Light Machine Corp. Benchman XTr (retrofit)

Hello!
Brand new to this forum, looking closely at a Benchman XT for a retrofit to a Centroid controller.
I have a few questions after spending the weekend going through this entire thread, things I didn't see or that I didn't quite understand. Any help would of course be appreciated, and you guys have done an amazing job of sharing info on your different projects.

The motor specs for the Servo's seem a little wonky:
Axis Drive Motors - From - www.machinetoolsonline.com/doc/benchman-...chining-centers-0001
Feed rate: 0.1 -200 ipm (2 - 5080 mm/min)
Motor Type: DC servo
Torque: 42 oz. in. (30 Ncm)
Z-axis thrust force: 80 lbf (356 N)
X, Y- axis thrust force: 50 lbf (222 N)

These N measurements seem odd, but my understanding of them is very limited. The torque of 42 oz. in also seems quite low, similar machines such as a Tormach often have around 1200 oz. in. for their motors. I am considering using the motors as they are clearly specced for it and are Servos so that is an excellent use for them. My alternate for them would be some Leadshine motors as they seem to be the best cost/value ration in this general size.

My understanding though is limited again, with the servos on the Benchman and the encoder system they have built in.
I also see a number of posts by Don, the X Engineer, remarking about the Laser Calibration, I would guess that by switching to a different controller that information would be lost although the marketing for the Centroid Acorn mentions (Screw compensation (laser or manual!)). My guess is that with tuning and compensation for backlash I should be able to hold 0.0002" or so with the machine as that is what it was designed for, but please let me know if you think that is a misinterpretation. Also, there is mention of the store.mesanet.com/index.php?route=produc...t_id=101&search=7i49 which may work to interpret the resolvers, rather than encoders on the servos?

The screw mapping file that was created at Intelitek is based on the axis home position. This position is the first index pulse back from the limit switch. If the encoder is moved/adjusted/changed, the file becomes useless.

Greetings folks. Trying to get back into my machine, as it has been a while and it would be nice to have this thing up and running, better.  After taking a few hours last night to read his thread, I noticed most of the problems were not with LinuxCNC, but the original OS... it would be nice to get this thread back on tract

My problems have come when trying to PID tune the machine. In the past, I had to use some tricky tuning, but it seems like things should be more so, straight forward. Hopefully, some insight will help wrap this up once and for all.

Just an recap:
I removed the original light machine motherboard and have all the individual pins going to the 7i77 and 7i84.  In the past, I was able to move the machine, but when increasing the numbers to full rapids, the machine would shutter and sometimes act like it was searching for an endpoint. Also, when trying to take a bigger cut, it seemed like a loss of power. All that said, I got it all setup again and the x axis moved really smooth before getting a joint 2 following error. 

As for the power setup, everything is OEM until we get to the bridge rectifier. I have assembled a board that has 3 capacitors (1 small and 2 big) and a 5.1K resistor in parallel to smooth the circuit.  Going into the board is about 68V and coming out is also 68V.  The schematic that Don offered up shows 50V coming out of the smoothing circuit board. The 68V goes to each of the 4 Copley Control boards. Anyone see a problem, so far?

One question that I cannot figure out is, were the Copley controls set to velocity mode or torque mode?

The mesa boards and power rails are powered by the 24v onboard PSU. The system has 24v relays to enable the copely controls. The 10v signal to enable were step-down from the 24V PSU. I rely on the 7i77 to send the correct 10v signal to the boards.

Here are the numbers that I have used in the past, but would like to come up with a better tuning.

                                   X                   Y                    Z                  A
P                              2048             2048              5000              300
I                                0.2225            0                    0                  0.1
D                               4.2                  4                  7.5                0.501
BIAS                            0                  0                    0                  0
FF0                              0                  0                      0                 0            
FF1                              5.5              5                        8               .505
FF2                             0.021          0                        0.25           .002
Deadband                    2e-5          2e-5                    2e-5           0.0001
Max_output                    10             10                      10              10
Output_scale                  10             10                      10              10       
Output_Min                    -10             -10                    -10            -10
Output_Max                     10              10                     10              10
Encoder_scale               50000    50000                  -50000         -1000

I had literally spent 50+ hours getting things to this point and need insight and advice.

Thanks!





 

Torque mode

andypugh wrote:

Halscope might help.
Moving smoothly until a following error can be an indication that the machine is pegged against the physical speed limit. That will be very smooth, right until the setpoint is too far ahead of the actual.

You _could_ consider cascaded control loops, with a PID that takes a velocity input to convert the drives to velocity mode and a second PID driving the velocity PID to do the position control.

But I would only suggest that if you really enjoy PID tuning
 

I agree with every thing Andy said except the tuning.  I thought my duel PID loop config was much easier to tune than tuning the torque commanded drives using only a single PID.  And it seems to perform much better under varying loads.

I guess there are a few comments and questions. The last time the machine had a working .ini file and conducted several operations. Since then, the machine was changed to the standard Debian wheezy 2.7 and upgraded to 2.8. Upon initially opening Linuxcnc, a command prompt came up with an .ini file conversion for the new build. At this point, after enabling the machine, the spindle drops and creates a following error. What is strange is that the .ini files are identical, so not sure what the problem might be or what exactly changed when "updating". Anyone more familiar with the change have an idea for a quicker solution to get back to the last working point?

This question is perhaps the more important. Can anyone recommend a more user friendly drive that allows an easier way to tune the existing motors on all 4 axes? At this point, I am ready to throw some money at this machine to start reliably making chips.

Mesa 7i25+7i77+7i84
Dc Brushed motors
High resolution encoders
67.5V

After upgrading to 2.8 then running a configuration for the first time it checks for a "VERSION =" line in the ini file and if it doesn't exist or is too small then runs the conversion script on it. If you did not let the conversion script run when prompted the config files will not be converted and they will not work with 2.8. The majority of the changes have to do with the Joints-Axis change that happened with 2.8. Before 2.8 the terms joint and axis were used interchangeably and meant the same thing. After 2.8 there is a distinct difference between them in the code, where a "joint" number only refers to a motor and may or may not correspond to a Cartesian axis letter. An "Axis" letter only refers to a Cartesian axis in space. This change makes setting up non trivial kinematics machines more straight forward and machine configs less ambiguous. There were a few other minor changes to some other pin names as well. The conversion script will usually automatically make all the the nessisary changes to a config, but there are always exceptions. For this reason the conversion script will also save the original files to a backup directory within the config dir.

(I believe Andy wrote the conversion script.)

Hi Guys,

I am listing two Benchman XTs for sale in Grand Rapids, Michigan. Both are operational and under power now but one is in need of some y-axis work as it seems like it's got a pretty gnarly groan to it. Would love to hand them off to an enthusiast like the folks are on here. Asking $3k but they will require pickup in Michigan from a residential address.

Ryan

grandrapids.craigslist.org/tls/d/grand-r...ning/7475023151.html