Amplify the hell out of it.

"Gregory L. Hansen" schreef in bericht news:cv357n$98k$ snipped-for-privacy@rainier.uits.indiana.edu...

Google for a RIAA pre-amplifier and remove the RIAA network. Or a microphone amplifier. The rest is cabling and shielding, much the same it is with record players and microphones. In a hostile environment, you want that amplifier practically on top of that pickup.

That looks different than what I was picturing. At 100 Hz the closed loop gain is still what I expected, but the signal coming out isn't what I expected by 10%?

I'll think in terms of a small aluminum box with bulkhead connector for the pre-amp.

Flicker noise is a new one for me. The specs give a standard coil resistance, in Ohms, of 150/240/380

I don't know what that means. There are three coils, I suppose they could have three different resistances, but it doesn't make sense to me that they should all be different. There's two lateral and one vertical, so I'd have thought two, at least, should be the same.

Can you recommend a low-noise bioplar op-amp? Would this be basic op-amp or still going with the instrumentation amplifier?

I'm not sure wha the "CURVE SHUNT DAMPING" box means, either, unless it's "shorting" the coils through resistors of various values and measuring the voltage across the resistor.

I should be asking the seller these things.

I'm getting some very specific information from you people, I'm saving some of these for later reference. Thanks.

I think I can attach a small aluminum box with bulkhead connector directly the device. That should be pretty close.

I'll have to get back to his later, but thanks for the reference.

The text immediately above is mental exercise. I wasn't sure what realistic numbers would look like, so I waved my hands and made some magic numbers appear.

But I started the thread with plans for the future building of something for the Mark Products L15B at the top of this page,

for which I have not yet placed an order but likely will in the next few days.

I also emailed the seller and verified that this interpretation is correct. After you have yours on the way, I may order one to play with.

Offset voltage drifts, so a single trim may not lass long. Using a very good low noise and low drift opamp would be better for the first amp. Something like an LT1028 or LT1007 has only microvolts of drift.

That suggestion separates the load resistor from the amplifier gain setting network.

Very low frequency high pass side constantly corrects for offset drift, low pass edge gets rid of high frequency noise. You could also add an in band low frequency boost to extend how low the frequency response would be linear (move the low frequency roll off corner down between an octave and a decade, perhaps).

After reading your link about the geophones, I did some Googling on the longer period Lehman horizontal pendulum seismometers. I noticed that none of the ones I saw had much technology in the pickup coil end, but put all the effort into the amplifiers and filters. I worked out how to make a shielded, hum bucking (and geomagnetic change canceling) coil with lots of output that would make the amplifier end easier and cleaner. But I think the way to go for multi second waves is optical (led beam to two photo diodes in a bridge amplifier, with the swinging weight bar that moves a shadow over the diodes). Flat response except for mechanical response of the beam.

Right. The seller got back to me, and he has the 380 ohm version. All coils at 380 ohms. And he'll make sure they have the appropriate shunt resistors.

Trim pot?

I'm going to have to think about that before I know whether it makes sense to me.

What is the bandpass filter for?

For the electronics, once I figured out the "issues", I had thought to make something on the cheap and easy side, with the understanding that I want something portable, I don't have anything like a laptop computer or PDA to plug it into, and my immediate application doesn't need a precision measurement. I mean sure, feeding an A/D into a laptop computer and throwing up a frequency spectrum versus time and saving it to disk would be nice, but not in the near future. So I'd thought of including a bandpass filter that could be switched in if I wanted some basic frequency information. I'd just know, having been the one that constructed it, that the dial selects a center frequency from A to C, halfway is B, etc.

I read in sci.electronics.design that Gregory L. Hansen wrote (in ) about 'Amplify the hell out of it.', on Fri, 18 Feb 2005:

To eliminate noise and out-of-band signals. Actually, since you want to go down to nearly DC, a low-pass filter is more practicable.

That sort of bandpass filter, with a narrow enough bandwidth (maybe 3 Hz) to do useful frequency analysis, is not what people have been suggesting you incorporate, and is not so easy to do in the very low frequency range.

Okay, that's sort of what I had in mind.

And it's not the first time in the past week or two that I've been bitten by the special needs of low frequencies. I've sort of been assuming that lower frequencies means easier. And in some ways it is, but not in all ways.

I think this isn't quite what you described, but the superspring jumps to mind. A spring is a mechanical analog of a capacitor, a mass is a mechanical analog of an inductor, and the position of a mass relative to the mount is a mechanical analog of a voltage difference... The supercapacitor!

Yes and yes.

Wow. LT1007, noise in the nanovolt regime, offset less than 25 uV (I assume that's 25 uV multiplied by my gain?), $2.42 each in small quantities. Capitalism works.

I read in sci.electronics.design that Gregory L. Hansen wrote (in ) about 'Amplify the hell out of it.', on Fri, 18 Feb 2005:

Per root hertz. OK, you don't want many hertz, but bear it in mind.

Yes, the 25 uV is referred to the input.

Both the noise and the offset depend on how you arrange the input circuit. The noise depends on the source impedance at signal frequencies, while the offset depends on the source impedance at DC. Hint: AC coupling affects that a lot!

Thanks for the report. That is about how big I expected it to be. But since talking with you, I have sort of gotten my heart set on making my own sensor coils and an Lehman type long period seismometer. I have a bunch of magnets selected for purchase, and am getting a quote from a machinist tomorrow, for the iron pole pieces. I think I can achieve quite a multiple of the signal to noise ratio (50, perhaps) compared to any sensors I have seen that hobbyists build. It has a hum bucking coil design that cancels the effect of large fields, like power line hum, lightning and variations in the geofield caused by aurora. And the whole coil sits in an approximately 10,000 gauss field. It would make one hell of a linear motor.

My goal is to be able to lie down on the concrete, next to my unit, and monitor my heart beat above the noise. If it will do that, I should be able to pick up any significant quake on the planet. Like the geophone pickups, this will be a velocity output.

If I get that to work, I want to add an optical detector that goes down to DC so the unit effectively becomes a tilt meter. I have no idea how sensitive I can make that, compared to the magnetic unit. Connecting them in a feedback loop, to hold the magnetic bob stationary with respect to the Earth will make the coil current proportional to acceleration.

I am very interested to hear how your amplifier works with the unit you have purchased. I need to get a few of those LT1007 opamps for the front end of mine.

After you get a signal, I would like to talk with you about filtering possibilities to reduce high frequency (road traffic) noise and extend the low frequency performance, if these things interest you.

Somehow it seems bigger than it looked in the pictures. The can is five inches in diameter, almost three inches high, the sensors mounted on the lid which has two arrows oriented with the two horizontal sensors and a spirit level in the middle. The large screws (5/16"?) on the bottom allow mounting to something. And the two fittings are rubberized compression seals that would allow wiring through conduit.

I haven't had the chance to plug it in to anything yet, but from the physical inspection I like it.

Got them on EBAY. About what you paid for the geophones worth of neodymium iron boron cylinder magnets.

That's it. Two coils who's signals add, but who's noise pickup cancels to the extent that they are identical. Took me a couple days to decide how to arrange the magnets and pole pieces and make room to get the coil support and wires out. It uses 4 of 1" diameter by 1/2" magnets down the center line, and 4 sets of 4, 1/2" diameter by 1/2" magnets around the outside, to close the flux path. One coil has a radial field with north in the center and one with south in the center. And there has to be no net external field, so the magnets (that hang on the pendulum) are not feeling the Earth's field. I think I got that all right. Now I am working out the structure that supports the pendulum and the bearings for the pendulum.

Saved you a lot of time and money, up front.

If I get it all to work as I hope before I lose interest, I might contribute an article to some one else's site. I have subscribed to the Public Seismic Network PSN-L mailing list to make some connections to some other amateur seismologists.

Understood.

Well, keep in touch as things develop.

That's quite a project. Where did you find the magnets?

What's a hum bucking coil? I had thought of making two identical coils connected in opposite senses with the magnets for each coil in different directions, so that e.g. line noise would be (approximately) equal and opposite in the two coils, but the pickup EMF adds together. But I went with the geophone instead.

You should do a web page.

I think I mentioned before that I might move kind of slowly on it. Partly because I'm learning as I'm going. I'm just starting to read about active filters in my Horowitz and Hill. And partly it's just hard to work around the fact that I'm living in a rented room on a student's budget. It's hard to build up much of a workshop.

I don't have the special op-amps yet, but I've been to Radio Shack to pick up miscellaneous electrical things. It doesn't seem too bad to spend 99 cents on an op-amp, 69 cents on a pair of sockets, $2.50 for a switch, etc. But when you run through a parts list, those nickels and dimes can add up. I wanted to try wire wrapping this one. They had a wrapping tool, but no sockets. Oh, well.