Setting up Dual 5i20's,7i49, 7i33 and7i44 setup

Hello to all,

My latest project is a Brown and Sharpe 1000vc mill from the 80's,it will be retaining the resolvers on X,Y and Z ,while having encoders for the A axis ,spindle and MPG. Im using two 5i20's ,a 7149, a 7i33 and a 7144 tied to a 7i84 for my control pendent, these will be on one of the 5i20,the second 5i20 will be solely for driving opto racks. I have some questions as how to set up the dual fpga cards

1 where do resolvers end and encoders start , X is axis 0,Y 1 and Z 2 on the 7i49 and identified the same in my hal file, next is my A axis encoder,is this encoder 1 or encoder 3 ?

2 How will I physically know which 5i20 is board 0 for axis feed back and one for gpio, ?

3 How exactly is the resolver scale set and does this negate the axis input scale on the ini file?

My attached hal file is a very ruff draft. Im still waiting on the hardware, so Im trying to get a jump start. Ill post some pics of the project in the machine section, its gonna be a beast....Thanks for your help!!

Theres one sizable fly-in-the-ointment here:
The 7I49 will not work with the 5I20

This is because the 5I20s FPGA is limited in BlockRAM
and has no built in multipliers so the resolver interface
firmware does not fit.

(with its 32 bit processor, lots of RAM and hardware multiplier)

Ok, so I need a 5i?

5I22,5I23,7I43,(soon 5I24)
If at least one of the 5I20s are not opened/used,
I would return it and ask Mesa to swap with a 5I23

Edit: I should mention that you can do all this with one PCI card by
using a sserial remote (7I69) to drive the OPTO-22 I/O, one 7I69 will
drive 2 x 24 I/O OPTO22 racks

I will do that Tuesday, as I havent received the hardware yet. How about the set up on the resolvers and encoders ? Im a little wary about using the sserial for machine i/o on such a big machine( this thing is HUGE) right now. I also need 3 24 space opto racks, this beast has a relatively complicated tool changer which eats up alot of i/o, a 2 speed geared transmission that has to be shifted in one spot only and a few other eccentric qualities. Im impressed with the 7i84 for a pendant board, as it does what I ask of it quickly. Off subject ,but how much is the 7i77 galvanic isolator board/thingie. I been fighting a noise problem on a machine with a 7i77,might solve it..

7I77ISOL is $59.00

What noise issue do you have?

edit:
For encoders+resolvers the encoder/resolver/PWM channels will not match since you will have 10 PWM gens (numbered 0 through 9) but 4 encoder channels (0 through 3) and 6 resolver channels (numbered 0 through 5) so if the 7I33 PWMs are first they will match the encoder channel numbers and the resolver PWM 4 will be 7I49 AOUT channel 0. Like wise if I put the resolvers first, the 7I33 encoder/analog outs wont match (7I33 AOUT 0 will be PWM channel 6)

Below is an excerpt from my hal file. Ill swap the first 5i20 for a 5i23. The 7i49 goes in P1,the 7i33 in p2 and the 7i44 goes in P3. Do I understand this correctly? The Z axis is the end of the resolvers and the A axis is the beginning of the encoders.

# ################
#z axis
# ################

# axis enable chain
newsig emcmot.02.enable bit
sets emcmot.02.enable true
net emcmot.02.enable => pid.2.enable
net emcmot.02.enable => hm2_5i23.0.pwmgen.02.enable
net emcmot.02.enable <= axis.2.amp-enable-out

###########

##Resolver feed back###

setp hm2_5i23.0.resolver.02.velocity-scale 1
setp hm2_5i23.0.resolver.02.scale [AXIS_2]RESOLVER_SCALE
net motor.02.rawcounts = hm2_5i23.0.resolver.02.rawcounts

# add pid.x.feedback

net motor.02.pos-fb hm2_5i23.0.resolver.02.position = pid.0.feedback
net motor.02.pos-fb = axis.02.motor-pos-fb #push copy back to Axis GUI
net motor.02.veloc = hm2_5i23.0.resolver.02.velocity
net x_index_en axis.02.index-enable = hm2_5i23.0.resolver.02.index-enable


# set PID loop gains from inifile
setp pid.2.Pgain [AXIS_2]P
setp pid.2.Igain [AXIS_2]I
setp pid.2.Dgain [AXIS_2]D
setp pid.2.bias [AXIS_2]BIAS
setp pid.2.FF0 [AXIS_2]FF0
setp pid.2.FF1 [AXIS_2]FF1
setp pid.2.FF2 [AXIS_2]FF2
setp pid.2.deadband [AXIS_2]DEADBAND
setp pid.2.maxoutput [AXIS_2]MAX_VELOCITY

# position command signals
setp hm2_5i23.0.pwmgen.02.output-type 1 #pwm on pin1, dir on pin2
setp hm2_5i23.0.pwmgen.02.scale [AXIS_2]OUTPUT_SCALE

net emcmot.02.pos-cmd axis.2.motor-pos-cmd => pid.2.command
net motor.02.command pid.2.output => hm2_5i23.0.pwmgen.02.value




#############
a axis
###########

# axis enable chain
newsig emcmot.03.enable bit
sets emcmot.03.enable true
net emcmot.03.enable => pid.06.enable #hal pin enables pid-true by poiting at emot.00.enable
net emcmot.03.enable => hm2_5i23.0.pwmgen.06.enable #mesa pin enables pwm
net emcmot.03.enable <= axis.06.amp-enable-out #hal pin enbles the amp

# encoder feedback
setp hm2_5i23.0.encoder.03.counter-mode 0 #mesa parm,set to 0 for quad encdr
setp hm2_5i23.0.encoder.03.filter 1 #mesa parm to time pulses
setp hm2_5i23.0.encoder.03.index-invert 1 #mesa parm to use riseing edge of channel z
setp hm2_5i23.0.encoder.03.index-mask 0 #mesa parm to use index input
setp hm2_5i23.0.encoder.03.index-mask-invert 0

setp hm2_5i23.0.encoder.03.scale [AXIS_3]INPUT_SCALE
net motor.03.pos-fb hm2_5i23.0.encoder.03.position => pid.0.feedback #hal reads the encoder
net motor.03.pos-fb => axis.03.motor-pos-fb #push copy back to Axis GUI

# set PID loop gains from inifile
setp pid.03.Pgain [AXIS_3]P #these are hal parm thta are acessed thru the .ini file
setp pid.03.Igain [AXIS_3]I
setp pid.03.Dgain [AXIS_3]D
setp pid.03.bias [AXIS_3]BIAS
setp pid.03.FF0 [AXIS_3]FF0
setp pid.03.FF1 [AXIS_3]FF1
setp pid.03.FF2 [AXIS_3]FF2
#setp pid.03.deadband[AXIS_0]DEADBAND
setp pid.03.maxoutput [AXIS_0]MAX_VELOCITY

# position command signals
setp hm2_5i23.0.pwmgen.06.output-type 1 #pwm on pin1, dir on pin2 #mesa parm
setp hm2_5i23.0.pwmgen.06.scale [AXIS_0]OUTPUT_SCALE #mesa parm

net emcmot.03.pos-cmd axis.03.motor-pos-cmd = pid.06.command
net motor.03.command pid.03.output = hm2_5i23.0.pwmgen.06.value #hal pin that connects the ##pwm command value

How is the resolver scale calculated? I get the encoder thing, but resolvers dont have counts per turn?




I may be hijacking my own thread with the 7i77 noise thing!! But, the 7177/5i25 is set up on a Cincinnatti Milicron lathe with Siemens ac motors and 6sc6100 drives. The X axis growls and oscillates if the Z axis Analog drive signal is plugged in the 7i77. Unplugged it is quiet and smooth. Simens ac drives of this era are known for being noisy, and this dog is no exception. The major problem is the oscillation causes a bad surface finish. admittingly, the ballscrew needs to be rebuilt and the spindle bearings are not good, but the oscillation and resulting vibration are I believe to be caused by noise. Suggestions?

Looks like you have the PWM numbering right, but 7I49 PWM type must be 2 and 7I33 is normally type 3 (PDM) at 6MHz, so PDM frequency needs to be set to 6MHz.

Resolver scale is easy, they are 1 count per turn (for 1 speed resolvers) so for a 5T/In ball screw, the scale would be 0.200

The 7I77ISOL will isolate the 7I77 from PC ground but the Z and X axis grounds will still be connected
together at the 7I77, but its probably worth a try anyway.

Its also possible to use a 7I83 to get a isolated analog out, This might be a last resort method

Thanks. I have the pwm set to 1 on my lathe with a 5i20/7i33 combo and it works, what is the advantage to setting it to 3? And thanks again!!

Mode 1 is fine. Mode 3 at 6 MHZ has a little better
performance in terms of lower output ripple. but its unlikely
you would see a difference