Hurco Bed Mill with Dynapath Delta 40, retrofit to LinuxCNC

Hello all,

This marks the beginning of my project. I expect my 5i25/7i77 kit to arrive today and I have been deliberating, reading, dreaming, etc. of this for some time now. My Dynapath system is giving me headaches - and I mean not to discredit Dynapath as they are a GREAT control and well supported - but a replacement control with them is a bit spendy for my humble budget. I have already replaced the PC power supply, which helped, and I've been keeping a 60w light bulb in the cabinet to keep it warm and dry, which helps as well, but every once in a while I have to warm it up a bit more with a heat gun before I can get a clean boot out of it. Once up and running it's stable and works well, except for the ~30k program memory, lack of drip feeding and somewhat non-standard g-code. Instead of splitting programs up and developing a post processor for an outdated control, I've decided that this is the right way to go. My use of the machine is mainly for hobbies and small side jobs, so I can't justify the cost of a newer, more up-to-date Dynapath control. I'll learn more by not going the plug and play route anyways.

So far my hardware list is as follows:
* Bed mill, looks like an older version of the 500 Bed Mill on Dynapath's web site
- X and Y manual handwheels on the leadscrews, Z is only power driven (at this point).
- SEM servo motors with 1000 ppr encoders and line driver output
- SD drives
- CAN interface to front panel (not sure if I'll make it all work using RS422, pull more copper, or scrap it and use a touch screen)
- X, Y and Z home and overtravel switches
- lube pump
- field power supply
- contactors for spindle control (this machine has a metric Bridgeport clone variable speed belt drive head with an NMTB 30 spindle)
* PC with Linux CNC already installed
- SSD
- Intel D2550MUD2 mainboard
* Full set of schematics for the machine, printed Mesa manuals, etc.
* 5i25/7i77 PnG kit
And the following items to make this kit as plug and play as possible:
* (3) DB15 female breakout boards from Amazon: smile.amazon.com/gp/product/B074HBDNMH/ (for encoder connectors)
* DB25 female breakout board from Amazon: smile.amazon.com/gp/product/B07435ZTM9/ (for op panel connector)
* IDC40 (think IDE hard drive cable) breakout board from Amazon: smile.amazon.com/gp/product/B074THMST3/ (to go to Dynapath IO board)

Hopefully I'll be successful, and hopefully I will help pave the road for someone else to do the same. More to come.

Please wish me luck.

Are those SEM motors DC servos?

What drives do you have for them?

Yes, brushed DC servos with encoder and tach-generator feedback. They are original to the machine, as are the drives. I'm pretty sure they're Servo Dynamics drives, but not positive. I'll have a look next time I am out there. On the subject of the SEM motors - I understand that the original EPC encoders are problematic - had to replace the one on Z axis because it was missing the index pulse, preventing the mahcine from homing, and it would occasionally run away as well. I was told I'm not the only one. Dynapath uses Heidenhein (sp?) to replace the originals. I found a rebadged replacement to the original with a different bore size on eBay for next to nothing and it's still on the machine with a brass bushing I made for it. Additionally, I bought 3 new ones from Automation Direct which will fit with little modification - I think I'm going to have to make reducer bushings for them as well. Nonetheless, I haven't had any issues since replacing the Z-axis encoder but wouldn't be surprised if it croaked eventually like the last one.

Just had a look at a photo of the drives. The following is silk screened on the board:

SDFPOC1525-17/CF 11/97

I have finally revived this project. I have a very nervous feeling about making modifications to a functional machine, as it is the only mill in my shop.

To make this easier on myself, as mentioned above, I have elected to make my conversion as plug-and-play as possible. This way I could plug it in to the retrofit controller for testing, and then switch back over to the Dynapath controller to use the machine until the retrofit is completely functional. I purchased breakouts to match the connectors used to interface the original Dynapath controller to the physical parts of the machine, and am in the process of wiring them to the 7i77.

The only part that will NOT be plug and play is the console. I was originally planning to do so, but it adds a lot of work to the machine and does not take advantage of the fabulous look and feel of LinuxCNC on the little low-res monitor that I have in there. I feel that a large, touch-screen display would make the machine very attractive and would allow me to use touch buttons in place of physical ones for many of the less critical functions. For example, I could use touch buttons jogging and replace the keyboard with an onscreen one, but for E-stop and manual spindle control, I prefer physical buttons. I will likely retain the existing "RETRACT" button, and physical start and stop buttons. As such, I purchased an Elo ET2243L open frame touch monitor. Optically, it is a great performer. It runs off 12VDC and the Linux driver on their page works well in . After installing the touch response was inverted on X and Y. Following the instructions on the README file included with the driver, this was remedied using the calibration procedure. The SAW touchscreen works great in Gmoccappy, and I'm hoping that that some future release will be able to take advantage of the display's dual touch for interacting with the graphical toolpath preview.

This also frees up some inputs on the 7i77 (not like my machine used a lot anyways) to install three handwheels. I originally was thinking of making some fancy X/Y freewheel DRO and encoder handwheel Z function to allow the mill to have better manual usability. Not being able to move the column manually without climbing up on the top of the mill with a ratchet has slowed me down significantly since I traded my Bridgeport for this machine. I may still try doing this, but I'm sure it will require some fancy coding to allow LinuxCNC to operate in DRO mode on two axes and handwheel operation on the third. Maybe I'll even expand on the work that has already been done on using an MPG in connection with Reverse-Run to add handwheel program jogging, much like what is available on a Trak machine that I once saw.

Soooo....enter @ThisOldTony with a nifty three axis handwheel that he used on his Maho retrofit. I like this idea VERY much, and would most likely use this MORE than the CNC function. Heck maybe I would even find myself making a guard for the machine and starting to use coolant! Of course the chip tray could no longer be used as a storage shelf, but what the heck. Life is full of compromises!

Photos of progress to follow.

I have always had the option to use my machines manually, but it is a facility that is _very_ rarely used.

My lathe has a pair of jogwheels in the usual lathe locations, and I sometimes use them to put in the undercut at a shoulder (when I forget to CAD one in)

But, in practice on a mill, if the machine is powered-up it is as easy to type MDI commands as to twirl a knob.