Fast Hole EDM from Hacked Parts

PCW,

Thank you.

Okay, as best I can tell I am running 2.9.0 (that is what is show at the top of my axis gui and what I see when I use linuxcnc -v).  You reminded me that Peter said something about needing 2.10.. Do you know if I need 2.10?

Can I update linuxCNC? Do I have to wipe the entire machine or can I just update? Will I loose any system configuration data if I do this or will that all be stored in my machine folder? 

Best,
-Dylan

It might be supported by 2.9.3 if not then you must use 2.10

You should be able to upgrade to 2.9.3 with the package manager
if you are running s normal Debian LinuxCNC release

PCW,

Okay, I am trying to start with 2.9.3 but am having some trouble, is this just done through the terminal? Do you happen to know the commands to do it? Do I need to find 2.9.3 and install the package on my machine first? 

If it is not too much trouble I could go straight to 2.10.  But I have found zero info on how to do that..

I am running bookworm.  I know that I tried to download whatever was best and most recommended when I set this machine up. 

Thank you.

Best,
-Dylan

There's some generic information here:

linuxcnc.org/docs/html/getting-started/u...e_to_the_new_version

But it might be better to just ask here on the forum with an appropriate topic, like
 "How do I update LinuxCNC to the latest"

etc



 

PCW, 

I was about to post a reply.  But I did so in a new thread per your suggestion www.forum.linuxcnc.org/38-general-linuxc...-do-i-update-to-2-10

Best,
-Dylan

Hi PCW,

I was able to get 2.10 installed which I have posted about on the other post.  The output of periodm pins is as follows (just copied and pasted and I tried to clean up some):

    38  u32   I/O            0x00000001     hm2_7i96s.0.periodm.00.averages
    38  float OUT               59.1607       hm2_7i96s.0.periodm.00.duty_cycle
    38  float IN                      0               hm2_7i96s.0.periodm.00.duty_cycle_offset
    38  float IN                    100             hm2_7i96s.0.periodm.00.duty_cycle_scale
    38  float IN                      1               hm2_7i96s.0.periodm.00.filtertc_us
    38  float OUT              102354.1      hm2_7i96s.0.periodm.00.frequency
    38  bit   OUT                  TRUE        hm2_7i96s.0.periodm.00.input_status
    38  bit   I/O                 FALSE           hm2_7i96s.0.periodm.00.invert
    38  float IN                      1               hm2_7i96s.0.periodm.00.minimum_frequency
    38  float OUT                  9.77          hm2_7i96s.0.periodm.00.period_us
    38  bit   OUT                  TRUE        hm2_7i96s.0.periodm.00.valid
    38  float OUT                  5.78          hm2_7i96s.0.periodm.00.width_us

I think the pin I am looking for is 'hm2_7i96s.0.periodm.00.duty_cycle'.  This should let me read the duty cycle of the input PWM cycle into pin 29 (which is actually TB2 Pin 13,14 as the naming convention is not complex enough here...).  Do you agree with the assessment? 

Best,
-Dylan

Here is an update.  I have plugged the positive from my EDM power supply into pin 13 TB2 and the negative into pin 14 TB2 on the Mesa. I then evaluated pin hm2_7i96s.0.periodm.00.duty_cycle using hal meter. The number is jumping around like crazy. It got a little better when I remembered to change the jumpers to TTL. Now it fluctuates only on the low end.  

I have tried shielded cable. I have tried a RC filter with a .1 uF capacitor and a 3.3 kOhm resistor.  I tried switching the red and black cables around.  I tried tying the black to ground.  The signal becomes 0 if either the red or black wire is removed from the mesa pin 13,14.  But I cannot get a signal that makes sense.  The signal jumps around even when the EDM supply is turned off and remains unchanged even when the EDM is cutting. 

I currently think that maybe the problem lies in the sampling rate.  Or perhaps it has something to do with the pin.  I think the issue is in the Linux side but I could be wrong.  I will try using the oscilloscope more tomorrow.  

If anyone has ideas please let me know.

Best,
-Dylan

Another update.  It seemed that I didn't have the Mesa tied to ground properly.  It is now tied to earth ground (I also tied the shielding into earth ground). The jittering all went away and now the 'hm2_7i96s.0.periodm.00.duty_cycle' is sitting at 0 even when the EDM is running.  Anyone have any ideas? 

Is 'hm2_7i96s.0.periodm.00.duty_cycle' the correct pin?

Yes, duty cycle is the correct pin.

Is the frequency reasonable and the valid bit set?

It sounds like the EDM output is single ended which means:

1. The EDM duty cycle output should connect to TB2 pin 13 (IDX+)
2. The EDM signal ground should connect to TB2 pin 15 (7I96S ground)
3. The 7I96S IDX input should be jumpered for TTL mode (W1 left)
4. 7I96S IDX- (TB2 pin 14) must be left unconnected

Note the if you connected the inputs to TB2 pins 13 and 14 without a ground
and ran the EDM system you may have damaged the EDM output or
7I96S input.
 

PCW,

I did that, why would it damage anything? Could a high voltage be carried through the ground wire to pin 14? This is the only thing that would make sense to me since the + wire is supposed to only be 3.3v..

Can you please explain what you mean by "Is the frequency reasonable and the valid bit set?"

Best,
-Dylan