Flux fux, and 10 Meg ohm

Oct 06, 2016 82 Replies

com, so we'll see in a couple of weeks.

n that some of them aren't very good at it. Paul Horowitz is a physicist, a nd Jan Hall (2005 Nobel) designed all his own stuff, just for starters.

.

These very high gain TIAs are intended for use in mass spectrometers, in pa rticular for geochronological dating (see e.g. Argon dating on Wikipedia). High sensitivity is required for low abundance isotopes relative to the maj or isotope and precision ratios are required. The original samples are usua lly also very small so the total number of relevant atoms is tiny. In this high gain region the primary source of noise is the Johnson noise of the fe edback resistor. The general argument is that the gain increases proportion ally with increase in feedback resistance while the noise only increases wi th root of the resistor value so the S/N ratio improves as the root of feed back resistor value increase. This improvement continues until the shot noi se of the IC amplifier bias current presents a limit and this occurs, at pr esent, around the 100Tohm mark (this is also the about the maximum value of resistor available). It indicates that the input IC needs to have very low bias current and the ADA4530-1 is the latest arrival with only femtoamp b ias current. It may be thought that ion multipliers would be better for the lowest abundances but they have a major drawback. The impact (and burying in the surface) of heavy ions changes the secondary emission coefficient so stable cross calibration of isotopic channel gains cannot be achieved and high gain TIAs must be used. It is also necessary that the response time of the amplifiers be short enough that measurement can be made in a few secon ds since the sample introduced into the spectrometer vacuum decays with tim e and the measurement has to be extrapolated back to zero time, a problemat ical procedure if the extrapolation is excessive. My book is aimed primarily at my physicist (and allied subjects) colle agues rather than EEs so there may well be as you say some ?weirdn ess? in the approach. As you slightly obliquely point out Horowitz and Hill is a very successful book indeed, but I wanted to encourage the us e of circuit simulation as an aid for those for whom circuit design was onl y incidental and not a primary skill. This is all getting a long way away from the original topic of cleani ng PCBs and components!

.com, so we'll see in a couple of weeks.

an that some of them aren't very good at it. Paul Horowitz is a physicist, and Jan Hall (2005 Nobel) designed all his own stuff, just for starters.

d.

Rev.Sci.Instrum. may arouse some bad memories but just for the record the a uthors of the paper I listed, Pelchowitch and Zaalberg van Zelst, were from Philips Research Laboratories, N.V. Philips Gloeilampenfabrieken, Eindhove n, Netherlands. I presume you would consider this source as sound. The auth ors of the PIRE paper were from Keithley Instruments and the Naval Research Lab., Washington D.C..

?In reflecting on these things one cannot help but be amazed at how short are the memories of many engineers and engineering faculties, and ho w often the battle to achieve these objectives needs to be refought. At th e beginning of World War II, when engineers were presented the problem of d eveloping radar, they were ( except in very few cases) found woefully lacki ng in an ability to cope with such an unconventional situation; and physici sts, both theoretical and experimental, had to be called in to do what was essentially an engineering job.? E.A.Guillemin, PIRE 50, 872-878, 1

962.

I am sure the training of engineers has changed immeasurably since the days referred to, but we should remember that we inhabit a continuum and the di visions between subjects are artificial so we should delight in enlightenme nt from wherever it may come. Regards, Scott.

Interesting. You'd certainly need something pretty sensitive for that job, and a fancy TIA is sure cheaper than a Channeltron. I think most EEs would use some sort of charge-dispensing loop for that sort of job, rather than a resistor. It's a tradeoff between e.g. contact wear, if one is using a relay to reset the cap, or nonlinear capacitance if you're using bootstrapped diodes. I don't think I'd want to use a FET at those current levels, but a sufficiently small one might work. Trouble is that all the small-signal discrete MOSFETs have gone away.

Some years ago I did a very low cost pyroelectric imager that used LEDs as biased switches way down in the fractional picoamps. (

formatting link
). Bootstrapping a couple of those would get you down into the low femtoamps, I expect. Have to keep them in the dark though!

One of those electrometer op amps might be interesting to combine with the sampling-bridge reset trick. I don't remember who invented it, but it's really cool: you reset the cap by inverting the op amp's power supplies briefly (with current limiting resistors, obviously), which turns the input protection diodes into the equivalent of a diode sampling bridge, Those diodes (or whatever they are) have to be low enough in leakage to work with the amp in normal bias, so you're golden.

(I love ju-jitsu tricks like that.)

Well, another approach would be to crash the ions into a thin scintillator on top of a CCD. That would spread out the problem, anyway.

We physicists are all weird. Eventually we learn to embrace it. ;)

Sure, why not? As I say, I bought a copy myself. There's room for a lot of cookbooks too.

Welcome to SED. ;)

Cheers

Phil Hobbs

Dr Philip C D Hobbs Principal Consultant ElectroOptical Innovations LLC Optics, Electro-optics, Photonics, Analog Electronics 160 North State Road #203 Briarcliff Manor NY 10510 hobbs at electrooptical dot net http://electrooptical.net

Join the Discussion

Have something to add? Share your thoughts — no account required.

Didn't find your answer?

Ask the community — no account required