Help with possible retrofit on Plasmatome B 20

[derekmccoy]

Hi there, Im new here. I've joined hoping to pick some brains and get somewhere with this plasma table I've been tasked with retrofitting. I chose this forum as this is the only one that I searched which had any results for Plasmatome and also quite a few users with Sanyo questions. I should say that although I'm quite technical, I know little about CNC but I know the basics (drawing is converted to gcode, software sends code to a motion controller which then sends correct signals to an amplifier/servo) I can pick up new concepts easily from that point if you guys are patient, haha.

Ok, that out of the way, on with this thing.

I have a SAF Plasmatome.B 20 table here. It has Sanyo Denki hardware (PY2 amplifiers driving P5 servos).

The Amplifiers are labelled "L.M", "L.E" and "T" (Looking at the table it might be two for X and one for Y, torch being controlled separately). These are connected to a very antique looking circuit board via the CN1 connector. This board looks to have some basic logic, some relays and a transformer.

At user end is a DIGISAF 100 Computer (if you can call it that). Its a big black box with some round connectors on the back labelled:
- Cutting oxygen and power input
- Drive
- Plasma Function
- Serial I/O B
- Y Encoder Input
- X Encoder Input
- Serial I/O B (again)

It looks like some of these are going into the old looking bare circuit board below the amplifiers.

Now call me silly but I thought encoders talked to the amplifier in a closed loop system, not to the computer (which I guess is the controller?), so why there's an X and Y "Encoder input" is confusing. If someone can explain why I don't need this going forward that would be of great relief.

To the right of the DIGISAF 100 unit is a Nertajet HP which is the plasma cutter which is ancient and will hopefully be replaced with a modern cutter.

The end goal (please stop me dead in my tracks if you think this is out of the realms of probability) is to rip out the DIGISAF 100 unit along with the old circuit board and the cutter. Then install a modern controller and cutter on the bridge (probably with a new THC). I'm really hoping that I can simply disconnect the amplifiers from the old circuit board and plug them straight into a modern controller.

I've read up a bit of the DIY forums and I was initially leaning towards UCCNC or maybe even LinuxCNC, however the Masso controller looks very appealing and sounds like it ticks all the boxes.

Any help is appreciated

Similar Threads:

• Need Help!- SAF Plasmatome HP
• Need Help!- Plasmatome B35 Default Sanyo controller program
• Problem- CNC Plasma Cutting Machines SAF PLASMATOME 20HP
• Newbie- Sieg X1 retrofit with Linux EMC2, CNC Fusion retrofit kit and Gecko G540
• Need Help!- camsoft retrofit to a anilam crusader retrofit

[Jim Dawson]

Looks like a fun project. If you can get the documentation for both the machine and the drives. If the documentation is not available, then rip everything out and start from scratch with modern hardware.

The Amplifiers are labelled "L.M", "L.E" and "T" (Looking at the table it might be two for X and one for Y, torch being controlled separately). These are connected to a very antique looking circuit board via the CN1 connector. This board looks to have some basic logic, some relays and a transformer.

I'm guessing X, Y, Z (torch height control is most likely integrated into the controls)

At user end is a DIGISAF 100 Computer (if you can call it that). Its a big black box with some round connectors on the back labelled:
- Cutting oxygen and power input
- Drive
- Plasma Function
- Serial I/O B
- Y Encoder Input
- X Encoder Input
- Serial I/O B (again)

It looks like some of these are going into the old looking bare circuit board below the amplifiers.

Now call me silly but I thought encoders talked to the amplifier in a closed loop system, not to the computer (which I guess is the controller?), so why there's an X and Y "Encoder input" is confusing. If someone can explain why I don't need this going forward that would be of great relief.

I assume that is the relay output board, also maybe the input buffer, along with an integrated power supply. Common in older equipment.

Looks like the loop is closed at the controller. Many older drives did not have the capability of closing the loop at the drive level.

To the right of the DIGISAF 100 unit is a Nertajet HP which is the plasma cutter which is ancient and will hopefully be replaced with a modern cutter.

Good plan.

The end goal (please stop me dead in my tracks if you think this is out of the realms of probability) is to rip out the DIGISAF 100 unit along with the old circuit board and the cutter. Then install a modern controller and cutter on the bridge (probably with a new THC). I'm really hoping that I can simply disconnect the amplifiers from the old circuit board and plug them straight into a modern controller.

It might be a bit more complex than just plugging the drives into the new control. I'm also guessing those drives are analog command input so this will require a controller that has analog command outputs as well as encoder inputs. You also may need the programming software for the drives to be able to adjust them for the new controller, you'll have to look at the drive documentation to confirm this.

[rodw]

Linuxcnc would be by far and away the best option for this retrofit for two reasons:
1. The LinuxCNC Plasmac controller is a complete plasma controller system incorporating high performance THC built in. See PlasmaC User Guide
2. There is some good hardware available to manage servo drives and encoders so You probably would be able to reuse your servos and drives.

I'm not the expert on servos but I think if you shared those images on the Linuxcnc forum, the right people will be able to make suggestions and come up with a retrofit plan.

Finally, there is a useful background document about plasma here Plasma Cutting Primer for LinuxCNC Users (disclosure: I wrote it.)

[derekmccoy]

Thanks for the replies, Jim & rodw.

I've had a look at the documentation and the Sanyo Denki PY2 units are digital and closed loop (optional):

https://www.sanyodenki.com/archive/d.../TR8e_h_en.pdf

After some mullng over, I thought the X and Y encoder inputs on the back of the controller were sensors on the table for its actual position feedback but can't see any sensors on the table itself.
I followed the X and Y leads and they go into the old circuit board, so now it looks like the amps are sending encoder data to the controller first passing through that board. I think this is because the two X motors have to be merged into one X signal before sending off to the controller. Unplugging these stops the table from working which makes me think the loop is open.

The other issue I've found is that it looks like the servos are CW/CCW so Masso might be out the window as its a Pulse/Direction controller.

LinuxCNC might be the first thing to try then as it doesn't require buying much kit. Am I right in saying that I can get an old laptop with an LPT port and wire it straight into the servos? No need for a separate motion controller board? This might be a question for the LinuxCNC board, so I'm getting a post ready for those guys.

[rodw]

No, you don't want to try a parallel port with servos. Don't take this as an authoritative answer without asking questions on the Linuxcnc board as don't have the experience with servos but I think you would have analog 0-10 volt control servo motors in which case you will probably start with a Mesa 7i77 7I77 card This will give you 6 axis with servo control and encoder inputs.

This board will connect to a PCI or PCIe PC bus using a Mesa 5i25 or 6i25 and they do have some bundled kits. (eg 6i25 and 7i77) More recently, there are also ethernet options so you could connect the 7i77 to a 7i92 ethernet board (which is a ethernet version of the 5i25/6i25). 7I92 picture

For plasma tables, we like to use ethernet becasue it isolates the PC from the noisy plasma environment.

To add more I/O you might add something like a 7i84 7i84 picture

Regarding your joint axis which has two motors. We make a distinction between joints (motors) and axes (XYZA etc) Assuming each joint (servo) has its own servo drive, you connect them both as if they are independent and it is a trivial matter to tell Linuxcnc that they are to be controlled together as one. We talk about a XYYZ machine (4 joints, 3 axes). Gantry alignment can then be done in software by changing the HOME_OFFSET. So once you get going if say the gantry is 1mm out of square after you put it together, you just need to make sure the home offsets are 1mm different in value.

We are continuing to improve the Plasmac control. Yesterday I got my first cut of a working void crossing feature which is something I've been working on on and off for about 3 years.

[PCW_MESA]

Thanks for the replies, Jim & rodw.

I've had a look at the documentation and the Sanyo Denki PY2 units are digital and closed loop (optional):

https://www.sanyodenki.com/archive/d.../TR8e_h_en.pdf

After some mullng over, I thought the X and Y encoder inputs on the back of the controller were sensors on the table for its actual position feedback but can't see any sensors on the table itself.
I followed the X and Y leads and they go into the old circuit board, so now it looks like the amps are sending encoder data to the controller first passing through that board. I think this is because the two X motors have to be merged into one X signal before sending off to the controller. Unplugging these stops the table from working which makes me think the loop is open.

The other issue I've found is that it looks like the servos are CW/CCW so Masso might be out the window as its a Pulse/Direction controller.

LinuxCNC might be the first thing to try then as it doesn't require buying much kit. Am I right in saying that I can get an old laptop with an LPT port and wire it straight into the servos? No need for a separate motion controller board? This might be a question for the LinuxCNC board, so I'm getting a post ready for those guys.

https://www.artisantg.com/info/PDF__...616E75616C.pdf

If this is correct, that drive can do CW/CCW Step/Dir or Quadrature (set with PMOD parameter) so almost any controller should work

[derekmccoy]

Thanks for that MESA, but the drive in question is this one:

https://s3.amazonaws.com/Icarus/DOCU...nuals_2000.pdf

Specifically the PY2A015. Is this definitely an analog servo?

On page 4-5 is the wiring diagram for the CN1 connector. About three quarters of the way down on the left I see "Forward Revolution Pulse" and "Backward Revolution Pulse". To me that means CW/CCW. I can't find anything that suggests Pulse/Direction. Many thanks

[PCW_MESA]

Thanks for that MESA, but the drive in question is this one:

https://s3.amazonaws.com/Icarus/DOCU...nuals_2000.pdf

Specifically the PY2A015. Is this definitely an analog servo?

On page 4-5 is the wiring diagram for the CN1 connector. About three quarters of the way down on the left I see "Forward Revolution Pulse" and "Backward Revolution Pulse". To me that means CW/CCW. I can't find anything that suggests Pulse/Direction. Many thanks

Its the same, you set the PMOD parameter to select the pulse input mode

Look at page 7-43

[derekmccoy]

Aha, just need someone that knows what they're looking for, lol. So I have to change Mode 2 Page 0 Bit 5 to 1. Now, Do I need one of those remote operator jobbies? They seem to be extremely expensive, like £800. Or can I plug a laptop in with some software?
Also, can I change the servo from open loop to closed loop? If it is indeed open at the moment..... How can I tell?

Sorry for all the questions.

[derekmccoy]

I think I have deciphered the LM, LE and T labels from the manual I found here:
https://www.rapidwelding.com/files/04081238.pdf

Longitudinal Maitre (master)
Longitudinal Esclave (slave)
Transversal

Now, longitudinal and transversal are not words I would choose to describe x and y but I guess its a french thing. I hope this master and slave setup doesn't affect the end goal of the retrofit. I will do more research.