Problems with multi-controller addressing

[kmpeters]

@se-hzb reported problems polling and moving axes when the controllers are daisy-chained:

Operation with two 8742 controllers with firmware v2.3 05/02/17 through the RS485 interface. Polling results in occasional wrong assignments of positions to the individual axes and move commands are executed by both controllers.

This is fixed by always also directly addressing the first controller by changing the lines 235 to 239 in newFocusApp/src/874xMotorDriver.cpp:

if (axisNo > axesPerController) {
      sprintf(axisName_, "%d>%d", (axisNo - 1)/axesPerController + 1, (axisNo - 1)%axesPerController + 1);
  } else {
      sprintf(axisName_, "%d", axisNo);
  }

to

sprintf(axisName_, "%d>%d", (axisNo - 1)/axesPerController + 1, (axisNo - 1)%axesPerController + 1);

If this is compatible with the other interfaces USB/Ethernet and doesnt affect any other functionality of the driver, I would recommend this operation.

source:
#9 (comment)

[kmpeters]

In the description for PR #13 @pkienzle said this:

This assumes the controllers have already been initialized with unique controller addresses numbered from 1 to the number of controllers. This can done sending "SC2" command to the controller, or by running Scan from the NewFocus configuration program on Windows.

@se-hzb, can you confirm that the controllers have been assigned unique addresses? It sounds like both controllers might have the same address: 1

[se-hzb]

The devices were set up using the Windows tool, which reassigns addresses to avoid conflicts. I get the following output for identification and controller address queries:

1>*IDN?
1>New_Focus 8742 v2.3 05/02/17 <SerialNumber1>
1>SA?
1
2>*IDN?
2>New_Focus 8742 v2.3 05/02/17 <SerialNumber2>
2>SA?
2

Monitoring one of the axes on controller 1, it gets assigned the value of the corresponding axis of controller 2 randomly, without anyone issuing motion commands:

PREFIX:Y.RBV 2026-04-23 09:49:30.572988 1400 STATE MAJOR
PREFIX:Y.RBV 2026-04-23 09:52:32.406279 0  
PREFIX:Y.RBV 2026-04-23 09:52:35.487029 1400  
PREFIX:Y.RBV 2026-04-23 09:52:44.577119 1400 STATE MAJOR

[kmpeters]

I think the "STATE MAJOR" messages in the camonitor output indicate that there are communication errors with the controller. Do you see communication errors on the IOC's shell?

Another user of motorNewFocus saw problems caused by the controller being slow to respond after some commands, which resulted in a misalignment of the replies parsed by the driver:

#16

The workaround was to add a small delay for write-only commands. Can you confirm the problem still occurs with the HEAD commit of the master branch of motorNewFocus?

[se-hzb]

The problem persists with the current HEAD commit.

I receive errors during poll causing the STATE MAJOR message.
2026/04/24 10:30:36.701 nf874xMotorDriver:poll: error, comStatus=1

With the following change in line 247 in newFocusApp/src/874xMotorDriver.cpp, operation works without errors. This change would maintain operation without addressing for cases without daisy chaining:
if (pC_->numAxes_ > axesPerController) {

[kmpeters]

Can you enable asyn traces by changing the following lines in the IOC's config:

asynSetTraceMask("$(PORT)_ETH",-1,0x1)
asynSetTraceIOMask("$(PORT)_ETH",-1,0x1)

To this:

asynSetTraceMask("$(PORT)_ETH",-1,ERROR|DRIVER)
asynSetTraceIOMask("$(PORT)_ETH",-1,ASCII)

I'd like to see which commands are causing the asynTimeout (comStatus=1).

[se-hzb]

It seems to be the TP? and MD? commands

2026/04/24 17:05:27.898 <IP>:<PORT> COM write 6
1 TP?
2026/04/24 17:05:27.903 <IP>:<PORT> COM read 4
_�8�
2026/04/24 17:05:29.905 nf874xMotorDriver:poll: error, comStatus=1

2026/04/24 17:05:23.645 <IP>:<PORT> COM write 6
3 MD?
2026/04/24 17:05:23.650 <IP>:<PORT> COM read 3
���
2026/04/24 17:05:25.650 nf874xMotorDriver:poll: error, comStatus=1

Here is the ESCAPE mask, but the replied bytes are not the same each time the error occurs:

2026/04/24 17:26:32.354 <IP>:<PORT> COM write 6
1 TP?\r
2026/04/24 17:26:32.360 <IP>:<PORT> COM read 4
W\xfe\xd6\n
2026/04/24 17:26:34.362 nf874xMotorDriver:poll: error, comStatus=1

2026/04/24 17:26:25.807 <IP>:<PORT> COM write 6
4 MD?\r
2026/04/24 17:26:25.813 <IP>:<PORT> COM read 3
\xee1\xf8
2026/04/24 17:26:27.814 nf874xMotorDriver:poll: error, comStatus=1

[kmpeters]

With the following change in line 247 in newFocusApp/src/874xMotorDriver.cpp, operation works without errors. This change would maintain operation without addressing for cases without daisy chaining:
if (pC_->numAxes_ > axesPerController) {

I created #19 with this change. If it looks good to you, I'll merge it.

[se-hzb]

I checked out this branch and it runs fine for me. Looks good! Thanks!