Statement
About eHam.net

   Home   Help Search  
Pages: [1]   Go Down
  Print  
Author Topic: Transistor Insulators  (Read 5085 times)
KB1GTX
Member

Posts: 462




Ignore
« on: August 06, 2012, 09:54:48 PM »

Which is better for heat transfer, Mica,silicon or the white ceramics?

For power supplies.
« Last Edit: August 06, 2012, 09:57:41 PM by KB1GTX » Logged
N4CR
Member

Posts: 1688




Ignore
« Reply #1 on: August 06, 2012, 10:16:52 PM »

I use the original mica and 'non-silicone' heat sink compound. The silicone based heat sink compound dries out after a few years and loses most of it's ability to transfer heat. Also, be aware that some heat sink compounds contain beryllium oxide and can be toxic so wash carefully when you are done.
Logged

73 de N4CR, Phil

Never believe an atom. They make up everything.
W0BTU
Member

Posts: 1794


WWW

Ignore
« Reply #2 on: August 07, 2012, 08:31:06 AM »

By itself, silicone heat sink compound is not an electrical insulator. It is only to improve the transfer of heat and make the device run cooler. If the circuit requires it, you still have to use a thin mica insulator (or something like it) to prevent a short between the device and the heat sink.

EDIT: Maybe I misread your post. I see you're probably referring to thin silicone rubber sheets.
« Last Edit: August 07, 2012, 08:34:20 AM by W0BTU » Logged

W4DRR
Member

Posts: 82




Ignore
« Reply #3 on: August 07, 2012, 09:28:17 AM »

Bergquist Sil-Pads are the way to go.
Logged

73,
Bob
KB1GTX
Member

Posts: 462




Ignore
« Reply #4 on: August 07, 2012, 09:37:01 AM »

What I'm playing with is shrinking a linear supply to as small as I can get it.
For the heat transfer paste I use the newer silver bareing artic freeze, it is by far better than the white silicon paste.

I guess I'm going to use a 2N5686 with no insulator on a copper based PC type heat sink with a fan and mount the heat sink on some insulated stand offs.


« Last Edit: August 07, 2012, 09:41:25 AM by KB1GTX » Logged
KD0REQ
Member

Posts: 1005




Ignore
« Reply #5 on: August 07, 2012, 09:42:50 AM »

Arctic Freeze is conductive, including the squeeze-out.  seriously conductive.

I'd consider cannibalizing one the transistors you intend to use to find out where the pellet is inside the case, and just put a dab of AF heatsink right where the pellet goes.  the truly anal builder, seeing this, might decide to lap-polish the heat sinks and transistor cases to insure totally flat surfaces and the best heat transfer possible.

but seriously, be careful where you put that stuff, it's flashover city.
Logged
W0BTU
Member

Posts: 1794


WWW

Ignore
« Reply #6 on: August 07, 2012, 10:11:10 AM »

Getting back to the original question, I found this: www.silicone.jp/e/catalog/pdf/rubber_e.pdf . Check out the thermal conductivity graph there. General purpose silicone rubber has only about 1/3 the thermal conductivity of mica, but thermal interface silicone rubber has twice the conductivity of mica.

If you can get it, I suspect BeO ceramic (toxic) beats them both hands down.
Logged

W9GB
Member

Posts: 2648




Ignore
« Reply #7 on: August 07, 2012, 10:54:58 AM »

Quote from: KB1GTX
What I'm playing with is shrinking a linear supply to as small as I can get it.
For the heat transfer paste I use the newer silver bareing artic freeze, it is by far better than the white silicon paste.
CAREFUL, you need to avoid the BIG mistakes that Pyramid and other low-cost linear power supplies made -- along that line of thinking.  
Later versions of those power supplies had to add a fan, to increase circulation, to address the design mistake -- and failure rates due to thermal issues.

You need a properly sized heat sink to transfer (radiate) the waste heat generated by the pass transistors.
You can not by-pass basic physics ... laws of thermodynamics ... out of desired design convinence.
Aavid (major heat sink mfg.) has many resources and specifications for the designer ... to select the proper heat sink.

I have been migrating from traditional Mica insulators/grease to Silicon-Fiberglass ThermoPads (by NTE and others)
available from major electronics distributors (Allied, Newark, Mouser, DigiKey, Future).

NTE's THERMO-PADS do away with the old fashioned mica wafer and conductive grease method of mounting power semiconductors.  
These thermally conductive insulators offer low heat transfer resistance while still providing high electrical isolation between the parts of the assembly.
The elastomeric material combines the electrical isolation of rigid insulators with the ability to conform to rough surfaces and reduce contact resistance in much the same manner as thermal greases.
Proper selection and use of these THERMO-PADS results in a securely-mounted power semiconductor and minimum resistance to the heat transfer between it and the heat sink.
« Last Edit: August 07, 2012, 11:00:06 AM by W9GB » Logged
AC5UP
Member

Posts: 3927




Ignore
« Reply #8 on: August 07, 2012, 11:39:37 AM »

Something To Consider:   If the actual 'pill' inside a TO-3 pass transistor is approximately the same diameter of a pencil eraser, and it could be, one of the challenges of proper heatsinking is how best to transfer this pinpoint of heat to a much larger structure where it can spread out and radiate.

How do the commercial dudes achieve this? Coincidently enough I'm working on an Acopian 12 volt rack mount power supply that had two issues: One slightly leaky electrolytic and a maximum output of 12.4 volts DC. When Acopian says it's a 12 volt supply that does not mean 13.8 VDC. I had a pair of good used replacement caps on hand and figured out how to bump the internal voltage reference a couple of volts higher. This is a 17 amp supply that uses eight 2N3055 pass transistors. I have a 19 amp Lambda sitting next to me that also uses eight pass transistors.

Wassup' with that? A 2N3055 is a 15 amp / 150 watt transistor and TWO of them should be enough to handle the draw from either supply. But......... Eight thermal contact points can dump heat more effectively than two and eight transistors cover four times the heatsink space. That spreads the heat mo'better so the heatsink is radiating more evenly, not just from the hot spot in the center. This means that a quad of smaller TO-220 transistors spread out over a heatsink could be more reliable than one Mongo transistor working near its capacity.

Running with a buttload of margin makes the chance of any pass transistor welding up to the point of overvolt slim to none. By Astron standards they're both overbuilt and that's fine by me.    Cool
Logged

Never change a password on a Friday                
KE3WD
Member

Posts: 5689




Ignore
« Reply #9 on: August 07, 2012, 01:50:34 PM »

Bergquist Sil-Pads are the way to go.

Seconded. No compound needed.  Wonderful. 

If you must use the older quartz insulators, by all means use the stuff one of my shopmates once labeled as, "Annoying White Compound" on the bench jars, the stuff made from Zinc Oxide, rather than silicone compounds, which do dry out over time. 

73
Logged
KB1GTX
Member

Posts: 462




Ignore
« Reply #10 on: August 07, 2012, 03:52:13 PM »

Quote
You need a properly sized heat sink to transfer (radiate) the waste heat generated by the pass transistors.
You can not by-pass basic physics ... laws of thermodynamics ... out of desired design convinence


The Idea here is to see just how far it can go and the hoped out come is to have a coffee cup heating power supply that'll make me the envy of my friends!
Logged
Pages: [1]   Go Up
  Print  
 
Jump to:  

Powered by MySQL Powered by PHP Powered by SMF 1.1.11 | SMF © 2006-2009, Simple Machines LLC Valid XHTML 1.0! Valid CSS!