AC servo motor for CNC mill spindle: Questions and answers

[@kentvandervelden]

Check out the video on cooling the AC servo: kzbin.info/www/bejne/pJzde4CBhNqYmbc A table of contents of time code links to each section and question is in the video description. The previous two videos generated several questions and I encountered several gotchas with integrating the DYN4 with LinuxCNC. With this video, I'm summarizing the experience and the answers with the hope these details will help others. Thank you

[@jimsvideos7201]

Thank you for sharing your experience in such depth; a servo on the spindle for my little lathe is a reach goal for now.

[@kentvandervelden]

Glad it was helpful and, to be clear, I'm happy with the DMM servo. On a lathe, the inertia mismatch between the servo and load will be greater than on a mill. Not sure this so important if the servo is not used for positioning. Probably should try to get a servo with a larger moment of inertia to be safe. Regardless, whatever servo you go with, check with the manufacturer's tech people to be sure.

[@internationalprecisionengi1737]

"Why did this video take so long?" ahh... cuz... life! Kent, you're the man. I could never keep up with all that blog, instagram, twitter, etc. What I like most is your attention to detail and willingness to put it all out there. Fantastic work, yep now we know why even a Tormach is 10K with no options...

[@kentvandervelden]

Hi Jacob, thank you. I've mentioned this a few times but I don't know it comes through well. I just don't see CNC conversions, as they are done today as far as I know, as reasonable activities. There are many items that are discarded and the bones of the machines were never intended for CNC. Many times, one will be an expert in one area (machining, mechanical design, electronics, programming, etc.) but not in all the areas and all the areas are needed to do a design. So, one is left trying stuff out, hunting for others who have tried stuff out, etc. and piecing stuff together. People vary their time far less then they should so end up spending lots of time on a conversion instead of buying a turnkey system. Now, for education, I completely believe a CNC conversion is fantastic, without any doubt, but so are many other mechatronics projects, but a CNC conversion leaves one with a tool that's useful. Until there is a conversion-killer option, it's going to be useful for people to report honestly and accurately what's working and what's not. But that's all just my opinion :)

[@JohnK8]

I did a similar upgrade to my Sieg X3. I used a DYN4 1.8kw (no brake) controlled via 0~10v. I used all of the recommend filters and regenerative resistor. I geared mine with a toothed belt at 1:1.8 so that I could get a spindle speed of 5400 spindle rpm. I can also do rigid tapping. I'm using LinuxCNC with Mesa 6i25+7i76+7i83. I had a little trouble getting setup but DYN's tech support was very helpful.

[@kentvandervelden]

Hi John. I completely agree about DMM's support. They went out of their way to help. Did you you need to mount a spindle encoder or manage to use the drive's encoder output? I switched to a Mesa 7I76E and it works well, and while it has differential I/O and a higher sampling rate than the Pico USC, unfortunately the encoder output of the drive is still unusable. Bummer

[@MichaelT_123]

Nice, in depth presentation...with the pleasant English :)

[@kentvandervelden]

Thank you Michael :)

[@heeder777]

I purchased all of my CNC conversion from DMM just before you made the two videos on the spindle motor replacement. I’m happy to say my weeks of research and talking to DMM wasn’t in vain. Your videos confirmed I made the right choices at least in the spindle motor. The only difference is I ordered a separate braking resistor instead of the one built into the servo and I did go with the 1.0 kw instead of the .75, DMM recommended the 1.0 but there wasn’t a large price difference at that time plus I was buying all the other components at the same time which bumped me into a higher discount percentage which may have made a difference. Thank you for the great info. I subbed and wish I found your channel sooner. I’ll be checking out all your other videos to get even smarter!

[@kentvandervelden]

Your 1kW motor, does it also have a 5kRPM limit? If so, what's the model number? And you saw a discount percentage for buying more? That might be a new offer. Thank you for all the feedback. Always wonderful to hear advice from others.

[@heeder777]

@@kentvandervelden Hi Kent, It’s a A11-DST-A6HK1. I haven’t installed it yet, I’m still milling the mounts for the XYZ servos. I have started the enclosure build. I was told if I used a 2:1 pulley setup I could see 6000 RPM. I have all new bearings for the spindle that I’m going to replace when I do the conversion. I haven’t even looked into the pulley setup yet, I know I’ll have to bore and broach an off the shelf one. I like the toothed belts but I’m also on the fence about a grooved V belt. Being an old school machinist, I went to school in 1979, it is hard for me to jump right in since I can crank out most any part on the mill and lathe manually right now. BUT! I’m having a blast learning this as a hobby. Especially with the information available on yours and others channels. Regarding the discount, I didn’t know there was one until I added the spindle servo. The rep told me that by adding that to the order moves me into the next discount tier.

[@kentvandervelden]

Unless you see a large price difference, I would go with the v-belt or multi-grove belt. I used a tooted belt because I wanted to use the encoder feedback, but that was a big time sink. Another small pulley to an encoder from the spindle would have been easier for the setup I have. Assuming our mills are similar, to use the stock pulleys all you would need to do is bore out and add a key way. If you do want to go with a tooted belt, drop me an email, I might have all you need, I know I have extra pulley and belt from Misumi and can get you the details. There are a lot of creative people on YT. I feel like a hack watching others :)

[@rickojames]

Kent - Thank you for all the detailed information in these videos. I just placed an order with DMM for the unlisted 1KW NEMA 34 servo and DYN4 drive. The motor's max speed is 5000rpm, with 3000rpm continuous. The DMM rep said it'll run all day at 5000rpm, it's just that the torque drops off after 3000rpm.

[@kentvandervelden]

Pretty nice! You'll be quite happy with it. I wish they would have told me about the 1kW motor when I ordered. The belt and pulleys the mill comes with might be just fine. I used them for a while, they were simple, cheap, and fast to get going.

[@rickojames]

@@kentvandervelden I want to do rigid tapping and am concerned about slippage. Do you have the info of which geared pulleys and belts you used listed in any of your videos?

[@kentvandervelden]

The Misumi part numbers are GBN390MR3-150 (belt), GPA60MR3150-A-P35-NFC (spindle pulley), and GPA60MR3150-A-N14 (motor pulley). Totals about $65. The belt may take a over a month to receive. Drop me an email about an alternative.

[@rickojames]

@@kentvandervelden Kent - Thank you for the part numbers. Email sent.

[@platin2148]

I’m still not sure on the ballscrews i wanted to get pretty high quality ones but sourcing them seems to be quite a challenge i did look into tdk ones. For the servos i wonder if i should go mains or 24 volts one is nicer to setup but isn’t so safe and the other one requires lots of power supplies. A complete other point is that i want to still keep the manual way too so i considerd getting a pice that does a 90 and i would place the motor underneath the base not sure if that is a great idea. For the servos i looked at lenze and leadshine not sure if there are cheaper ones. I would also like to control everything with ethercat as it’s easier to setup at least i think it is compared to modbus etc.

[@kentvandervelden]

Thank you for sharing, that's going to be an impressive machine! At 24V response might be slower as overcoming inductance will take longer and thicker cable will be required for the higher current. Don't know if anyone would notice. Grounding well will create a foundation of safety.

[@stefanhertweck]

Wow, truely an in-depth coverage :)

[@kentvandervelden]

Thank you Stefan, you are very kind to be so consistently supportive :) A cornerstone of support

[@michaelbrocato7535]

After watching this video Im truly impressed ... subscribed right away

[@kentvandervelden]

Thank you Michael, I cherish gems of support like yours. Hope you continue to find value in my posts. Suggestions are welcome.

[@darkobul1]

I tried few motors on my g0704 and ended up with china servo motor 1.8kW. Price was close to what I have spent on used treadmill motor and driver. Motor is rated 3000rpm and I was impressed with torque on low speed compared to previous motors. Motor is flush with mill head so it does not look too large. I use quadratic A/B pulse signal to control it as it is most resistant to noise. At first I found problems with optocouplers not being able to generate fast signal of 500kHz but when I figure out that motor driver has opto isolation i did it directly and had no issues since then. I saw they started selling 8000rpm servo spindle motors but they are much more heavy and are not suited for this small machine. I am happy with it. But I didn't test limits of it as I don't look for record of speed and depth of cuts. Its sufficient for me that motor don't stall like other did and ruin my work.

[@kentvandervelden]

If you have a reference to the supplier, please include it here. I looked at manuals of a few of several servos on eBay. If all works out, they would be a great value. They have lots of additional features too. Your last three sentences especially sum up my priority and initial experience, so the real tests will come when I'm back to projects. Thank you

[@ItisImarko]

Good video!! you seems the right pearson to ask.. I have old lathe (small one) and want to make cnc .. and 1 question which i cant answer is that i want to use main spindle as deviding head.. now the question is will be ordinary servo good enough or do I need to buy servo with brake? can even be programed in linux cnc to use brake when you have souch operation ?

[@kentvandervelden]

Many thanks! Great question, there's a lot to it... Using a servo you could use the spindle as a dividing head. With rigid power transfer between the servo and spindle and good gearing you may be able to perform machining. You'll need to consider the holding torque vs. machining torque. The brake will help to hold position but likely with power no greater than the servo is rated for if matched. The brake does slightly degrade the servo performance and without a good gear ratio between the servo and the spindle you may already have inertia mismatch. This is not a terrible problem if the servo is strictly use for velocity and rapid exact position is not required. You can engage the brake from G code using a digital out, using M62-M65. Please ask if something is unclear.

[@ItisImarko]

@@kentvandervelden Wau that was fast:) I got the basic out.. you see.. Lathe is small a watchmaker lathe.. very accurate one ( tos 80) Not that todays čing čong kwlity. While reading your answer I get an idea of quick modification. Brake on servo is 2x the price than without a brake.. so I will use rearly machene and brake on servo dont justify the price ...but a new mount for servo needs to be made and why not build a part with small fly wheel and rig up there a breake.. simple and efective- and cheap. Original motor is 330W and there was model with 550W and im little worried that 750W will be overkill on that smal machene. Thank you for answer!

[@kentvandervelden]

@@ItisImarko The flywheel would have an electro-magnetic clutch? While I said inertia-mismatch is not as serious with non-position mode, adding a heavy flywheel might push the mismatch too large. Might require some experimentation. I like small machines... we don't cry as much if the experiment does as hoped.

[@ItisImarko]

@@kentvandervelden Ja correct el breake.. But i think that breake in servo will be more than enough to hold position while you will make small gears like very small. Just justifay the double price. Thank you nfor answers!!!

[@kentvandervelden]

@@ItisImarko Should work really well. You may check Alibaba for a stand alone brake. I purchased a NEMA34 brake recently for ~$110USD. It mounts to the front of motor, and there are a couple couplers for attaching the shafts.

[@ScotY808]

I watched the entire video and understood maybe about 10% of it.😂 I kept hoping that you would explain how you got rigid tapping to work. To refresh your memory, it was working fine for me (Centroid Acorn control) but the motor would do a weird fwd/rev movement when reversing out of the tapped hole. You seemed to have got it to tap without that weird behavior.

[@kentvandervelden]

Only 10%?!?! I didn't do my job well :) I've tried so many different methods and had such varied results I don't remember all the combinations that create that you describe. The most recent one though is if the motor is commanded to a particular speed over the serial port, then commanded to stop (speed = 0 command because there does not seem to be a stop command), while the motor is coasting to zero following the acceleration profile set in the motor's drive, the enable line is toggled off, all looks OK. But, when the motor is enabled, the motor will continue decelerating (which means it snaps to the speed it was at before being disabled and continues coasting to zero) even if a command to run the opposite direction has been sent immediately after the enabled. I remember us talking about this but I don't remember if you said you use the enable line... If so, perhaps there's something there? Give the motor time to ramp down to a stop or increase the deceleration time? Given the problems I've had, and that the serial protocol does not seem to have them, I'm switching to the serial protocol and I may not be able to rigid tap because the encoder signals are so full of noise. Once I have a working mill again (down for nearly three months) I'll see if I can make the encoder pass-through work better. Thank you

[@banefsej]

am nat expert or any thing but i love the details in your vid . i feel like i learn something and worth the time . gj and can't wait to see next video . keep going

[@kentvandervelden]

Hi Habeeb, I really appreciate your words. I'm no expert but I work hard to learn and sharing is a great way to organize the thoughts And meet others. I wish we had more hours. Peace

[@SteelCogs]

I'm only 15 minutes in, but so far it seems like you've covered a lot of the issues I've had with the same motor/driver (I don't have the brake). I've never been able to get the RPM output to match what it should be from 0-10v and I don't think I've ever got it tuned correctly without getting some kind of overvoltage error from time to time. However, I never went through with using a line reactor or noise filter, so I'm wondering if that would maybe solve some of my issues. I've never been able to get the encoder to talk to my Masso controller either, but I think that's more the fault of the Masso than the servo encoder.

[@kentvandervelden]

Thank you for confirming that I'm not alone. Like Fox Mulder, "I want to believe", but I can't imagine everyone out there going through the effort to workaround some of the problems we've had. The encoder signal might be too noisy. On the USC board I have, I believe the encoder lines go right into the FPGA and the spikes that I measured could damage the chip. I have to assume that my drive has problems and other copies are fine.

[@stephan9976]

Thank you Kent for sharing your experience.I think it's so great that you finally have an electrical control cabinet.Great Your path shows exactly the path of an electrician.You always have to be able to measure everything.I admire that about you.You can improve everything on the electrical system.Great.Also how you set it all up in Linux.I just look there and have no idea how you do it. You know I actually did exactly the same thing only with Mach3 and a slightly larger servo. But who I look at all you did to get it to work in Linux.Incredible.I would never get it done.My electrical setup took 1 day maybe 2 days. I think you are a very smart man.That the installation was so difficult(linux installation) for you shows me that Linux is not for me or Michael.Too difficult.It's unbelievable to me how you did it all.I'm surprised that your pm 25 can't make coffee yet. It was just a joke. After all the work I congratulate you on the good servo.I don't want to give up mine anymore.I wish you a lot of fun with yours. Now you only need linear guides?They would be good.Then your workpieces will be even better.Let's say that:""""you can sell them""" I wish you all the best best regards Stephan

[@kentvandervelden]

Hi Stephan, thank you so much for all the support and kind words! This was ridiculously difficult and maybe there was just something wrong with the drive. DMM stopped replying when I suggested that. So, were the difficulties because of Linux, the particular drive, my breakout board, me? I don't know. I am so glad you didn't have the difficulties I had. My previous mill had a 1.5kW 3-phase induction motor and VFD. That was super simple to setup even though the VFD had many more settings. I've written a serial interface to the drive and that seems more reliable. I'll demonstrate that soon. Just send a serial command "run at 500RPM" and it just works. It'll never work for rigid tapping, the feedback from the drive is a mess and there are pulses lost at even low RPMs, doubled checked against the drive's own position information, but that's OK. Today, spring is near, I just want a working mill, and to leave the basement. Have a great day my friend :) Kent

[@kyleanderson3442]

Excellent documentation on the DYN4 spindle setup for the PM25! I'm currently researching AC servo's to replace my stock BLDC setup and found this info to be very insightful. Regarding the noise problem found coming from the drive, Could you attenuate the 20KHz noise by using an inline RC style low pass filter installed on the encoder common to filter the noise causing the encoder error?

[@kentvandervelden]

Thank you Kyle, glad this was useful! I can't comment on using the filter. It's worth a try. The best mechanical and electrical solution is likely to use a second encoder. Some time has passed since this video. I have enjoyed the DMM servo. There might be a 5kRPM 1kW special order available, which I would get if starting over.

[@iainsquelch6223]

Kent. Fantastic information. I have been following your PM25MV videos as I am in the process of modifying mine. Questions: Did you consider replacing the spindle itself and/or the stock spindle bearings on the mill in an effort to improve performance for surface finishing, or generally the performance of the mill? Did you have any thinking on that to share? Also, I see guys replacing the stock spindle bearings to angular contact bearings when chasing higher spindle speeds. The AC bearings being rated for higher speeds than roller bearings. How fast have you been running the stock bearings in the mill and what can you say about their performance. Thanks so much.

[@kentvandervelden]

Hi Iain, we all make do with what we have :) The people you mention are honestly more skilled in the mechanics than I am. If you love the conversion process most of all, both ideas you mention may be worth pursuing. Some people use very fast 20k RPM spindles with integrated motors or BT30 cartridges with a big external motor. These are often inspirational conversions. Sometimes functionality is lost because of the specialization they are seeking. Before getting into that expense and time like that I would thoroughly investigate purchasing a turn key machine. One of the best PM25 conversion threads on CNCZone is from Starleper1 and he states the bearings on his machine are rated to a minimum of 5000RPM. The particulars may have changed since then. With this reference, with no intention of going beyond 5000RPM, with no measurable play, and with minimal heat, for me, little may be gained by replacing the bearings. The splined shaft has had a greater negative impact on surface finish. www.cnczone.com/forums/benchtop-machines/101231-pm25mv-bf-series-9.html#post859477 Best wishes

[@joat_mon]

Hi Kent. Great information. Thanks for the video. I'm using an ABS enclosure, but planned right from the start to place any spindle motor driver/VFD in a separate metal enclosure in hopes of reducing EMI/RFI noise. I know what you mean about finding a reasonably priced enclosure. The large metal cans sure are pricey. I was looking at wall mount server racks a while back. Maybe 6U to 12U? Might need to mod it a bit, but volume per dollar they seem attractive, and mostly metal. Do you have any experience with wall mount server racks?

[@iainsquelch6223]

Hi Kent. Can you comment on your thinking in selecting the 0.75kW servo over the larger servos that DMM offers? The 1.0kW and 1.8kW variants are more costly (although their marginal cost isn’t especially large) but they do provide considerably more torque than the 0.75kW motors, albeit with a maximum speed of only 3000rpm. What do you feel you traded off, with respect to the performance of the mill, by selecting the 0.75kW motor over its 1.0kW and 1.8kW counterparts? I am considering a similar upgrade for my PM25MV I am having difficult time framing my thinking on the motor selection. For me cost is a consideration although it isn’t clear what the benefits could be realized with the additional cost of the more powerful motors. My feeling is that the benefits may not be realized because the mill itself may not be sufficiently rigid to perform operations that require the torque they can deliver. Although that just my feeling.....which I’m looking to substantiate. Any insights would be appreciated.

[@kentvandervelden]

Hi Iain. I wanted more torque, speed, and measurements. I managed to get these with ok success and caveats. DMM has a special order 1kW 5000RPM motor for about 50% more. Starting over, I'd consider that instead of their 750W, larger if I could find one that still runs at 5000RPM. I got the 750W because I only had 120v, but then found later that 240v is needed to go beyond 3000RPM. Consider installing a 240v service if you don't already have one to increase your choices. I have decent evidence from drilling that to match the max torque of the original BLDC, the 1kW servo is needed. However, even the 750W exceeds the original BLDC over a wider range of speeds. The next question would be so the servo be a DMM... If just trying to turn the tool, they are probably fine. If trying to do the fancy stuff I tried, looking for a drive with more functions, or looking to save some money, I'd investigate the eBay options. Hope this helps

[@MarioGarcia-yd5yj]

Hello friend, thank you for taking the time to upload videos that help the community to carry out their projects and make better decisions to be able to choose well when buying a component in this case between one engine and another, I would like to know if you could guide me on how to connect a 3.7kw servo / spindle that I have acquired in the Chinese pages, my idea is to connect it by the step / direction mode to obtain a better performance I have a servo motor manual in case this help, my question is; do I have to sacrifice an axis to be able to connect my servo controller? Yes that's how it is; The manual asks me to configure a frequency of 500khz. Is there a card that is capable of sending that frequency? I would greatly appreciate any guidance you could give me, in advance thank you very much and congratulations for your good contributions

[@kentvandervelden]

You'll need to have an additional axis if controlling the servo with position control, whether the mode is step/direction or a/b phase. The pulse generation with the Mesa 7i76e is 10MHz, and that's a LinuxCNC card. I had to use position servo control, instead of velocity control, to get good performance but you may not need to. Velocity control should have worked well for me, but something prevented it. Thank you for the kind words of support. I wish you the best of success!

[@MarioGarcia-yd5yj]

@@kentvandervelden I appreciate you took the time to respond, I tried to make the connection by sacrificing an axis to be able to operate M03 and M04 I appreciate your support, blessings for you and your family, greetings from Mexico

[@durangotang1681]

Hi Kent. I’ve been a big fan of your progress so far. I have a question. Last year I purchased a 4 axis closed loop kit from Longs motor as my intentions were to use it on a cnc lathe project. The specs are 2.2nm 4 amp holding torque. Think that’s 313oz inch. Rotor torque is 480gm cm2. I changed my mind to converting the pm 25 using Arizona vids ball screw kit and figured those motors would just be reserved for the lathe, until I went back and noticed you are using what appear to be similar sized motors. I was really surprised that you were using one for z. Do you think the motors I have will get me by for now until I can afford the clearpaths? Thank you in advance.

[@kentvandervelden]

A huge thank you for being interested in the progress :) Your motors may have higher specs than mine. I have an early video when I monitor position error while moving the head at 100ipm or so with an extra 40 to 50 lbs. These are servo loop stepper motors. The video is titled something like Pm25 conversion test. So, there's evidence that small motors may work fine. The stock spindle will give out before the steppers fault. The lower power may have prevented tool breakage. Without a doubt, David's kit is worth the cost. I considered the Clearpaths but the cost climbs as options are added and I couldn't balance the cost with advantages. If you have the kit and the motors can easily be mounted, might be great to get some experience to find out what the best area for improvement is.

[@durangotang1681]

Kent VanderVelden Thank you for the quick reply. It’s good to know and makes my day knowing I won’t be waiting as long to get this thing up and running. That allows me to afford centroid acorn now and a good vice. Thank you again.

[@simonconvey9645]

I'm setting up exact same spindle on my PM25. Based on you videos, and other research, I figured that it was a good option, and if for some reason it doesn't work out. I have a small lathe that I can use it on. One question.... Did you find a good source for poly-v J series pulleys ? I'm whaving trouble finding one .... Or did you bore the original pulley ?

[@simonconvey9645]

Duh, I watched the video again, and you made up a set of timing pulleys to suit....Oh well, I'll macine up some pulleys...

[@kentvandervelden]

The original pulley was bored up a couple mm. The servo works really well, a great upgrade from the BLDC. I wouldn't get the brake option and I especially wouldn't mess with the drive's encoder feedback (HUGE waste of time.) So, that means adding an external encoder with a small timing pulley and probably keeping the original two-speed pulley and belt setup. Check this video out for a very cool belt selection system. kzbin.info/www/bejne/oYjCaXWAgdx9gpI

[@BondmeJesus]

@@kentvandervelden im trying to hook up the encoder to the acorn. I got the 1.8KW servo I hope it works ok. you say it's a waste of time to even try? I ordered 2 of the 1.8KW servos one for my g0704 and one for my g0602 and I ordered 2 db9 screw in plugs ans 2 db25 screw in plugs and 2 acorn boards and ive had so much problems with other BOB's that it's time to go with something rock solid, and I thought this was the way to go. the 1.8KW servo works awesome on my g0602 its so quiet with your multi rib v belt 3 pin spindle setup, I love it, I made the pulleys myself, and they are really really close to the same size and I can tell when i put an index mark on them and turn the motor on, and see how long it takes for the marks to come un-alligned from each other. i really hope the encoder works, I dont want to add an external encoder to it. still need to figure out how to hook the encoder to the acorn. too many years of trying makes jack a dull boy LOL

[@kentvandervelden]

@@BondmeJesus For sure try using the DYN4 encoder passthru. If it works, you have a very clean setup. In my case, I had lots of trouble. Maybe the actual advice is, give it a go, but if it's not work, don't keep trying to make it work. Wishing you the best of luck!

[@BondmeJesus]

@@kentvandervelden not sure what the DYN4 "passthru" is

[@SteelCogs]

I've come back to messing with my mill again and was attempting to drill a 3/8" hole in steel and the calculated HP was .446, which is only half the rated HP of this motor. Unfortunately it stalled (even though it seemed to be making a pretty good cut!) and I remembered I had this issue in the past with endmill cuts as well, even in aluminum. My question is, how did you tune your servo to work ideally with your spindle? Using autotune seems somewhat useless since there is no cutting load on the spindle/motor as you tune it, and the same would apply even if you were tuning it manually.

[@kentvandervelden]

Sorry to have missed your question. DMM does not recommend auto-tune for spindles. Are you using velocity or position mode? I had poor performance with velocity modem despite tuning attempts, but position mode worked (for a spindle) out of the box. More power is needed just before breakthrough, I think because of spring back and tool grab following deformation. Assuming that's not what's happening, since some time has passed, you may have solved this. If not, let me know and I'll get you the drive parameters.

[@Ale_Lab]

Hi Kent, I have watched all your videos. I didn't find it mentioned but I want to add the power meter for the spindle of my CNC (BF20L modified to R8 spindle, belt drive eccc so similar to your mill). I want to install a 1.8Kw AC servo spindle. In case of AC motor where would you measure the current, power flow? Before the drive or between the drive and the motor? I am not sure where the 1.8kw are referenced at, if for the current supply to the driver from the wall outlet or from the driver to the actual motor. Thanks for the help!

[@kentvandervelden]

Hi Ale, this is a great question and has several answers of various accuracy and complexity. The sort answer: For a simple relative load meter, an amperage meter clamped to one wire of the input should work fine. A max-hold or average function on the amperage meter would be nice. More complicated... For the DMM DYN4 and LinuxCNC, I wrote some code to get the current assigned by the servo. This has very fast response and must be filtered to read. For a hardware meter, you could measure the current either at output or input of the drive. Measured at the output will have quicker response but that might not be useful. Measured at the input will be smoothed and easier make sense of. The complexity comes into this if you want to measure power (will need to multiple by co-measured voltage), care about potentially imbalanced legs, 3ph vs 1ph, phase angle, and on and on. Hope this helps. I'm making a video on a new spindle now and after that plan to make one on instrumentation which should be relevant. github.com/kentavv/kvvcreates/blob/master/linuxcnc/modules/dmmserial

[@Ale_Lab]

@@kentvandervelden Thanks Kent for the quick reply. Yes, indeed measure DC is easy but AC can be done in different ways and give false result if you do not understand it properly. I will then wait your video to shed some lights on the subject :-)

[@kentvandervelden]

@@Ale_Lab Sorry, I didn't fully answer. The power rating is of the motor. The drive will limit power delivered up the rating (plus maybe some boost.) Output power = Input power x efficiency. For a relative rating, either place should be fine, but efficiency depends on load. So, for practical machining operations, an amperage clamp on the inputs should be fine. I used a wired-in amp meter for a long time. Oh, and some drives have an analog output for reporting current that could be pretty convenient.

[@Ale_Lab]

@@kentvandervelden Thanks a lot aI will have a look at the driver then. If the analog output is present then I will use that one instead. Otherwise the imput will be plenty for my application.

[@jaseng7697]

You can change the Speed mode 0-10v via the gear number in the DYN4 setup. Basically the gear number is what speed you want at 10v. So if you are running 0-5v for 5k RPM, then you would set the gear number to 10,000 as at 5v it would be 5k RPM. I don't think that was very clear in the manual, but I have a tech bulletin that explains it better. I also use torque filter to set the gains higher (it's a give and take between them) and of course low integrate to give a longer response time for position to catch up. Sorry, I have just not had time to put out a video on my 1KW, 5k RPM spindle, but I did get it installed and tuned....it's a beast. I think the R8 spindle is just too much for the 750w. My autotune wouldn't work with the 750w, but works great with the 1KW.

[@kentvandervelden]

It's great to hear you had an easier time at this. I hope no one else has the difficulties I experienced. I didn't have much success with the torque filtering vs gain settings. The documentation is so-so. Including pseudo code of the control loop the meaning of the parameters would be clear even if the English description is not. DMM said autotune was only for positioning applications and gave recommendations to start with. Apparently I have the old firmware to which some differences might be attributed. I'll document the serial interface to wrap up this series of videos. Have a wonderful week!

[@jaseng7697]

@@kentvandervelden Yeah, the Autotune does just adjust to position mode (even if you pick speed). But I guess my point on that was it always faulted with the 750W and actually made it through on the 1KW. So I think the 750w just wasn't a good match with an R8 spindle (at least not with the large pulleys I used). Yeah, all the current DYN4 come with a new higher current firmware. It's worth swapping it out.

[@kentvandervelden]

@@jaseng7697 When you run autotuning, does the always motor move? Out of the few times I've tried autotuning, only once did the motor obviously turn. All other times, the same tuning duration was required but the motor did not turn. Given we have different firmware, who knows. Thank you

[@jaseng7697]

@@kentvandervelden Yes, it moves for each tune test and ramps up within each test until it gets noisy.

[@traviswetmore7366]

@@jaseng7697 I realize with different machines, different gain settings will be required, but can you share what worked with you? I keep getting an over-current fault. Im running a stock taig spindle (for now) with a 1kw DMM DYN4. I have the motor and spindle connected via a timing belt (for now). Im trying to figure out where to start the tuning process at. Also, what firmware do you have?

[@duediligence791]

The induction motors on my fadals all have encoders. I’m sure where a will and way is concerned even a less expensive motor could have it retrofitted to the back side. Our robodrills are all servo, our fadals all induction. Frankly they all perform alike and the inductions have more torque. For your little mill I’d never use an induction motor for a few reasons you mention however depends on the application where I’d suggest one versus the other. Overall machining an induction motor tends to out perform however ours all have rigid tapping capabilities it isn’t common for a hobby grade machine. By the way if you decide to try and broach I believe you’re going to need a larger z axis servo. Most broaches will require 7-8NM of torque or more. Perhaps the weight of the head will assist but I suspect you will get an axis error if you try with the one you have.

[@kentvandervelden]

Thank you for this great information! My last mill had a 3HP induction motor and I keep wondering what-if, so it's very useful to hear your experience and recommendation. Also, I use the same axes motors and similar ballscrew pitch on my lathe and while broaching, slots definitely need to be narrow and processed slowly. I'm waiting on a MesaNet card which can sample the encoders faster and has built-in differential signaling. It's the last thing I'll try to get the DYN4 working well. Serial command of the DYN4 works well but that's not fast enough for encoder feedback. Btw, in school I a great friend shared your name, years pass without passing him though, so when I saw you message I wondered if he is now a machinist :) Best wishes

[@zulfqarali8603]

Very well presented, you did answered most of my question came up during watching the other videos. BTW: Which CNC controller do you with Linux CNC?

[@kentvandervelden]

Thank you very much. I have a simple i7 PC and Pico Systems USC running the latest stable LinuxCNC. I'll be trying a MesaNet board soon for a CNC router design using analog servos and will also try the MesaNet hardware on the mill. Let me know if this doesn't answer your question. Have a great weekend!

[@robertpeters9438]

Have you tried A dual VFD drive motor? High torque at low rpm.

[@BondmeJesus]

thank you for your videos i followed your belt drive 3 pin pulley and its wonderful absolutely quiet. now I'm trying to hook up the encoder to the acorn and what I don't know is where to hook the 5+VDC from the acorn encoder plug into the servo driver. what I do know is that on page 2.3.2 of the manual A+ is pin 20 A- is pin 19 B+ is pin 19 B- is pin 17 Z+ is pin 15 Z- is pin 16 GND is pin 1 and the only remaining pin is pin 22 and that is called OZ (Z open collector output) so I would thing that since that pin 22 is the only remaining pin; that it must be the 5+VDC for the encoder?????? oh man its always the last bolt you un-tighten that breaks off LOL on the acorn encoder plug it has pin 9=+5VDC pin7=A+ pin 8=B+ pin5=B- pin 6=Z+ pin 3=Z- pin 2=COM so it seems all so clear that A+ A- B+ B- Z+ Z- GND (COM) all have their places to connect to, but it's that 5+VDC pin 9 on the acorn that I dont know where to plug it in on the db25 connector on the servo driver, do you know if it plugs into the pin 22 OZ (Z open collector) please help, I am going crazy LOL

[@kentvandervelden]

The +5vdc supplied by the Acorn could power a standalone encoder. If you're using the DYN4 encoder output, you'll only need to connect GND between the two boards. Maybe pin 22 is on the Acorn? I don't know about that. Hope this helps

[@BondmeJesus]

@@kentvandervelden I'm using the 1.8KW servo spindle with the m110_06030 driver basically it's the same Chinese servo kit that is most common on ebay right now. this link shows someone trying to figure out the answer, and someone responds that they dont need to hook up the power to the servo driver. this is a link of what servo I have. and a link to the person asking the same question. I tried to connect only A+ A- B+ B- Z+ and Z- to the pokeys 57E M16D setup that I am using right now, and even though the connection was shotty and i didnt change the PPR from 10,000 down to 500PPR like what I have configured; it did show a fluctuating rpm of 13,000 and when i turned on the motor it went down to like 500 rpm. so i wonder if I am getting a reading even though i didnt hook up GND or 5+VDC but only A+ A- B+ B- and Z+ and Z- then maybe it supply's it's own power. I would hate to burn up the new acorn. I have been trying to get rigid tapping for years and just blowing all kinds of money its awful. www.ebay.com/itm/392911210777?var=661819044574 forum.linuxcnc.org/27-driver-boards/37608-wiring-a-chinese-servo-to-a-7i77-setup-for-yet-another-newbie?start=0

[@kentvandervelden]

@@BondmeJesus At first, tying GNDs would not necessary with differential signals. After all, it's the difference between the pair that matters for noise rejection. Further, the Acorn may use optical isolation on the input, increasing the allowed difference in ground levels. However, there will be a limit after which non-linearity and then damage occurs. Safest to tie the grounds, and attach the twisted pair cable shield at the receiver with a short pigtail. No need for the +5vdc. Some practical, some theory, hope this helps.

[@BondmeJesus]

@@kentvandervelden so connect only a+b+z+ and nothing else? I am not too good at understanding encoders much. I wish I knew someone who has hooked these Chinese servo encoders to the acorn. also I should check to see if acorn encoder plug is opto isolated? or hook up opto isolation? I don't really understand. I really don't want to fry another cnc controller I have a very limited income, and I spend every penny i get my hands on to make this stuff work so I will have something for my son someday if i ever get to see him again. the only thing i have going for me is not giving up. i may not be that smart, that's why not giving up is so important LOL if I cant figure out how to plug in this encoder, then I guess I will just get another encoder, and hook it up externally. it's hard when there is no information available and i've been looking everywhere and people just don't know how to hook them up to acorn that's all I find is that no one knows how. I love a clean shop because I hate spending 2 hours looking for a tool to do a 5 minute job. now that's how I feel about some things in this process. can you believe I got a 4.0 GPA for the 1st year of machine tool tech? and I cant even hook up an encoder. LOL

[@kentvandervelden]

Please send me a few pictures and manuals of what you have and I'll do my best to help. [e-mail] In general, if you have differential output and inputs, use them (to reduce impact of electrical noise), connect digital grounds grounds (to not blow receiver), connect encoder cable shield at receive (to reduce electrical noise.)

[@mitchellmayberry8126]

What spindle head are you using? Is there anything out there for under 800 $?

[@kentvandervelden]

I'm using the stock spindle and head that comes with the PM25. I recently purchased a BT30 with power drawbar and housing for $1400 where ~$500 is the heavy housing and shipping. Just make a spacer for between the Z saddle and housing. If you only wanted the BT30 spindle, they can be found on eBay and Aliexpress for under $800 and free shipping. Add another $300 for the air cylinder. These spindle are seemingly rebuildable, good up to several HP, and 6k to 8k RPM. If you only want high speeds, 18-24k RPM, and nothing under 9K, you'll have many options under $800.

[@Ryan-dz3jo]

Simpl fantastic, thank you.

[@kentvandervelden]

Glad the information was helpful

[@traviswetmore3601]

Any tips on tuning the DYN4 on velocity mode?

[@kentvandervelden]

Hi Travis, I wish I could help. I've not been successful in using velocity mode. When in velocity mode, my servo turns, but the torque is terrible. Increasing the gains much results in instability. DMM said that a belt drive, being less mechanically stiff than a direct drive, will be limited. Regardless, position mode has worked well. I can't explain the poor velocity mode performance. Unfortunately, maybe the drive is more sensitive to faulting under high load if in position mode than in velocity mode, but I don't know.

[@traviswetmore3601]

@@kentvandervelden That is actually really helpful. I have been struggling trying to get my 1.0kW DMM servo motor w/ DYN4 to work properly as a spindle in velocity mode. How are you implementing position mode? I too am running linuxCNC (using a Mesa 5i25 & 7i76). I currently have the spindle setup as analog and velocity control, but there is no torque to speak of... A screen capture of your final DMM Servo Settings would be wonderful if that is not too much to ask (unless they are what is posted at 5:15 in your video).

[@kentvandervelden]

That's very interesting! I attached the fourth motor channel of a Mesa 7i76e to the A-B phase inputs of the DB25 connector on the DYN4. You can find my configuration at the address below. Indeed, those at 5:15 are what I'm using, and I doubt they've changed. My DYN4 is from ~2 yrs ago, and there are newer firmware options, but the meaning of the parameters are likely the same. I wish velocity mode would work better because in position mode the servo faults when ~90 deg off, but in practice, it's not been a problem with a 750W except beyond 3/8" drills. github.com/kentavv/kvvcreates/tree/master/linuxcnc/configs/pm25mv_mill/mesa/configs/my_LinuxCNC_machine

[@blazini]

I would doubt that you're actually losing counts due to noise if the controller you use has differential inputs. The noise barely effects them because the way the differential receiver chip needs to be pulled in opposite directions to change the output. Only way to really see those missed counts in LinuxCNC is to set the spindle up as an axis and test positional movement over a certain distance. Velocity feedback isn't a good indicator because it's calculated over cycles. Even the cleanest velocity signal needs to be software filtered to look smooth. I't sort of a common misconception that LinuxCNC really cares about a spindle encoder for ridged tapping. I just so happened to be doing some tapping with my "worst-case-scenario" DMM setup and thought I'd take some quick video. kzbin.info/www/bejne/bp-lmayNabh8j7s I'm using the same DMM servo (minus brake) with step+direction upto 5000rpms. I had a bit of a fiasco with the DMM drive but I can comfortably say it was the terrible auto-tune thing they have baked into their software, I've actually had it kill a drive. Otherwise the drive has worked great for me, but I'd tell anyone asking about a DMM setup DO NOT USE AUTOTUNE!

[@kentvandervelden]

Thank you! Your video is useful and I left comments. Another couple explanations for my setup is lagging the encoder are that 1) the Pico USC is not sampling fast enough, though the documentation says it's good up to 1MHz; and 2) the opto-couplers don't have sufficient bandwidth, though at low speeds there should be no problem. After comparing the DYN4 internal encoder position, read by serial, to the encoder state sampled by the USC, I'm confident that pulses are missed even at low speed. At high speeds it's so obvious after a few seconds that no fancy stuff is needed :) Best wishes

[@blazini]

@@kentvandervelden If that is actually the case I can strongly suggest that your next upgrade be to Mesa hardware, you may also be able to ditch the optos. I pretty much use Ethernet cards exclusively, there really is no performance penalty because of the way the counting is done directly in the FPGA, and the hal driver only reads a position feedback at the servo-thread rate. Mesa claims 10mhz for stepgens and encoder counters, I've never seen anyone dispute that. I'd recommend either a 7i76e or 7i96, I use both on different machines

[@blazini]

actually yeah, a quick look at the Pico card and it is not using differential I/O for anything and I cannot for the life of me understand why not for the cost of the board. Stepgens and encoders are always differential on Mesa cards. If you look at the difference between rs232 and rs422 in terms of speed, it is exactly the same difference as single ended encoder inputs and differential. www.contec.com/support/basic-knowledge/daq-control/serial-communicatin/ You could likely find some to rs422 to rs232 adapter to use on the pico board side, to convert the DMM differential outputs to single ended inputs for the Pico at the card, or you can make one it's pretty easy. The DE10 nano PCB on my git has an example of of differential hardware for stepgens and encoders, just hook it all up with a piece of unterminated cat6 cable any you wont need the optos.

[@SmallShopConcepts]

sounds like you are finding every wrong way along the quest for the right one...lol most of your assumptions on why the surface finish is poor is incorrect. the spline drive separates the misaligned lower cartridge from the fixed upper bearings. the single set screw is sufficient holding the cartridge in place, but over tightening it will deform the cartridge so don't do that. the spindle and upper quill barrel should remain completely separated except for the spline drive contact points. this stops the misalignment and won't affect the spindle shaft from non concentric rotational loads ie.. force the spindle to bend as it spins which causes a significant amount of vibration. As a safety to the single set screw holding the cartridge in place, use a top hat on the threaded spindle shaft top that will not allow it to fall through the pulley (the pulley should ONLY be driving the upper barrel and not in contact with the actual spindle shaft. lastly, replace the stock bearing with Nachi Tapered roller bearings and use kluber grease. the timing belt adds a lot of vibration and your recommendation to use a timing belt given other options in my opinion is a poor one on this machine. you would be better served to drive an encoder off the spindle shaft with a very lightweight small belt that will not influence the spindle such as a larger main drive cogged belt would. I found the sweet spot for superior finishes and rigidity and minimal runout at the tool long ago and have been making gorgeous mirror like finishes for half a decade now on the G0704. so you may want to re-evaluate your design and approach to your perceived issues from a different perspective. it will help you achieve the end results you are looking for. happy to help if you would like, many i have helped are now happily making gorgeous parts with impressive cycle times on parts! Chris

[@kentvandervelden]

Hi Chris, I can't thank you enough for you taking the time to write a detailed recommendation which clearly comes from lots of experience. And thank you for offering to help. I may take you up on that. Now, I'm waiting on a MesaNet card. Much has been delayed. Stay safe and healthy!

[@SmallShopConcepts]

@@kentvandervelden going linuxcnc? \o/ have you seen the interface I developed and built on top of qtpyvcp in conjunction with Hazzy and Turboss, it's called probe basic and had a lot of really cool features along with an updated feel in layout and functionality! Let me know if you would like to join a group chat on Skype with several other high performing g0704 and mini mill owners. Shoot me an email I'd you would like!

[@mastermoarman]

next upgrade should be a tool changer

[@kentvandervelden]

If I had not already been working remotely, the last few months might have been more productive in the shop. Now, I'm without work, so we'll see what comes from the new found free time :)

[@johnrafols133]

May I know your FB Messenger, WhatsApp, Viber for my technical queries about servo motor application as both fast communication.

[@kentvandervelden]

Sorry, I don't have accounts on these platforms. KZbin, in case others will be interested, and email are the best options.

[@mikeissweet]

Gotta have high speed and twerk