DIY giga-ohm resistors???

You don't kill an electrometer sensitivity with a resistor. Use a capacitor.

Thanks, excellent advice. Certainly cuts the gain. Works great for what it is.

Next experiment is to build a field mill. Don't think I can stand the increased capacitance in a field mill. And need it to come off the peg back into the dynamic range of the amplifier...eventually. I think I need a resistor. But this is all new territory for me.

Still want ideas on making BIG resistors. mike

On a sunny day (Tue, 28 Feb 2012 18:24:12 -0800) it happened mike wrote in :

I did put 10 110 MOhm SMDs in series to get 1.1 GOhm, here:

On the left side:

Works fine, you are limited by the maximum voltage those things are specified for.

OK, what would you suggest?

Strike voltage is about 70V in the preferred direction. That limits the gate current to 4 microamps reverse. Suppose that's excessive?

I bit the bullet and put back-back zeners at the gate. I can see the added leakage, but I wanted some leakage to ground anyway.

Digikey stocks 10G 5% resistors for less than $4. HVF1206T1008JE, and others.

Or, if you really feel the need to make something, maybe your pencil line on something like this:

(ceramic base).

Best regards, Spehro Pefhany

When I was but a stripling in my first engineering job (before grad school) one of my first tasks was to figure out the right bias resistor for a coax-in, waveguide-out, SRD doubler using small ceramic devices with gold plated end caps. (IIRC they had been using a Microwave Associates one and wanted to qualify a second source, Gigahertz Devices.)

The doublers were self-biased, i.e. rectification provided the right bias to get the best snap, so the bias network was just a parallel RC across the diode. The diode mount and waveguide provided the capacitance, and the resistor in the production units was a small glob of thick-film resistor goo right on the side of the ceramic diode case.

To find out what that resistor should be, the trick was to scribble on the side of a clean diode with an HB pencil, measure the resistance, plot the 2f vs 1f power conversion efficiency, scribble some more, repeat. It worked pretty well, except that I had to reduce the input power range to avoid burning the graphite during data taking. (The new resistor was something like 200k vs 120k for the old one.)

So there are actual uses for pencil-on-ceramic resistors.

Cheers

Phil Hobbs

Not much.

Maybe distilled water.

Or a gas gap exposed to ionizing radiation. Water vapor and a UV LED might work, but I don't know the required wavelength. Should be adjustable by the amount of ionizing radiation or the electrode distance. Pressure should be high enough to avoid avalanche break through. You could try to use a smoke detector that uses americium as source of ionizing radiation. Uranium is fairly easy to acquire, but has a low activity. There are professional radiation sources available, too. Betalights come to my mind. They contain radioactive tritium. You would have to break open the vial, though, to make use of the tritium, as the radiation won't penetrate the vial.

A 1N4148 diode has 0.8 GOhm reverse resistance at 20 V and 25 degC. You can certainly find diodes with higher resistances. The resistance depends on voltage and temperature, though, and is DC only.

Bernhard

mike wrote in news:jikf1r$f9a$ snipped-for-privacy@dont-email.me:

Take a 4-8 inch high transmitter tube, power up the cathode keating, and use a negative voltage on the grid to tune for the resistance you want. In a pinch an audio endstage tube will do.

If the gate oxide breaks down at 15V (typical) youall aint gott no perrtektion. STM makes a number of FETs with integrated back-to-back zeners for protection. Added leakage can be in the nanoamp or less region at 25C.

OOooh! Shades of the old tube SexTronix 'scopes!

Suggest you experiment with the plate voltage at a given grid voltage..

I once did this using india ink on painted acrylic. The ink is a carbon resistor, and you can dilute it to produce large values.

One problem was with the junction from ink to electrodes. Better use conductive silver ink as terminals, then run the india ink up over the silver.

Here's one I haven't tried: neon glow lamps with low turn-on voltage, where thorium in the electrodes provides ion leakage.

As they note, you'll need to clean the glass, heat it to de-sorb any water, then cover it with a coat of insulating paint.

Another possibility, from Phil Hobbs' $10 thermal imager paper: two LEDs facing each other, covered in something opaque. One acts as a (dim) light source, the other is a photodiode. If you pump some uA into the emitter, will the PD give some pA of bias current?

On a sunny day (01 Mar 2012 01:48:10 GMT) it happened Sjouke Burry wrote in :

This only works for triodes (resistor like Va/Ia curve). Penthodes have a constant current behavior. Most audio output stage tubes are penthodes.

Back-to-back LEDs are much better. Keep 'em dark.

That's normal for an abs. max rating, but I wonder what the typical actual breakdown voltage is. For example, a Calogic 2N7002 has an abs max rating of +/-40V, I would guess that the typical breakdown is more like 100VDC, so a neon might be okay.

A nA can be huge. According to measurements made by skilled practitioners (eg. Win Hill) actual leakage of a 2N700x is a few fA.

I started by asking about making big resistors. This has morphed into a party where every drunken cowboy is shooting his gun in all directions.

Let's put a target on the wall and encourage everybody to AIM.

The question has become, "what does it take to damage the gate of a JFET. (2n700x is a MOSFET)

Neon lamp is across the input.

Neon fires at about 70V (in the dark) if you stick it in the right direction. About 90V in the other direction Gate breakdown voltage spec is 30V.

In the forward direction, I've got about 9 microamps of gate current. That shouldn't be harmful???

In the reverse direction, I've got about 4 microamps of gate current at -30V. So, the basic question is, just that; "is 4 microamps at -30V on the gate harmful in this application?" I'm not worried about increasing the noise figure. I am worried about significant increases in gate leakage current at ~0V.

For now, let's just ignore all the nasty parasitic and transient things that happen if you actually strike an arc at the input. That's a whole other bucket of worms.

Ready...AIM...

Yee Haw! Hey if you want a STABLE high value resistor it ain't so simple. First you have to be very careful of the insulator. Glass (which includes ceramic) is OUT. The absolute best is polystyrene. For quick and dirty a pencil line on mylar drawing paper isn't too bad.

In a case where I needed very high resistance with reasonable accuracy, we built a circuit board with hundreds of high value surface mount resistors in series. Handles voltage, low leakage (with teflon circuit board) and accurate. We only had one blow itself to kingdom come. Don't know why it failed.

Yes, I've tried the neon "protection" thing. Basically worthless. Problem as noted above is transient response. And anyway, don't you know that the gates of FETs of ANY type make better fuses than ANY protection circuit?

That's how SED works--we all discuss what we feel like discussing, just the same as you. If you want people to do something specific for you, pay them or else be helpful and friendly yourself so that they want to do it for free.

Cheers

Phil Hobbs