7i77 encoder split to amp and servo

Hi,

I have an encoder on a servo motor that is being read in single ended at the servo drive (A, B, velocity feedback) and want to use all the channels on the mesa 7i77 side (A, /A, B, /B, Z, /Z).  The 7i77 is sourcing the +5v going to the encoder and +5v  is not connected at the drive. Is it okay to do so or should the 7i77 also be reading in single ended? In other words, is it okay to mix open ended with differential?
For testing, if the connector has all the wiring in place, can I just move the 3 jumpers to go into single ended mode on the 7i77 or do  I need to disconnect the /A, /B, /Z wires?

TIA

The thing I would worry about is that you would lower the single ended swing
considerably because of the differential mode termination.

Thanks for the response PCW. I will try single ended as tuning has been very difficult. 
Do I need to pull the inverse wires or can I leave them in place and just switch the jumpers?

Besides noise issue on my other post, I'm trying to see if encoder is working correctly. Initial encoder counter mode 0 resulted in rough running motor. I switched to mode 1 and is smoother. A and B channel look to have closer to 90 quadrature shift than in mode 0 which had the rising edges almost in line. Is that possible? If incremental encoder outputs step/dir, (mode 1) wouldn't the direction channel be flat?

Tuning questions. I'm running the analog drive in velocity hall feedback and not reading in the encoder to the drive. Drive was configured internally to match best motor response.
Hall velocity mode in drive and pid tuning in linuxcnc does not follow regular encoder velocity tuning with I. It seems to like high P term. Why?
Not really grasping the difference between velocity tuning and torque/current tuning as linuxcnc still reads in a velocity from the encoder regardless of how the drive is set up.

Last, in hal scope, f-error plot should follow vel-commanded. Correct? Mine is mirrored. Can I flip the channel view somehow or is something misconfigured?

 

The hostmot2 encoder can accept step/dir or quadrature
I would expect quadrature to be more standard. Step/dir
output may have a constant value on the direction pin but
it's not necessary as it's only important that the direction pin
be in the correct state on the rising edge of the "A" input

A true velocity mode drive need encoder feedback to the drive.
Without this feedback you need higher D and P terms in a single
control loop.

The following error direction relative to velocity is dependent on the FF1 term
Though if you have a pure torque mode drive you normally do not need any FF1

Wouldn't it be wise to use something like the Encoder Y for this? I assume Mesa still sells it, I've had one for years and never got to use it.

IIRC it has a buffer/redriver......might help with some of those other issues.

blazini36,

Thanks for pointing out the Encoder Y! Of course Mesa makes a solution for everything!

PCW, will the encoder y board offer isolation? Not sharing the grounds? Could I read in differential from the encoder and split(Y) to drive single ended without lowering the voltage swing? Sorry, I could not find the product doc on your site. Do you have a link?

I need to further scope the encoder as the A and B channels look flaky on halscope. I have an inconsistent pulse width both in mode1 and 0.
I also enabled  quad-error-enable true but it didn't trigger. I assume in 2.8 version, axis gui would display a warning? Or does the error checking need further hal connection?

The EncY is not isolated by will function as a re-driver
to lower loads from termination. We have considered an
isolated version but have not made it yet.

You have to monitor the encoder error bit youself
and it requires using 2.9 or later as there were errors
in handling errors in multiple channels before that.

OT-CNC wrote:

blazini36,

Thanks for pointing out the Encoder Y! Of course Mesa makes a solution for everything!

PCW, will the encoder y board offer isolation? Not sharing the grounds? Could I read in differential from the encoder and split(Y) to drive single ended without lowering the voltage swing? Sorry, I could not find the product doc on your site. Do you have a link?

I need to further scope the encoder as the A and B channels look flaky on halscope. I have an inconsistent pulse width both in mode1 and 0.
I also enabled  quad-error-enable true but it didn't trigger. I assume in 2.8 version, axis gui would display a warning? Or does the error checking need further hal connection?
 

I don't fully understand your setup but i've never considered sharing a single encoder between 2 devices as a particularly good idea, it's probably espescially bad if one is single ended as the load on the inverting and non-inverting outputs will be different besides however it affects RS422 termination.

If you actually need isolation on the encoder you might try looking for a magnetic or "digital" isolator. I think they'd be fast enough to handle differential inputs directly, pd is <8ns. I whipped up a few in case I ever needed them, don't think I've bothered to test them for that. I use the same family of isolator for a bunch of serial stuff since they work as simple 3v-5v level shifters as well.