LVDTs are position sensors that get used in tough environments. They are simple, robust, and can be very linear. They could be used as an alternative to potentiometers where linearity is an issue or where increased sensitivity to small movement is needed.LVDTs are rather less common than potentiometers, and rather more expensive.Given their basic simplicity, I decided to have a go at building one. After an investment about two dollars and as many hours, I had an LVDT that works surprisingly well.Details, pictures, etc. are here: http://www.mikesflightdeck.com/lvdts.htmMikewww.mikesflightdeck.com

FANTASTIC work Mike! You've certainly intrigued me. I find this setup would be ideally suited for toe brakes and pedals which tend to have very little motion and IMO not well suited for pots.Looks like I'll have to brush up on my PLL theory. Check out the 74HC4046 and use only the Phase comparator portion of it. This should output a dc voltage "PC1" representing the phase shift between primary / secondary coils. "PC1" is usually fed back into the VCO portion, but we don't need a locked loop in this case. You may need an amp at the output to make it full range 0-5V...not sure about all this, its been a while. How

Hi Leo,Thanks for the feedback. This was a fun little project. I'm getting a little brain dead from editing and needed a break. (Of course, I should have been working on the bathroom remodelling, but...)The output phase doesn't vary smoothly as the core moves. It's the voltage level that varies. Still, now that you mention it, if you added to the output signal a fixed amplitude, same frequency signal, 90 degrees out of phase with the exciting signal, the net output phase would move from 45 degrees lagging the driving signal to 45 degrees leading. Then a phase detector could come into play. Actually, you could just use the timer capabilities of a micro controller.I didn't look at jitter, per se. I looked at repeatability in a crude way. I moved the core against the background of a machinist's scale. To the extend that I could accurately move the core back to the same position, I got the same output voltage. I wasn't using the best equipment. The signal source is a surplus HP function generator I picked up a a swap meet some years ago. It has a flakey pot on the output, so if you knock it, the level shifts. I used a Radio Shack DVM. (Calibration history? What calibration history?) Analog Devices has a few specialty chips that both drive LVDTs and condition their output. Their website has some docs on LVDTs that cover different approaches to wiring them. I'm not one to use expensive chips when I don't need the performance, (not that I don't WANT to, I'm just broke) so I think pragmatically that a few general purpose op-amps and diodes will probably do well enough.If you're interested, I could probably get a 0-5 volt signal if you could get it from a DC signal into a USB device. USB is on my list, but if I don't get my book together, my wife's going to uuhhh, disrupt my sense of well being. So USB study has to wait.Mikewww.mikesflightdeck.com

Please hurry and get that bathroom project done already. This community needs 100% of your creative input. :DYes, I'm fully aware of the dangers of using Analog Device$ ICs! I've inadvertently designed stuff around a sample chip only to later discover it sells for over $80 per chip!!!!Nowadays, I make it a point to do a quick price check on Digi-Key before deciding to use a chip that just sounds too good to be true.Now that I think of it, I did find a source for LVDTs but starting prices were about $100. There might be something cheaper out there, but I didn't look further into it. Besides, your solution looks practicable and definitely something I'll revisit once I (re)build my rudder pedals.If you can get a 0-5V DC signal out of it, then I can most definitely get that into a USB adapter no problem. I just completed this past week a totally new USB adapter design that does just that.The old ones were fine, but they essentially used the same gameport approach of charging a capacitor. Max resolution attainable with that design was 64 steps and was very limited in IO pins.The new one uses a Microchip USB chip with onboard ADCs. If digital filtering is activated, I've managed to get roughly 230+ steps jitter free using a standard pot. I loose a bit off the top and bottom due to Vref on the ADC but it saves me from having to use an external power source instead of the USB bus voltage.So if your home made LVDT needs some signal conditioning, I still have about 25% of ROM left and some 15% RAM for extra code and I might be able to squeeze in some basic cleanup code if required. I also have a few spare pins left over that I could use, as you suggested, to output a timer signal to the LVDT, thus possibly reducing the need to have additional external components to drive it.-Leo

Leo,I have a couple of ideas for cheaply interfacing an LVDT. I should be able to try them over the next several days.First idea is to separately convert the outputs from the secondaries to DC using the precision rectifier circuitry I mentioned on the web site. The resulting voltages are then subtracted, scaled and offset by another op-amp. This gets 0 to 5, or 1 to 4, or whatever's convenient.The second is an extension of the first. Using the net DC signal, generate a variable frequency pulse output ranging from perhaps 50KHz to 100KHz. A PIC could simply count this signal for a tenth second interval. A range of 5,000 gives you about 12+ bits, No ADC required. Of course, whether or not it's a REAL 12 bits remains to be seen.Mikewww.mikesflightdeck.com

Mike, sounds good. I'll leave it in your capable hands. See which approach looks most promising to you and we can take it from there. Either way is fine by me and I can easily add support into my existing design. If you can get it to 7 bits with LSB as error, the filter should be able to boost it to a usable 8 bits. I'd have to do the math but I may have enough room in the current USB buffer for 10-12 bit resolution for each LVDT axis. I wonder how the Game Controller driver will handle more than 8 bits per axis? I never tried that before.Sounds like we'll have a trade on our hands. How

If anyone ever wants to translate this thread into English and repost it ......:-(

Leo,I think we can reach some sort of agreement. Best to wait and see how the LVDT electronics performs. I've a little too much vaporwave odor here at my end. Let's make sure I've something that works first.Mikewww.mikesflightdeck.com

Guess Leo and I are getting a bit obscure...Here's a recap of the high points.LVDTs are accurate, robust sensors that get used in aerospace applications instead of potentiometers.They are inherently simple, but even the economy models cost too much for casual use.Their accuracy makes them useful in sims for applications like throttles. Their robustness makes them useful all around but particularly in rough applications like rudder pedals and toe breaks.They are simple enough to make cheaply with simple tools.At question is how to interface an LVDT to a flightsim PC. Leo, our resident USB expert, can handle getting a 0 to 5 volt signal into a PC. We have been discussing options for getting from the LVDT to a USB micro controller of Leo's design.And, yes, all this could be accomplished to some degree of performance with potentiometers at a fraction of the effort.But, where's the fun in that??Mikewww.mikesflightdeck.com

Take as long as you need Mike. I'll see if I can dig up those old chokes or tuning coils, whatever they're called and try a few things out myself when I get some spare time.FYI - You may want to use the AD623AN from Analog Devices instead of the usual OP Amp. Its perfect for this low level signal stuff and performs much, MUCH better than any OP amp I've ever designed or used. Extremely linear response and uses a single resistor to set your gain. No Rf or Rin to worry about.You can get samples of it and it only costs $2.90 each:http://www.analog.com/Analog_Root/productP...AD623%2C00.htmlRichard, maybe Google translate might work. :-lolWell, let me see if I can translate. Try and think of an LVDT as a sliding pot, but it makes no contact internally so it can never wear out. Unfortunately, the output is not as simple as a standard pot since its not resistive in nature so it can't be used in the same way without additional circuitry.-Leo

Iwas with it as long as the actual LVDT's were in discussion.... When you guys moved on to the ways to turn two out of phase AC signals into one DC one..... Whoosh! :-zhelpRichard(Good cabinetmaker, competent mechanic.... electronic illiterate)

You followed better than you think. When we got to the signal conversion part, we were lost too.Mikewww.mikesflightdeck.com

Truth is neither did I. I was just humoring Mike with technobable hoping he didn't see through me so he'd keep working on perfecting them. ;)I just want a few working LVDTs that

Excellent work Mike!Fun to see an achivable solution to the everlasting "the only pots i can afford are crap"-problem. Will hang on to this project Since it would be great for my soon to be throttle quadrant.I have showed your page to a collegu of mine (same guy who came up with the improvement to REDEC that im currently working on). He immediately put his glasses on, which is a good sign. :-)If I get it right the need-to-be-done things are:1. A simple driver oscillator, square wave? (probabvy easiest) sinus wave?(harder to do, but probably more efficient) Which frequency is best?2. Testing different construction techniques of the LVDT. Materials? Dimensions? 3. Decoder circuit. As i have understood it, if the core is in the neutral position, the current should be 0 in the secondary windings. If it is pulled in any direction, sine wave will be shifted 180 degrees depending on the direction of movement. The amplitude of the sine wave will also change proportionally to the movement.4. Adapt the decoder output to a PC interface, preferably USB.Quite interesting project! Lets share!Found a good explanation and description of LVDTs by a producer of LVDTs, can help explaining what it is and give sime hints on construction.http://www.msiusa.com/schaevitz/pdf/lvdt/LVDT_Intro.pdfCheers!/ Olle

Good morning, Olle,This project started out as a "let's see what we can do" sort of effort. It appears that may indeed become an alternative to pots for those willing to do a little building.So far I've been using a 1 KHz, 2 volt sinewave signal to drive the LVDT. Initially this was from a test oscillator. I have since build a dedicated oscillator. The current core material is a piece of steel rod from the local hardware store. It's not an outstanding material for a magnetic circuit, but not all that bad. If the frequency goes too high, losses go up. If the frequency goes too low, there is not enough inductance in the winding and the current demands go up. I simply SWAGed (scientific wild #### guess) the frequency, and as it seems to work, I stuck with it.Construction techniques are important not only for functionality, but for choice of materials that are reasonably available to the hobby crowd. Part of the fun I get from the hobby is challenging myself to build something using common materials and items. For example, the case of the LVDT should be a material that offers magnetically shielding. The next prototype will be built to fit into a length of narrow diameter threaded pipe. This will also offer a way to mount the LVDT, because the outside will be threaded.In an LVDT it's the amplitude variability that is most important in its operation. The phase flips as the core moves through the null position. The decoder can subtract the AC signals from the two output windings, in which case the phase must be used to determine which side of the null the core is on. If the decoder converts the AC to DC first, it can subtract the converted signals without reference to the phase. The circuit I am working with takes this second approach. It's working, putting out a little more than 1 volt change as the core is moved throughout its range. There are a few minor changes I will make. When the circuit's performance is verified I will post the circuit on my LVDT page.I think Leo already has the USB interface working. Mikewww.mikesflightdeck.com