Thread Rating:
  • 0 Vote(s) - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
DIY inductive scale for less than 50€
Here's some info about scale's which I have been working on for some while.

The sensors are built out of a tube filled with ballbearing ball's and a set of induction and pick-up coils.
The theory which the sensors use is that of an LVDT.
By inducing a sine wave voltage in the induction coil, the pick-up coils generate an other sine wave which you can shift 360deg. per ball which they move over on.

The total cost for 3 sensors of +-1meter + board is approx. 100€ at this moment. This does not include the necessary tools (soldering iron, scope, ...).
The downside of this sensors is that at the moment every single sensorcoil needs a callibration file which takes houres to make. I'm hoping that by increasing the stability of the board I will be able to overcome this backdraw.

This weekend I was able to make my fist stable reading and calculation. This with a resolution of .008mm. By fine tuning the claculations in the mcu I should be able to get this up to .002mm.

Since I'm not an electronics engineer everything was based on trial and error, and a lot of searching and reading on the internet.
If there is anyone who can provide some feedback on how to increase stability of the board, simplify the schematics, ... etcetc I would be very grateful.

Attached Files Thumbnail(s)

.pdf   sensorbordV1.0.pdf (Size: 100.73 KB / Downloads: 37)
After a long discussion on a Dutch forum we have come up with a solution for the stability and accuracy of the bord.
To improve this we are gonna make use of a C2000 Delfino mcu wich will provide the drive signal as well as the differential pickup signal in 16 bit.
With the provided DSP librarys form Texas we should be able to significant improve the signal.

@yuriy: Would you consider sharing your source code for the ccs? This way I can implement the quadrature logics immediate in the soft.

I'll keep you posted and when the project is finished I'll post a write up of it...
After working on the project during the past hollidays I've made some progress.
After changing almost the entire analoge part of the bord the signals comming from the detection coils can be measured stable with a resolution of 14bit/8mm.
The osscilator have been changed to a christal driven one buffered by some darlingtons and the préamps of the coils have been changed to cascode amp's with all the nessecary filters.
This makes that every 8mm we can measure a voltage shift of 400mV steady without any countable noise or transformation of the signal.

Programming of the mcu is done and I hope that in the next week I will be able to do a first test together with an industrie encoder. Resolution of 0.05mm should be within reach.

In annexe you find a picture of the first sensorhead.

Attached Files Thumbnail(s)
I am VERY interested in this project! I have a friend who is wanting to put scales on his mill and use the TouchDRO app. Do you have any details on this project so that he can make his own?
(01-29-2017, 03:18 AM)Eskimo2k Wrote: I am VERY interested in this project! I have a friend who is wanting to put scales on his mill and use the TouchDRO app. Do you have any details on this project so that he can make his own?


Today I'm finalizing the hardware as my test of two weeks ago showed me some room for improvement.
After this I will adapt the schematics and upload them (maybe I will be able to do this this week).

The finalization of the software will take me a few more weekends since I still need to adapt the algorithms to work with universal adaptable calibration parameters. The draft of the firmware works fine but this one is written for my one specific sensor only.

If you want to build this, however I can only encourage this, do realize that it is not a project to be underesthemite and that the results of the final build will be the outcome of your owne capabilities. Knowledge of analogue electronics as well as code composer studio are absolutely necesarry to realize a working sensor.

In annexe you can find some pictures of the mounted sensor and the first result graph on the automatic feeder.
The hysteresis was the result of a wrong post processing and this is fixed in the new firm ware.

Attached Files Thumbnail(s)
Looks promising!
Is there any news on this project?

regards, Bob

Forum Jump:

Users browsing this thread: 1 Guest(s)