Why do you need 210 VDC?

Normally you would use rectified 220V for the DC bus to run for 220V  motors

Are the motors 150 VAC?

I'm hoping to catch the eye of some of the experienced retrofitters here so I can learn some things. I'm in the process of retrofitting a Deckel FP4a with STMBL drives and LinuxCNC, it was originally 3 phase, but I want to run it on single phase 220, but not being an electrical engineer, I'm stumbling on the power supply for the drives.

I know I need to supply my motors with ~210VDC via a capacitor-smoothed DC bus, I'm just trying to find the best, safe option of doing so without breaking the bank, causing a fire, or killing myself and all my equipment, so I need to learn more.

From what I understand I have 3 options:

• An isolation transformer to drop the 220VAC to 150VAC, rectified to ~210VDC, but this needs an odd ~5KVA transformer that is likely very expensive and may be hard to source. I have the original, monstrous 3 phase transformer, but I have my doubts that it would work effectively in a single phase capacity.
• A skookum switch mode power supply that can handle the load of the 3 axis motors (2x 1.5KW, x1 1.75KW) and spindle (1.5KW). This is also likely to be expensive and hard to source.
• Direct rectification of the 220VAC to ~310VDC, very simple, but potentially quite dangerous, though I don't fully understand the risks and potential failure modes.

How are others handling this situation for their machines?


For a little background, the machine came to me free in a broken state, the original drives were questionable and the controller was non-functional, the machine is physically in great shape. I chose the STMBL drives because they can handle the resolvers in the stock Industrial Drives brushless axis motors and spindle motor, I couldn't find a decent, off-the-shelf drive that was cost competitive and they seemed like the ideal solution, provided I can power them.

I use DeskCNC for toolpathing, and it does do that for circles, but I do so many repetetive shapes, although not necessarily the same amount each time, that I simply write my own code I can reuse just by changing the number of repeats. I love it this way cause i can change stuff in seconds, but when it comes to circles in 4-axis mode, it's tougher

For the 7I97/7I97T, you need to set offset mode true for all PWMgens:

setp hm2_7i97.0.pwmgen.00.offset-mode 1
setp hm2_7i97.0.pwmgen.01.offset-mode 1
setp hm2_7i97.0.pwmgen.02.offset-mode 1
etc
also dither should be enabled:

setp hm2_7i97.0.pwmgen.00.dither 1
setp hm2_7i97.0.pwmgen.01.dither1
setp hm2_7i97.0.pwmgen.02.dither 1
etc
 

I've finally got my 7i97T running and I am at the point to test driving the spindle of my Lathe. The Simodrive 611 LT Module expects +-10V for running the spindle cw or ccw at certain speeds.
im not sure about my wiring, since I always get -10V and the spindle only turns one direction at full speed.

On TB3 analog0 i have enable +- on the first two pins. Then 10v+- on the third and ground of my 24V Powersupply on the fourth pin.

I did similiar retrofitting before and every time i ran into this, problem is: I cant remeber the solution.

I read somewhere about "spindle mode" and dithtering etc. Where can I read that up and how would I know what to look for ?

I have previously used 485 dongles but I think the mesa board connection is a cleaner wiring solution. I have used mb2hal but hm2-modbus is much easier for me.

I also found hm2-modbus/mesa refresh to offer much faster refresh than hm2hal/USB dongle which was important with some spindle calculations I was using to calculate gear ratios.

My advise is get the correct firmware and enjoy the cleaner wiring, easier message setup, and faster response with the mesa connection.

Having done this on many occasions but on other CNC machines, its just using what you got to make what you need.
Ive turned near full spheres on the end of a stick using a 3/8" round button insert lathe tool clamped on whatever angle in a vice, probably easier with button inserts- easy to touch off on.

The more difficult part is coming up with the NC code to do it- its not really a linuxcnc thing as a CAM thing. I was using mastercam x9 at the time, created a program using an inverse offset chain to drive the spindle/x/z in the right motion to achieve the part. Depending on the accuracy you require- you may need to include cutter geometry/radiuses to do it. In my case above the button insert radius was added to my offset via telling mcam the tool was 3/8" ball nose.

Was also able to hold a tolerance of around ±0.0002" with some fiddling around in A2 tool steel. 
 

If you are using a USB to RS485 adapter you are not running a USB cable to your device. These adapters generally don't have a USB cable at all, and are just a dongle you plug into your computer's USB port. You then run a RS485 cable from there to your device. As to speed, I'm pretty sure USB is quite a bit faster than RS485.

It seems to me that we are all wandering around like blind people in the fog...
I suspect that the cause of the problems lies where no one tries to look for it...
Let me tell you a story.
Some time ago I bought integrated Lichuan Ethercat servo steppers.
I set up LinuxCNC and basically everything worked.
After some time, I decided to buy identical drives for a second machine, essentially identical, only slightly larger, so I used the configuration from the older one.
Everything worked, except the engines did not rotate and the drives did not want to enter the correct mode.
I started looking for the cause, changed various parameters and restarted LinuxCNC until they worked properly, but after another restart they did not work again.
I already knew that the engines were working, but something must have been misconfigured, so I started writing the configuration file from scratch, first with one drive, then adding others...
It worked and everything has been working as it should since then.
Except that these two files (working and not working) are basically identical, and the only difference is in the block order.
If Ethercat is so sensitive to syntax that changing the order causes it to fail, then you can look for tiny errors until you die...

sorry many work, simply late publish source code

@MaxEkb77 is it possible to get the source of this?

Unfortunately this does not run my Linux box.

Thank you.

I used the code from the link above; you just need to add your own pins in config section(yours may be different). If it helps clarify things, I've attached my version of the code.

As for the difference between the SK4 and SK6—I've held both versions in my hands; the button layout is the same, but the circuit boards are different. So if you have the SK6, things might be different there.

Yes, I replaced the original rotary switches with encoders. I find encoders much more convenient—I don’t have to check their position every time; I can just turn them in the right direction. And they have a button; I’ve mapped pressing it to set FRO and SRO to 100%. Plus, if you use the stock ones, you won’t be able to adjust the settings from the screen (to avoid confusion in the readings, you’ll have to disable the on-screen sliders and leave only the current value displayed). That’s why we’ll always use only these rotary switches. With encoders, there’s no such problem—you can adjust the FRO and SRO both from the screen and using the encoders.

i can, if you ask a specific question. obviously i don't have a written specification to share.
but maybe it is possible to adapt your 4th axis to use a mitsubishi dtive/motor? would be much easier..

Thanks. A couple test runs of this look promising. I will try this out on an actual servo as soon as I find time.
I extracted the zip file then sha256summed it:
sha256sum servo_tuning 
9f2f5cb11561d13a6a3a4a62ac4788c9b2ea742844b31f8aeb7327cae0cb9cb9
next I marked it executable:
chmod +x servo_tuning
Then I started it:
./servo_tuning