unexpected negative voltage from JFETs

Could your voltmeter be injecting pulses into the gates, which then get rectified by the gate-source junction? Some digital voltmeters do that.

Jeroen Belleman

Hi,

HP3478A multimeter, 10 GOhm for < 3V DC. Voltcraft M8660 shows the same. I just tried my analog Metrix MX225A, 100KOhm/V: also the same

Gerhard

Flash memory cells get erased by 'hot electron injection' using a 'high' current through the FET. A mechanism like that might be present.

What is the voltage on the drain, are they conducting or cut-off? Does it vary when measuring the gates? What is the short-circuit current from the gates to GND? Is the effect the same with an extra 100nF over the DMM inputs?

Arie

I suppose you scoped the gates? I suppose the JFETs must be completely pinched off. Are the drains at Vdd? Is there some possible interference source nearby? Maybe you made an accidental oscillator by connecting long instrument leads to the gates.

Jeroen Belleman

Phil Hobbs has noted a "venturi effect" in depletion phemts, where they bias themselves off.

It sounds odd. Sometimes, when circuits behave in an odd way, they turn out to be oscillating at frequency that you aren't set up to detect. It doesn't happen often, but it's a pest when it does.

ATF38143 depletion pHEMTs self-bias to about -0.3V if you leave the gate open.

Cheers

Phil Hobbs

On a sunny day (Wed, 2 Dec 2020 12:08:04 +0100) it happened Gerhard Hoffmann wrote in :

What happens when you connect say a 10nF poly capacitor between gate and source, short it for a moment, wait a while and measure again? That should rule out any external RF detection by the gate-source diode.

That seems to be the same effect.

I have looked into Cobbold, Theory & application of Field Effect Transistors, Wiley-Interscience, but this is like drinking from a fire hose. :-) And in paper form, it's hard to search.

But it seems that gate current is dominated by recombination-generation effects in the channel and not so much by diode leakage.

Effective Vdd in my case is abt. 2.5 V, there is absolutely no RF or oscillation involved, at least below 2.5 GHz. Seems to be a pure DC effect.

I know that even the BF862 can show nasty gate current effects at Vdd > 5V. They can even spoil the noise behavior, but @ 2V5?

Cheers, Gerhard

In a JFET, you'd think that in the presence of free carriers, the built-in E field would forward-bias the junction the way it does in photodiodes. It's the sign of the effect that's strange.

I imagine it's because there aren't many free holes, just electrons--the chemical potential makes electrons diffuse against the junction potential into the gate. In a photodiode there are the same number of electrons and holes.

Cheers

Phil Hobbs

What happens in a p-fet?

And that's probably the explanation for Pease's effect too: the B-E avalanche collects generates a lot of carrier pairs. In a saturated BJT, the frozen-in field is the reason the collector goes below base potential, but that field gets zeroed out before the collector goes below ground.

In a B-E avalanche, the collector collects charge in the same way, but because the base is grounded, the collector goes negative from there, producing the negative open-circuit voltage.

Cheers

Phil Hobbs

When you reverse polarity? The engines canna' take it, Captain.

(There are still not a lot of minority carriers around, so it ought to bias itself positive.)

Cheers

Phil Hobbs

You've not seen a two stroke UAZ. The rumour has it you just switch the battery terminals and turn the ignition key.

You could make a 2CV run backwards IIRC. There was a Car Talk puzzler long ago that hinged on the ability to bump-start a two-stroke in reverse while moving forwards, resulting in one forward and four reverse speeds.

Cheers

Phil Hobbs

Big ships have reversing diesels now. Reversing a steam turbine plant was a little messy.

We can do things with fets and inductors that the mechanical boys can only dream of.

Do hybrid cars have actual reverse gearing? Seems like it could be done electronically.

It cannot be the four-man overall (Citroen 2CV), as its engine is four-stroke, with valves. The valve mechanism does not allow backward gas exchange.

An example could be the national car of the DDR (East Germany), the Trabant and its predecessor the IFA, whih were two-stroke, as was also the Saab 92.

Sounds like bullshit, that won't crank the engine in reverse. and if it did it would burn up the generator.

I had a Morris Minor and reversed the battery so I could install a new radio. The only mod I did was put some red tape on the (now) positive battery lead.

the the starter motor was a univeral motor (like almost all starter motors) and the generator used a commutator and brushes for rectification, so with the field polarity now reversed it adjusted to the new supply polarity.

The heater fan and wiper motor were also wound field motors.

I think it was the Saab they were talking about.

Cheers

Phil Hobbs