Fanuc 10 Position scaling-calibration

[madkins]

Hi all,
Through an untimely event of ignorance I managed to clear the parameter memory on my Fanuc 10 CNC controller (see other posts).

So we are on the road to re-establishing the parameters but have hit a snag on calibrating the positioning of the 3 axis so that for example 1.000" desired movement is physically achieved.

Can anyone help me to understand the Fanuc method to scaling the positioning? I have an EE degree and many years of experience in digital logic design and machine automation, so I would like a very technical explanation.

Here is what I know (or think I know) the servo motors (DC brush) have a 2000M in their part number which I have been told is the encoder PPR. The x-axis ball screw appears to have a 5mm pitch (I can manually turn and measure this to be certain).

So how does the DMR, CMR and counter maximum value all relate? I understand that we are trying to achieve at least two things: 1- calibration, while 2- maximize resolution and thus ultimately accuracy.

Thanks for any response, this is my last ditch effort to deal with the Fanuc, otherwise it will be ripped out and replaced with a PC (really I am not bitter).

Mike

[Al_The_Man]

I have the Tech notes on this in PDF they exceed the limit here so if you want to send me an email ad by PM.
I can send a copy.
Al.

[Dan Fritz]

Fanucs have two sets of parameters that affect the postion scaling. They are called "DMR" (Detect Multiplying Ratio) and "CMR" (Command Multiplying Ratio). On the Fanuc 10 control, these parameters are #1816 (DMR) and #1820 (CMR). There are some other paramters that come into play when the CMR or DMR are changed, such as the servo-loop gain multiplier (1826) and the capacity of the "reference counter", which affects zero-return settings (1816).

Basically, you have a pulse coder on each axis, which sends either 2000 or 2500 pulses per revolution. Whether you have a 2000 or a 2500 line encoder should be stamped on your motor tag. A 10mm pitch ballscrew, or a 5mm pitch screw will probably have a 2500 line encoder, where other pitches like 4mm, 6mm, and 8mm will have a 2000 line encoder. This is because the distance traveled by the axis must be somehow be evenly divisible by the number of pulses per rev.

Now, each pulse coder sends in a train of "A" pulses, and a train of "B" pulses. With a Detect Multiplying Ratio of "1", your CNC is only using the RISING edge of the "A" pulse as a detect increment. This gives you either 2000 or 2500 pulses per rev of detection. A DMR of "2" reads the RISING and the FALLING edge of the A pulse, or double that number (4000 ppr or 5000 ppr). A DMR of "3" uses the RISING and FALLING edge of the A pulse plus the RISING edge of the B pulse (3 times the number on the pulse coder). The highest DMR setting is "4", which reads the RISING and FALLING edge of both the A and the B pulses. By manipulating the DMR, you can set the size of what Fanuc calls the "Detect Increment". This detect increment will be the smallest linear error that the CNC can detect through the pulse coders.

Generally, you would think that the smaller the detect unit, the more accurate the machine. This is true up to a point. On a typical metric machine (metric ballscrews), the DMR should set the Detect unit to .001mm, while the typical inch machine will be .0001 inch. Making the detect unit smaller (with a higher DMR) is usually not beneficial because of limitations to the maximum feedrates, etc. Also a problem here is that the servo-loop gain settings are affected by the DMR settings, so if you change the DMR, you have to re-calculate the gain multiplier (parameter 1826)

Now, for the CMR. The CMR is set to "1" when the Detect unit is .001mm AND you want to program motions with a finest increment of .001mm. If your detect unit is smaller (say, .0005mm), then you must use a CMR of "2" so that TWO command pulses are used to move your finest increment of .001mm. Changing the CMR ratio does not affect servo gain settings like the DMR does, but it does affect how far you move for every command increment.

If you have a metric ballscrew, you MUST make all these settings in metric. The Inch/Metric conversion software in the CNC will handle the inch motions. When testing your motions, be sure to have inch/metric conversion turned off so that the position display shows increments in your NATIVE increment system, otherwise you'll go crazy trying to figure out what you're seeing on the display.

Good luck!
Dan

[madkins]

Thanks to everyone who has provided information on this thread...

Here is an update: I verified that the motor encoder is 2000PPR based on the servo motor model number. The screw pitch is 0.2 inches per revolution (used the my dial caliper to measure the table movement for 10 screw revolutions). We are plenty happy with 0.0001" counting detection unit. Seems straight forward right...? Referencing a data sheet marked page 59, featuring an imperial system table it appears CMR=1, DMR=1 and counter capacity=2000.

So referencing my manual I went to parameter 8000 and set the PWE bit to 1, set 1820 X-axis = 1 (CMR), 1816 X-axis = 00010001 which should be the code for the DMR and counter capacity values listed above. I also checked that 1835 = 2000. Set PWE = 0, cycled power.

Upon power up I used my dial caliper to measure the physical distance moved by the table, 1.250" versus 1.000" displayed on the position screen???

So now I would like to characterize my position error as I change the parameters above to try to figure this thing out but when I try to set PWE=1 I get a WRITE PROTECT message. What is up with that???

Any suggestions would surely be appreciated, as in my old age I have very little patience for things that don't make sense and are totally non-user friendly. Thanks, Mike

[madkins]

So I found another thread on the forum that told me how to navigate through the menu structure and set the PWE bit... so that was one hurdle overcome.

I then went through all the values of DMR with parameters 1820 = 1, 1835 = 2000 and the counter capacity at 2000. The closest I came to 1.000" of actual movement was 0.844" with DMR set a 100 binary.

I remembered reading somewhere that the CMR must be doubled when entered into its 1820 parameter, so setting DMR back to 1 and 1820 to 2 I measured 2.535" of table movement with 1.000" displayed. (looks a lot like 25.4mm/inch).

So not knowing any better I checked parameter 0010 and the IN1 bit was set to 1, I changed it to 0 (again 1816 = 00010001, 1820=2, 1835=2000). After the power down/power up sequence I noticed the position data was one digit shorter #.### versus #.####. Checking my 1.000" of displayed movement I measured 0.10" of actual movement. I feel like I may be onto something.

Can anyone confirm that I am heading down the right track for inch units with 0.0001" resolution. My Fanuc manual is pretty general, what parameters should I be looking at to remove my 10X error factor?

Thanks in advance. Mike

[madkins]

Just a quick update:

With my current parameter settings when I move the X axis, 0.3937 on the position display the table moves 1.000", has to be a metric conversion going on that I just don't want or need....

Mike