Un-lobotomizing a Maho MH600T

Nice work, that cabinet looks really tidy!

Given how large it is I can imagine that you must have put some serious effort into it.

Thank you. Yes, it's been a fair bit of work with planning, creating the schematics and of course wiring everything up as well. I will post the updated schematics are well once I'm a bit further along.

LS703 scales are 20um 11uApp and the interpolation is set to 5x
what ENCODER_SCALE must enter to ini config 1000 ?
20um/5um=4um / 4 for quadrature =1um -> 1000 ENCODER_SCALE

Yes, that is correct and what I used in my tests.

I have tested the pcbs and work fine except they have burned 3 times 0.5A fuses
I have 3 pcbs in mill and all times fuses burned in all pcbs.
I have replaced with 3A fuses and no problem.
Your scales work with 0.5A fuses ?
Because i didn't found tic116 scr i have used tic106 also not populated quad differential transmiter
www.farnell.com/datasheets/39877.pdf
www.farnell.com/datasheets/63251.pdf

tic106 is more sensitive so maybe this is the reason for burning fuses because conducts more easily and shorts fuses.

My apologies for the tardiness in the reply. Cool that you are using these as well! Did you order PCB:s based on the manufacturing files? And do you use normal terminal blocks or the ethernet jacks?

I designed the power supply with a crowbar circuit, because I wanted to have decent protection for the bulb in the scales against overvoltage. The overvoltage "margins" are intentionally low, meaning that the crowbar may trigger already at a voltage slightly above 5.0V. Maybe that's what happening for you? But it really should blow the 3 amp fuse as well, if that is the case. Are you able to measure the current consumption in normal operation (with the 3 A fuse)? And the voltage between pin 3 and 1 on the TIC1x6 may also provide some clues.

I have been using my scales with a 200 mA fuses. That should be enough, as the bulb in the scale is the main "consumer" with a nominal current of 100 mA.

My voltage is exact 5V.
Can some small leakage to zener 5v1 trigger the tic106 that is more sensitive than tic116 and short the fuses ?
If change C2 to 1uf and R2 to 470R maybe is better(is more difficult to trigger tic106)?

Yes, it's possible. The TIC116 needs approximately 100 times higher gate current than TIC106 to trigger. The 1000 ohm resistance to ground means that a leak current of at least 0.5 mA is required to reach approximately 0.5 V for triggering the SCR. And with TIC106, only 0.060 mA "more" is needed to trigger, while for the TIC116 it's about 5 mA more.

But I still don't understand why a 3 A fuse would help. Unless maybe it's more than you power supply can handle, so the voltage drops to zero, the SCR unlatches, the power supply comes back up and the cycle repeats. But in that case, I would imagine that not much would work.

You could disconnect one leg of the zener diode and check with a 500 mA fuse again. Then we at least know if it's the SCR/zener combination that is causing the issue or something else.

I have done again the circuit without supplying ic-nv and the linear scale
just only the protection circuit and does not burn the fuse it draws 0.00A from 5v psu so i suppose that the scales draw more than 0.5A current..
Other reason maybe is the electrical noise from servo-inverters
that triggers the tic106

Tests with 5v supply:
5v1 zener 1k resistor -> 351mV
5v1 zener 470R resistor -> 240mV
5v6 zener 1k resistor -> 101mV
5v6 zener 470R resistor -> 55mV

So because tic106 is more sensitive i will keep 5v6 zener 470R resistor...