Rotary Encoder ?

[Craig]

Hello

I need to start a new project which I was thinking of using a Rotary Encoder on a Small Stepper Motor to give Very Accurate Distance Measurements on a 1 Meter Liner Guide.I would like to ask for a bit of advice amongst forum members who might have some experience in this to guide me in choosing the right unit. I have looked at Absolute Optical Encoders which look like they could be the best Choice as they can re home if there position is lost?

Thank you for your input and advice!

Regards
Craig

 

[John Drew]

Hello Craig,
I've had a fair bit of experience with absolute encoders. If you have a look on my website: http://www.vk5dj.com
then menu items: projects/beam controller
You'll see the encoders I have used. They include 10 bit (1024), 12 bit(4096) and 16 bit (65,356) devices.
I'd consider 12 bit as a minimum and with some averaging you can expect 0.1 deg resolution.
I've written code for all of those encoders I refer to on my website. Contact details on the manufacturers is there also.
Cheers
John

PS I've only ever wanted 1 revolution on antenna positioning so if you use multiple turns you'll have to keep count of the revs and save them during a power outage. Most of the encoders (except 16bit) are based on the AS5040/45 chip under a rotating diametral magnet (2 pole). Accurate readout up to nearly 10 revs per sec.

[Giuseppe MPO]

hi, you can insert a quadrature encoder, much cheaper, managed with a PIC with CIP peripheral,
for example PIC18F2431, which I used for this purpose and it works flawlessly even with high
speed, without using the PIC resources.

[John Drew]

Hi Craig,
I notice you used the word "re-home". There is no home for an absolute encoder, I think you probably know that. It always reads the correct position (+/- accuracy) when power is applied. It is the disadvantage of quadrature encoders, but as Guiseppe notes the latter are much, much cheaper. If you use the 16 bit absolute devices and they are very expensive, with one turn on the encoder the best readout you could achieve with one turn would be +/- 0.02mm. A 12 bit encoder would give you 1mm resolution on a length of 1 metre.
 
In my use of the absolute encoders where I only require one rotation it is easy to create a calibrate function using simple maths. Users just peak the antenna dish on noise from the moon and press "calibrate". The routine corrects for installation errors by comparing the antenna position (Azimuth and Elevation) with the calculated position of the moon. The error correction is saved to EEDATA then applied to position and the system is thereafter calibrated. Turn power off at the end of the activity and then power on later, the error correction is read from EEDATA and all readings remain accurate and calibrated. Unless there are physical changes to the antenna mount "calibrate" is not used again.

Alternatively, if you're happy to re-home each time then I think you'd be able to track the position of the stepper by counting the pulses and avoid an encoder completely.
John

[Craig]

Hello John and Giuseppe Thank you very much for your feedback very much appreciated, On second thoughts I think a quadrature encoder would probably work fine as 1mm resolution on 1 meter would be accurate enough.
John I could look at a good quality Small stepper as an option, Which would make things easier.

Regards
Craig