Yaskawa usarem motors, Amc be25a20 making it work?

  • JR1050
  • JR1050's Avatar Topic Author
  • Offline
  • Platinum Member
  • Platinum Member
More
24 Nov 2018 04:57 #121285 by JR1050
I have two Yaskawa ac servos, usarem 05c and 03c , they ac servos that use an 8192 absolute encoder. They don't have hall switches for communtation. The Yaskawa drives are out of my price range, so I'd like to use Amc be25a20 drives, as I already have them. These are intended for bldc motors that have hall switches and are analog ,+- 10v. Would it be possible to use the bldc component to create the hall signals and still feed the drive an analog signal? Present set up is 7i90 and 7i33

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

More
26 Nov 2018 13:25 #121364 by andypugh
It might be possible, if you can find a way to convert the encoder feedback into an angle.

A better solution would be STMBL, except that currently there are none available due to the main power IC being discontinued.
github.com/rene-dev/stmbl/blob/master/do...tting%20Started.adoc
It "talks" Yaskawa absolute and can be commanded by Mesa smart-serial.
The following user(s) said Thank You: JR1050

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

  • JR1050
  • JR1050's Avatar Topic Author
  • Offline
  • Platinum Member
  • Platinum Member
More
27 Nov 2018 19:10 - 28 Nov 2018 17:25 #121483 by JR1050
From reading thru massive amounts of Yaskawa literature, it seems the actual encoder has standard differential signals and the serial absolute signal is requested by the Yaskawa drive/host controller. This being the case, and this is an educated guess, I can use the bldc component to generate the hall signals by setting it up as an incremental encoder and use the marker channel for home?I attached the Yaskawa paper on the absolute encoder.

This browser does not support PDFs. Please download the PDF to view it: Download PDF

Attachments:
Last edit: 28 Nov 2018 17:25 by JR1050.

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

More
30 Nov 2018 15:36 #121636 by andypugh
It isn't immediately obvious from scanning that document if there is an index signal.
even if there isn't you might be able to home the motors using bldc in direct current mode. How accurate that is depends on the static load on the motors. (ie it is fine for a spindle, but unlikely to work well on a moving-knee Z axis)

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

  • JR1050
  • JR1050's Avatar Topic Author
  • Offline
  • Platinum Member
  • Platinum Member
More
01 Dec 2018 05:04 - 01 Dec 2018 06:41 #121671 by JR1050
Here is a short video of where I'm at, encoder counts, index channel appears to work. I'm not getting any hall signals though . The bldc component is rather busy, on the man page there are several references to personality, what do these pertain to? What would cause the lack of hall signal generation? I'll attach the Hal and ini files.



File Attachment:

File Name: hnc112517_...2-01.hal
File Size:16 KB

File Attachment:

File Name: bldc.ini
File Size:5 KB
Attachments:
Last edit: 01 Dec 2018 06:41 by JR1050.

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

  • tommylight
  • tommylight's Avatar
  • Away
  • Moderator
  • Moderator
More
01 Dec 2018 06:11 #121673 by tommylight
Incremental encoders have A, B, and Z channels.
Yaskava has A, B, and C channels, all of them give a pulse train that changes the density while rotating and that change is shifted on all channels from each other.

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

  • JR1050
  • JR1050's Avatar Topic Author
  • Offline
  • Platinum Member
  • Platinum Member
More
01 Dec 2018 07:02 - 01 Dec 2018 07:06 #121674 by JR1050
The paper from Yaskawa above states the a and b channels are standard incremental quad channels and the c is the marker pulse(you can see the index enable go low to high in the video). The serial data is sent thru channel A on start up, on a 15 bit encoder, then it reads incrementally. There are several manuals on these motors and drives, TSE-s800-2.4e and TSE-s800-15c do a good job at explaining how the encoder works. The wiring scheme is the same for both their incremental and absolute encoders with the exception of the battery backup being excluded on the incremental version, they call the marker/home/Z channel C on both. Ill bet they are they same .They are not gray code output.They count the specified distance on the dro based on the input scale.These are also older (1995 vintage) motors.

I am by no means an expert on this, as its new to me, but based on the info ive read and what Ive observed, Id say its looking like this will work if i can get the hall signals to output. Id bet im missing a parameter or don't have the component setup right...

This is worth figuring out, if it works it will save anyone who wants to use Yaz motors and not have to pay the big yaz drive cost.
Last edit: 01 Dec 2018 07:06 by JR1050.

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

More
01 Dec 2018 10:55 #121676 by andypugh

Here is a short video of where I'm at, encoder counts, index channel appears to work. I'm not getting any hall signals though .


"qiH" needs you to send an "init" signal:
4  bit   OUT         FALSE  bldc.0.hall1-out
     4  bit   OUT         FALSE  bldc.0.hall2-out
     4  bit   OUT         FALSE  bldc.0.hall3-out
     4  bit   I/O         FALSE  bldc.0.index-enable
     4  bit   IN          FALSE  bldc.0.init
     4  bit   OUT         FALSE  bldc.0.init-done
     4  float IN            0.2  bldc.0.initvalue
     4  float IN             90  bldc.0.lead-angle
     4  float OUT             0  bldc.0.out
     4  float OUT             0  bldc.0.out-abs
     4  bit   OUT         FALSE  bldc.0.out-dir
     4  float OUT             0  bldc.0.phase-angle
     4  s32   IN              0  bldc.0.rawcounts
     4  bit   IN          FALSE  bldc.0.rev
     4  float OUT             0  bldc.0.rotor-angle
     4  s32   OUT             0  bldc.0.time
     4  float IN              0  bldc.0.value

That starts a search-for-index.
I would change your enable section of HAL to look like
# axis enable chain
net emcmot.00.enable <= joint.0.amp-enable-out 		#hal pin enables the amp
net emcmot.00.enable => hm2_7i90.0.pwmgen.00.enable	#mesa pin enables pwm
net emcmot.00.enable => bldc.0.init       #magnetically home the motor 
net enable-X-control    <= bldc.0.init-done
net enable-X-control => pid.0.enable			#hal pin enables pid

The bldc component is rather busy, on the man page there are several references to personality, what do these pertain to?

All references to "personality" are auto-generated and should be ignored.
(the q, i, a etc flags become a bitmask that becomes "personality" and that controls what pins are created, but the compiler doesn't know that the user shouldn't be setting those and auto-documents them)
The following user(s) said Thank You: JR1050

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

  • JR1050
  • JR1050's Avatar Topic Author
  • Offline
  • Platinum Member
  • Platinum Member
More
21 Dec 2018 22:23 - 22 Dec 2018 01:13 #122735 by JR1050
Some info, progress , issues

It took a significant effort to get info on these motors. Yaskawa motors with an FJ suffix were made for Fuji and Yaskawa USA has little to no info , but.... I do have this
All Yaskawa ac motors are 8 pole
Fuji used incremental encoders , they will be utop i or h and the number after x 10 is the count. As far as I can tell FJ11 is usually a 2000 ppr incremental. All Yaskawa encoders are the same hardware wise, they are programmed to count and absolute/incremental. That was from Yaskawa direct.
Yaskawa aligns the marker pulse with motor zero.

In the video below is the drive running the motor in current mode, it is be run off a watch battery.
My present issue is I'm not getting any analog voltage out of the 7i33, I get pid, as putting some P and D on it stiffened the shaft up. I Could the bldc component be clashing somehow with the pwm out put? Since the marker and motor zero are already aligned, does it still need magnetic homing?


Addendum.... first jogs. Motor cogs some. On the right track.
Last edit: 22 Dec 2018 01:13 by JR1050.

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

More
22 Dec 2018 01:13 #122741 by andypugh

Since the marker and motor zero are already aligned, does it still need magnetic homing?


No, you should use "qi" mode.

But just bear in mind that "motor zero" has more than one definition. A+ B- C- is a common one (and the Mesa one) bit it is not universal.

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

Moderators: PCWjmelson
Time to create page: 0.271 seconds
Powered by Kunena Forum