For a test we need a simple capacitance meter that works at a frequency between 1-10kHz, offers low amplitude sine wave (100mV or so) and most of all has a nice long lowpass function or even better an analog or at least pseudo-analog display.
None of that cheapo charge-discharge kind and no science project job, something with two terminals or prefreably BNC that just gets plugged in and works. The capacitance range needed is zero to 2000pF, later maybe slightly higher but that would do for now.
Does it exist? Where can one buy that? Or maybe it did exist in the days of Methusaleh and there is a chance to catch one on EBay?
--
| James E.Thompson | mens |
| Analog Innovations | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| San Tan Valley, AZ 85142 Skype: skypeanalog | |
| Voice:(480)460-2350 Fax: Available upon request | Brass Rat |
| E-mail Icon at http://www.analog-innovations.com | 1962 |
I love to cook with wine. Sometimes I even put it in the food.
Not sure what you are asking for that is hard to find. My volt meter has a capacitance range built in. I've never put it on a scope to see how it works, just assumed it used a low level sine wave. No?
A Boonton 72B does all that, except the sine frequency is higher. Nice low excitation voltage, synchronous detector, 2 and 3-wire measurement, nice pale green color. I got several on ebay.
--
John Larkin Highland Technology, Inc
picosecond timing precision measurement
jlarkin att highlandtechnology dott com
http://www.highlandtechnology.com
The capacitance meter that has been recommended here from time to time, and I bought on that recommendation, is the Almost All Digital Electronics L/C Meter II2. There's a web report that the guy that designed and sold it - N eal Heckt - is dead.
I haven't got a clue about the frequency and amplitude of his test signal, and the output is digital.
What you ask for would be easy enough to build around a sine wave oscillato r and a Blumlein bridge.
formatting link
If I was doing it, I'd get the sine wave out of a Baxandall class-D oscilla tor, and use a bifilar output winding as the two inductive arms of the brid ge.
1000pF at 10kHz is about 16k so which ever side of the bridge you monitor, you'll want to do it with a relatively high input impedance amplifier - pro bably configured as current to voltage converter.
You have to run the output into a phase sensitive detector, but what you ge t out is then an output signal that's linearly proportional to the impedanc e imbalance in the bridge. You've always got to filter the output of a phas e-sensitive detector, and you can make the output time constant as long as you like. If you put in a Sallen-and-Keys-with-gain output filter - more ac curately a VCVS - page 8 here
formatting link
you can use the gain to get a well-damped response without having to find f unny capacitor values - E96 resistors will get you close enough to the righ t gain (around 1.586 for no peaking)
It is excellent and I use it a lot, for capacitors and inductors. It uses a LC oscillator, which I built in a separate out-board module so that I could locate it right at the DUT. I buffered the oscillator output to drive the cable to the display box, and I fitted a local voltage regulator at the oscillator, and low-battery indicator.
Still you said you don't want a project, so what I would do is borrow or hire one of the low-frequency HP VNAs from the 1980s, I can't remember the model but perhaps HP3577A. You don't need the S-parameter test set, actually the 1Mohm inputs are probably better for what you are doing since your impedance will be far from 50 Ohms. I remember that you could enter complicated trace mathematics into that VNA so that you could make it read directly in capacitance, and do whatever error correction you want, to make up for any non-idealities in the cable or any measurement bridge, DC blocking caps, etc. that you connect up. What I would suggest is run three coax cables from the network analyser to the DUT, one for the source and one each for the R and A receivers, (or you could use scope probes for the receiver cables). Right at the DUT, connect some resistors (chosen to be similar magnitude impedance to the DUT), in a sort of voltage divider configuration, so that the ratio of the receiver voltages A/R is a sensitive function of the DUT impedance. You then enter some formulae as trace maths that figures out the impedance from the ratio of the receiver voltages and some "known" resistor in the measurement circuit. When I did something like this, I had transformers isolating the source and receivers so I had a calibration routine where I connected up a known impedance (in your case a precision 1000pF capacitor might be suitable) and stored that trace. Then I could use that trace in the trace maths to figure out the complex bridge reference impedance (including correction for the transformers, cable capacitances etc.), to give me a well calibrated answer for the DUT impedance later on. It won't give you a truly analogue readout, but it would give you a nice graph of capacitance vs time, which could be even better.
By the way, I got started reading the HP Journal article about the HP3577A, and was wondering why they went with double conversion receivers rather than a single conversion right to 10kHz. Either way they have no image rejection, so it will be just as sensitive to the image frequency (offset by double the IF) as it is to the source frequency. I am wondering why an image response at 500kHz offset from the source is that much better than an image response at 20kHz offset from the source frequency. The only thing that springs to mind is that perhaps the source has enough phase noise that steep filters would be harder to measure, due to source phase noise in the DUT pass-band corrupting the measurement of the DUT stop-band where the source frequency is highly attenuated.
There are capacitance bridges that will resolve attofarads at that frequency. It's basically narrowband-detected RF, and a lot of gain will extract really tiny currents.
--
John Larkin Highland Technology, Inc
lunatic fringe electronics
Consider: An oscilloscope might have 15 pF input capacitance. Connect a 20 volt p-p sinewave generator to the scope input through Cx. Run the scope in external trigger signal averaging mode. Resolving 5 mV p-p is easy, and that corresponds to 4 aF.
Use a fet buffer/amp to get another 100:1 or so sensitivity. What comes after "atto" ?
--
John Larkin Highland Technology, Inc
lunatic fringe electronics
--
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
You'd be surprised. Boonton 72BDs have a 3-1/2 digit display and a 2 pF full scale range, which means 1 fF resolution. Once it warms up a bit and you zero it out, it'll sit there at zero fF all afternoon. Wave your hand at it and it registers a few fF. (That's 100 kHz, but 1 fF is still 1.5 gigohms at that frequency.)
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
Yeah, a small-gap air capacitor is a really interesting position transducer. I've got 90 nm resolution out of an LVDT, but a capacitor could do a lot better.
--
John Larkin Highland Technology, Inc
lunatic fringe electronics
So can an LVDT, and it doesn't respond to air pressure or humidity changes.
Capacity change in vacuum tubes was routinely detected at low levels - it was what made the tubes microphonic. Thus, the 'preamp' was born, a box of tubes unconnected from the chassis of your hi-fi with the humming power transformer.
I don't think a cap would be much affected by atmosphere changes. The big problem with the LVDT was thermal. We had to mightily insulate the setup to get the drift down into the nanometers. An LVDT self-heats a little too, whereas a cap doesn't.
--
John Larkin Highland Technology, Inc
picosecond timing precision measurement
jlarkin att highlandtechnology dott com
http://www.highlandtechnology.com
Good cap sensors are differential, so all that changes is the scale factor. Air is more or less an ideal gas, so the dielectric susceptibility is proportional to density, i.e. P/T. Of course, epsilon-1 is only about 0.00056, so a 10% atmospheric pressure change changes the scale factor by about 100 ppm without moving the zero. Similarly the tempco of scale factor due to the air is about -2 ppm/K.
The materials are probably less stable than that.
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
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.