Advanced Search

Hi All, 

I'm trying to add a 7i78 and 7i85 card to P1 and P2 of my 7i76e. 
When I plug the 7i78 into P1 and mesaflash --readhmid, the usual 7i76e pins show up "on card" and I also get a list of the 7i78 pins on the "IO Connections for P1" list, an encoder and some stepgens, so that looks positive.

However I connected a bunch of test nets in my hal file, something like this but I'm not by the machine and the SSH isn't connecting now so I can't check: 
net test017 <= hm2_7i76e.0.7i76.0.0.input-17
net test018 <= hm2_7i76e.0.7i76.0.0.input-18
net test019 <= hm2_7i76e.0.7i76.0.0.input-19
net test020 <= hm2_7i76e.0.7i76.0.0.input-20

all of my test nets stay low whether or not I apply 5v to the pins. 
Do I need to change something in the FPGA firmware (bitfile?) to define these pin directions? If so, is this a matter of downloading the Xilinx tool or is there a repo somewhere with combinations of mesa bitfiles in there? I've seen PCW on here post them but I'm not sure if they live somewhere! 
Do the IOPort numbers from --readhmid directly map to the hm2_7i76e.0.7i76.0.0.input-nn pins in the hal or am I missing something here? 


When the 7i85 arrives, I'd like to get it reading a glass scale to close the loop on the axes (I could also test this sooner with the encoder input on the 7i78 if I get the scales mounted before the 7i85 shows up, depends on DHL!) presumably the standard firmware shipped in the 7i85 is set up for the right encoder IO so I only need to set up the hal file plug it into the 7i76e, or will that require an update to the 7i76e bin? 

Finally, once the axis position feedback is working and tuned I'd like to get index homing working from the glass scales on top of my existing home switches. From seeing posts on here It looks like i'll need a new bitfile to enable the home-seeking with a limit switch input and home-latch off an index, but the homing config page seems to say I just need to set HOME_USE_INDEX in the .ini ? 

I tried to install tigervnc and it destroyed my system after logingb in with the correct password it just keeps taking me back to the login.So I reinstalled linuxcnc und updated it.now I can not connect to the mesa Ethernet card 7i76Ei get the following error. 

 

Did I do something wrong during installation?

I would make some changes in the hal file.

Change:

setp hm2_7i76e.0.dpll.01.timer-us -50

to:

setp hm2_7i76e.0.dpll.01.timer-us -200

and change all joint PID "P" terms to 666



 

Networking is not set up
forum.linuxcnc.org/27-driver-boards/3559...ethernet-mesa-boards

How did this work out? I also have a 7i76e and are sick on my USB-RS485 converter that resets itself sometimes in the middle of a program... Might be interference from my HY VFD. Want something more robust.

pktUART bin file is not available from MESA. Do not really know if this is the way to go or if there are something simpler?

You are welcomed, always, and welcome to the "dark side" !
Used them once, worked MUCH better than advertised, got unbelievably good support from PCW, used them ever since, a lot.

Thanks for confirming this, I haven't used multiple Mesa boards at once before. What a great system.

Yes, you can use 7i76E with 7i84 with a cut in half LAN cable (see manual for wiring), and 7i85 with a IDC26 to DB25 cable (available on Mesa stores and even some older PC's have them, look for a parallel port on the back slots). BTW, 2 of 7i85 could be used as 7i76E has two IDC26 headers, or similar boards, and you can add more 7i84 or 7i79 or 7i71 or 7i72 to 7i85.

Is it possible to run all these things together: 

Mesa 7i76 E  ---- 7I84 (Isolated remote field I/O card). Will use it to output quadrature signals to the Ultra3000s, it will generate this quadrature output from the Fanuc encoders [is this possible?], remaining I/O used elsewhere. 
                               
                         ---- 7I85 (4 Channel encoder 5 channel Serial RS-422 interface). Will hook up the 4 Fanuc redcaps to 4 of these RS-422 ports. And the encoder inputs are good to have around incase I want to add linear scales (2 axis) and maybe two MPG wheels. 

Sure would be great if this system worked because it happens to be two fairly inexpensive boards. Canadian dollar isn't so great right now 
I would be pleased to have those encoder inlets, too. 

Progress pic from a little while ago: 
 

I'm under the impression that the encoders are 1 million counts per rev. I did a crude test where I rotated the pully associated with the motor one full turn as marked by a sharpie line on the face of the pulley. Seemed to deliver around 2 million raw counts, and the pulley appears like it could be 1:2 ratio. 



I also observed 4 cycles of the commutation output from the HAL Fanuc Encoder pin for one rotation (if I remember correctly!). 

These computer-side details can be adjusted as I setup the system. Now I need to figure out what extra hardware I need to talk to 4 of these at once, and also how to output either commutation to the Allen-Bradley Ultra-3000 drives or incremental encoder signals. I'm fully out of inputs and outputs on my 7i76e used on the machine so I will need to get some expansion hardware. 

Expansion needs: 
-4x serial ports (maybe just 3 if I can also use the one in the 7i76e, but Mesa seems to sell them in breakout boards of 4 or 6). 
-8 or more outputs to communicate encoder or commutation information to the Ultra3000 (still figuring out what the drives will need) 

Wondering what the most cost effective Mesa solution would be to add onto my system. 

Or maybe someone knows if there is there setup that does not require sending encoder or commutation information to the Ultra3000s. Can't seem to find whether or not these drives can be run 'open-loop' or/ with feedback only being taken care of on the LCNC side (doubtful?). 
 

I have not found any formal documentation on the encoders or any note in the machine’s schematics, so I don’t know.

You did say in a previous response that you thought my encoders are 1M units per revolution, in response to a post where I provided the model number.

Tomorrow I will do more testing - I will try to verify one complete rotation and see how the counts check out.

What is the counts per turn value for the encoder?

Back from holidays, finally got to dig a little deeper with the oscilloscope. I managed to figure out the polarity of the connections and alas, I am now receiving the response from the encoder!

Some values are showing up on my HAL Show and now I need to figure out if everything is working correctly and whether or not I need to create a custom setup string.

After a slight turn (a few degrees) of the motor shaft in the reverse direction my HAL show reads:

 

Suggestions for next steps are encouraged.

fully_defined wrote:

spumco wrote:

fully_defined wrote:

 
"Help" is a strong word for what's happening here.

You seem to be getting frustrated.  Maybe I can help.

Correct me if I'm wrong, but your situation is that you have a Mesa 7i96s and Clearpath SDSK motors.  You want to:

• Set some rational values for max velocity and accel
• Connect the HLFB pins to the 7i96s
• Make LCNC do something with the HLFB signals
  
  
  Have I got it right?
  
  
  
  
  Yes, and thank you for engaging.
  
  Those points are just what are directly in front of me, right now. It's mostly because I am at the pncconf stage, and there isn't a selection for setting the pins for drive fault there. I am not hung up on moving on, as long as I know where I need to set this function. I have LinuxCNC forum people telling me they have no idea how Clearpath HLFB works and Teknic saying they have no idea how LinuxCNC works, and here I am, stuck in the middle. That's why I thought maybe seeing an existing configuration would skip the middlemen and I could just copy someone else's homework.
  
  Beyond that, I also need to know what to expect when homing the dual-Y gantry axis. There is more I'd like to learn about autosquaring too, but that's just one thing out of a long list.

Some of these features will likely not be supported by pnccconf

It seems this post gives you some options for HLFB. 
www.forum.linuxcnc.org/10-advanced-confi...with-hlfb-mesa-7i76e
I think I would use servo-on and use  a charge pump which I think pncconf supports

The drive error is connected to 
joint.N.amp-fault-in 
ref: linuxcnc.org/docs/stable/html/man/man9/m....9.html#JOINT%20PINS

Squaring the gantry is dealt with when homing
linuxcnc.org/docs/stable/html/config/ini-homing.html
Paired joints (eg your gantry motors) will have a negative HOME_SEQUENCE of the same value
Usually you will also use a HOME_OFFSET
You can  square the gantry by editing the HOME_OFFSETS. eg  if one side is out by 2mm, the HOME_OFFSET needs be 2mm different to the other side.


 

OK, went through those, and i have to agree with PCW on number 2 here
forum.linuxcnc.org/10-advanced-configura...fb-mesa-7i76e#112557
It is pretty much useless as it does a lot of things but the thing it should do, it does not, namely amplifier fault!
I would not use it.