Mazak Impulse - Oak retrofit build log

All things related to Centroid Oak, Allin1DC, MPU11 and Legacy products

Moderator: cnckeith

ashesman
Posts: 395
Joined: Thu Dec 03, 2020 4:54 am
Acorn CNC Controller: No
Allin1DC CNC Controller: No
Oak CNC controller: Yes
CNC Control System Serial Number: none
DC3IOB: No
CNC12: Yes
CNC11: No
CPU10 or CPU7: No

Re: Mazak Impulse - Oak retrofit build log

Post by ashesman »

Jqmce wrote: Sat Feb 06, 2021 8:20 am If you don’t mind me asking what kind of screen did you order?
The screen is one of https://zhixianda.en.alibaba.com/produc ... 5726aoCjGY.

I know some people are anti "import" parts, but where I come from, everything is import! There are heaps of other industrial screen options for five to ten times the price of this one. If it doesn't work out then I will look at expensive ones then. The touch screen works really well. I opted for capacitive instead of resistive touch screen so I could run other applications like CAM software on it and pinch to zoom etc. A resistive one would be better for use with dirty hands and gloves. I don't like the soft surface a resistive touch screen has. This one has a glass surface like a tablet. The picture quality degrades a bit at wide viewing angle but fine for a CNC screen! Hint: make sure to ask for a "High quality" screen with all working pixels.
martyscncgarage
Posts: 9912
Joined: Tue Mar 28, 2017 12:01 pm
Acorn CNC Controller: Yes
Allin1DC CNC Controller: Yes
Oak CNC controller: No
CNC Control System Serial Number: none
DC3IOB: No
CNC12: Yes
CNC11: Yes
CPU10 or CPU7: Yes
Location: Mesa, AZ

Re: Mazak Impulse - Oak retrofit build log

Post by martyscncgarage »

ashesman wrote: Sun Feb 07, 2021 4:01 pm
Jqmce wrote: Sat Feb 06, 2021 8:20 am If you don’t mind me asking what kind of screen did you order?
The screen is one of https://zhixianda.en.alibaba.com/produc ... 5726aoCjGY.

I know some people are anti "import" parts, but where I come from, everything is import! There are heaps of other industrial screen options for five to ten times the price of this one. If it doesn't work out then I will look at expensive ones then. The touch screen works really well. I opted for capacitive instead of resistive touch screen so I could run other applications like CAM software on it and pinch to zoom etc. A resistive one would be better for use with dirty hands and gloves. I don't like the soft surface a resistive touch screen has. This one has a glass surface like a tablet. The picture quality degrades a bit at wide viewing angle but fine for a CNC screen! Hint: make sure to ask for a "High quality" screen with all working pixels.
Screen is nice looking. Nearly everything is imported. What suits you is what matters.
Marty
Reminder, for support please follow this post: viewtopic.php?f=20&t=383
We can't "SEE" what you see...
Mesa, AZ
ashesman
Posts: 395
Joined: Thu Dec 03, 2020 4:54 am
Acorn CNC Controller: No
Allin1DC CNC Controller: No
Oak CNC controller: Yes
CNC Control System Serial Number: none
DC3IOB: No
CNC12: Yes
CNC11: No
CPU10 or CPU7: No

Re: Mazak Impulse - Oak retrofit build log

Post by ashesman »

Minor progress made this weekend. I have been away riding so didn't achieve much.

I removed the control panel and reconnected it to the controller and powered it all back up. Still works! I did this so that it can be powered every now and then to get the memory backup battery charged. I don't want it to lose all its parameters and want to be able to demonstrate it working in case I can sell it, or make it into an ornament on the wall!
20210205_110029.jpg
Next thing I did was cut all the cable ties on the wiring loom. It turns out that there is a heap of cable bundled up inside the machine behind a panel between the control cabinet and air valves. That gave a heap more cable which will make my life much easier.
20210205_093050.jpg
Next job was to start mocking up how the control panel is going to go together. I usually just like to place all the parts and build this sort of stuff by feel. But this time as I want to incorporate the heat sinks and existing panel layout I decided to very quickly rough it up in CAD.

I didn't manage to find a place for the PC in the cabinet yet but I may try squeeze it into the front panel. The jury is still out on that one!
Capture.PNG
Next I will physically mock it all up on a board and make sure all the important cables that I don't want to cut or extend (like servo and encoder cables!) can be routed nicely.
cncsnw
Posts: 3763
Joined: Wed Mar 24, 2010 5:48 pm

Re: Mazak Impulse - Oak retrofit build log

Post by cncsnw »

A minor detail: I would flip the PLCADD1616 around so that all the inputs (Oak plus PLCADD1616) are on the same side, and so that the I/O cable does not have to cross the board.

You might also consider putting the drives at the bottom, and the Oak in the middle. This way, your servo motor power cables, and your spindle motor power cable, can be routed out of the cabinet without passing close by any of the more noise-sensitive components.
Jqmce
Posts: 36
Joined: Sat Feb 06, 2021 8:17 am
Acorn CNC Controller: No
Allin1DC CNC Controller: No
Oak CNC controller: Yes
CNC Control System Serial Number: A900979
DC3IOB: No
CNC12: Yes
CNC11: No
CPU10 or CPU7: No

Re: Mazak Impulse - Oak retrofit build log

Post by Jqmce »

Thanks for the info on the screen Ashesman.
Really looking forward to your next instalment.
Eric S
Posts: 28
Joined: Mon Feb 17, 2020 3:36 pm
Acorn CNC Controller: No
Allin1DC CNC Controller: No
Oak CNC controller: Yes
CNC Control System Serial Number: none
DC3IOB: No
CNC12: Yes
CNC11: Yes
CPU10 or CPU7: No

Re: Mazak Impulse - Oak retrofit build log

Post by Eric S »

Was your Mazak using the SSCnet drives? Do the encoders on the original Mazak motors not work with DMM drives without your board?
ashesman
Posts: 395
Joined: Thu Dec 03, 2020 4:54 am
Acorn CNC Controller: No
Allin1DC CNC Controller: No
Oak CNC controller: Yes
CNC Control System Serial Number: none
DC3IOB: No
CNC12: Yes
CNC11: No
CPU10 or CPU7: No

Re: Mazak Impulse - Oak retrofit build log

Post by ashesman »

Eric S wrote: Sun Feb 07, 2021 10:09 pm Was your Mazak using the SSCnet drives? Do the encoders on the original Mazak motors not work with DMM drives without your board?
I am not sure what a SSCnet drive is but I don't think it used them. The pictures in my first post show the drives. It is all integrated into one big control unit made by Mazak.

DMM drives use a proprietary encoder protocol that only works between their motors and drives. They offer an adapter from some particular encoders to their protocol. Each of the little boards standing up in my PCB is supplied by DMM. My encoders are 2500 pulse (10000 count per RPM). DMM make a 2500 ppr adapter PCB. I powered it all up and tested it and it seems to work OK.
ashesman
Posts: 395
Joined: Thu Dec 03, 2020 4:54 am
Acorn CNC Controller: No
Allin1DC CNC Controller: No
Oak CNC controller: Yes
CNC Control System Serial Number: none
DC3IOB: No
CNC12: Yes
CNC11: No
CPU10 or CPU7: No

Re: Mazak Impulse - Oak retrofit build log

Post by ashesman »

cncsnw wrote: Sun Feb 07, 2021 5:21 pm A minor detail: I would flip the PLCADD1616 around so that all the inputs (Oak plus PLCADD1616) are on the same side, and so that the I/O cable does not have to cross the board.

You might also consider putting the drives at the bottom, and the Oak in the middle. This way, your servo motor power cables, and your spindle motor power cable, can be routed out of the cabinet without passing close by any of the more noise-sensitive components.
Thanks for the feedback. The ADD1616 was just placed how it landed. I would worry about its orientation when I know what is wired to it. I wont be using any inputs and probably less than three outputs. Ideally I would like to put it below the Oak as I dont really like the look of it!

I fussed over whether to put the Oak in the middle. My thoughts were that if it is in the middle, then the extremely noisy and unshielded power cables to the drives would be passing the Oak. Actually the main driver for putting them in the middle is that the big holes in the door panel are where the big heat sinks go. These are the only way of transferring heat out of the cabinet. I figured the best thing to mount to them was the parts that make the most heat.

The spindle and servo drive cables will be heading out to the right then down to the bottom. Then then do a loop and come back up again. As all this is on an opening door, the cables need to be able to hinge. I started building a mock up, screwing everything to wood so I can see how the cables look. I can then easily swap it all around as I please until I like it.
ashesman
Posts: 395
Joined: Thu Dec 03, 2020 4:54 am
Acorn CNC Controller: No
Allin1DC CNC Controller: No
Oak CNC controller: Yes
CNC Control System Serial Number: none
DC3IOB: No
CNC12: Yes
CNC11: No
CPU10 or CPU7: No

Re: Mazak Impulse - Oak retrofit build log

Post by ashesman »

OK, so its been a while since I did an update. I feel like I have achieved nothing but have been spending almost all of my (small amount of) spare time on this project. I have had an absolute battle with trying to find a suitable spindle motor drive and get working servo drives. I have also had a fight to try to get everything to fit on my limited space control panel. I have however got most of my PLC code sorted while I have been waiting for parts and tested all my IO and wiring. I will post some pictures of the control panel once it is tidied up.

So, what I have done that is relevant to this machine build is determine some of the characteristics of the spindle servo motor. This motor has no data available. The OEM controller had data in it for the axis motors but zeros for the spindle motor. After testing the motor on several VFDs and doing a ton of research on motors and drivers, it became apparent that this motor has some unique driving requirements. It is only 1.5kW rated but has a huge OEM driver capable of 75A. As it turns out the motor is very low impedance and capable of pulling a heap of current (~60A) during hard acceleration. It can do a speed reversal from -6000 to +6000 RPM in around half a second on the OEM controller. The VFDs were not able to come close to that sort of acceleration due to current limits, even one ABB 5kW rated drive. I have just received a Hitachi SJ-P1 so will see how that goes, but to be honest I am not holding my breath.

You will probably be thinking who needs acceleration that fast. Well, not me, but I have taken it upon myself to try make a new control achieve what they did 25 years ago. But, even I could accept a slower acceleration!

Anyways, motor characterization. The first and easiest thing to determine is the winding resistance R. As it is in the low ohms range, a multimeter is not going to be able to measure it. Passing a known current through the winding and measuring the voltage drop is a better method. This yielded a resistance per phase (including cable resistance) of 0.1242 Ohms. Approx 0.1 ohms if you take out the cable resistance, but the drive sees the cable resistance.
20210504_202736.jpg
Next is to measure the back EMF Ke. This is done by measuring the motor windings with an oscilloscope while rotating the motor. I drove the motor at 2500 RPM. After the maths and conversion to appropriate units, this motor has 40mV/RPM (RMS line to line) or 39.265 mVs/rad (peak phase, electrical angle).
20210504_210805.jpg
Spinde back EMF U-V at 2525 RPM.png
The final important parameter is the winding inductance L. This can be measured by two methods. I tried both and got similar results. One method is to drive a sine wave into the windings and find the frequency that the motor impedance matches the driver impedance. At this frequency the measured sine wave will be 50% of the unloaded driving signals amplitude. The other method is to inject a square wave and measure the time constant of the current rise. These gave me an inductance of 0.7uH and 0.5uH. A bit of variance, but much closer than guessing at it!
20210504_205326.jpg
Spindle motor step response to 10Vp-p square wave on U-VW.png
Anyone who knows their motors will know that there are usually two inductance values required. Ld and Lq. On motors with internal magnet rotors, these values are quite different. On surface magnet motors they should be identical. They are the inductances measured at the pole and 90 deg from the pole. You can determine the type of motor by confirming that the measured inductance does not change while the motor is rotated. In my case it stayed constant so is a SMM motor.

Finally you need to do a bit of maths to convert the results. Then type the numbers into the drive. Any self respecting drive should have these parameters required for a permanent magnet motor. Although some claim to be able to self measure them.
20210506_210535.jpg
This weekends mission if I can find time will be to try run the motor on a new Hitach SJ-P1 drive with encoder interface and see how it goes. If I can get this spindle to work I have probably overcome my biggest hurdle. Fingers crossed.
Last edited by ashesman on Thu May 06, 2021 5:39 am, edited 1 time in total.
ashesman
Posts: 395
Joined: Thu Dec 03, 2020 4:54 am
Acorn CNC Controller: No
Allin1DC CNC Controller: No
Oak CNC controller: Yes
CNC Control System Serial Number: none
DC3IOB: No
CNC12: Yes
CNC11: No
CPU10 or CPU7: No

Re: Mazak Impulse - Oak retrofit build log

Post by ashesman »

Another thing that has been plaguing me is tripping of the main circuit breaker when the control panel is first powered after a while of being off. Only maybe one in five times, but too often once it is locked up inside the back of the machine.

Of course the answer is that the circuit breaker is incorrectly rated. I can already hear Marty saying you should use fuses. Fuses are not really accepted practice here any more and personally I prefer circuit breakers.

So, I did my normal anal(ytical) approach to things and measured the current to see what is going on. There are about eight switch mode power supplies driven off this circuit breaker (four motor drives, Oak PSU, 24V PSU 12V PSU, PC PSU and 12V AC transformer). Under normal operating the total current draw is 0.7A at 200V. This was measure with a clamp on current transformer and an oscilloscope.

I then turned the panel on and off until I managed to capture the current on the scope when it trips the circuit breaker. Inrush current for that many PSUs is quite high and I knew it would be but was still surprised. Typically inrush current is around 30-40A for 4-6ms, but when it trips is 72.5A for 3.4ms.

Mainly I am posting this to share what really goes on when you first power up cheap AC-DC flyback converter power supplies.

I am using a 4A 2 pole C curve MCB. I chose C as I knew inrush would be an issue, but I slightly underestimated! The picture below shows the typical startup (yellow) and tripping (red) current/time. Borderline! I expect actual current is slightly higher than what I am measuring as the CT has a maximum response rate.
Capture 1.PNG
The D curve variant has a bit of headroom and should do the job just fine. Although it does increase the trip time and current for a real fault.
Capture2.PNG
In fact I could probably go to a lower rated D curve breaker and still be well under the line.
Post Reply