Signal diode markings

[DJPhil]

I finally sorted through my heap of salvaged diodes and organized them while I listened to Amp Hour #3. I tested them all at ~5mA/20V and ~3mA/40V (I know it's a bit low) and separated all the small signal diodes out. Many were 1N4148, but many were unmarked.  All of these tested with 1V to 5V voltage drop at 40V forward and 0.65V +/- 0.10 reverse biased. All are unmarked with a glass body and a single cathode marking band.

I noticed the cathode band was either black, yellow, or even white, but I can't find any information about what (if anything) the different colors might mean. I figured they're pretty much interchangeable, but I thought I'd ask in case anyone had any insight.

Edit: I should have put a sense of scale in the photo. The diode bodies are ~3.5mm long, so these are the standard-ish small through hole diodes.

[djsb]

There should be no voltage drop when the diodes are reverse biased and approx 0.7v when forward biased. Not sure about the markings. Are you testing for reverse breakdown or just testing if they're working?

[Zero999]

I know what you mean but to be picky that statement isn't accurate: there will be a high voltage drop (close to the full supply voltage) when the diode is reverse biased.

[djsb]

If there is a very small current flowing when the diode is reverse biased there will be a very small voltage drop (I was effectively calling this zero for practical purposes). So, for example, if 3 volts is applied a multimeter should measure maybe 2.98 or thereabouts (I'm guessing at the actual reading as I havent tried this) across the supply. So there has been a voltage drop of 0.02 volts. If theres a drop of a few volts I think the diodes are close to reverse breakdown.
I usually just use my multimeter for testing diodes (using the diode check function), simple as that.

David.

P.S Just did this as a quick simulation using Pspice and the reverse current is 5.67nA. Voltage drop across a 1 ohm resistor is (obviously) 5.67nV.Not that it proves anything. I dont trust simulation results without question.

[alm]

If there is a very small current flowing when the diode is reverse biased there will be a very small voltage drop (I was effectively calling this zero for practical purposes). So, for example, if 3 volts is applied a multimeter should measure maybe 2.98 or thereabouts (I'm guessing at the actual reading as I havent tried this) across the supply. So there has been a voltage drop of 0.02 volts. If theres a drop of a few volts I think the diodes are close to reverse breakdown.
I usually just use my multimeter for testing diodes, simple as that.

The drop in supply voltage is because the non-zero output impedance of the voltage source, and is likely to be way lower than your ~1% estimate for most voltage sources. This voltage is not dropped over the diode, but over the resistance in series with the theoretical voltage source (which has a zero output impedance). The voltage over the diode in 3V (minus the drop in the power supply, so maybe 2.99 something volts), and depends on the supply voltage. For most practical purposes, you can say that the voltage over the diode when reversed bias is only limited by the reverse breakdown voltage. As long as you stay under this voltage, it's for all practical purposes equal to the supply voltage.

[djsb]

Agreed, my figures where made up purely for illustration.

David.

[DJPhil]

I may have my terminology flipped, it's something I do too often and can't seem to stop. 

I used a power supply set to first 20V then 40V to test all my salvage diodes. I put a series resistor in to limit current, and checked the voltage across the diode with a meter. The goal was to test for dead parts and ferret out zeners. All of these unmarked parts tested ~0.65V one way (I assumed that was reverse biased) and at 40V the voltage would droop a bit, sometimes as low as 35V. I didn't think they'd be zeners because the action was slow (a few seconds) and I assumed I was nearing or exceeding their breakdown voltage.

I ended up just packaging them together as 'Signal Diode Mk. 1' and will likely use them for nonessential applications.

The color bands are still a mystery to me. I have this nagging feeling that there's meaning to the colors, but I'm almost sure it was just whatever paint was cheapest at the time.

[alm]

The 0.65V is forward biased, and depends somewhat on the current. This indicates that it's a silicon diode, not a Schottky (it would be more like .35V in that case) or germanium (even lower I believe). The 40V is with the polarity reversed, I assume? That is probably because the leakage current is increased when it's close to reverse breakdown, which increases the voltage over the resistor.

My guess is that Signal Diode Mk. 1 (which is almost what 1N4148/1N914 means) is pretty close, as long as they're not zener (you would see a much sharper knee at reverse breakdown), Schottky or power. They might be faster or slower, but as long as neither speed nor high current are critical, they're probably fine. I tend to throw these away, since the 1N4148 is really cheap, and I dislike working with unknown parts without datasheets.

The color might be an internal marking or selection by the company that used them, they may have been selected for eg. forward voltage or reverse leakage, or may be different parts altogether.

[DJPhil]

The 0.65V is forward biased, and depends somewhat on the current. This indicates that it's a silicon diode, not a Schottky (it would be more like .35V in that case) or germanium (even lower I believe). The 40V is with the polarity reversed, I assume?

Yep, I just flipped it over on the test rig. Thanks for straightening me out, I'll stop mixing things up someday.

That is probably because the leakage current is increased when it's close to reverse breakdown, which increases the voltage over the resistor.

This is what I suspected, though I didn't remember the exact mechanism. When I was running at 20V I'd get strange behavior on a few that made me suspect I had higher voltage zeners in the mix. Bumping it up to 40V ferreted out a few low 20V zeners.

My guess is that Signal Diode Mk. 1 (which is almost what 1N4148/1N914 means) is pretty close, as long as they're not zener (you would see a much sharper knee at reverse breakdown), Schottky or power. They might be faster or slower, but as long as neither speed nor high current are critical, they're probably fine. I tend to throw these away, since the 1N4148 is really cheap, and I dislike working with unknown parts without datasheets.

Understandable. I've got a fair bit of guitar pedal crafting in my future, so I figured I'd put them to use in clipping circuits. In that application even matching doesn't matter much, and I'll be able to recycle them to some end. I've got an equal quantity (30 or so) 1N4148 marked for anything slightly more demanding, and heaps of various 1N400X rectifiers and high current Schottkys. I've even got an oddball set of four teeny tiny black epoxy cased diodes that registered less than 0.2V forward biased at ~5mA/20V.

The color might be an internal marking or selection by the company that used them, they may have been selected for eg. forward voltage or reverse leakage, or may be different parts altogether.

I guess I may never know, but I'll find them a home.

Thanks to everyone for their input on this.