On Mon, Dec 2, 2013 at 2:40 PM, Tom Kerekes <tk@...> wrote:

 

Hi Andy,

I would hate to see you switch out servos for steppers.  Do you work with Hugh? 

We are working on a bug which is likely causing your Z Faults with particular GCode Arcs.  Do you have other problems?

SnapAmp is rated for a maximum of 1KW continuous.  This is well over 1HP.  Peak power is 2KW max.   55Amps at 80V is well over 4KW.

If your motor is a very low voltage motor then higher than normal currents may be needed and not be a good match for SnapAmp.

What type of system do you have and what type of problems do you have?  I would expect limited currents to work ok but somewhat limit max acceleration.

Regards

TK

Group: DynoMotion	Message: 8763	From: andysontag	Date: 12/2/2013
Subject: Re: 660 Watt BLDC Motor with SnapAmp

Tom, Yes, Hugh and I work together.  It sounds like the bug you are working on will make an improvement although the main issue we are running into is that the SnapAmp will fault and shutdown due to over current unless the feed speeds are set at a snail’s pace. We might be missing something but it appears that it is extremely difficult to prevent our motors from drawing more than 25 amps and faulting a SnapAmp.  In addition to the over-current faulting, tuning each axis has been difficult for Hugh. It’s my assumption that the reason tuning has been so difficult is because easy faulting has made the exercise more about preventing over-current than about optimizing performance.  Our motors are rated at a nominal voltage of 48 volts. They have a peak torque of 6.3N.m (892oz-in) at 55A. We have been running the motors at 56 volts and we’ll have the ability to drive them at 80 volts with the addition of a new power supply that will arrive soon. 80 volts x 12.5 amps is indeed 1kw, although it's my understanding that current is what creates torque. The torque constant for our motors is .117 N.m / Amp, and therefore at 12.5 amps our motors will only be able to output 1.46N.m (207oz-in) of torque. If we were able to reliably limit the current output to the motors at 25 amps we could potentially have a peak torque of 2.93N.M (414oz-in). At the moment, getting so close to the 26 amp fault isn’t possible. 207oz-in of torque appears to be lower than what is required to move our cnc router (CRP PRO4896) at a reasonable pace with reasonable performance.  The stepper motors that we are considering offer a peak torque of 920oz-in with an inductance of 2.4mH, and current of 7A. The stepper motors are attractive because it eliminates the possibility of faulting the SnapAmps due to over-current, and theoretically will provide more usable torque and higher feed rates. Andy ---In dynomotion@yahoogroups.com, <tk@...> wrote: Hi Andy, I would hate to see you switch out servos for steppers.  Do you work with Hugh?  We are working on a bug which is likely causing your Z Faults with particular GCode Arcs.  Do you have other problems? SnapAmp is rated for a maximum of 1KW continuous.  This is well over 1HP.  Peak power is 2KW max.   55Amps at 80V is well over 4KW. If your motor is a very low voltage motor then higher than normal currents may be needed and not be a good match for SnapAmp. What type of system do you have and what type of problems do you have?  I would expect limited currents to work ok but somewhat limit max acceleration. Regards TK Group: DynoMotion Message: 8765 From: Tom Kerekes Date: 12/2/2013 Subject: Re: 660 Watt BLDC Motor with SnapAmp Hi Hugh, I looked at those links and it appears the Z axis is designed for a Stepper.  Steppers are fairly low speed/high torque devices.  Servos are usually higher speed/lower torque. Is that Z axis a 5mm pitch lead screw direct coupled to the motor?  If so something like 3000mm/minute speed would only be: 3000mm/min /60 = 50mm/sec = 10 rev/sec = 600 rpm This is pretty slow for a 3000RPM motor.  Compounding this is the low voltage motor.  At 600 RPM the back emf is only: 0.6 x 11.5 = 6.9V If the motor was wound to require 69V instead then the same torque and power could be achieved with 1/10th of the current. This motor also has very low inductance 0.3mH.  The current ripple might be high with a 50V supply.  Increasing the Supply voltage further would increase top speed rather than torque and probably not help or be a good idea. Any way you could add something like a 3X belt reduction on the motor.  What speeds do you require? If you wish we could arrange a time for me to connect in to your system and check the tuning. Regards TK Group: DynoMotion Message: 8786 From: andysontag Date: 12/9/2013 Subject: Re: 660 Watt BLDC Motor with SnapAmp Hi Tom, Rather than switch the snapamps out for high current drivers we decided to order 4 nema 34 stepper motors which should work well with our current setup. They'll have a rear shaft so we can use the same encoders we have been using on our BLDC motors.  The new steppers should allow faster acceleration and faster overall feedrates although we'll still need to control the Z axis at its own feedrate instead of defaulting to the x/y feedrates. Have you had any luck tracking down the Z axis feedrate bug? Andy Group: DynoMotion Message: 8787 From: Tom Kerekes Date: 12/9/2013 Subject: Re: 660 Watt BLDC Motor with SnapAmp Hi Andy, That sounds like a reasonable plan. The Z feedrate bug was more complex than I thought.  There were several things that were not being handled correctly.  The issues mainly showed up with different axis constraints (velocity/tangential acceleration/radial acceleration) in different directions.  When traversing a large angle arc/helix the directions can change a lot.  We need to honor the worst case constraints throughout the arc.  We also added an option to replace arcs with many multiple line segments within the specified Collinear Tolerance.  This allows more optimized motion through the entire arc as the constraints vary.  See the description here: http://dynomotion.com/Help/KMotionCNC/ToolSetupScreenTP.htm#Trajectory_Planner The new Test Version here: http://dynomotion.com/Software/KMotion431p.exe Please try it and let us know how it works for you. This Version also includes support for attaching Konnect IO boards to either JP4 or JP6. Regards TK Group: DynoMotion Message: 10844 From: bknighton28 Date: 1/15/2015 Subject: Re: 660 Watt BLDC Motor with SnapAmp Was this ever resolved? I am building the same router and was trying to avoid steppers too.  The default way seems to be steppers and geckos but I really prefer servos and snapamps from previous projects.  Maybe I should be open to redesigning the belt drive for more reduction. But if there is a way to get steppers with feedback working well using snapamps that would be great. We're you able to? We're you able to get the high Rapids of 1200ipm like the recommended motors deliver?    I am trying to find a dual shaft stepper that resembles cncrouterparts stepper.   but I'd much rather servo.  Group: DynoMotion Message: 10855 From: Andy Sontag Date: 1/17/2015 Subject: Re: 660 Watt BLDC Motor with SnapAmp Hi Bill, We found that the snapamps weren't able to provide enough current for our BLDC setup. You might be able to get by with snapamps and smaller BLDC motors, but if your goal is to get BLDC motors with similar torque as a Nema 34 stepper you'll need to use a driver other than the snapamp. The main issue we ran into was that the motors had a max current that was well above what the snapamp could provide and would thus trip the snapamp with regularity. Our fault for not knowing any better. We decided to switch out our Nema 34 BLDC motors for Nema 34 steppers from PrimoPal. We contacted them directly and requested a custom dual shaft version of their 7-amp PHB86S150. The new motors and snapamps combined with around 73 volts makes for a powerful, quick and stable machine although our rapids are only set at 600ipm. We have some work to do on improving rigidity (our kit was the original, not the current version) and when done we very likely will be able to increase our rapids, but 600ipm is plenty fast for our needs.  We have plans to build another large cnc router although this time it will be 5 axis. Going the BLDC route is tempting although for simplicity and cost we'll probably go with steppers again. If you solve the BLDC driver issue you'll still need to provide incredible amounts of power in order to satisfy the BLDC's max current draw (in order to get max torque you'll also need to provide max current). When setup correctly the snapamp + stepper combo works really well and very smoothly.  If you end up going the stepper route you might want to consider using glass scales as your encoders. You might not need them if you have ultra premium hardware but for a cncrouterparts kit they should allow you to achieve very good accuracy and repeatability. Not just repeatability. You also wouldn't need dual shaft steppers. Otherwise I recommend the CUI AMT11 encoders. These were released recently and are a much improved version compared to the AMT10 encoders. They have higher resolution, better accuracy, and the index pulse works with stepper motors (essential for high accuracy homing). They have some other features but they work well and can be purchased from Digikey for cheap. Good luck. Andy On Thu, Jan 15, 2015 at 11:00 PM, bill.knighton@... [DynoMotion] <DynoMotion@yahoogroups.com> wrote:   Was this ever resolved? I am building the same router and was trying to avoid steppers too.  The default way seems to be steppers and geckos but I really prefer servos and snapamps from previous projects.  Maybe I should be open to redesigning the belt drive for more reduction. But if there is a way to get steppers with feedback working well using snapamps that would be great. We're you able to? We're you able to get the high Rapids of 1200ipm like the recommended motors deliver?    I am trying to find a dual shaft stepper that resembles cncrouterparts stepper.   but I'd much rather servo.  Group: DynoMotion Message: 10857 From: Hugh Sontag Date: 1/17/2015 Subject: Re: 660 Watt BLDC Motor with SnapAmp We resolved the problem by replacing the servo motors with NEMA 34 1600 oz-in that we had made for us by a company in China. We had the company put a rear shaft on the stepper for a CUI encoder. As I recall, 4 of these motors cost around $400. Once we built a custom optoisolator for the encoder signals, we had excellent performance and no hiccups. Without optoisolation, there was significant noise on the encoder lines. The servo motors we bought were a poor match for the Snap Amps. The motors required too much current, much more than the Snap Amps could deliver, so the available torque from the servo motors was insufficient. I recommend you use stepper motors and Gecko or Leadshine stepper motor drivers. Our installation with Snap Amps sometimes has a problem which appears to be that the KFLOP is unable to correctly communicate with the Snap Amps. When this happens, the control of the stepper motor fails and the KFLOP and Snap Amps have to to powered off and back on again. This happened with two completely separate sets of hardware, so I believe it's inherent to the design. Hugh Sontag On Fri, Jan 16, 2015 at 1:00 AM, bill.knighton@... [DynoMotion] <DynoMotion@yahoogroups.com> wrote:   Was this ever resolved? I am building the same router and was trying to avoid steppers too.  The default way seems to be steppers and geckos but I really prefer servos and snapamps from previous projects.  Maybe I should be open to redesigning the belt drive for more reduction. But if there is a way to get steppers with feedback working well using snapamps that would be great. We're you able to? We're you able to get the high Rapids of 1200ipm like the recommended motors deliver?    I am trying to find a dual shaft stepper that resembles cncrouterparts stepper.   but I'd much rather servo.  Group: DynoMotion Message: 10860 From: bknighton28 Date: 1/17/2015 Subject: Re: 660 Watt BLDC Motor with SnapAmp are you using Mach 3? I had a similar problem with the Kflop and mach3 on my cnc mill. If I got a following error a reset would not be enough to get going again. I had to restart both. I did that for a long time until noticing that while apparently all input is disabled it turned out jogging in one direction on the offending axis was allowed. I think it might be a Mach feature that let the user back out of a problem. after that all is well. Group: DynoMotion Message: 10861 From: bknighton28 Date: 1/17/2015 Subject: Re: 660 Watt BLDC Motor with SnapAmp I was thinking that if I did go stepper I might just get the stock Nema 34 motors from CNC router parts and use glass scales instead of worrying about a double shaft and for superior accuracy. Looking at the specifications of that motor it looks like they did a really good job. Extremely low inductance, very low idle power. It appears that would really scream with a snapamp pushing it to 80 V. I think from the thread you said your motors had a torque constant of 1.17. I wonder if some of the motors I've seen with a torque constant of 1.8 would be well driven by the snapamp? Group: DynoMotion Message: 10862 From: bknighton28 Date: 1/17/2015 Subject: Re: 660 Watt BLDC Motor with SnapAmp were there any problems with the recommended motors that CNC router parts sells other than they are single shaft? I like the linear encoder idea a lot. thinking more on it now I think I might just make things easier on myself and get the stock nema 34 motors and linear encoders in the future. but I believe I will still use kflop and snapamps even though I will have to get two of them. I have them on both a Bridgeport/interac Mill and a Hardinge turret lathe. I mill a lot of aluminum impellers that require 4 hours of program and tens of thousands of lines of code with few motion control problems. The snapamp has been better for me than gecko drives, although I was using the gecko 320 X servo drive and not a stepper drive. Group: DynoMotion Message: 10865 From: Tom Kerekes Date: 1/18/2015 Subject: Re: 660 Watt BLDC Motor with SnapAmp Hi Bill, One thing to keep in mind regarding Steppers vs Servos is that Servo Motors typically are designed to run at higher RPM (ie 3000RPM) where Steppers are usually limited to much less (ie 1000RPM).  Power = Torque x Speed.  So to swap a Stepper out of a system and replace it with a Servo, without changing the gearing, is going to put the Servo at a big disadvantage and only make available a small fraction of its available power. Regarding linear encoders: keep in mind that to close the loop around linear encoders it is usually necessary to also have some feedback from the motor shafts to help stabilize the servo.  Either a tachometer or encoder.  Although Steppers provide some inherent feedback as they are more like position devices. HTH Regards TK Group: DynoMotion Message: 10914 From: bknighton28 Date: 1/25/2015 Subject: Re: 660 Watt BLDC Motor with SnapAmp Tom- I intend to get another Kflop and two more snapamps to control the 7amp 48 volt stepper motors for the router. I have never used a snapamp for other than servos. are there surprises I might encounter doing this compared to using gecko stepper drives? in the future I would like to experiment with closed loop stepper operation using linear scales but you said it might be necessary to include feedback on the motor. is there a known way to use the differential inputs on the snapamp(since there are enough for two per stepper) to include both a rotary and linear scale feedback or is this custom software territory? Group: DynoMotion Message: 10919 From: Tom Kerekes Date: 1/25/2015 Subject: Re: 660 Watt BLDC Motor with SnapAmp Hi Bill, Stepper motors provide a somewhat inherent feedback from the motor shaft as they can move in a somewhat controlled manner without active feedback (when the generated torque overcomes stiction and the motor begins to move and accelerate the torque will drop, unlike with a servo motor).  I'd suggest leaving the option for an encoder on the motor shaft open, but trying with only Steppers + Linear Scales for feedback first. To use KFLOP with two feedback encoders (dual loop) requires the use of two KFLOP Axis Channels connected together with a line of C code.  Not sure if you would call that custom software territory or not. Regards TK Group: DynoMotion Message: 10922 From: Andy Sontag Date: 1/25/2015 Subject: Re: 660 Watt BLDC Motor with SnapAmp Hi Bill, To answer your question from a while ago, we started using kmotion CNC from the beginning and never found a need to use mach3. The problem we experience does happen with some regularity although not frequent enough to track down exactly what is going on.  Tom is right in that if you were to gear down a BLDC motor that has a max current draw of less than 25 amps you could be better off than a stepper based system. The trouble would be engineering the gear down and making it work with the cncrouterparts kit.  Tom's advice of going with a dual shaft stepper is probably wise. While I agree the cncrouterparts Nema 34 steppers are great motors, not having a rear shaft is a bummer. If you're interested in getting the same motor with a rear shaft you'll need to contact the manufacturer directly and request a custom motor. I'm 90% sure that cncrouterparts purchases their motors from LDO Motors Co, as they were the only company to provide a quote with the exact same specs as the CRP motor. The price would have been $400 shipped for the custom dual shaft steppers although we ended up going with the larger Primopal motors.  I also thought that the CRP motors at 80 volts would be an awesome way to go although thankfully I learned about "Iron Losses" before attempting it and completely melting the motors down. Gecko has a great post describing the phenomenon but it basically comes down to the lower your inductance the less you'll be able to "overclock" the stepper with higher voltage. With an inductance of 2 mH you're already over the recommended voltage if run at 48 volts. 80 volts would likely be melted motor territory. That said, I would interested in seeing what happens if you're willing to give it a shot.  We paid a total of $430 including shipping for the 4 custom dual-shaft PHB86S150 stepper motors from PrimoPal. At 73 volts these steppers don't overheat, are much more powerful than the cncrouterparts steppers at lower RPMs (1700 oz-in vs 960 oz-in) and also maintain a superior torque curve at higher RPMs. At 48 volts the larger PrimoPal motor would eventually create less torque than the cncrouterparts motor at higher RPMs.  Something I didn't mention in my last post was that 600ipm is our max feed rate and our max cutting rate. What I found while tuning the system was that at feed rates above around 600ipm the following error would rise above around .002" on the Y axis and would trip the max error cutoff I had setup. If it were possible to either turn off the max error during G0 (rapids) or set a different max error for G0 movements, we could easily move at speeds greater than 1200ipm. I just never had the chance to look into it.  Andy On Sun, Jan 25, 2015 at 12:42 PM, Tom Kerekes tk@... [DynoMotion] <DynoMotion@yahoogroups.com> wrote:   Hi Bill, Stepper motors provide a somewhat inherent feedback from the motor shaft as they can move in a somewhat controlled manner without active feedback (when the generated torque overcomes stiction and the motor begins to move and accelerate the torque will drop, unlike with a servo motor).  I'd suggest leaving the option for an encoder on the motor shaft open, but trying with only Steppers + Linear Scales for feedback first. To use KFLOP with two feedback encoders (dual loop) requires the use of two KFLOP Axis Channels connected together with a line of C code.  Not sure if you would call that custom software territory or not. Regards TK Group: DynoMotion Message: 10923 From: bknighton28 Date: 1/25/2015 Subject: Re: 660 Watt BLDC Motor with SnapAmp the cnc router parts motors just arrived. I see an easy option for a rear encoder though. they are open on the back and I have access to the recessed end of the shaft. the shaft is center drilled where it was turned between centers. I could easily turn a short adaptor for the encoder disc that glues to the shaft, kept concentric by the center drilled hole. no need to open the stepper, likely ruining it. but I will try linear encoder first if I feel the need for feedback.