[SWATH]

Quick update and progress report:

Circuit is working with some added safety features that prevent the spindle from starting on its own or if there is a bad cable connection. The motor sounds great with the new bearings but I noticed when it was running on the analog signal from the C41 that it sounded rough, so I scoped the 0-10v output signal and it was pretty ragged.
Attachment 215012

So I cleaned it up with smoothing circuit built with a bank of capacitors (which I may improve with bigger and more capacitors later if needs be):
Attachment 215014

This significantly reduced the analog signal oscillation and the motor sounds so much better and smoother when running.
Attachment 215016

Now the problem is the speed control. After much testing and experimentation I have discovered that there is nothing wrong with my VFD and it in fact does have a very linear frequency to RPM relationship. The problem is the C41 output. It is not linear through most of its range but it is really pronounced at the higher RPM range
Here is the actual motor speed vs commanded speed:
Attachment 215018

And here is output voltage of the C41 vs commanded RPM:
Attachment 215020

If you look the RPM matches the output voltage really well though they are straddling the linear reference lines at different gains. So I have sent a message to CNC4PC to ask how linear this output should be then I will try to get a better card or maybe even make one.

Still waiting on another db9 cable, the supplied one is bad and after countless headaches chasing phantom anomalies in my circuit I discovered it was the cable making bad and intermittent contact.

[SWATH]

Art thinks the problem is the 10hz PWM signal coming from the Mikini. The C41 only goes down to 300hz. I went ahead and ordered a PMDX-106 since it goes down to 5hz. Hopefully that will clear up that issue. Now suddenly the VFD won't recognize a reverse command so I have to figure that out or call tech support again on Monday. For those of you using the pmdx-106 can you tell me if the F/R relays can handle 24v.

Edit:
I fixed the reverse issue. I reset the VFD back to its default parameters and went through them again and now it works. Must have been some conflicting parameter I must have set at some point.

Now I'm just waiting for the pmdx-106.

Sent from my SCH-I545 using Tapatalk

[SWATH]

I told myself that I was going to get this machine fixed by the end of THIS year. I did it today and with one day to spare!

Behold:


I forgot to also mention in the video that the E-stop works and immediately stops the spindle when hit. The Spindle Halt will also be fully functional in the next revision of REX84 ver.2.0. I just mis-wired it in ver.1.0. This essentially triggers an E-stop and halts all axis motion in the event of any VFD drive fault, motor overloading, spindle stall, or over heating etc. Very useful provided the motor doesn't get overloaded from cutting air...like some motors I've had the misfortune of using.

Turns out part of the recent problems was the C41 card itself not producing a clean and linear output voltage. After consulting with Arturo, the likely culprit for the non-linearity was the PWM frequency. The Mikini outputs a 10hz signal whereas the minimum frequency recommended for the C41 is 300hz. So I bought a PMDX-106 which can receive as low as a 5hz signal.

I got the filter capacitance issue figured out for the RPM feedback but it was very tricky, partially because I uncovered something sneaky in the Mikini electronics. It has a tendency to try and display the real RPM as whatever the commanded RPM is set for regardless of what the actual rpm is as long as it's close to within a few hundred RPM. That drove me nuts trying to figure out.

The PMDX-106 produces a very clean analog voltage
Attachment 216264

and a very linear output with no sag at the upper range.
Attachment 216266

This corresponding to a very linear RPM scale
Attachment 216268

Allen I think you may have experienced a similar problem with the RPM being off at one end of the spectrum, this may the your analog output voltage sagging and not rotor slip.

I'm working on Ver.2.0 of the REX84 card
Attachment 216270

If anyone has any questions feel free to ask. If anyone is interested in the VFD conversion and wishes to make it seamless with the existing Mikini you NEED the REX84 card. Please let me know if you want one so I know how many to have made. I estimate they will be around $70

Here is a rundown of everything it does:
All function of the front panel is retained
-Spindle on/off from front panel or Mach
-Spindle forward/reverse from front panel or Mach
-Accurate speed command from front panel or Mach
-Real RPM feedback in both front panel and Mach (new feature as the old BLDC did not provide feedback to Mach)
-Both instant (dial) and average (LCD) spindle load metering on front panel
-E-stop immediately halts spindle as well as axis motion
-Spindle Halt alarm function retained. (stops axis motion should the VFD fault from numerous programmed events such as a spindle halt or overload).
-Status LEDs on card itself
-Digital isolation (better than opto-isolation) on VFD encoder signals.
-Uses existing db9 connector other connectors are included (such as for power, load, PWM/dir out).


The Rex84 is designed to work with a 1024ppr encoder for RPM feedback but it is not necessary. I used an Automation Direct VFD (GS3-23PO) with a 70Ohm 300W braking resistor but again these aren't necessary they are just what I used when making the card.

Hopefully my next videos will be of actual machining for a change and from my nice camera not my phone camera. Still trying to get a good editing software that does HD like adobe premier or something.

[slowtwitch]

That's great news !!!!! I want to see the video of this thing finally running

[millersson]

Swath,

Great job. I look forward to seeing chips.

[FannBlade]

Thanks for the update. Good to hear the PMDX board works so well. I will be using it along with the PMDX 107 for my Hurco build.

Great job,glad to see your nearing the end and is all working.....FINALLY!!!

[SWATH]

Another video of boring the 2nd pulley. The first was done when the spindle was tilted forward about .007@6in. out of tram. This not only resulted in a pulley not bored to the center of mass but also a somewhat poor surface finish. After mounting this pulley I noticed a vibration intensify as the RPM went up. You could hear it and see the monitor on the machine vibrate, looking at the belt as it tuned also displayed a wobbliness to it. I measured the TIR of the pulley to about .007. When I bored out the first one I indicated off of the center of the existing bore.

After tramming the mill to within 0.0005 or less in X and Y and verifying again before boring and tried it again on another pulley. This time I indicated off of the outside diameter of the pulley splines to insure the new bore would be centered to the body of the pulley (the important part). First with the 0.5 inch endmill then later with a the carbide insert. Both cuts went very well and the insert did not chatter as I thought would happen based on my previous experiences with it. So I can account my lack of ability to bore with the insert to the mill being out of tram. Also the cut just sounded soo much better like the motor was under no load whatsoever. With the previous motor you could hear the motor loading down even though it didn't stall. The TIR of the new pulley was about 0.001 or less and the vibration totally disappeared. I set a glass of water on the machine and ran the speed up to max, there was no perceptible disturbance of the surface whatsoever.

[SWATH]

Finally got a suitable editor and got part 1 made of machining the motor mount with the old motor. This was also with the old lighting scheme and pre-rain-xed Lexan so it kind of sucks but I'll work out camera techniques to make my videos better in the future. I keep forgetting to turn the damn auto focus off. There will be many more videos to come now that I can edit them but now I need to invest in a mass storage device like a 6TB external drive to hold all of this HD video data, all of my computers are full. Looking into that right now. Part 2, the motor mount base will be uploaded as soon as I can get to editing it. I also have a tramming video, the original pulley boring, and machining a key for the new motor rotor. Please subscribe to my channel, share my videos, and like them or whatever. There will be much more to come and it will be better. I'm still learning the ins and outs of youtube so criticism is always welcome. I'm continuing to post in this thread because I think it is good data concerning the spindle conversion and these are the last bits of machining done with the original BLDC motor and drive. I'll post some new videos of the new spindle cutting when I get a chance, but after that I'll start another thread for just video and cutting data.

[mcphill]

See if you can edit from the GoPro Studio: GoPro Official Website: The World's Most Versatile Camera

You can edit, speed up/slow down, etc. and it is free... Sorry if you like it now that you bought something!

Also, have/will you post the drawing files of the parts you have made?

[SWATH]

Thanks Mike but I didn't buy anything. A friend of mine in film school gave me premiere pro. It's very nice. Sure absolutely I'll post the files but I need to make a couple of modifications to it since I changed them. Mainly the back of the bottom piece I had to modify from the stupid measuring error I made. When I was getting measurements off of the spindle head, on one of them I accidentally wrote a 4.xxx instead of a 5.xxx. I modeled in on the bad dimension but ended up fixing most of it. As I recall there was a little more tweaking I had to do that I didn't update the model with. I'll look into it and post them here. What format? I use Rhino, can you import .3dm?

[mcphill]

Solidworks appears to be able to import .3dm. STEP is my preferred if it exports easily from Rhino.

BTW, I redid the "sardine can lid" for my Dad for Christmas. You may recall I posted about them in 2011. I got really in to making the mold and forgot that the measurements were all "guesses" (I had intended to go back and put in real dimensions and forgot)! So I made a 4 piece aluminum mold and cast the parts, and they... didn't fit.

Well, redid the mold this year, and it works great! Even seals air-tight.

So you're not the only one with "off" dimensions, trust me!!!

[SWATH]

Here is part 2 of the machining the motor mount with the original BLDC motor and driver:

[SWATH]

While running a program the other day I decided to take some temperature measurements of the new motor and the spindle. The previous motor would get very hot, too hot to touch even after a short program. I also noticed the spindle would get very hot as well. Sometimes I would just stop a program and let everything cool down awhile before continuing to cut. I had a program run with the new motor for a little over 40min straight at 5000rpm so I took readings every 10min with an infrared thermometer.

Attachment 217932

This motor run sooo much cooler than the old one and even the spindle is cool. I'm not sure why the spindle is cooler but maybe the heat from the previous motor would heat the spindle as well.

The program also had some 3D profiling and a ton of Z rapid retracts. There was not a single hiccup so I'm confident the weight of the new motor won't be a problem. Video to follow.

[SWATH]

Here are the first CNC operations with the new motor. I finally got around to editing it. It's in wood but I mainly wanted to test the Z axis motor to make sure it wouldn't stall under the added weight of the new spindle motor. I did some 3D contouring with a lot of rapid Z retracts and it worked fine without any hiccups.


The next video will be cutting titanium alloy

[slowtwitch]

It's great to see you finally making chips !!!!!! The mill sounds great..... fantastic job !!!!!! Can't wait for the metal cutting video If only Phil would have listened to us...he would have had a fantastic machine.

[mcphill]

That machine shop looks WARM!!

[SWATH]

Yeah it is! I put in an infrared heater. Is toasty.

Sent from my SCH-I545 using Tapatalk