Bridgeport Series II Interact 2 Conversion

More
14 years 5 months ago #20198 by andypugh
andypugh wrote:

KPA wrote:

The servo's have both a resolver

In that case I think that the 5i23 / 7i49 combination is the one to use.


I forgot to say that Pico Systems also have a Resolver interface board for the PPMC system which Igor used, so both Mesa and Pico can be used with Resolvers.
If you have an oscilloscope it is easy to see if the transducers really are resolvers, there will be a sine wave going in ,and two sine waves going out, and te two ouptut waves will vary in relative amplitude as the motor shaft is turned, once per rev (or possibly twice). If you see the amplitudes change hundreds of times per rev, then it is an analogue encoder.

Please Log in or Create an account to join the conversation.

More
14 years 5 months ago #20210 by andypugh

Rendering Error in layout Message/Item: Kunena\Forum\Libraries\BBCode\KunenaBBCodeLibrary::renderEbayLayout(): Argument #1 ($ItemID) must be of type int, string given, called in /var/www/html/libraries/kunena/src/BBCode/KunenaBBCode.php on line 320. Please enable debug mode for more information.

Please Log in or Create an account to join the conversation.

  • KPA
  • Topic Author
  • New Member
  • New Member
More
14 years 5 months ago - 14 years 5 months ago #20212 by KPA
They are Hendenhain ERO 115 125 D12. I've taken one apart and it has a copper disk with slots in it around the circumference - so I suppose the term Analog Encoder would then be what it is. I'm borrowing a scope from a friend in a week or two and will be able to see the signals at that time. My alternate approach is to try and re-use the existing A-D Board. The Heidenhain manual calls it the Analog Component Board. Encoders and handwheel feed into it, and supposedly digital signals come out. Again the scope will help me determine that.

So given that they are Analog Encoders, I'm now not sure if the resolver boards from Mesa of Pico would work with it. If that's the case then my choices narrow down to:

A. Try to re-use the Analog Component Board like John did with his conversion
B. Replace the Analog Encoders with Digital Encoders

From John's Web page where he documented his build

For the initial setup I just used halrun with a test hal file. This way I could test to see that the motors moved according to the command signal and check the position encoder feedback. This was the first setback in the project as I found that the Heidenhain encoders produced analog sine and cosine 125 cycle/revolution output signals. So to handle the encoder output problem I took the encoder interface board from the old Heidenhain controller and mounted it on the back of the cabinet door close to the PC. The Heidenhain board then outputs 625 A quad B digital cycles per revolution giving a resolution of 2500 counts per revolution. The ball screws have a 5mm pitch and are driven by a 2:1 timing belt reduction from the motors. So 2500 counts of the encoder represents 2.5 mm of motion or .001mm per count.




Jay
Last edit: 14 years 5 months ago by KPA.

Please Log in or Create an account to join the conversation.

More
14 years 5 months ago #20214 by andypugh
KPA wrote:

So given that they are Analog Encoders, I'm now not sure if the resolver boards from Mesa of Pico would work with it.

No, they won't, they are for resolvers, you have encoders.

If that's the case then my choices narrow down to:
A. Try to re-use the Analog Component Board like John did with his conversion
B. Replace the Analog Encoders with Digital Encoders
y

C. Get an interpolation box from eBay (3-channel, or one for each axis)
D. Make an interpolation device out of a microcontroller (such as an Arduino)
E. Square the pulses up with a simple comparator circuit.

E is the easiest, and will also boost the 1V signals to 5V TTL. It might not give you as much resolution as you would like.

Please Log in or Create an account to join the conversation.

  • KPA
  • Topic Author
  • New Member
  • New Member
More
14 years 5 months ago #20220 by KPA

Rendering Error in layout Message/Item: Kunena\Forum\Libraries\BBCode\KunenaBBCodeLibrary::renderEbayLayout(): Argument #1 ($ItemID) must be of type int, string given, called in /var/www/html/libraries/kunena/src/BBCode/KunenaBBCode.php on line 320. Please enable debug mode for more information.

Please Log in or Create an account to join the conversation.

More
14 years 5 months ago #20223 by andypugh
KPA wrote:

My original control board provides a 5 fold increase so given the ERO 115 gives 125 pulse per revolution (ppr) * 5 Fold * 4 quad digital = 2500ppr.

That is probably enough.
Assuming 4tpi leadscrews (and they are probably finer than that) then 2500ppr gives you a resolution of 0.0001" (or, 5mm pitch gives you 0.002mm). You need more resolution than you expect to have machine accuracy, but that sounds good enough for 0.001" or 0.02mm positioning accuracy.

Also the EXE PDF Andy provided talks about clock frequency (?) not sure what this refers to.

It is the number of counts per second, and limits the motor speed. Your 2500 ppr system, if it has one of the 2MHz boxes then means that it can only count up to a motor speed of 2M / 2500 = 800rps (48000rpm). Which is probably not a problem at 5-fold interpolation. 25-fold and 2MHz would be 9600rpm. None of these seem like significant limits.

Please Log in or Create an account to join the conversation.

Time to create page: 0.142 seconds