diyAudio (
-   Parts (
-   -   unpotting, solvent, melting agent, removal agent, for transistor or IC cases?? (

bear 3rd October 2012 11:57 PM

unpotting, solvent, melting agent, removal agent, for transistor or IC cases??
Ok, I have a large quantity of recently acquired transistor (TO-92 case, if that matters) that I am convinced are bogus versions. Yes they function, but the curves don't match the mfrs at all.

I would like to un-pot via chemical means the die + leads from the plastic case. What can be used that works??

Mechanical means are not an option.

I can then compare the known good NOS devices to the recently acquired devices and probably know for sure... maybe. Well, if it is night and day, then for sure. We have microscopes.


PS. a search or two did not reveal this answer... if there is one.

wrenchone 4th October 2012 12:10 AM

Glacial acetic acid is a good starting point. The more drastic alternatives should be avoided, as they are really nasty (not that acetic acid is a treat). A better alternative would be to check the parts out and see if they've been remarked, the usual conterfeiter's strategy ( actually making parts is too much work, so just get some inferior parts/castoffs and mark them with a hot-selling number).

Are the parts laser marked, or with ink?

WSJ 4th October 2012 12:18 AM

I used un-potting chemicals, it expanded the epoxy and made big mess.
The PCB inside had the parts ripped apart. I found it to me useless.

Conrad Hoffman 4th October 2012 01:11 AM

Why aren't mechanical means an option? I've split TO92s well enough to see the die and confirm what it was. If you have a lot to experiment with, crush a couple in a vise from the sides and see if they'll split. Don't know what type you're dealing with, but I always found the old National Semi book on discretes useful as it had a die image for each type. I was able to ID many surplus house numbers that way.

WSJ 4th October 2012 01:39 AM

I used the failure analysis lab at Fairchild Semiconductor in 1966,
where epoxy devices were analyzed. We would grind the epoxy down and
use polishing wheels with different grit compounds until the die was visible.
We did this with hundreds of transistors and ICs.

bear 4th October 2012 12:10 PM

I can try the mechanical method(s)... would prefer a straight melt away... no diamond wheels here right now...

...Conrad, I will try the vise crush method, do you mean the flat & curved sides , or the wide side in the jaws of the vise?

These are not Nat Semi parts. I will have to compare NOS to these... regardless they do *not* look the same when curve traced.

Found online in a paper that commercially strong acids are used, and that the die remains impervious... doesn't seem like "plan A".

wrenchone, I am unsure if they are laser marked, they well may be. they don't look quite right under a high power loupe. I'd say, not ink, now that you mention it. I can test for that with a bit of acetone, that ought to remove the ink and not the laser etch, I would expect.

What effect will the glacial acetic acid have on the plastic??
How long in the acid??
Have you done this??

Also, it might be easier to get "battery acid" than glacial acetic.
Is there an easy source for glacial acetic (I used to get Kodak, but that was when there were still stores that sold photo supplies for the home darkroom)?

bottom line here is Grrrrr...

SY 4th October 2012 01:38 PM

Epoxy won't melt, alas. Crosslinked thermoset. You're down to acid or mechanical.

Elvee 4th October 2012 02:32 PM

There are a number of specially formulated solvents for thermosetting plastics, including epoxy.
A typical example is the Panasolve range, but there are many alternatives now, including (relatively) green ones.

hweb 4th October 2012 05:38 PM

Fuming nitric acid (you get it to fume by boiling it) will disolve the mold compound.

Use very little acid, a fume hood and other safety equipment. It's very nasty business.

A much better idea would be to engage the services of an independant semiconductor failure analysis lab (let them play with the acid).

bear 5th October 2012 01:05 AM

at this price point, the independent lab will cost more than the parts... :(

regular nitric acid won\'t do a thing??
wish I could hit it with a laser and heat it locally to boil a teeny tiny bit...



I think I am stuck with being screwt, as they say.

They are jfets, fwiw, and they do function, just that they are very unlikely to the be ones that were claimed. Bummah...

All times are GMT. The time now is 06:09 PM.

vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2017 DragonByte Technologies Ltd.
Copyright 1999-2017 diyAudio

Content Relevant URLs by vBSEO 3.3.2