zener sorting

[watchmaker]

I wound up with a bunch of stuff that looks like it once came home in a lunch box.  Aside from about 300 boxed ECG, there was a large cache of unlabeled diodes; over 500.  Maybe 1000.  Turns out they were blue banded so at least I knew they were zeners.

Thanks to a post elsewhere (I forgot, sorry), I decided to sort them with my CCCV power supply.  Set the current to 3 mA.  Ultimately I figured out to tie maybe 20 anodes together at my test probe and then test each cathode.  Started at 7v and determined the clamping voltage.  Removed and bagged em.

Then I raised the voltage to 13V.  Repeat.  Then 20, eventually up to 40.

Not only was this fairly fast, but I was able to classify zeners having non standard voltages.

Beats curve tracers for sorting.  Still will verify before using though.

[jwet]

Clever- It doesn't up come up a ton but I did this recently in a more barbaric way.

It reminds me of those brain teasers where you have a 2 oz cup and 6 ounce cup and you want to make 1 cup with fewest moves...

[Ian.M]

Risky.  Many bench PSUs have enough output capacitance to zap a Zener (or LED) if you rely on the current limit and connect the diode while the PSU is switched on.  A better approach is to set the PSU to lets say 30V, and use a series resistor to limit the current.  A 5.7K 1/4W resistor straight on the + out binding post will do the job.  Croc clip the other end of your resistor to your DMM's red probe and - out to the black probe, or better, if you've got a lot to sort, make a test jig with two L shaped brass strips screwed to an insulating base, with their ends V notched to fit the leads, across the diode body so you don't have to fiddle with the probes!

[RoGeorge]

Important pitfall that might happen with some lab power supplies:  Don't count on the constant-current mode when the voltage is set too high.

For example, my Rigol DP832 has a few hundreds uF inside, right before the output connector.  These capacitor are placed AFTER the current sensor.  Therefore, the current regulator won't see the current discharged by the capacitors directly to the load.

If I set the voltage for 32V and 20mA, the output capacitor will charge to 32V.  Now, if I let it on and connect at the output, let's say an LED, then all the energy stored in the output capacitors will suddenly discharge through the LED I've just connected.

The LED will die instantly, with a pop sound.
Did that once, and I didn't see it coming. 

[MarkT]

A series resistor to limit current is the way to defend against modern crap SM bench supplies - Get a linear bench supply with proper current limiting if you can, some even have foldback limiting.

[Zero999]

The output capacitance is less of a problem if the zener is connected first, before switching the power on.

[watchmaker]

You would think by now I would know to give complete details. 

 Yes, I am using an Instek GPS 4303 linear power supply.

[CaptDon]

Those ECG parts in the little yellow boxes could have a very great value, especially the discontinued hard to find devices. They are numbered identically to the NTE parts and cross reference easily to Motorola HEP, RCA SK and GE replacement lines! If you inventory the lot and how many of each you could sell the whole lot. I just bought new-in-the-bag ECG923 regulators (ua723 in the TO-5 can) and the ECG778A's are always in demand. Some of the physically big ECG hybrid audio output modules are always in demand as replacements for the STK Toshiba parts.

[watchmaker]

I did inventory them and am using the ECG system as my standard.  I have the NTE app and I bought the paper versions of the Replacement Guide and IC Guides even though they are available in PDF.

Pretty good run from 100 to 300, then another 100 or so up to the 5000s.  The system makes it very easy to select and find things. I keep unneeded boxes for future additions.  Put a load of 1n4148s into an ECG519 box for example.

A system for organization is everything.  In my business, my staffs, jewels, pinions, mainsprings, etc are organized in a way that I can easily pick out what I need.  I am amused by watchmakers who either have things all jumbled together and have to through a pile and then order a part they have but could not find, or those who develop sophisticated computer-based schemes to keep track.  I am very visual and I find physical inventory control is the best for me.

While it is not my goal to be DigiKey, I learned in watchmaking the importance of buying large lots of components when others deemed them "worthless".  I spent hours sorting pinions by OD and leaf count and today several colleagues are waiting for my widow to call.  Beats the Hell out of making them.  (Just doing the setup can take an hour, the cutter has to be dead center with "no" tolerance.)  https://www.historictimekeepers.com/Omega%20Jump%20Seconds.htm


May not get to the point where I am as proficient at electronics as I am at watch restoration, but at least I am planning for it.

[jonpaul]

in 1990s used TEK 576 curve traces with diode axial test adapter to sort 5000 special 150V DO=4 Zeners for GO/NOGO , looked for votage at 1 mA and amount of Zener noise at knee.

Last year we found some fine 1980s Hameg analog scopes in Paris at stree/flea mkts.

These HM103, 203, 204 have a handy component checker, a simple 3VAC current limited curve traces, good to ID diodes and transistor junctions.

Simplest solution


mains><24V transformer>>Series R>< DUT.

Scope looks across series R for I and across DUT for V.

Jon