7i76E + THCad 10V

I would trace the encoder velocity signal with halmeter
if its always 0 you likely have a connection issue on the encoder inputs
Note that you may have to set the B input to the proper state or the velocity will be negative

Finally get it working. My issue were connected to wrong switch on THCAD board - it's been set to unipolar encoder signal but on 7i76e were set to bipolar. Now seems it work but I have weird behavior of THCAD.

Initially I've used battery cell 3V to calibrate the velocity scale and set correct values for 'thc' module parameters. Also to see how the system works against the stable voltage source. The problems start appear right from the beginning: when using 1:1 default divider halmeter showed me that the signal from ecoder.velocity pin is jamping like a random in range from 0 to about 200000+ HZ. It looked like the encoder on mesa board is out of time to count the frequency in correct way (which I don't beleave because it's hardware encoder). Then I've switched of the reference voltage to measure frequency for 0V, and it given me 0Hz??? (according to sticker should be 120000 Hz). Really strange...

Next: I've used 1:32 divider to see if encoder interface will cope with the divided value and find that the frequency counted more or less stable. But it still jamping between 23785 and 23809 Hz. Another surprise was to see the frequency of 0V input, which was not 120000 / 32 = 3750Hz but 16000+??? another surprise

So my general impression the THCAD board seems to work but totally different way that it explained in documentation, and I have feeling that the board of mine is probably either misconfigured or broken. Guys do you have any similar issues with it and suggestions how to get it work? Could you recommend me contact from MESA to fix the issues I have?

The unipolar/bipolar jumper on the THCAD has nothing to do with the encoder interface
it determines if the input voltage range is bipolar (-5 --> +5 ) or unipolar ( 0--> 10 )
(you want unipolar)

If the THCAD is jumpered for bipolar mode you would expect around 16 KHz at 0V
when the 1/32 output frequency option is selected

The reason you cannot count the output frequency correctly at 1:1 is likely that you have the input filter enabled
on the encoder so it cannot count single 1 usec pulses (you can turn this off in hal)

EDIT pulse width at 1:1 is approximately 0.5 usec, too fast for the encoder if the filter option is enabled
in the hal file

Thanks for the prompt reply,

It seems strange because I have connected to Hypertherm PM65 remote control cable, and according to Hypertherm it's unipolar...
anway I'll investigate this more deep..

Now I'd like to increase the precision of the counts, so would like to clarify where the filter settings appeared in my config.
From the first look I definitelly have no such explicit setting in the configs, but it might be caused by other settings I believe. Could you check my config if I have them to get rid and try to use 1:1 divider?

at first my servo period is:
SERVO_PERIOD = 1000000

hostmot 2 driver config:
loadrt hm2_eth board_ip="10.10.10.10" config="firmware=hm2/7i76/7i76e.BIT num_encoders=1 num_stepgens=3 sserial_port_0=20xxxx"

then I have couple lines which was in scratch file and I have just left them there:

setp hm2_7i76e.0.watchdog.timeout_ns 5000000
...
# latch 50 usec before nominal read time:
setp hm2_7i76e.0.dpll.01.timer-us -50
setp hm2_7i76e.0.stepgen.timer-number 1

and the only lines related to the encoder config are:
# Set up the Encoder for the THC
setp hm2_[HOSTMOT2](BOARD).0.encoder.00.counter-mode 1

full config files are in the attachment

If you want to improve the resolution, you can install the lowpass filter component
and filter the encoder velocity output, this will not improve the precision however.


If you want to count at 1:1 you will have to turn off the encoder noise filter,
but I suggest you do not do this but use the 1:32 ratio, for two reasons:

1. The resolution will remain the same (about 10 bits at 1 KHz)
2. You will have improved noise immunity by leaving the encoder noise filter on

well, it explains a lot of "magic" parameters in my config file )))
I'd like to clarify if the .counter-mode=1 (step/dir mode) is the best to work with the TCH card?

thank you in advance,
Oleg

The standard encoder counter only has 2 counter modes: quadrature and step/dir

only step/dir (counter mode 1) will work with the THCAD

According to the manual^

"THCAD outputs a 1 MHz reference frequency that is synchronous with the
output pulse train. This may be used as a frequency counters gate reference source. The
advantage of using the THCADs reference frequency as a gate source rather than a local
crystal clock is that it is synchronous with THCADs frequency output and therefore
eliminates the +- 1 count uncertainty that comes from using a asynchronous gate."

Is it possible to use this FREF signal with the 7i76e board?

regards,
Oleg

No, its not possible currently