300Meg resistor?

[scott wurcer]

Of course, if you bootstrap the gate bias resistor, you can make it look much higher in value. Some of the trickier condenser microphone circuits not only bootstrapped the gate resistor but also the drain of the input FET (to reduce input capacitance). Are you doing this with silicon or valves?

The noise is the issue. Bootstrapping helps everything but that. I found an article by Scott Dorsey today with some DigiKey part numbers, they might fit my application. The article is on moding Oktava mikes and addresses similar issues.

 

[EC8010]

Tug those bootlaces!

The noise is the issue. Bootstrapping helps everything but that.

Actually, bootstrapping does help noise. The capsule is a Norton source loaded by the input impedance of the pre-amplifier. If you convert the capsule to a Thevenin source, you soon see that reducing input capacitance reduces attenuation and improves S/N. You also then see that the input resistance forms a high-pass filter in conjunction with capsule capacitance, but I expect you were already aware of that.

 

[scott wurcer]

Yea, I meant the current noise of the resistor is still a problem. The Nakamichi head has the resistor bootstrapped, but in this case the capsule to FET capacitance ratio is pretty high and the improvement is small. I bet this would help those little Panasonic modules a lot more. Another issue is noise in the channel and inactive gate overlap capacitance looking like input current noise that goes up with frequency. This is why I think the smallest FET that gets the electrical noise out of the picture is the easiest to design with. I have a couple of radical charge amp ideas from a long time ago that might be of DIY interest if they pan out.

 

[EC8010]

I bet this would help those little Panasonic modules a lot more.

Yes, bootstrapping helps little capsules far more than, say, a large Neumann capsule.

I bow to your knowledge on semiconductor design, but you'd be amazed how often system design is poorly implemented. For instance, why on earth would you take a high impedance wire 2" from the centre of the capsule (to a semiconductor) when it's possible to put that semiconductor on the back of the capsule? And if you were forced to, wouldn't you use tensioned wire with air dielectric to earthed parts? OK, it's a minor point, but there are lots of them, and they all add up.

I'm intrigued by your mention of channel noise and inactive gate overlap capacitance. Could you expand, or suggest good references?

 

[johnferrier]

The opposite of no resistance...

...not for audio, I use 1-100 Giga-ohm resistors for calibrating a nanoammeter. I don't use them, but I've got two 1 Tera-ohm resistors in front of me (+/- 50G ohms). These are, of course, 1,000,000,000,000 ohms. And I don't have a way to measure them. I'm sure higher resistances are manufactured and used somewhere.

And here is one source for such devices: http://www.ohmite.com/cgi-bin/showpage.cgi?product=v_maximox

Okay, well this is certainly over the top, but here is a source for 1 Peta-ohm calibration standards: http://www.guildline.ca/products/Guildline9337Datasheet.pdf


JF

 

[scott wurcer]

I bow to your knowledge on semiconductor design

I'm intrigued by your mention of channel noise and inactive gate overlap capacitance. Could you expand, or suggest good references?

That's not necessary, I would hope that here I could talk off the top of my head and make stupid mistakes sometimes

My copy of Van der Ziel's book has taked a walk again, but that is where to go for noise theory. Think about what the short circuit noise would be of a capacitor with parasitic series resistance. In making large FETS some of the interconnect metal can be over a high sheet resistance region. Thermal noise is coupled out of the gate in this way. As Van der Ziel shows a portion of the channel resistance does this through the normal gate capacitance causing any JFET to have an input current noise that increases with frequency at some point. I have measured this effect (it is not that hard if you have a JFET with a big enough Cgs). In a charge amp (purely capacitive source impedance) the equivalent input voltage noise flatlines at a higher level than you might expect if you are not very careful.

 

[EC8010]

Thanks for that - it all makes sense. Although we are forced to consider current noise in bipolar transistors, voltage noise usually dominates when using a FET, except, of course when driven by a 30pF microphone capsule.