15HP motor drive, vectorless at 300hz? 1

[fastline12]

We are still trying to decide an economical solution for a CNC machine spindle to get better torque regulation from it. Right now it runs a 7.5HP 4 pole motor with a 15HP Magnetek drive in V/Hz. When cutting material, it pulls down very bad and will stall easily because the drive has no clue other than the current to the motor but does nothing about it. We purchased an older but new Mitsubishi drive for it planning to run it in sensorless vector but then found out that drive is only good to about 150hz in sensorless. It has the encoder feedback card so we were going to do that but then found out it can't do that either....


Is sensorless or encoder feedback at 300hz a specialty application for drives? I am still trying to determine if we can mount an encoder direct to the motor but I do know there is already an encoder that is on the other side of the transmission used for rigid tapping. Why it is not direct on the spindle, I will never know. However, I would sure like to use the output of that encoder for the drive BUT there seems to be a question if that will be a problem for the drive or control? I can't see why a pulse train of 1s and 0s from the encoder could not be shared but maybe I am missing something. However, I DO know the encoder is NOT 1:1 with the motor shaft and not sure if that would be the main issue.


As well, I am trying to determine drive sizing for this application. I can tell you that typically machine tool spindles have a bigger drive to get better accel and decel which is what I am after as well. Faster in every way. I will have to install a thermal sensor on the motor though. Might be easiest to run that back to the drive.

 

[mikekilroy]

I'll take a stab at some comments as this is the stuff we have been doing daily for the last 37+ years ....

We are still trying to decide an economical solution for a CNC machine spindle to get better torque regulation from it.

Know you are not trying to get better TORQUE regulation; u r trying to get better SPEED regulation - since the torque is changing.

Right now it runs a 7.5HP 4 pole motor with a 15HP Magnetek drive in V/Hz. When cutting material, it pulls down very bad and will stall easily because the drive has no clue other than the current to the motor but does nothing about it.

If you are not totally set in buying a new drive, perhaps you could 'fix' this one's shortcoming - in someone's programming.... yes, the drive knows the current draw. the speed droops a small amount due to normal slip but drops big time when out of current. Ie., your drive knows the current, YOU know the motor current rating, why not set a drive output to turn on when current hits some max - the point before it starts to stall. Most CNCs have a feed hold feature that can be triggered if this output turns on (called 'upto speed sig'); hence you do not stall the motor but let up on the load some until current goes down then begin feeding again. is this an option? even with the new drive you really shud have a current overload output to tell the cnc to slow down.... I sometimes program a logic funtion on open loop drives like this using upto speed sig OR overcurrent so you get feed hold from both functions.

Is sensorless or encoder feedback at 300hz a specialty application for drives?

no

I am still trying to determine if we can mount an encoder direct to the motor but I do know there is already an encoder that is on the other side of the transmission used for rigid tapping. Why it is not direct on the spindle, I will never know.

typically because there is gearing between the output of spindle and actual motor; if you need orient (ie., 1/rev marker from encoder) then it MUST be on output.

However, I would sure like to use the output of that encoder for the drive BUT there seems to be a question if that will be a problem for the drive or control?

there are considerations doing that - mostly if the gearing slips: then the encoder counts are no longer correct for vector control. the good news is a little slip every now and then is usually ok since vector control of an induction motor has slip anyway!

anyway, to do this you must have a CONSTANT GEARING - ie., not multispeed gearbox between motor and load. (eg., if you have say 100ppr encoder on output and 10:1 gearing then you simply tell vfd drive the encoder is 1000PPR instead).

I can't see why a pulse train of 1s and 0s from the encoder could not be shared but maybe I am missing something.

hmmmm... some people ARE cheap and simply parallel the encoder to both devices but this is bad engineering practice & has potential issues not worth going into here. Luckily there are devices you can buy that will take 1 enc in and output 2 isolated or buffered outputs to make it right.

However, I DO know the encoder is NOT 1:1 with the motor shaft and not sure if that would be the main issue.

see above.

As well, I am trying to determine drive sizing for this application. I can tell you that typically machine tool spindles have a bigger drive to get better accel and decel which is what I am after as well. Faster in every way.

not generally so; accel & decel is most of time NOT a deciding factor in vfd spindle drive size (I am not discussing a screw machine that must accel/dec every 5 seconds 24/7). what may be is the motor's S6 rating (eg., 10hp motor can be rated say S6-30 (minute) at 13hp) which of course is more current, so a larger vfd since they generally have an overload capacity rating of only 2-60 seconds.

I will have to install a thermal sensor on the motor though. Might be easiest to run that back to the drive.

Not sure how this is relevant to the above discussion, but of course a thermal switch is a good idea to help prevent overload.

 

[mikekilroy]

my bad: last words should have been to help prevent overheating.

rather than to help prevent overload.

 

[fastline12]

That is a thorough reply Mike! I will verify a constant rpm on the encoder relative to gear selection.

The current drive is well built but lacks a LOT of functionality of newer drives. I will PM you with details of the rather small book to see if there are any ideas. I have been through it thoroughly and tried a few setting changes most for accel/decel and only ended up with a LOT of slip on accel and alarms on decel and drop of load from trying to stop too fast. I think the OEM mostly set this thing where it might need to be but I HOPE not...


Regarding sizing, I typically see m them bigger mostly to handle the bigger accel/decel currents, and to take advantage of the service factor area of a motor or 150-200% for X minutes. I learned over the years about a "constant HP" spindle which is the cats meow. They use a bigger than normal motor. The spindle in question is no such animal. Basically a slow slug... I want to ensure I am able to get as much current to the motor as possible to push it into the 150-200%.

The thermal sensor note is acknowledging that by pushing harder on the motor, I realize heat is the byproduct that will fail it if not protected. Since it is crazily NOT protected right now, it only makes sense. It has the TEFC fan removed and a constant flow electric fan added but it needs thermal protection IMO if we are going to push harder on the motor.


Basically what it does right now is as you indicate, does NOT regulate speed worth a darn and since the speed drops, the load goes up in a cut thus ramps down to a stall. Actuall anything over 60% spindle load is asking for a stall. Irritating that we cannot even program for 100% capacity. The CNC contrl does have "at speed" handshake but I am not sure it will slow down since it will go right into a stall with spindle overload. I believe 170% load is alarm level. We really need everything we can get from this machine.

 

[fastline12]

I apologize for my error but our Magnetek is in fact sized at 7.5HP to match the motor. I know that later machines however use a 10hp motor and 20hp drive. And stamp "20HP spindle" all over their machines.. That ain't anywhere close to 20HP all day long..

The drive we purchased is a 10HP which we figured would be a great match but may be that we need a little more education in this. Looking for overall better performance. I know later machines use this 7.5HP motor with a bigger vector drive and the performance difference is night and day!

 

[mikekilroy]

I will verify a constant rpm on the encoder relative to gear selection.

words like 'gear selection' tell me you have more than 1 gear range on this machine so you cannot use load side encoder as the PPR is different in different gears.

The CNC contrl does have "at speed" handshake but I am not sure it will slow down since it will go right into a stall with spindle overload.

you can be sure your 'at'speed' sig is NOT changing as the spindle stalls: it cannot - there isnot actual speed feedback. in v/hz or sensorless vector, the drive will not tell you it is slowing down due to overload as it does not know the actual speed. It is of course possible to design a zero crossing detector circuit to count speed (freq) but none of the drives you will find have it.

 

[fastline12]

I believe the accel "at speed" is nothing more than a accel delay. I know the control uses the encoder to display a tach at the display but it does not seem to do much more unless rigid tapping.

However, it does seem that the encoder is mounted on the motor side of the trans which really makes no sense for rigid tapping. I know if I was designing one, I would want to take all belt deflection and gear lash out of the equation and have that encoder mounted right at spindle speed but.... Regardless, we have run down to a 4-40 tap and seems to pass the go-nogo gauges... I believe the control uses a multiplier depending on gear selection.

 

[fastline12]

I guess there are no PMs here so I uploaded the drive manual for anyone that might have a minute to take a peek. The manual is rather short and brief. I also have the factory settings for the drive. I tried to increase the accel rate and decel rate and neither worked out. I am not sure if there is anything we can do to push this thing but I would be happy to try with some setting changes. I reviewed the manual in depth earlier in the year and could not find much that might help. I would however think that accel could be improved if the drive knew the rotor speed. It seemed to blast right past the rotor and slip like mad causing a motor stall condition.

 

[fastline12]

Woop, here we go.

 

[jraef]

When you go above the base motor speed, you can no longer maintain the proper V/Hz ratio that the motor was designed for, because you will hit max V at the programmed motor base speed. Therefor as you increase frequency above that base speed, you are now at a fixed voltage and losing torque.

A "vector" drive is able to overcome that somewhat, because it can (depending on quality) separate the flux producing current from the torque producing current and tweak the output to the motor to get Break Down Torque out of it at any speed for short periods, without saturating the windings. That means as you begin to LOSE torque above base speed, the vector drive can compensate for SOME of that loss, but only within the current and overload capacity limitations of the drive, and only to the limits of the BDT of the motor (typically 200-225% of FLT). For cutting tools that may be enough briefly, but you may not be able to sustain long cuts at full torque, and at above about 2x base speed it can't happen, the motor is no longer capable of it at all. 300Hz is 5x base speed on a 60Hz design.

Doing any of this however requires a fairly high speed microprocessor in the drive. Your G3 drive is not vector capable at all, it is way too old for that. Don't know what Mitsi drive you have but if it too is used, even though it has SVC it too might be too old to have a fast enough mP to handle the number cruching it takes to get to any more than 150Hz. The newer the design, the better the processing power. So when Mike Kilroy says 300Hz is not a specialy application for drives, I'm sure he is thinking of CURRENT models of drives being sold now. But even then, unless the motor is DESIGNED around a 300Hz base speed, you are not going to get FLT from it at that speed. You would need to start off with a with a higher base motor design frequency. My apologies if you have that and I failed to see it, but your initial post implies it is a standard motor. Worth checking though, "spindle" motors are sometimes high speed designs like that. If so, you should be able to get there with a NEW drive that says it can do it.

"Will work for (the memory of) salami"

 

[fastline12]

Thanks for the reply. I did not mean to imply that I could mod the old G503 for vector duty, I was simply wanting to look at settings to see if we can optimize it though. However, I do realize I believe we are entering mostly into the area of limitation on the motor. I know from chatting with the OEM that the motor is literally an off-the-shelf Lincoln motor. I don;t think they even rebalance them but may have had Lincoln tune them a little sweeter. This motor runs at a max of 202hz right now which would be 6060rpm minus slip. I will increase the speed to approx 6800 since I know for certain that the gear box and spindle were offered from the OEM with motor speeds of 268hz but with the the next size up in drives.

The other drive we had was a nice A500 with vector and network card from NASA. I do not remember the book ever stating max frequency in the different modes. A tech told me that. The bottom line and has been the reason for our delay, is if this is a diminished gain due to the motor and we are ultimately limited by that. I am not real sure why I see a lot of 4 pole spindle motors but I am also referring to off-the-shelf stuff. I hear some Fanuc spindle motors are up to 10pole which puts them at WAY over base speed? What gives. What kind of motor is desired for a high speed spindle application?

As well, I would be real interested to quantify the performance difference between our current drive and another drive? The OEM can't spec anything other than it will be "better". I would hope so for a $6k upgrade to a turd. Calculating our power in a cut, it looked closer to about 3HP worth of cutting at full speed.... That is embarrassing for a VMC.

We had this machine apart before and neglected to get a pic of the more plate but we did realize the drive gear on the shaft is actually heat shrunk on the shaft and motor is custom fitted into the transmission. The OEM actually will not even sell just a motor. They want to sell a motor/trans assy. If it was not for that, we likely would be looking for a motor upgrade as well. The whole machine is the same as ones that run over double this power so the axis servos and ways are good for it. Sorry to wonder there or a sec.

 

[mikekilroy]

You are bouncing around a lot of different ideas and thoughts, some don't fit together with each other. For instance, I answered what I thought was the question "do vector drives today go to 300hz' with absolutely, just about all of them - this has NOTHING to do with the idea of not having full rated torque at 300hz, etc. Anyway, maybe you should look for some other places that build spindle assemblies to tell you what they could offer you? I know 2 in particular who could offer you a new spindle that would far exceed what you have so far described (I have no monetary connection to either). Anyway, you have bounced a lot of disconnected ideas around here that when intermixed confuse a lot of the answers. I am not trying to sound hard, just realistic. I am not sure how you proceed from here without detailed help based on your specific hardware.

 

[fastline12]

Mike I apologize for the scatter brain posts. We are pretty torn on what to do with the machine right now. We put too much into it and we are better off buying something else.

Bottom line is we have an off the shelf 7.5HP 4 pole motor being used as a spindle motor, we have a matching 7.5HP v/hz drive, and a spindle orient that uses a shot pin system that could benefit from vector positioning. I am just not sure which aspect should be targeted (if any) for improvement. We truly need to quantify any upgrade which is proving impossible since the OEM cannot even provide a spindle performance chart.

Accel, decel, orient speed, and cut torque all really need improved but we might have to live with what is an economical solution. $8-10K and we can probably buy something better and newer.

What do you recommend Mike? I am not sure getting full motor torque at full RPM is practical even for newer machines with high speed spindles but the idea is that the smaller the cutter, the faster you need to spin it but you need less power to run it too.

I realize putting the motor on a dyno along with the drive is probably a good starting point but just not practical. I am really trying to determine how much improvement we can get with just a drive upgrade, and if we really need to do motor changes along with this?

We are rewinding another motor right now and I am not above gutting the motor coils and reconfiguring but I will admit that might be getting a little over my head.

 

[mikekilroy]

You have a 7.5hp motor and 7.5hp drive. You want more performance in some areas. You ask what to do to get those performance improvements but you have not listed your present performance in detail so someone can decide how to suggest more than general
ways to get the performance improvements you ask for.

I would suggest the first step is to empirically measure, quantify, and list what your present performance is, and then list what improvements you wish to see to each.

from your comments I see various areas you would like improvements; so how about a chart like this:

ITEM PRESENT WANTED COMMENTS
-------------------------------------------
Orient Shotpin "IMPROVEMENT" What does this mean?
Accel 5sec?10?20? what wanted?
decel 5sec10?20? what wanted?
orient speed 10rpm?20?100? what wanted??
cut torq? 22.5#-ft what wanted?
cut torq above base speed what wanted?
top speed what is it now? what wanted?
$ to do it what is available?
gearing 1 pass?2?3? keep em? change em?

You do NOT need a dyno: ur 7.5hp motor on a 7.5hp drive has a very definitive speed torque curve that any competent motor engineer can draw for you - but you gotta give specifics of what you got! so far i see smatterings of data like it is a 4 pole motor, you go to 202hz top speed today (6060rpm). I can go to 268hx/6800rpm cuz the gearbox can handle it.

You previously said you presently stall the system; so it seems one of the improvements you want is to not stall so much? at what speeds do you stall? In what gear? it is sorta importance to know if you are stalling above or below base speed - in motor rpm - or no valid answer can be given. See the points?

"rewinding" will NOT change the torque u cvn get from your present motor; that is not an option.

If you want real life details on what you can get from what you have or how to improve by changing the motor or drive or both, you gotta LIST what you got today as a starting point & what EXACT improvements you want to each.

 

[fastline12]

I would be happy to start gathering data for this upgrade. I will try to answer a couple questions as well as add a couple of my own for understanding.

1. "shot pin orient". Many machine tools with a tool changer need to orient the spindle for tool change. This one is no different. The drive is programmed to decel to stop, then told to turn the spindle at about 10rpm while a lock pin is fired with an ir cylinder. When the pin falls in place, this sends an "all stop" signal to the drive while the tool change is completed.

A much better way to orient is with flux vector obviously. Faster, more precise, less to go wrong.

2. I would have no way to quantify the torque requirement other than determine how we want to run tools. With that in mind, I very much doubt that is realistic. What we want and what can be reasonably done are two different things. I would rather be able to better predict what the spindle can do now and determine a percentage gain wtih upgrades.


3. Motor rewind - I was referring to modding the wind to 10HP. Near as I can tell, this motor is likely a 213TC frame in which seems to be the exact same as a 215TC which are commonly 10HP. Obviously it might be cheaper to buy a motor and I would still have to see if the stator has room for more wire. This all assumes we can find a way to carefully remove this shrink fit gear on the motor shaft.


MY QUESTIONS

1. In V/hz mode, the voltage and frequency should be at a linear ramp up to base speed determined in the programming? If a motor is loaded down and speed drops, current would increase but voltage might not? Obviously once over base speed, voltage is maxed out. This may make little difference since the motor commonly operates over base speed 99% of the time.

2. In researching vectoring vs v/hz, below base speed, it looks like vector has a large advantage in torque available for acceleration?

3. Braking seems to be solely a function of the braking transistor capacity, brake resistor value and capacity, and the motor size?

 

[mikekilroy]

1) Not sure I totally agree a shot pin orient is so bad compared to position orient with the drive. Our time spent repairing shot pin orient is a lot less than drive orient issues.... faster? say 100rpm to position instead of 10, ok. but is that significant in the whole cycle? more accurate? we typically have a +/-3 to 10 degree in position window with drive orient; shot pin is similar not worse so not much difference there. CAN we make in pos window down to 1/3 degree? often yes, but it increases settling time so it takes longer and tuning is more critical - usually not worth it.

2) my point exactly. today you have 22.5@-ft avail from around 150-1800rpm, then this reduces linearly with speed to around 4000 then drops off quickly to 6060rpm. It can be shown clearly on a plot today with the info you have. then you need to tell what increase you want/need.

3) my point was you cannot 'rewind' a 7.5hp motor into a 10hp motor. not even if it is in a 10hp case. give this thought up.

---------------------------------------

1) current would increase but voltage might not? WILL not. It is open loop. there are games that can be played to increase voltage as current increases (slip comp) etc, but yes, voltage does not change in a v/hz drive.

2) Nope; torque avail is function of current available. vfd drives typically are rated 150 or 200% overload (more current)max. so no matter what kind of drive it will only produce this much additional torque. a good vector drive will produce that torque faster (say 20msec instead of 100msec) but not more.

3) and inertia it is stopping, so yes.

So you say here 99% of the time you run your machine's spindle motor at 1800rpm; this is the kind of stuff you need to quantify and write down in order to improve your machine. I will post a typical retrofit study from one we did on a multi gear head machine a month ago; you must do the same or you will not be able to quantify what you want to improve and how much you do or do not get.

 

[mikekilroy]

fastline, here is the type of data gathering you need to do in order to quantify what you presently have and then be able to define what it is you want to do in order to improve it.

 

[fastline12]

Thanks Mike. Actually I meant the motor spends most of its life way above base speed. Usually around 200hz for Aluminum. We have had an instance where we needed to run a manly 5" saw. Machine stalled multiple times. That was about 7hz to the motor in low range. Motor was about 230rpm and spindle was 70rpm. Next to zero torque. Vector drive could assist here?



Question, reference to "base speed" is common but how do you apply a base speed to a motor that runs on a VFD therefore can be anywhere in the vfd range? Why is base speed determined by 60hz?

 

[jraef]

Base speed is the speed that the motor was designed for, as listed on the motor nameplate.

Sccalar (V/Hz) drives can perform admirable over about a 4:1 speed reduction, maybe 6:1 with slip comp and torque boost, befor the sloppiness in the motor performnce becomes an issie in maintaining torque. But you were at 7Hz, well below where those things can help. Vector control, even open loop (Sensorless Vector Control), would definitely improve that. SVC is considered good down to about 1Hz, better ones down to maybe .5Hz, but below that the internal current sensors in the drive run up against accuracy issues. Adding an encoder feedback can improve that to where you can get 100% torque at zero speed.

"Will work for (the memory of) salami"

 

[fastline12]

Thanks Jraef, could I reasonably assume that the frequency reference for base speed is also utilized for power and amperage ratings?

Not to change the subject here but I am trying to understand the core differences in motors (fanuc for instance) designed specifically as "spindle motors"? I have a couple and they do not have a base frequency. If I understand right, they usually have a lightened rotor for reduced inertial loads on accel/decel constantly? Also, I notice these motors are square and have no cooling fins but do have an electric fan. How can they dissipate all this heat?