Diode voltage drop stability/accuracy

[Infraviolet]

Cn I quickly check how stable the 0.7V drop across a typical small signal diode is, and how accurate between diodes of the same model. If one is in a situation where several separate analog signals are all passed through diodes* then can one rely (to within what percentage accuracy 1%, 5%, 10% ,20%...) on all the signals being dropped by the same amount?
Thanks

*for my situation I can definitely assume small currents, so no signficant heating of the diodes from this, and all physically close together so all the diodes would be getting the same rise or fall in ambient temperature if the environment around the PCB gets hotter or colder

[exe]

I'd say passing analog signal through a diode will cause distortion. Anyway, what's the use-case, do you have schematics?

Diodes have quite some tempco (2mV/C or so), they self-heat when passing signal, and they are highly non-linear (i.e., voltage drop depends on current). So, without a feedback look it's bad to have in a signal chain.

[TimFox]

When matching 1N4148 diodes from measurements using the ohms function on a DMM (2k\$\Omega\$ full-scale, probably 1 mA source current), I always have to wait for the diode to heat up to get a stable rating (3.5 digits).
I have a couple of old-style Grayhill double-banana (0.75 inch centers) to component clips (on 1.75 inch centers) adapters that I use for repeatable measurements.

[bdunham7]

While the datasheet might indicate that individual units can vary +/-50mV, units from the same batch will likely match quite a bit better.  Typically you still have to characterize them if you need them to be very close.  I grabbed 5 at random from a bag that probably all came from the same lot and I got values from .609 to .615@1mA and .499 to .503@100µA.  If you match V_f under one set of conditions, the matched units should track all of the parameters (V_f vs current and temperature) pretty closely. 

[David Hess]

I'd say passing analog signal through a diode will cause distortion. Anyway, what's the use-case, do you have schematics?

Diodes have quite some tempco (2mV/C or so), they self-heat when passing signal, and they are highly non-linear (i.e., voltage drop depends on current). So, without a feedback look it's bad to have in a signal chain.

That is true when the signal is a voltage, but before JFET and MOSFET switches became available, diodes and bipolar transistors were used to accurately switch currents where their change in forward voltage drop is irrelevant.  This works especially well when the signal is the output of a transconductance stage, which naturally produces a current.

Some applications still use diodes or bipolar transistors this way because it works well at high frequencies.

[Infraviolet]

I have noticed that datasheets for the likes of 1n4148 diodes say just "maximum 1V" for the drop, rather than giving a range and mean. Does this mean they're manufactured to a pretty loose tolerance in this regard, or is it mroe reflecting the way that according to the datasheets the forward voltage changes a fair bit with current being passed and with temperature (for currents of mA and less one assumes temperature is purely controlled from ambient effects not internal heating by the current)? Are there any diode types which do have a highly consistent voltage across manufactured devices and which keep it pretty constant with temperature and/or current?

[Benta]

Diodes like the 1N4148 are miserable in this respect, and most others as well.
If this is something you want to pursue, transistor V_BE usually has much better definition (BC547, 2N2222 etc.).

[TimFox]

In the diodes similar to the general-purpose 1N4148, other units (e.g., 1N4152) were considered "controlled conductance".
(The 1N4152 has a lower PRV rating than the 1N4148.)
https://datasheetspdf.com/pdf-file/133339/FairchildSemiconductor/1N4152/1
The datasheet gives max/min voltages for six discrete current values from 0.1 to 20 mA
At 2.0 mA, the range is 0.62 to 0.70 V, at ambient temperature of 25^o C.
The 1N4150 has a similar set of voltage values, but the 1N4149 and 1N4151 have the usual broad maximum values.
 

[Kim Christensen]

If you really wanted them matched and to track with temperature, you could use a diode array. Don't know if you really need to go that far. It really depends on what you are trying to do.

[David Hess]

I have noticed that datasheets for the likes of 1n4148 diodes say just "maximum 1V" for the drop, rather than giving a range and mean. Does this mean they're manufactured to a pretty loose tolerance in this regard, or is it mroe reflecting the way that according to the datasheets the forward voltage changes a fair bit with current being passed and with temperature (for currents of mA and less one assumes temperature is purely controlled from ambient effects not internal heating by the current)? Are there any diode types which do have a highly consistent voltage across manufactured devices and which keep it pretty constant with temperature and/or current?

One difference is that switching diodes are gold doped.  Maybe that leads to greater variation in forward voltage drop.  It sure increases their reverse *and* forward leakage current.

[EPAIII]

One fast and inexpensive way to check diodes and see their forward Voltage drop is with one of these automatic parts testers.

https://www.amazon.com/Mega328-Digital-Transistor-Resistance-Capacitance/dp/B07WT9VVZB/ref=sr_1_6?keywords=electronic+part+tester&qid=1686535191&sr=8-6

This is just ONE ad. There are a number of variations available. Just search for "electronic part tester". Most are priced under $20 USD. They are a great value.

They are very handy and I have three of them. Wish I had something like that 50 years ago or at any time since.

[Sredni]

I have noticed that datasheets for the likes of 1n4148 diodes say just "maximum 1V" for the drop, rather than giving a range and mean. Does this mean they're manufactured to a pretty loose tolerance in this regard, or is it mroe reflecting the way that according to the datasheets the forward voltage changes a fair bit with current being passed and with temperature (for currents of mA and less one assumes temperature is purely controlled from ambient effects not internal heating by the current)? Are there any diode types which do have a highly consistent voltage across manufactured devices and which keep it pretty constant with temperature and/or current?

I believe it is related to the maximum current they can pass before giving up the magic smoke.
The concept itself of 'forward voltage' is kind of ethereal because that 'knee' that seems so sharp in the VI characteristic of a diode... is not real. It's just an artifact of the scales of the V and I axes (more the I axis, actually).

Here's a post on SE where I put some pretty pictures
https://electronics.stackexchange.com/questions/655076/forward-bias-voltage-of-diode

What you are asking about, in essence, is the dispersion of the characteristics, not of that nonexistent knee.

[magic]

I have noticed that datasheets for the likes of 1n4148 diodes say just "maximum 1V" for the drop, rather than giving a range and mean. Does this mean they're manufactured to a pretty loose tolerance

It means that the datasheet reproduces specifications written 60 years ago. Also, didoes have some parasitic series resistance, so they lose their exponential characteristic at high currents and Vf starts to rise linearly with If - this is particularly seen in power diodes.

In practice, you will probably find that most specimens from one production lot are within a few mV of each other. I recall building a bridge of four diodes (two strings of two diodes, all forward biased) and the difference between voltages on both sides was a few mV. YMMV.

If you want guaranteed matching specs, look for transistor pairs. There are relatively cheap dual 2N3904/BC547 type transistors (two dice in one plastic case) with guaranteed 3mV or so matching, and there are much more expensive monolithic duals (from AD, THAT corp) with sub-mV matching IIRC and guaranteed thermal tracking, being made on one die. Beware that diode-connected transistors will only stand ~5V in reverse.

Hardly anyone uses diodes for passing signals or switching anymore, I think it's only a thing in RF at this point. You may be able to find some matched dual diodes for RF, but they may be pricey too and have low Vr and If specs, ESD sensitivity and other weird issues.

[armandine2]

I did the exponential ideality factor test on a IN4001 diode over the weekend - following Millman and Grabel Microelectronics

Came out pretty much dead on two

[Terry Bites]

It won’t be very reliable and not a parameter you want drifting about in your design. The last thing you need is to have to match them up.

Diodes are frequently used as temperature sensors, so go figure.
Vf varies with temperature and current. ("Shockley diode equation"). Light shining on a glass body diode like a 4148 will also affect Vf. For a given diode type (even better from the same batch) the temperature and current dependencies will be fairly well matched, it’s hard to put an exact figure on it though.

I'd make up a string of diodes and put a constant current through them. Shield your diodes from ambient light, get out your DMM, protect them from today's coffee breath, tongue set to 24.076’ and measure the Vfs. Attached, measure 10x (12x at a stretch) diodes at once at a set current.

A low Vgs mosfet is ideal but just about anything will work well enough for this.

Share your findings.

[MrAl]

Cn I quickly check how stable the 0.7V drop across a typical small signal diode is, and how accurate between diodes of the same model. If one is in a situation where several separate analog signals are all passed through diodes* then can one rely (to within what percentage accuracy 1%, 5%, 10% ,20%...) on all the signals being dropped by the same amount?
Thanks

*for my situation I can definitely assume small currents, so no signficant heating of the diodes from this, and all physically close together so all the diodes would be getting the same rise or fall in ambient temperature if the environment around the PCB gets hotter or colder

Hello,

If you want better matching then go with an IC that has several diodes inside.  They are close together too so less temperature differences.

A couple things to think about...
[1] First, if one diode has 1ma and the other diode has 2ma then the forward voltages will be different.
[2] Second, if the temperature of one diode is 25 degrees C and the other is 27 degrees C the forward voltages will be different.

For #1 above, ideally that means that the signal currents have to be the same, and that means if they carry different signals then they could have differences.  It depends on how different the signals are.
For #2 above, ideally that means the temperatures have to be the same, and when you use two separate diodes there is a space  between them and that means that with the slightest air current there will be a temperature gradient which means one diode could be at a different temperature than the other.  Moreover, if the air current changes direction it could mean the role temperature plays gets swapped between diodes.

These are probably the most important parameters to think about.  As above, the distance between diodes is important too they have to be very close to each other and ideally wrapped in insulation to help keep the temperature gradient low.  This is why using an IC chip with multiple diodes works better, and they will probably be matched better also.  It won't help with #1 above though because if the currents are different the voltage drops will always be different too.

If you describe your application, there may be better solutions.

[David Hess]

I did the exponential ideality factor test on a IN4001 diode over the weekend - following Millman and Grabel Microelectronics

Came out pretty much dead on two

Bob Pease published results from a bunch of different diodes.  The 1N4001 he tested had lower conductance than a good diode.

[mawyatt]

I did the exponential ideality factor test on a IN4001 diode over the weekend - following Millman and Grabel Microelectronics

Came out pretty much dead on two

Bob Pease published results from a bunch of different diodes.  The 1N4001 he tested had lower conductance than a good diode.

Yep, what Bob's classic paper shows is what old folks have known for multiple decades, the simple 2N3904 Vbe is about as good a PN junction regarding following the classic diode equation as it gets:

Ib = Is[e^Vbe/Vt) -1], where Vt = KT/q
deltaVbe = Vt[ln(ratio Ib)]

Ironically almost all the diodes are really poor in this respect

If you look at the 2N3904 Vbe at <10na and >1ma (L), it's perfectly straight with the correct slope of ~59.2mv/decade. We've used this characteristic in so many ways over the years for all sorts of useful functions from log amps, to temp sensors, even log domain filters and many applications we can't remember.

Edit: The important point here is the slope of the 2N3904 Vbe vs. Ib will be identical for all devices, regardless of the highly device and process variable Is or Vbe. So one should be able to take devices from various lots, manufacters, time frames and achieve identical results wrt the Vbe slope. Theory predicts this and our experience confirms such (altho we haven't measured any devices in decades), quite an amazing parameter that few folks realize.

BTW this is one of the most important aspects of bipolar transistors that analog IC designers utilize. Bob Widlar utilized this "feature" in his brilliant Bandgap Reference, very valuable feature indeed

If one looks at the various plots, the K, L, M show the ideal junction behavior with proper slope, and all are bipolar Vbe junctions!! Also, note the old HP SBD HP5082-2811 plot C, this shows good junction behavior up to ~100ua.

Best,

[Infraviolet]

"If this is something you want to pursue, transistor VBE usually has much better definition"

I can just use the base and emitter of a typical transistor as a diode then? leaving the collector floating? So long as I never expose the base to emitter pins to reverse voltages of 6V or more.

I'm doing some analogue peak detection on multiple signals, each has:

signal_source----diode >|------cap_and_res_in_parallel_to_gnd

I'm in a situation where I can't use op amps to make an active peak detector, and want to make sure each separate channel of signals I do it on has the same amount of voltage lost to the diode.

[Benta]

It's recommended in several places to connect base and collector for a good result. I'm not certain why (didn't pursue the theme).

[Infraviolet]

Base and collector together serving as anode. I can give that a go, thanks.

[Kim Christensen]

Or connect the collector to the +supply rail and have a hFE buffered "diode" for a positive peak detector with an NPN... (You could do the same with a PNP & negative rail for a negative "peak" detector)

[David Hess]

If one looks at the various plots, the K, L, M show the ideal junction behavior with proper slope, and all are bipolar Vbe junctions!! Also, note the old HP SBD HP5082-2811 plot C, this shows good junction behavior up to ~100ua.

HP sold a whole series of "high conductance" silicon diodes but Broadcom discontinued them all.

I can just use the base and emitter of a typical transistor as a diode then? leaving the collector floating? So long as I never expose the base to emitter pins to reverse voltages of 6V or more.

You can and the base-emitter junction has other advantages; besides being high conductance, it is also fast and has low leakage.

[TimFox]

Just be careful:  as noted above, the base-emitter junction of a silicon BJT has a relatively low reverse breakdown voltage (6 or 7 V).

[vk6zgo]

I'd say passing analog signal through a diode will cause distortion. Anyway, what's the use-case, do you have schematics?

Diodes have quite some tempco (2mV/C or so), they self-heat when passing signal, and they are highly non-linear (i.e., voltage drop depends on current). So, without a feedback look it's bad to have in a signal chain.

The diodes are "biased on" with a DC voltage to act as a switch----remove the bias, & the "switch" is off.

The analog signal is normally a much lower voltage level than the bias, hence, it cannot cause the diode to cut off or saturate as this voltage differential between signal & bias ensures it appears in a substantially linear part of the diode transfer function.

Such diode switching allows a transceiver or similar device to use the same special modules (such as filters) in both the transmit & receive modes.
The ham transceiver I am currently fixing uses this technique extensively.