![]() |
Tuning my Spindle
I have my 2.2 kw water cooled spindle up and running under Mach3, with full speed control, using the PWM method through the G-540. No major problems to report.
As to the tuning of this spindle, I would like to ask the spindle owners a few questions. 1) Is it normal to see rapid fluctuations of frequency(133hz min-400hz max)/RPM(8000 min -24000 max) of up to 3% +/-, on the VFD display when using PWM? e.g. Mach3 MDI code M03 S12000 = VFD display frequency 'dances' between 216-230hz with the median being 220hz most of the time. This is with the spindle idleing. I would think that the display values would be fixed and constant. 2) What are you spindle owners using for RPMs for most of your work? Would you use the full 24000 rpms for cutting MDF and Melamine along with the proper feed rates? Or would you slow down the spindle and feeds? 3) Has anyone had success using a 3/8" compression bit and cut through a 3/4" MDF/Melamine board in one pass, using the 2.2k spindle / 1ph power to VFD? Thanks in advance. |
1) Somewhat. It implies that you have some noise on the analog speed control line. You might want to consider adding a little bit of capacitance. Too much and you'll notice that you are limiting how quickly speed ramps up or down after commanded. Depending on how long and isolated the wire is between the g540 and the VFD, you might want to add a cap between the VFD voltage and ground at either end, or both. You can get it dead solid if you want to experiment with values.
2) Depends on the bit characteristics, and how consistent your machine is. You want to make sure you're getting good chip clearance. A down cutting bit in a given material may be less tolerant of warp speed than an up cutting bit; however, if you can manage to balance feed, speed, and bit strength, faster is better. If your machine has lots of little variances, it's tough to nail that perfect balance, because conditions change when you hit a variance. 3) Dunno. Expect someone else to chime in with a yes and some settings. |
3) How are you holding down your material if you are going to cut through with a 3/8" bit? I would start with cutting half way through and see what kind of results you get. If you are in a big hurry, you can try cutting all the way through. I use a 1/4 bit and I cut half way through.
|
|
Quote:
I'm sure 'noise' is the problem, as the VFD is the infamous Huanyang. Could you be a little more specific with the sizes and types of capacitors you had in mind? I should also mention that in addition to the VFD 'hunting' for the proper frequency while running, after a few tests changing speeds, The VFD fails at times to even come close to the requested speed after a 'S' code change?? |
Quote:
I guess I will try it and let everyone know. |
2.2kW is 3 horses.
My 3kW can not do 3/4" MDF in one pass. Barely do 1/2" MDF with 3/8 straight flute. Sometimes, it breaks the bit. |
Surely cutting depth is beside the power of your spindle also affected by your feed , speed and cutting tool. ( and the proper set up of your VFD )
I think you can cut 3/4" in one pass if you really need to, if you tune the parameters above. But even when it is possible or not ... you also would have to think if you want to ( unnecessary putting higher stress on bearings, tools, machine, finish on work etc. ). 'Shortcuts' are not always quicker in the end. The MM basic setup is also limited to what it can take in terms of side loads. So even with lot's of HP in you spindle you will run into the limitations of your guides and frame. That's why the big boys have BIG HP Spindles combined with superheavy BIG frames. My 3,5 KW will do 3/4" 'comfortable' in one pass, but despite I use multiple passes more often. |
Well I was able to clear up why I was having issues with the PWM 'hunting' frequency problem. I had my grounds crossed on the VFD. The PWM ground was on the digital side (very noisy!). Changed it to the ACM slot, cleared up the frequency.
Now for other issues. I get proper frequency rates on the low end of speed calls (S10000=168kw) and lower than normal frequency rates on the high end of the speed calls (S18000=289kw). Also, when I make a M03 S12000 call to start the spindle, the frequency ramps up to 268kw and stays there. If I resend the S12000 code the frequency drops to the expected 201kw. Both ends of the frequency range respond correctly with a S8000/S24000 call. Speed calls below the halfway mark seem to map correctly to the proper frequency. Above the halfway mark speed calls are below the expected frequency range. Any suggestions? |
What version of Mach3 are you using?
|
|
After more testing I find that 'S' calls between 18000 rpm-21000 rpm do not produce frequencies from 300kw-350kw. The frequency barely moves, from 289kw-320kw range. I'm in the US so 60kw is our input power frequency. A 'S' call at 21125 rpm causes the spindle to 'hunt', with low RPM at 19000 for a few seconds, then a increase of speed to 21300 for a few seconds, then back down again, and so on. Once I do an 'S' call above 21500 rpm then the spindle stablizes at a little above 21500. Top speed (24000) is always reached.
|
Try deleting the file "linearity.dat in the Mach3\macros\yourprofile\ folder, while Mach3 is closed.
|
Also, you need to measure the voltage from the G540 and see if it's changing, to see whether the issue is from The G540 or Mach3, or if it's the VFD.
Most people with those VFD's seem to use Modbus with one of the modbus plugins. |
Quote:
THe VFD is putting out 10v. The wiper is putting out voltage, but I only have a analog meter for DC and the range is 0-75v, so I cannot tell exactly the wipers voltage, except that it is there and changes when I change speeds. |
If you command S18000, is the VFD outputting 10V, and is the G540 sending 7.5V back to the VFD?
I'm wondering if the VFD voltage is fluctuating? First thing I would do is try a different version of Mach3. Looking back though, it may be that Brad has the answer in his first reply. |
Quote:
|
Ask Geckodrive, they have great support.
My understanding of that circuit is that it's simple Pulse Width Modulation through an optoisolator, and then integrated into a voltage with a capacitor on the output side. I can't think of any ways that circuit can be bad in just one voltage range. I could believe that the computer isn't outputting the right pulses in that range, although that seems pretty weird. Can you throw the voltmeter on the +10v output from the VFD and ensure that it stays at 10v all the way through the speed range? |
Quote:
A solid 10.04 vdc across the speed range. :) Any way to confirm Mach3 PWM send proper signal throughout the speed range without having to use special equipment? Sent a email to Gecko customer support. Awaitiing an answer. They have the G540 on sale for $255 with another discount at checkout that paid the shipping. I bought one just for backup. I got my G540 used from another MM builder who didn't follow through and sold his parts. I remember having issues with it from the start, and contacted Gecko. They had me send it to them and they took care of the problem, shipped it back, no charge at all. Pretty amazing, as I wasn't the original purchaser. I tried a few capacitors in the 0.1-.001 range, between the input and ground of the VFD speed control with little success. |
Sounds like the G540 swap will be the next diagnostic step :)
If you can get at the parallel port pin, you can try to see if it moves linearly from 0 to 3.3 or 5v as you go up the range. You'll probably have to add a capacitor to ground to get a reading on a digital meter. |
Marcus from Gecko returned my email today and had me test the G540 with the VFD off. This meant looking for another power source. I found a 9vdc power supply putting out 9.6 vdc. Fired up Mach3 and the G540 and read the voltages on pin 8 as I input different speed ranges. The output from the G540 was spot on, so no bad G540. Marcus thought that the VFD was robbing current and causing my problems. So, back to the drawing board.
Each of the below adjustments improved the speed response in incremental stages: 1) Added a ground line on the left bottom corner of the VFD in addition to the ground I already had installed on the VFD's power terminal strip(read that that cleared someone else's problem). 2) Added ferrite chokes to the signal lines going into the VFD(already had ferrite rings on the power cables). 3) The VFD has two locations on the two terminal strips to attach analog grounds. I originally had my ground attached to the front terminal strip and changed it to the back one. This restored the lost voltages between 6-9 vdc on the higher end, but also increased the vdc on the lower end? Must be magic!? :rolleyes: 4) Changed the minimum speed in Mach3 Pulleys from 0 to 8000 to match the lowest speed the spindle should turn. Changing minimum pulley speed changes the speed ratios in the VFD throughout the entire speed range? 5) A call out of 18k would get me 17,500rpm. I added a 68 pf capacitor between analog input and ground to increase speed to 18k, my sweet spot for most machining. So, now I have a speed range between 8k-24k, with the lower speeds being slightly faster than call out(100-200 rpm), 18k my sweet spot and the higher ranges (22k+) within 50 rpm. Speeds between 20k-22k are not tracking well. They start off low(19k) and jump to the call out speed and back low again. If I need to cut in this speed range, I will have to use the controls at the VFD. Except for the 20k-22k range, the speed is stable and within a few rpm of call out. Way better than when I started. A big thanks for the suggestions on what to do. I'll be getting that compression cutter next week and will report on how it handles cutting MDF in a single pass. |
Nice job Steve, and thanks for writing up those great diagnostic steps.
You've reconfirmed my opinion that those HuanYang VFDs leave a lot to be desired. I now regret having resisted ranting at you to replace it the instant you reported a problem. |
Regarding the 19mm or 3/4" melamine particleboard I've cut it with 2.2kw spindle with 8mm and 12.7mm compression single pass no problem, little on the slow side but after a z-axis mod a bit faster. My VFD says 60% load with a brand new sharp cutter at around 100ipm with a 1/2" compression (12.7mm) , don't remember the 8mm load because I used it only once or twice on melamine, most times on phenolic coated plywood and it had no trouble.
3kw would be prefered but for you to satisfy the manufacturers feeds/speeds 7.5kw or 10HP is a must for anything 1/2" or above on melamine. So ... I live by with shorter endmill life by using unapropriate feeds and get the job done when needed. I am sorry to hear so many bad experiences with HuanYang VFDs , I installed 4 or 5 of them and they are still in service around 3 years with no problem what so ever and giving all the same results as other VFDs. But I use Modbus to control them (USB-485 cheap adapter) |
I ran some 3/4" particle board with my new 3/8" single flute compression @18000 rpm @200 ipm single pass with no issues yesterday. Manufactures suggested feed rate for proper chip load would be 300 ipm. I will continue experimenting and see how close I can get to 300 ipm.
"But I use Modbus to control them (USB-485 cheap adapter)" Could you tell me about the Modbus method. I hear that the VFD is not compliant to Modbus, so a work around needs to be used. Also, exactly which USB-485 adapter did you buy, because it seems that some people have found problems using some of them. Thanks Steve |
This one
http://i01.i.aliimg.com/img/pb/598/5...589598_516.jpg you can use this plugin, just make sure the VFD is ON before starting Mach3 so it can communicate and find it on start http://royaumedeole.fr/informatique/...-huanyang-vfd/ |
Quote:
On our industrial router (Morbidelli), we cut at about 700-800ipm at 18,000 rpm with a 3/8" compression spiral. |
Gerry,
I would guess that you are using at least a 2 flute cutter, if not a 3 flute. This would necessitate a much faster feed rate than my single flute cutter. |
My 1/2" compression asked for 500 ipm on 19mm particleboard (16m/min or around 650ipm on 1x Diameter and lower for more depth per diameter) its something unimaginable for a 2.2kw spindle and mechmate
That single flute compression must look awesome :) |
I cut 19mm marine ply in two passes 6mm dia single flute compression spiral (Onsrud 60-152), first pass -12mm 17000 rpm 3000mm feed 0.5mm offset, second pass -19.5mm 17000 rpm 3000mm feed zero offset.
2.2kw watercooled, I can do this all day long. . I have also cut a lot of 12mm faced marine plywood with a single pass with above rpm and feed. Works the same on Birch ply, sounds different but same result. been running since july 2013 only missed a beat when I tried to edit a file mid point, re started and forgot to restart the spindle.... lol hit the estop, and lucky not the break the cutter, this is fun eh. Long may it continue. Have fun, Regards Andrew . |
Today I achieved 300 ipm cutting 3/4" two sided melamine, single cut and chip free! With the 2.2kw spindel @18000 rpm, single flute 3/8" compression bit. Ampere draw was 3.1. That is suppose to give me a .020" chip load. But charts say that .023" @ 360 ipm is also acceptable. So, I went back I tried a 350 ipm cut. Same results! I'm very pleased!
|
All times are GMT -6. The time now is 19:00. |
Powered by vBulletin® Version 3.8.3
Copyright ©2000 - 2024, Jelsoft Enterprises Ltd.