LVDT help please

Analog Devices would be a good place to start:

Get a couple samples and start playing.

Depends how fussy you are trying to be. I just checked over the notes for a project I did with an LVDT, and for all intents and purposes an AC voltmeter on a secondary would do the job. Preferably one that you can easily open up and put a paper face on with markings that have something to do with your reading. The primary coil input needs to be consistent for this to work...

Excellent idea but you need to check to see if the output is DC and what the range is.

Unless the thing has a rectifier built in it's AC.

It looks like that might work. And all that needs to be done is to use gauge blocks to move the probe tip and measure the voltage change per each .0001". Thanks, Eric

If you can get some specs, it would help greatly. First, you need a sine wave oscillator, to excite the driving coil. Then, you need two simple differential amplifiers. You could use regular op amps, but the AD620 makes a much better instrumentation amp. You then compare the amplified signal level from the two sense coils. A good circuit would be an active rectifier (look in any op-amp circuits book) on each of the AD620 outputs, after a DC blocking capacitor. Then, just feed the two active rectifiers into a difference amplifier, or rig one of the active rectifiers to provide the revers polarity, and sum them. Feed this to a range selector switch and a big analog meter.

Finding the right frequency and drive voltage would be real helpful in making it work best. The frequency is likely to be somewhere around 800 Hz to 3 KHz for the smallest LVDTs.

Jon

you will get your best accuracy with a synchronous demodulator - basically you want to cancel out variations in the modulation voltage - if you can find the schematics for any of hte old Collins flight control stuff you will see the circuit. You will need to know the excitation freq for your unit - many are 400hz, some are higher, and drive it with the proper freq - make a simple sine wave source to do that. You can get really excellent accuracy and linearity, they are great for servo systems

I've seen a digital system that was pretty ingenious.... it started with a clock. The clock drove a free running counter maybe 0-1024. The output of the counter fed a pair of roms, one for Sine lookup, and one for Cosine lookup. The roms fed a set of dacs and those fed the sine and cosine coils, as drive coils.

The pickup was off the center coil.... Depending where the slug was, the pickup coil is going to be closer to Sine or Cosine in terms of phase relationship.

The pickup coil fed a zero crossing cicuit. The zero crossing time captured the value of the free running counter. If the slug was near the sine drive coil, the zero cross would be closer to the sine wave output for zero crossing. If it was closer to the Cosine drive coil, it would be closer to 90 degrees phase shift before the zero crossing would occur, and the value captured from the free running counter would reflect the position of the slug.

This circuit is fairly clever for a noisy environment. Since the sensing circuit does not have a varying amplitude, it is not as sensitive to noise.

The traditional method is to drive the center coil, and measure the _realtive_ amplitudes of the Sine, and Cosine coils, so with one coil delivering a low level signal, noise immunity is a concern.

The method here is an all digital approach.

And NOT that complicated really. It could likely be all done inside a small micro processor.

Pete

If it is an aircraft LVDT, the exc. voltage will be 26vac, 400 hz. There are LVDT's all over Boeing, Lockheed, and Airbus autopilot flt. control actuators. But they are_never_used for anything but feedback to the flight control computers. All position indications to the cockpit surface position indicators are always from RVDT's. Rotary variable displacement transformers. I don't know why this is, it just is. One of those questions I'd like to ask a Collins engineer someday. But I see no reason that an LVDT couldn't be used in a linear position circuit.

Garrett Fulton

Depends on the accuracy you need. For a "Science Fair" demonstration you can do it with an oscillator and an AC voltmeter. For real measurement, find, scrounge or buy an LVDT module from Burr-Brown or Analog Devices. They appear simple, but they're prone to all sorts of inaccuracies -- best dealt with by someone else's clever work on a hybrid circuit.

I don't know squat about the electronics, but did use an LVDT in a servo-hydraulic circuit to control position within .001". Worked great. It required a scaled down servo valve for the low flow. The Vickers sales guy said they were originally developed for the Disney animatronic critters. By the way, I really didn't like dealing with Vickers. If I ever need to do that again (unlikely), I'll go with a Moog.

Pete Keillor

Hi, Eric:-

(sorry about threading- my ISP changed their news setup)

Check out these:

The PLCC version is USD 27 1-off the Cerdip version is USD 80, so you'll want to pick up a PLCC socket and use the former.

Best regards, Spehro Pefhany

Thanks Spehro, It looks like there are many solutions to this. I am going to see if Edmunds, the gauge company that made it, will give me the exitation frequency for the thing. Their web site doesn't list this probe. I think it is now obsolete. Cheers, Eric

Hi Pete, I don't know enough about electronics to build what you are talking about. But I am learning about microprocessors so maybe someday... Eric

It's harder for me to explain than it is... The LVDT has three coils. the center coil, and the outer coils. The slug in the middle moves, and either couples more from the left coil to the center or, at the other end couples better from the right coil to the center.

You could drive the center coil, and then compare the feedback from the left and right coil to see which one is getting a stronger signal. The coil with the strongest signal is where the slug (movable part) is. A quick and DIRTY way to cheat is to drive the center coil, and look at the signal level from just one coil, to guess where the slug is at. Without comparing the relative strengths of the two end coils, you are subject to noise, and varying signal levels. The accuracy, and long term stability might not be real great.

Both of those methods require you to measure the signal level of the sensing coils.

In my previous post, a completely different method of finding the position is described. In that method, the two outer coils are driven with a sine/cosine signal. The center coil is used for sensing position. Driving the two outer coils with sine/cosine signals will genreate a signal response from the center winding that will vary from a Sine to a Cosine depending on where the slug is located. If the slug is to the left, the sine drive coil will couple more strongly to the sense coil, and the sense (center) coil will match the sine wave in phase angle. If the slug is to the right, the cosine drive coil will couple into the sense coil, and the sense coil will have a phase angle closer to the cosine drive signal.

This sensing method does not rely on the signal level, but rather the phase angle.

So, all that is needed is a sine/cosine gernator, such as a ring oscillator if done in analog, and a zero crosing detector to find where the sense signal crosses zero.

The zero crossing detector could drive a sample/hold circuit to capture the sine wave drive value, to give an indication of the position.

I hope expaining it another helps to clarify the concept.

Pete

Eric,

FWIW there are LVDTs that only require a DC input and they in turn deliver a DC output swinging from a minus level through zero to a positive level, thus the output is calibrateable/readable on a DV voltmeter.. These types contain the AC oscillator and the rectifier/demodulator built in. The Series-240LVDTs work on input voltages from 6V DC to 30V DC. I'm sure tightly regulated input voltages are required. For example, one of the Series-240 LVDTs has a working range of +- 0.050" and with an input voltage of 6V DC, it gives a full range output swinging from - 1.2 VDC to +1.2 VDC. Another in the same series powered with 15V DC input has a working range of

+- 1.00" and the output swings from -10.8 V DC to +10.8 V DC.

I suggest you find some data sheets on LVDTs. My information on the Series-240 LVDT is over 10 years old so you may find better versions in new data sheets.