[MrWild]

I've been wanting to upgrade a very dated control to something faster. I'd originally thought to use UHU drives and EMC or Mach, but problems with their power capabiilities has really tossed an anchor into that plan. Looking at other options all seem to be either pricey or low power.

While reading threads, I came across Mesa Electronics. They offer a four axis controller(?)/converter(?) that accepts PWM and direction signals and converts to analog + - 10v for analog drives. Here are the particulars.

7I33/7I33T Quad Analog servo interface



The 7I33 is a 4 axis analog servo interface intended for operation with MESAs Anything I/O cards when used for motion control applications. The 7I33 takes the PWM and direction signals from the Anythng I/O card and converts them to +- 10V analog output voltages for direct connection to analog input servo amplifiers. The 7I33 also conditions the encoder input signals with input filters for TTL inputs or a RS-422 recievers for differential encoder inputs. The controller connection is a 50 pin header that matches the pinout of the 4I34M, 4I65, 4I68, 5I20 and 7I60 Anything I/O cards. Another 50 pin header is used for servo amp/encoder connections. The -T version uses 3.5 mm screw terminal compatible plugs for motor I/O.


Although they say it works with their controllers, is it possible to hook it up to a BOB some way? Four axis for $70 seems cheap for something that y ou need fur Pixies ($66 ea.) to do. If I remember right, a Pixie is also a step direction controller that reads the encoder signal. This negates the ability of EMC to close the loop in the computer.

So, I guess I'm asking:

1. If the closing the loop in the computer is that much better, why is it better?

2. With good high power drives on hand, which would be the way to go? Pixes, or the Mesa 7133 (if it will work).

I'm really getting tired of this Crusader M control It has a 2,000 lines of gcode limit and I keep hitting it. 3D would never work even with drip feed.

more Mesa 7133 info

CONTROLLER CONNECTOR
50 pin header connector J1 connects to the anything I/O card/motion controller.
This can be a male 50 pin header on the top of the 7I33 card or a female 50 conductor
header on the bottom side of the 7I33 depending on 7I33 model. The controller connector
pinout matches the 4I27 (2 axis 7I33-2 only), 4I34M, 7I60, 5I20 (2 or 4 axis) pinouts.
Controller connector pinout is as follows:
PIN FUNCTION DIRECTION PIN FUNCTION DIRECTION
1 QB1 FROM 7I33
25 QB3 FROM 7I33
3 QA1 FROM 7I33
27 QA3 FROM 7I33
5 QB0 FROM 7I33
29 QB2 FROM 7I33
7 QA0 FROM 7I33
31 QA2 FROM 7I33
9 IDX1 FROM 7I33
33 IDX3 FROM 7I33
11 IDX0 FROM 7I33
35 IDX2 FROM 7I33
13 PWM1 TO 7I33
37 PWM3 TO 7I33
15 PWM0 TO 7I33
39 PWM2 TO 7I33
17 DIR1 TO 7I33
41 DIR3 TO 7I33
19 DIR0 TO 7I33
43 DIR2 TO 7I33
21 /ENA1 TO 7I33
45 /ENA3 TO 7I33
23 /ENA0 TO 7I33
47 /ENA2 TO 7I33
49 +5V PWR TO 7I33
Note: all even pins are grounded.
AUX 5V POWER
4 pin header P1 can be used to supply 5V power to the 7I33 if the controller cable is
too long and voltage drop too high. P1 has the following pinout:
PIN FUNCTION
1 5V
2 GND
3 GND
4 5V
7I33 5
CONNECTORS
SERVO AMP/ENCODER CONNECTOR
The servo amplifier / encoder connector (P2) is a 50 pin latching header. P2 has the
following pinout:
PIN FUNCTION DIRECTION PIN FUNCTION DIRECTION
1 GND
26 ENCA2 TO 7I33
2 ENCA0 TO 7I33
27 /ENCA2 TO 7I33
3 /ENCA0 TO 7I33
28 GND
4 GND
29 ENCB2 TO 7I33
5 ENCB0 TO 7I33
30 /ENCB2 TO 7I33
6 /ENCB0 TO 7I33
31 VCC FROM 7I33
7 VCC FROM 7I33
32 IDX2 TO 7I33
8 IDX0 TO 7I33
33 /IDX2 TO 7I33
9 /IDX0 TO 7I33
34 GND
10 GND
35 AOUT2 FROM 7I33
11 AOUT0 FROM 7I33
36 GND
12 GND
37 ENA2 FROM 7I33
13 ENA0 FROM 7I33
38 ENCA3 TO 7I33
14 ENCA1 TO 7I33
39 /ENCA3 TO 7I33
15 /ENCA1 TO 7I33
40 GND
16 GND
41 ENCB3 TO 7I33
17 ENCB1 TO 7I33
42 /ENCB3 TO 7I33
18 /ENCB1 TO 7I33
43 VCC FROM 7I33
19 VCC
44 IDX3 TO 7I33
20 IDX1 TO 7I33
45 /IDX3 TO 7I33
21 /IDX1 TO 7I33
46 GND
22 GND
47 AOUT3 FROM 7I33
23 AOUT1 FROM 7I33
48 GND
24 GND
49 ENA3 FROM 7I33
25 ENA1 FROM 7I33
50 GND

[samco]

You will need the 5I20 FPGA based PCI Anything I/O card to plug the 7I33/7I33T Quad Analog servo interface card into. There is direct support for the 5i20/7i33 in emc2.

As far as having emc2 close the loop.. with step and direction servo drives - there is no feedback to the controller (emc2/mach). having the loop within the pc - emc2 knows the position because it acutally gets the encoder counts from the mesa board. So if the drives are off - and you move the table manually emc2 will still know where it is. Plus all configuring is in one place - instead of having to tune the drives - then setup the controller. (not that that is a big deal)

that is all I can think of at the moment.

sam

[MrWild]

You will need the 5I20 FPGA based PCI Anything I/O card to plug the 7I33/7I33T Quad Analog servo interface card into. There is direct support for the 5i20/7i33 in emc2.

As far as having emc2 close the loop.. with step and direction servo drives - there is no feedback to the controller (emc2/mach). having the loop within the pc - emc2 knows the position because it acutally gets the encoder counts from the mesa board. So if the drives are off - and you move the table manually emc2 will still know where it is.


sam

Exactly what i needed t know. I was confused ho[w EMC spoke with perepheries. I assumed just the 7133 was needed. The real knowledge with an AH HA! to it was

"with step and direction servo drives - there is no feedback to the controller (emc2/mach). having the loop within the pc - emc2 knows the position because it acutally gets the encoder counts from the mesa board. So if the drives are off - and you move the table manually emc2 will still know where it is."

I like to use manual moves to align and zero. Electronically, unless I use a MPG I am at the mercy of input a move, and creep up for zeroing, and indcators can be real fragil if you think you are moving .010 or less and move an inch or rapid. Now that this is evident, I will look more into EMC. For ease f use, I'm still looking at Mach. Although no home brew control is turn key, some are a lot less painku l to get past the teething stage.

[unterhaus]

EMC is getting pretty painless if you have a spare computer. Mine is running fine on a 750MHz Pentium 3 with a 10 gig hard drive that I got for free. The only limitation is that it really likes 512Meg of ram, which is more than most computers like this are equipped with. Newer computers can actually cause problems, particularly some P4's and laptops are very unlikely to work.

[AdamM]

MrWild, I am doing a 2-Axis Retrofit on a "First" Knee mill.

The old control is a CrusaderM, The new control is EMC with the 5I20, 7I33 and the 7I37.

It seems to be fairly easy, I am using the old drives, servos, scales and power supplies. And I am running EMC off of an Nvidia Nforce2 with 1gig ram, 3200+ AMD, Nvidia FX5500 128mb graphics card.

I purchased a VFD from automationdirect.com (the GS1 series) to power my 2hp induction motor.

A guy around here by the name of RogerN has also done this mod to anilam crusader controls.

Any questions ask.

Adam

I've been wanting to upgrade a very dated control to something faster. I'd originally thought to use UHU drives and EMC or Mach, but problems with their power capabiilities has really tossed an anchor into that plan. Looking at other options all seem to be either pricey or low power.

While reading threads, I came across Mesa Electronics. They offer a four axis controller(?)/converter(?) that accepts PWM and direction signals and converts to analog + - 10v for analog drives. Here are the particulars.

7I33/7I33T Quad Analog servo interface



The 7I33 is a 4 axis analog servo interface intended for operation with MESAs Anything I/O cards when used for motion control applications. The 7I33 takes the PWM and direction signals from the Anythng I/O card and converts them to +- 10V analog output voltages for direct connection to analog input servo amplifiers. The 7I33 also conditions the encoder input signals with input filters for TTL inputs or a RS-422 recievers for differential encoder inputs. The controller connection is a 50 pin header that matches the pinout of the 4I34M, 4I65, 4I68, 5I20 and 7I60 Anything I/O cards. Another 50 pin header is used for servo amp/encoder connections. The -T version uses 3.5 mm screw terminal compatible plugs for motor I/O.


Although they say it works with their controllers, is it possible to hook it up to a BOB some way? Four axis for $70 seems cheap for something that y ou need fur Pixies ($66 ea.) to do. If I remember right, a Pixie is also a step direction controller that reads the encoder signal. This negates the ability of EMC to close the loop in the computer.

So, I guess I'm asking:

1. If the closing the loop in the computer is that much better, why is it better?

2. With good high power drives on hand, which would be the way to go? Pixes, or the Mesa 7133 (if it will work).

I'm really getting tired of this Crusader M control It has a 2,000 lines of gcode limit and I keep hitting it. 3D would never work even with drip feed.

more Mesa 7133 info

CONTROLLER CONNECTOR
50 pin header connector J1 connects to the anything I/O card/motion controller.
This can be a male 50 pin header on the top of the 7I33 card or a female 50 conductor
header on the bottom side of the 7I33 depending on 7I33 model. The controller connector
pinout matches the 4I27 (2 axis 7I33-2 only), 4I34M, 7I60, 5I20 (2 or 4 axis) pinouts.
Controller connector pinout is as follows:
PIN FUNCTION DIRECTION PIN FUNCTION DIRECTION
1 QB1 FROM 7I33
25 QB3 FROM 7I33
3 QA1 FROM 7I33
27 QA3 FROM 7I33
5 QB0 FROM 7I33
29 QB2 FROM 7I33
7 QA0 FROM 7I33
31 QA2 FROM 7I33
9 IDX1 FROM 7I33
33 IDX3 FROM 7I33
11 IDX0 FROM 7I33
35 IDX2 FROM 7I33
13 PWM1 TO 7I33
37 PWM3 TO 7I33
15 PWM0 TO 7I33
39 PWM2 TO 7I33
17 DIR1 TO 7I33
41 DIR3 TO 7I33
19 DIR0 TO 7I33
43 DIR2 TO 7I33
21 /ENA1 TO 7I33
45 /ENA3 TO 7I33
23 /ENA0 TO 7I33
47 /ENA2 TO 7I33
49 +5V PWR TO 7I33
Note: all even pins are grounded.
AUX 5V POWER
4 pin header P1 can be used to supply 5V power to the 7I33 if the controller cable is
too long and voltage drop too high. P1 has the following pinout:
PIN FUNCTION
1 5V
2 GND
3 GND
4 5V
7I33 5
CONNECTORS
SERVO AMP/ENCODER CONNECTOR
The servo amplifier / encoder connector (P2) is a 50 pin latching header. P2 has the
following pinout:
PIN FUNCTION DIRECTION PIN FUNCTION DIRECTION
1 GND
26 ENCA2 TO 7I33
2 ENCA0 TO 7I33
27 /ENCA2 TO 7I33
3 /ENCA0 TO 7I33
28 GND
4 GND
29 ENCB2 TO 7I33
5 ENCB0 TO 7I33
30 /ENCB2 TO 7I33
6 /ENCB0 TO 7I33
31 VCC FROM 7I33
7 VCC FROM 7I33
32 IDX2 TO 7I33
8 IDX0 TO 7I33
33 /IDX2 TO 7I33
9 /IDX0 TO 7I33
34 GND
10 GND
35 AOUT2 FROM 7I33
11 AOUT0 FROM 7I33
36 GND
12 GND
37 ENA2 FROM 7I33
13 ENA0 FROM 7I33
38 ENCA3 TO 7I33
14 ENCA1 TO 7I33
39 /ENCA3 TO 7I33
15 /ENCA1 TO 7I33
40 GND
16 GND
41 ENCB3 TO 7I33
17 ENCB1 TO 7I33
42 /ENCB3 TO 7I33
18 /ENCB1 TO 7I33
43 VCC FROM 7I33
19 VCC
44 IDX3 TO 7I33
20 IDX1 TO 7I33
45 /IDX3 TO 7I33
21 /IDX1 TO 7I33
46 GND
22 GND
47 AOUT3 FROM 7I33
23 AOUT1 FROM 7I33
48 GND
24 GND
49 ENA3 FROM 7I33
25 ENA1 FROM 7I33
50 GND

[MrWild]

I've had my head buried in the EMC manuals. A question that concerns me is, do I need to learn all of that HAL stuff? What does the MESA HAL do, and would i need to do more than PID, step per inch, and lead screw tracking? That right there is as much as I want to fight with it. If there is building things in HAL too...

AdamM,

What you are doing is exactly what I'd been thinking about. I started in the UHU direction and was told the drives were good for 200v and 30A. Way more than i need and a big step up from the old drives. Until people had problems with ithem blowing up at far less power. The fix isn't a patch, but a whole new revisied PCB with new and different componants. At some point i have to stop bleeding money. Three more drives and Mach, or the MESA boards and EMC.

So, I'm real interested in the problems you run into and ease of adapting. You're using the scales and not encoders? One of the problems I've noticed with the crusaderM control is a notchy finish as thie table reacts to the larger steps of .0005. Might be slightly out of tune PID with the original drives too. The mechanics are tight. An encoder can give tens of thousands of impulses per inch, but the scales give 2,000 with 8k in quadrature. To be honest, if you get your scales working well, I'd follow suit. Mounting and wiring three encoders is a bit of work and surgery on my servos I'd like to avoid.

[unterhaus]

I've had my head buried in the EMC manuals. A question that concerns me is, do I need to learn all of that HAL stuff?

You don't have to learn anything about Hal, although it's pretty easy once you look at it. It has a lot of capability built in, but you don't need most of it to make a machine work.

[AdamM]

I am happy with the glass scales, I am getting a consistant 0.0004 repeatability. I added a Z-axis to my quill a few days ago and it uses a 1000PPR rotary encoder. I am getting less than a ten thou of error on that axis.

The hardest part of my entire build was the PID, I found that if you use mostly P with some FF0, FF1 and FF2 in there you can get some awsome results. PID tuning is made easier by HALSCOPE, an oscilliscope included with the EMC release, you simply plot your PID error and make adjustments accordingly.

I suggest the Mesa Boards and Gecko drives if you need drives. then some decent servos. I am using SEM servos. I used the anilam/servo dynamic/glentek drives that came with my machine. They work fantastic, no problems. You should have no problems with these "blowing" as long as the DC voltage coming in is fused before the Bridge Rectifier.


If you do resort to mounting rotary encoders for some reason, the wiring is the same as the glass scales. You just switch them directly and just change your input scale in your 5i20.ini file. They don't have to go on the servos either, you could mount them on the other end of your ballscrews...


From start to finish the build took me about 3 weeks.

I think the problem with the crusaderM is just that it is old Technology and things have come a long ways since then. More advanced trajectory planning for one.


Last but not least use EMC, its opensource. Its easy enough to use and is supported by a great group of people. Questions asked are answered quickly and very well.

AdamM

AdamM,

What you are doing is exactly what I'd been thinking about. I started in the UHU direction and was told the drives were good for 200v and 30A. Way more than i need and a big step up from the old drives. Until people had problems with ithem blowing up at far less power. The fix isn't a patch, but a whole new revisied PCB with new and different componants. At some point i have to stop bleeding money. Three more drives and Mach, or the MESA boards and EMC.

So, I'm real interested in the problems you run into and ease of adapting. You're using the scales and not encoders? One of the problems I've noticed with the crusaderM control is a notchy finish as thie table reacts to the larger steps of .0005. Might be slightly out of tune PID with the original drives too. The mechanics are tight. An encoder can give tens of thousands of impulses per inch, but the scales give 2,000 with 8k in quadrature. To be honest, if you get your scales working well, I'd follow suit. Mounting and wiring three encoders is a bit of work and surgery on my servos I'd like to avoid.

[MrWild]

I am happy with the glass scales, I am getting a consistant 0.0004 repeatability. I added a Z-axis to my quill a few days ago and it uses a 1000PPR rotary encoder. I am getting less than a ten thou of error on that axis.

The hardest part of my entire build was the PID, I found that if you use mostly P with some FF0, FF1 and FF2 in there you can get some awsome results. PID tuning is made easier by HALSCOPE, an oscilliscope included with the EMC release, you simply plot your PID error and make adjustments accordingly.

I suggest the Mesa Boards and Gecko drives if you need drives. then some decent servos. I am using SEM servos. I used the anilam/servo dynamic/glentek drives that came with my machine. They work fantastic, no problems. You should have no problems with these "blowing" as long as the DC voltage coming in is fused before the Bridge Rectifier.


If you do resort to mounting rotary encoders for some reason, the wiring is the same as the glass scales. You just switch them directly and just change your input scale in your 5i20.ini file. They don't have to go on the servos either, you could mount them on the other end of your ballscrews...


From start to finish the build took me about 3 weeks.

AdamM

I have a "working" CrusaderM three axis now, but it leaves quite a bit to be desired. The UHU drives were because I was told the Crusader drives will eventually fail, but if they have lasted this long, I'm pretty sure they will continue to last as long as I don't short them out.

Because I have the entire Crusader M system, I have the servo power supply for the drives. My motors are SEM 142v. examples. I think I'd be happy with .0004 accuracy. Not building space ships here.

Only three weeks? As I have most things mounted already and just changing the brain in actuality I wonder if I can half that. My time will be caught up in fixing some spindle power transfer problems. Sticky mechanical speed control due to a hanging pulley on the motor. While everything is apart, I also want to look into the brake as that doesn't work well either.

What do yo use for CAM? Do you write it in windows then save as text and open the text from EMC? Loading programs into Crusader's very limited memory is a PITA I want to rise from.

[AdamM]

I installed VMware on my ubuntu box, which lets me run windows within ubuntu splitting system resources. Its a virtual PC within a PC basically. I installed windows XP pro in it and in that Mastercam X.

I save my posted G-code to a pen drive then access the pen drive from ubuntu.



Your build should go quick, if you have all the anilam parts. Anilam used high quality drives, servos and scales you should have no problems with them. do you plan on using a VFD?

I installed a VFD on my induction motor. It works really well with EMC.


For your estop chain... wire your limits and button in series. Then through an output switch on the 7i37. I used that to power three relays that then powered my 3-phase VFD and coolant pump and my seingle phase power to my drive circuit. so when a limit or estop is switched everything is off and the 7i37 knows aswell. the 7i37 is also used to power up the machine with that output switch. It also knows to trigger a reset state onthe control of you are off the estop....


I could send you a diagram when I scan it.

AdamM

[MrWild]

I installed VMware on my ubuntu box, which lets me run windows within ubuntu splitting system resources. Its a virtual PC within a PC basically. I installed windows XP pro in it and in that Mastercam X.


do you plan on using a VFD?

I installed a VFD on my induction motor. It works really well with EMC.


For your estop chain... wire your limits and button in series. Then through an output switch on the 7i37. I used that to power three relays that then powered my 3-phase VFD and coolant pump and my seingle phase power to my drive circuit. so when a limit or estop is switched everything is off and the 7i37 knows aswell. the 7i37 is also used to power up the machine with that output switch. It also knows to trigger a reset state onthe control of you are off the estop....


I could send you a diagram when I scan it.

AdamM

Working in reverse, YES! If you could send me or post a diagram, I'd be indebted. I want to control my mill exactly as you have set yours up.

I want to go with a VFD, but will deal with the manual speed control and my rotory converter for now. Slow $$$ time.

My piggy bank is full and I think the quarters and dimes are enoiugh to finally afford a MESA set up. All I really need and want to hear is that your machine has been woking with no problems. A working imbecile controller is better than a fitful genius.

Buddy finally had the time to tear the spindle drive apart and replace the bad parts with ones I bought off eBay. Sounds sweet, and spins up like a top now. Ready for round two.