Basic Diode Physics Revised?

[mhuss]

All my life a standard silicon forward-biased junction has dropped .65 to .7 volts when conducting. The Vishay 1N400x data sheet (https://www.vishay.com/docs/88503/1n4001.pdf) says Vf is 1.1v - perhaps at one amp?

We have a batch of 1N4007s that are forward-dropping 0.45v at meter test levels (very low current), and a similar amount when in circuit conducting ~16mA.

Has anyone else seen this and can anyone explain it?

This came up because we have an application that uses three series Si diodes in parallel with an LED to protect the LED from over-current, but this recently stopped working (3 diodes dropping a combined 1.4v end up stealing current from the LED!)

[g1]

data sheet (https://www.vishay.com/docs/88503/1n4001.pdf) says Vf is 1.1v - perhaps at one amp?

Yes, pg.2 of pdf says 1.1V is at 1A test current. Pg.3 has a graph showing Instantaneous Forward Voltage correlated to Instantaneous Forward Current.

[mhuss]

Yes, the graph data drops off at low current, but it looks like there should be at least 0.6 Vf at ~20mA.

[Helmholtz]

Data sheet forward voltages are "typical" values, so not guaranteed.
Lower forward voltage is typically considered better, as it means lower losses in power applications.
For the diode manufacturer (and typical customer) it's most important to specify a guaranteed max. value.

Some years ago I measured forward voltages of a number of diodes at low currents. For 1N4005s I found 0.43V@0.1mA and 0.53V@1mA.
1N4007s were around 0.04V higher. But that may not be representative.

There's no law of physics stating that the forward voltage of Si diodes must be between 0.6V and 0.7V, though these are typical values at diode-typical currents.
Rather diode current and voltage are related by an exponential law (+ a constant resistive term) depending on several diode-specific parameters.
Generally forward voltage increases with current, drops at higher temperatures and tends to be lower with larger diode chips.

Not sure if clamping the (forward) voltage across an LED to protect it is a good idea, as LEDs are diodes themselves.
Clamping the reverse voltage makes sense, though, if required.

Maybe we could come up with a better solution if you don't mind to explain the exact problem/application.

[mhuss]

Thanks for the thoughtful reply. It seems odd I've never run into this before, even 'diode check' settings on DMMs typically report .6 or more, and you have to think that's a fairly low current.

I totally agree, the perfect diode would have a zero forward drop, and in the real world, the lower Vf the better. Perhaps the manufacturing processes are just getting better/more ideal...

[Enzo]

One lesson Bob Pease stressed was you cannot extrapolate the data sheet. If they have a graph down to 5ma, you cannot assume the curve continues down to 0.1ma.

[mhuss]

Absolutely true.

I pulled out my 'diode drawer' and did some measurements - none were higher than 0.62 Vf, and the UF400xs were as low as 0.41 - only the Ge diodes were lower.

I guess you learn something new every day.

Thanks, all.

[SoulFetish]

One lesson Bob Pease stressed was you cannot extrapolate the data sheet. If they have a graph down to 5ma, you cannot assume the curve continues down to 0.1ma.

yeah, but, what if we only did it every once in a while. special occasions?

[kg]

I totally agree, the perfect diode would have a zero forward drop, and in the real world, the lower Vf the better. Perhaps the manufacturing processes are just getting better/more ideal...

ironically, thermionic (tube) diodes have zero forward voltage drop.

however, they do have internal resistance, so that never manifests with any real current load.

schottkys are closer to ideal Vf with very low R, but low max inverse is a hindrance.

[Helmholtz]

ironically, thermionic (tube) diodes have zero forward voltage drop.

Tube rectifiers (diodes) have considerable forward voltage drop.

[kg]

Tube rectifiers (diodes) have considerable forward voltage drop.

thermionic current flows whenever there is an electric field gradient, even infinitesimally small.

they do NOT experience any bandgap threshold.

[mhuss]

You're splitting physics hairs here. At anything near normal current usage levels, the large internal resistance will ensure a healthy Vf. Citing a .001mv gradient condition is just being pedantic

Let me amend my original statement: the perfect diode would have a zero forward drop at a useful voltage and current level,

[kg]

of course, that depends on what you consider "useful current," and what tube you're using.

there is absolutely nothing stopping you from using a transmitting valve with grid(s) strapped to anode to get a "perfect diode" with very very low internal resistance and zero Vf.