RE: Experiment to Test Tandem LM395 IC Darlington Modules

D_Hersey — Thu, 25 Dec 2014 18:31:02 GMT

Here is a useful application note regarding the LM195:

ftp://ztchs.p.lodz.pl/PRP/PRP_2008_2009/AN-110.pdf

RE: Experiment to Test Tandem LM395 IC Darlington Modules

DAB — Mon, 24 Nov 2014 20:05:51 GMT

Excellent post John.

You can also set up a small circuit to better balance the current flow between the devices, but as long as they behave, it looks like you have a good solution.

Also, all of us have a burn scar somewhere on our hands. We engineers just have to touch stuff, you know!

DAB

RE: Experiment to Test Tandem LM395 IC Darlington Modules

shabaz — Mon, 24 Nov 2014 19:03:42 GMT

I'd not heard of this device before, but such a bullet-proof transistor replacement seems very interesting. Thanks for the investigation!

RE: Experiment to Test Tandem LM395 IC Darlington Modules

D_Hersey — Mon, 24 Nov 2014 18:52:09 GMT

I'm thrilled that you tried my suggestion! IIRC these were developed by Widlar et al at the same time as other Linear Monolithic devices such as voltage regulators of the day. It uses the same thermal shutdown circuit, again IIRC, leveraging the design development.

Just as in science, in electronics we need to make sure that we have done our research. And I don't mean this personally whatsoever, directing it at the both of us if anywhere, typically, common problems have common solutions. So we have to spend some of our design effort just trying to see if someone else has done it already and put the solution in the public domain. This is true in the software domain as well, there is a lot of available code out there that we can re-use rather than write ourselves.

When we use large transistors (esp in linear mode) we do well to incorporate thermal feedback or at least model our situation well so as to be sure to avoid thermally calamitious regions of our control space. With modern microprocessors we can cheaply and easily implement escalating regimes prior to cold shutdown, such as speeding up a fan.

In the world of one-offs and few-offs, using the LM195 or a big mosfet on a bigger heatsink may be the way to go because it may be less compute-intensive to implement than an ideally-scaled solution. With production volume, design effort can sometimes be leveraged over more units. The marketing ideation behind this part is aimed right at someone like you, in the field, needing to implement something that works quickly. I think the LM195, due to its relative ease-of-implementation, would be a great boon to hobbyists if they were made aware of it. A classic in the classic sense.

If you have some down-time I recommend reading the application notes of the guys from this early analog era. Their rendering was typically fresh, vital and market-driven.