7i95T newb advice

So all your modbus config files are related to using this adapter?
I thought I'm gonna use the serial port from Mesa to control the VFD, was there a limitation on why you could not do it that way?

vmihalca wrote:

So all your modbus config files are related to using this adapter?
I thought I'm gonna use the serial port from Mesa to control the VFD, was there a limitation on why you could not do it that way?

It should be possible, but I did not have a looked into it yet

Ok so I have a bit of time off work again, changed my IP to 10.10.10.10 and able to flash my mesa 7i95T with mesact. I was having alot of trouble figuring out how to fill in most of the tabs, and my head is still spinning as to what is going wrong now.

I successfully created a ini and HAL (i think) and when I load up the UI gmoccapy I get hm2/hm2 7i95.0: error finishing read! iter=3324 <--- that number at the end was alot larger before I read about IRQ affinity here ---> forum.linuxcnc.org/38-general-linuxcnc-q...irq-affinity?start=0
The UI AXIS would get an error I can't remember, but does not do it after IRQ affinity steps, but gmoccapy still does sometimes. Does this mean my latency is too high? Even when it is working without an error code, i cannot get any motors to move. I am doing this off an old gaming laptop.

What information should I supply to ask for help better? I only reach out when I start thinking I might not be able to solve this

Ok, so looking back, yes flashing the card isn't necessary. But for context, being very new to this all, and the mesact config tool, I didn't think about how the mesact tool is just there to help you get an .ini and HAL file. Having the Firmware tab there I assumed flashing was a part of the process.

I reached out to mesa to get a swift reply which firmware version was stock 7i95t_d.bin
Many thanks to Peter if you ever read this.

So having that sorted, the latency issue was still there, changed the .ini to 2mill and there was an improvement. I had a thought, and the cat 6 ethernet cable i was using was old, maybe it was damaged, so i tried another newer one, but was cat5e. Both gave less than a minute of the gmoccapy gui runtime. --->Then I thought i'd go buy a new cat6, and bam, no more error messages. Restored the .ini file to 1mill and runtime would range from 5 mins to 10mins before a error message of the same iter.

All throughout this, still no movement of my steppers (the dream, like the first "hello world" when you start coding). I'm starting to question my wiring and can spend a bit of time making some clear wiring illustrations if anyone has a familiar setup. Laptop --> mesa board --> closed loop stepper drivers --> steppers

These are the configs for the exercises I did above. I have not put the time into pairing the dual axis and joints together as nothing was moving, but I should be able to figure that out from other threads. I went through BIOS and disabled any power saving, think I done the IRQ affinity. 

To be proactive about the latency I want to source a desktop. I read in the IRQ affinity thread some components have more compatibility with linuxcnc, ie i7 processors , and network cards, realtek I think? Maybe referring to the code used though. 

What is ideal or known to work well with linuxcnc as good components in a desktop anyone?

Edit: I am keeping notes on what trouble I run into doing this project since I'm a "learn by failing" type of person... would it be helpful once I succeed to envelope this thread into a new one with a summary of the 'trials of a newb' for others to reference?

If you have no motion, a common issue is too short step pulses
For initial testing I would set:

Step Length 5000
Step Space 5000
Dir Setup 20000
Dir Hold 20000

Latency issues:

Make sure you have all power saving, C1 states >C1, EIST, Turbo modes etc disabled in the BIOS.

If you have a Intel MAC make sure you have IRQ coalescing disabled (hm2_eth manual page)
(what does "lspci | grep -i eth" report)

Finally, using IsolCPUs  and pinning the Ethernet IRQ to the isolated CPU makes a big difference


 

Tried those steps still no movement.

Since this is my first setup, I havn't been able to find a wiring guide for going to a mesa control board to a closed loop stepper driver. Here's what I got

Control board  -->  Closed loop stepper
GND     unconnected
Step0- --> Pul-
Step0+ --> Pul+
DIR0- --> Dir-
DIR0+ --> Dir+
+5VP unconnected

It can't be right to leave the gnd and +5vp disconnected and could only find vids of people with mach 3 boards with 1 step and 1 dir output, then wire the +5vp output from the mach3 controller to the stepper driver  Dir+ and loop that to Pul/step+. Is that what i'm supposed to be doing? Lol i've spent days reading and too scared to try for fear of blowing up my board

Edit:The driver is optically isolated, high level 3.3-24v in and low 0-0.5v. www.omc-stepperonline.com/closed-loop-st...pper-motor-cl57t-v41
 

For differential drive, 5VP and GND are unconnected.

Can you jog  one axis in both directions and measure
(with a voltmeter) the direction signal for that axis
to make sure that its changing based on the jog direction?

I got hold of a cheap but newish Dell optiplex 7070 today. massive improvement on latency.
:~$ lspci | grep -i eth <--returns
00:1f.6 Ethernet controller: Intel Corporation Ethernet Connection (7) I219-LM (rev 10)

Latency test is Max Jitter (ns) around 50,000 (it was 500k before).
Tried the IRQ affinity again but something not quiet right with the commands i put into terminal. I'll have another go at the later, AXIS and other GUI's do not crash.
irqbalance is not running OR there are only 2 cpus - setting up IRQ affinity script

Put the settings in for the steppers again along with everything else and tested with volt meter over every positive/negative terminal and power supp/gnd - while jogging the corresponding axis in the GUI AXIS. All read 5v and a few less and from the power supply to drivers 42v. None changed while jogging except for Direction+/- on the driver and control board.
So where the control board and wires into the stepper driver at the Direction +/- will read 5v while jogging + on AXIS GUI and read -5v while jogging - on AXIS GUI.

Hope that makes sense. I'm actually happy it showed some feedback despite where I am. Any help is a God send.