Difficulty measuring diodes opinions ?

[strassercaster]

Ok i am building a gain pedal  .Any way i am trying to get the forward voltage on several different makes of diodes. i have three meters . For instance my first meter is always the lowest it says my d9e diode is .267 the second meter says .320 and the third says .412. my third meter is the most expensive and its still a cheapy mk-168 transistor testor.no matter what i measure one is always the lowest two is about twenty percent hight and meter three is always 50-70 percent higher . which one should i go with? ha ha funny how they all measure resitors the same . i know they all three probably use a different charge voltage to test them. Is there a jig or test rig out there that will be exactly accurate so i can do a few do the math and know where i stand? you would think its a no brainer and i am probably over analysing but i actually have a real centaur here to test against and i think mine was a little too compressed compared to the real one. for my first set i used 1n34a that measured .340 on the mk168  but on
y .196 on the low meter . thanks in advance

[Keppy]

Diode forward voltage increases with current. It's not just one value...

Is there a jig or test rig out there that will be exactly accurate so i can do a few do the math and know where i stand?

... so no, unfortunately. To really know where you stand, you'd have to measure the forward voltage at a variety of currents and plot the results as a graph, which you could then compare with graphs for other diodes. Even then, germanium diodes have leakage, which is a separate measurement that also varies with conditions (particularly temperature).

I'm afraid I don't have any real, constructive ideas. Just an outline of the problem.

[MaxPower]

Is it possible to use a second DMM to test the current being output by the DMM that is being used to test the diode? That is, having two DMMs going at the same time. One testing the diode and the second DMM checking the current being output by the first DMM.

If not, then just use the same DMM to test all your diodes and while you may not know the exact voltage drop at what current, you can at least tell which have a relatively lower/higher voltage drop.

[teemuk]

...And it's not like the diode is "off" when voltage is lower than Vf, and "on" when the voltage exceeds it. No. There's actual gradual transition from not conducting to forward conduction, and this characteristic varies all over the place with diodes. Most diodes usually start to conduct few hundred millivolts before their nominal forward voltage and you will find diodes with "harder" and "softer" knees in their characteristic curves.

See for example:

AA112 and 1N34A are both Germanium diodes and have about the same nominal forward voltage. But excluding this characteristic 1N34A resembles more a traditional silicon diode than the AA112 which "slopes" much more gradually. Unfortunately there's only one silicon diode included in the graph but I trust that alternative types would show similar variation where other characteristics than nominal forward voltage differ quite a lot.

People who just want to rectify with diodes (the most popular application) will not care about forward characteristics excessively. In generating harmonic distortion with diodes (niche application) these characteristic are very important because they directly effect on harmonic pattern of the created distortion.

For what you are interested in to measure I would suggest a "Curve Tracer" tool instead.

[slacker]

Measure the diodes from the real Centaur with any of your meters and then use the same meter to find matching diodes, that way the actual value your meter gives doesn't matter.

[GibsonGM]

Is it possible to use a second DMM to test the current being output by the DMM that is being used to test the diode? That is, having two DMMs going at the same time. One testing the diode and the second DMM checking the current being output by the first DMM.

If not, then just use the same DMM to test all your diodes and while you may not know the exact voltage drop at what current, you can at least tell which have a relatively lower/higher voltage drop.

Good question, Max.  YES, you can!  You just have to decide if it's *worth* it.  As an experiment, it might be cool to do this, and plot the results on a graph as implied by Teemuk.  For all the reasons above, tho - it's more practical to just match the diodes using one meter and be done with it.   

Using Ohm's law, you can 'find the curve' by placing a resistor in series with the diode and measuring the voltage drop across it - convert it to current. One would want to be careful not to exceed any parameters and burn out the diode, tho, so a low voltage source is probably a good idea (maybe 3V).     This is much like how the curves are obtained for FETs and such...

[duck_arse]

.... and while you have your voltage source, resistor and diode jig up and running, measure the voltage across the diode with each of your meters (and each chosen resistor value), and see how far apart the meters are.

[Quackzed]

i've done a 'relative threshold' test before by setting up the dist box with the knobs set for high gain and moderate volume and recording a clip on my pc... then swapping diodes, recording with same box settings for each pair of swapped diodes then just comparing the height of each waveform to the others for a 'relative' threshold comparison... you cant adjust vol or gain or recording level etc otherwise its apples and oranges, but if you record the same settings with each pair, you'll get a different threshold with each diode pair if you slam em with some sustained power chords... quick and dirty...

[Kipper4]

Just an idea but have you tried measuring a known value diode such as an LED which will have a "normal" Vf
Somewhere there a chart with differant colour led and there forward voltages. Google is your friend.
Then test with your differant meters and see which is closest.

[midwayfair]

I just keep notes about the Fv measured against each other and group diodes together in my parts container. It doesn't REALLY matter most of the time and if you think it might in any particular build, then you can use sockets and record A/B tracks.

[armdnrdy]

Is there a jig or test rig out there that will be exactly accurate?

Yes...it's called a good meter. 

Just for arguments sake...I took my two meters (Fluke 87 & Extech EX540) and measured different diodes.

I measured 1N60P, 1N4001, 1N270, 1SS133, and 1S1588.

After reading your post, I was surprised to find that there was a variance of less than 100mv between the two meters.

You can take measurements with 100 mediocre meters but...unless you have any sort of accuracy...you might as well make up the number.

[PRR]

As said, Vf depends on current.

For simple Si diodes, Vf increases about 60mV for every 1:10 change of current.

So you may have one meter measures at 5uA, another at 1mA.

There's no standard current. The meters are for rough go/no-go work, not fine tuning.

Do you have *two* meters? (Sounds like you do!) Put one in Diode but measure a 100 Ohm resistor. Put the other in Volts (milliVolts) and see what voltage the "diode test" puts across the 100r resistor. Now you know the ballpark current of that diode-meter. If you bother to play round-robin with all three, you probably find considerable difference in current.

If you are measuring Ge diodes, a meter which tests at 5uA may read extra-low because a Ge diode may leak more than that.

[strassercaster]

i really appreciate the in depth answers knowledge and wisdom . THANK YOU ALL