Go Down

Topic: welder control (Read 1 time) previous topic - next topic

joeaverage

i want to use a uno to control output of a 3 phase scr contolled welder.

i see two different aspects to the problem
1) read user input (required output current, mode)
    read feedback (output voltage and current)
    compare input and feedback to produce a variable that controls scr conduction angle
2)accept 3 inputs of 50hz as timing marks
   calculate delay required from variable above
   using timer or counter pulse 3 outputs to scr's

reading user input need only ocurr 10/sec, feedback 100/sec but scr pulses 300hz

the 'c' reference dosen't show direct access to timers/counters.

can i write assembler routines within 'c' sketches that i might be assured that the scr outputs
run without interuption/delay?

Grumpy_Mike

I did build something very similar years ago for controlling a plaster discharge in a mass spectrometer. Unfortunately it was not very successful, basically controlling the phase angle over every cycle was not fast enough for the discharge. With an SCR you only get one chance every 1/25 th of a second to do anything, once it it on, it is on for the rest of the half cycle.

joeaverage

thanks for your reply mike

if i run the micro just as a firing board ie a second board for input/feedback
do you think it might work.

i was planning to use a transformer type blocking oscillator to give multiple
repeat trigger pulses to each scr, something like 1us pulse repeated every
10-20 us. thus the output pin need only go high after the appropriate delay
then reset low after some 10ms or so. i dont think the scr's will mind  after all
they must commutate when they get reverse biased.

one thought i had was to set three interupts, one per phase,on rising zero crossing
to reset phase angle delay counter and phase no. flag. would probably use the same event to reset the
previously fired scr output, ie no wasted calculation time.

Grumpy_Mike

I don't think you need a seprate board if you want to do this. But once an SCR is on it is on for the rest of the cycle. Therefore all the feedback in the world is not going to limit the current. That was the problem I had.

joeaverage

again thanks for your coment mike.

this project is for an existing welder having worked hard for 20+ years
using analogue boards to control the scr's.indeed most of the major
welder manufacturers have similar units, a huge transformer and 3 phase
conduction angle controlled bridge using puck scr's, this particular unit 400A
output.

this machine has a faulty board. 50Hz boards are no longer available so my
options are:
1)repair existing boards with only sketchy schematics
2)fit still available 60Hz boards and try to retime the delay circuits
3)roll my own design using a microcontroller
4)write the machine off as uneconomical

a new machine in the region $8000 NZ.this price gives me a little scope to explore!

joeaverage

should add a few design aims
1)i plan to use 2 V to F ics, one to monitor the oem 50mV shunt and the other to moitor
   welder output voltage, about 80V open circuit.optiacl signalling back to controller.weld
   output must be isolated from controller!
2)i will retain the pot for output control as this machine is used in submerged arc welding
   where the remote wirefeeder unit controls output. existing remote control would therfore
   still work.
3)i will retain the existing mode switch (constant voltage for MIG welding, constant current for
   manual arc and gouging, and lift start TIG welding)
4)if capacity remains in the controller i will incorperate the temp sensors into the controller
   otherwise will retain oem scheme where over temp drops the main 400V input contactor

Go Up