Spindle Controller to VFD Replacement/Retrofit?

We are updating a Mazak VMC (VQC 15/40), and have a 7.5HP Mitsubishi spindle motor - 5.5kW / 32A max. A 1024 ppr shaft encoder is mounted on the spindle.
The current spindle controller is a Mits Freqrol FR-SE 5.5kW, and handles orienting for toolchange, and closes the speed loop (these components were all part of the original Meldas/Mazatrol M2 control system). The spindle controller is currently non-functional due to a power supply with a fairly significant repair cost from Mitsubishi.

We are looking at replacing with a Hitachi VFD - either a WJ200 or a SJ700 series - both are 'sensorless vector drives'.

I've spent a fair amount of time digging through the forum, but have a few questions I was hoping some of you might be able to help with:

1) Is there a problem with swapping a 'spindle controller' for a VFD? It seems basically the same thing, but I'm not sure if I'm missing little details

2) The output of the Hitachi SJ700 7.5HP VFD I'm looking at is 230V/25A. The motor is 200V/32A... Based on some discussion I saw by MotionGuru, it appears I would need to upgrade to a 10HP VFD 230V/35A to get full motor capacity?

3) Any idea on line chokes / line reactors - how do I know whether I should put them on the VFD input? output? We're planning on mounting the VFD in the cabinet on the machine, if that makes a difference.

4) I have the option of buying an encoder feedback card for the VFD ... is there any benefit (besides config ease) to allowing the drive to handle the velocity loop and spindle orient instead of (saving money) by allowing the new control to close the velocity loop & handle orientation (LinuxCNC)? Stability? Speed of response?

5) Any gotchas or anything else I'm not thinking of or should be aware of?

Thanks in advance!

Not to put a damper on your idea, but, no, it is very unlikely you would make it work - for MANY MANY reasons...

let me just ramble and hit some of the ones that stick in my head without reading your post line by line again...

we are hitachi distributor. we have had folks try to do similar and none made a go of it. I also sell and commission very sophisticated Refu drives that will do this with the special programming required.

from the top...

You get a hint that motor is unique when it requires significantly MORE current than a generic 7.5hp model. and hitachi, like others, gives EXTRA current for their 7.5hp drive; yet it is still too low.
Yes, you would need a 10hp drive. there is a REASON that current is high: the 'base speed' rated voltage is very LOW. So when you went to commission the hitachi, putting in 60hz and 460v, it will be faulting out on too much output current since your v/hz curve is way way off. how will YOU find the right voltage? how will YOU find the real base speed? how will YOU find the correct REAL much higher speed real base speed and voltage? tricky stuff!

going along with above: does your m otor have a nameplate? does it SHOW ampps, speeds, freq? I would bet you $ 100 that is a HIGH SPEED motor - goes a lot faster than 3000rpm tops! based on the amps you gavee, I will guess about 7500rpm top speed. few sensorless vector cheap drives will tune right by themselves to this 5-6-7x constant hp range motor. but assume you get this far...

There is ZERO way to make a sensorless vecttor drive orient. next.

Add that encoder card option - what does that do for you? I do not recall any orient routine in this or any other cheap drive.

even if you find a cheap drive with orient, how are you going to adjust it? It no doubt will be some position other than plain zero...

If you have prints for your machine, you will see there is upto speed, zero speed outputs. But also Orient input and IN-POSITION output. Where are you going to get that signal on the cheap drive?

If you are up for lots of experimenting, then you can make it work with encoder feedback and letting your CNC control do the orient. If the control was the mitsubishi that came on the machine, and it did not do the orient, I doubt you would make it work.

Agree..

Is this thing a tool to make money? If so, why screw around? If you DID get it going, with many many hours of non-billable time, you'd still have the only one like it in the world, with full responsibility for fixing it. And you would have paid more than the repair cost to get there.

The cheapest solution does not always cost less.

Do you make parts for customers? Or do you make CNC machines? Choose.

Thanks for the feedback - I appreciate the dose of reality and detailed potential problems.

I probably should have mentioned a couple of things to clarify the situation...
1) We have already fitted this VMC with a custom controller built around LinuxCNC. Integration isn't a problem for us. It makes economical sense for us because we are a family prototyping firm with a prototyping shop, not a job shop, and it pays better to spend extra time on integration than paying loans. We also get to save on tech support because we can quickly fix problems ... which has been born out on other equipment.

2) This is not our first retrofit, and we are fairly comfortable tuning PID loops (on servo drives), etc.

So ... we're not (overly!) concerned about a challenge/experimentation, and not totally new to the concepts, but have little practical experience with all the details to take into account specific to a spindle drive setup. THANKS Mike, for the start at that!

Updated answers/questions:

1) Forget the encoder card. We are planning now on just bringing the encoder feedback directly into the controller, and sending FWD/REV & +/-10V to the VFD (at this point, probably a Hitach SJ700 10HP). The controller can handle orientation just fine, and has been used in numerous identical setups.

2) Speed on the nameplate is in three increments: 1500, 4500, 8000 - I assume this is low/medium/high speed ranges. These seem to correspond to Cont (kW): 3.7, 3.7, 2.0 and 1/2Hr (kW) ratings of 5.5, 5.5, 3.0
Max Volt: 200V
Max Amp: 32A

The spindle motor unit is connected directly to the shaft, with a toothed belt driving the encoder (1:1).

In the Mitsubishi documentation for the motor, BASE SPEED: 1500RPM, MAX SPEED: 8000 (pretty close, Mike!)
Rated for 200/230V 50/60Hz. Working it out based on 50Hz, it looks like a 4-pole motor.

3) I want to make sure we get a VFD with proper frequency range. Based on the 8000RPM max, would the maximum frequency needed be about 4poles x 8000rpm / 120 = ~267Hz? I reckon there are some factors I'm neglecting here, and would appreciate any education... At this point, I'm basing this plan off of similar builds I've seen/read about, some intuition, and lots of reading, including here on Practical Machinist by Mike, MotionGuru, & JRaef.

4) We're in the situation that I'm suspecting our whole controller is fried, & a drop-in replacement is out of budget. (@JST ... if some of the projects this machine is meant to test take off, we won't have time to rebuild another one and I will be first in line to buy a brand-new machine!)
Is there another, preferable option for spindle control that is not exponentially more expensive?

Thanks again!

SRDC said:

The spindle controller is currently non-functional due to a power supply with a fairly significant repair cost from Mitsubishi.

Click to expand...

Why make a problem out of a solution?

Dunno why others missed this, but Power supplies as clans and tribes are about two if not three orders of magnitude less work to rev engineer and field repair even with no manual and 100% obsolete parts than the rest of what you believe its "death" is forcing you to do.

If nothing else, you are REPAIRING to a known-good design that all matched-up before it broke, rather than inventing new or having to engineer integration out of genetic incompatibilty.

Sure would have a go at fault detection and repairing it or duplicating it first. And/or sourcing a used one, even two, - and even if known in advance that they also needed repair.

Bill

SRDC said:

Thanks for the feedback - I appreciate the dose of reality and detailed potential problems.

I probably should have mentioned a couple of things to clarify the situation...
1) We have already fitted this VMC with a custom controller built around LinuxCNC. Integration isn't a problem for us. It makes economical sense for us because we are a family prototyping firm with a prototyping shop, not a job shop, and it pays better to spend extra time on integration than paying loans. We also get to save on tech support because we can quickly fix problems ... which has been born out on other equipment.

2) This is not our first retrofit, and we are fairly comfortable tuning PID loops (on servo drives), etc.

So ... we're not (overly!) concerned about a challenge/experimentation, and not totally new to the concepts, but have little practical experience with all the details to take into account specific to a spindle drive setup. THANKS Mike, for the start at that!

Updated answers/questions:

1) Forget the encoder card. We are planning now on just bringing the encoder feedback directly into the controller, and sending FWD/REV & +/-10V to the VFD (at this point, probably a Hitach SJ700 10HP). The controller can handle orientation just fine, and has been used in numerous identical setups.

2) Speed on the nameplate is in three increments: 1500, 4500, 8000 - I assume this is low/medium/high speed ranges. These seem to correspond to Cont (kW): 3.7, 3.7, 2.0 and 1/2Hr (kW) ratings of 5.5, 5.5, 3.0
Max Volt: 200V
Max Amp: 32A

The spindle motor unit is connected directly to the shaft, with a toothed belt driving the encoder (1:1).

In the Mitsubishi documentation for the motor, BASE SPEED: 1500RPM, MAX SPEED: 8000 (pretty close, Mike!)
Rated for 200/230V 50/60Hz. Working it out based on 50Hz, it looks like a 4-pole motor.

3) I want to make sure we get a VFD with proper frequency range. Based on the 8000RPM max, would the maximum frequency needed be about 4poles x 8000rpm / 120 = ~267Hz? I reckon there are some factors I'm neglecting here, and would appreciate any education... At this point, I'm basing this plan off of similar builds I've seen/read about, some intuition, and lots of reading, including here on Practical Machinist by Mike, MotionGuru, & JRaef.

4) We're in the situation that I'm suspecting our whole controller is fried, & a drop-in replacement is out of budget. (@JST ... if some of the projects this machine is meant to test take off, we won't have time to rebuild another one and I will be first in line to buy a brand-new machine!)
Is there another, preferable option for spindle control that is not exponentially more expensive?

Thanks again!

Click to expand...

I think you are over reaching on this one. Some of the questions you are asking indicate you have a serious limit to your knowledge to get the spindle to work and get the tool changer to ever work again. Nothing wrong with not knowing and asking but it does put you at a serious disadvantage.

The tool changer requires that you can precisely rotate the spindle to a home position and do this with repeatability. You're current approach sort of ignores this which indicates not understanding the depth of the problem. Machine tool builders that did take this approach usually used a shot pin for zero that would lock in the spindle orient position. This is a lot of work to pull off.

The second issue is that you do not understand motor characteristics and the V/HZ relationship. Depending on what rpm you are going to run the spindle at will determine the torque output. Much below 30HZ and the motor is subject to overheating unless derated. Much above 120HZ, you will enter core saturation with a drop off in torque, an increase in winding current and temperature.

Being that it is only a bad power supply, as others have suggested, I would fix that. There are a multitude of ways to skin that cat without spending a fortune and even if you did spend for the Mitsubishi repair, you would still be money ahead of the VFD approach with all of the added complexity to get the tool changer to work and the compromise in spindle performance.

Another alternative is to find another identical used machine and take the needed control components off from it if your machine is mechanically in good shape.

I would proceed with the VFD approach only if you are willing to not use the tool changer anymore and are willing to have a compromise in spindle performance.

More good input on why you should consider repairing the present drive. (only wrong comment was overheating below 30hz and no torque above 120hz as YOUR motor is designed to do these things).

If at the end of the day you decide to go forward with new hardware, then some info for you:

- the sj200 is as good a choice for cheap vfd as any, and better than a lot.
- you WILL end up requiring the enc option card as your spongy control below around 20rpm will prevent orient working IMO.
- now you should buy an encoder emulator card from someone to take the encoder in and then buffer and send it out two outputs - one for the sj200 and one for the cnc.
- your base speed setting will NOT be 50hz but rather 150hz, with 267hz as top speed.
- you will set voltage gain variable to 200/240v or use the preset for 200v - do not miss this setting.

Ziggy2 said:

I would proceed with the VFD approach only if you are willing to not use the tool changer anymore and are willing to have a compromise in spindle performance.

Click to expand...

Well, we are not willing to do either. Thank you both for your input - we are used to overreaching a bit & then catching up, but try not to overreach so much we fall over! I was hoping that we could find out some of the biggest things we didn't know to think about, which we did. Thanks!

Does anybody have recommendations for:
a) a good electronics repair outfit (besides Mitsubishi) that could repair the power supply?

b) a spindle drive to look at that (if this whole drive ends up being fried) might be a good compromise between reliability & economy?

SRDC said:

Well, we are not willing to do either. Thank you both for your input - we are used to overreaching a bit & then catching up, but try not to overreach so much we fall over! I was hoping that we could find out some of the biggest things we didn't know to think about, which we did. Thanks!

Does anybody have recommendations for:
a) a good electronics repair outfit (besides Mitsubishi) that could repair the power supply?

b) a spindle drive to look at that (if this whole drive ends up being fried) might be a good compromise between reliability & economy?

Click to expand...

many have good luck with PLC center.

why n ot give the complete mitsubishi drive part number here? might find someone has one or can help you get a used one! I have a friend who does a lot of Mitsubishi drives/control work; he might have something to add to help find one. but we cannot help if you don't list the model!

do you have a wiring diagram of the machine? if so why not post the spindle drive wiring page here? it would go a long way to answer your 'does anyone know a good replacement.'

Sorry - I didn't even think to post the part #. It is a FR-SE-2-5.5K-A.

That said, based on all the feedback, and after a call to our local CNC service company who recommended a new power supply rather than repair (he said the repaired ones don't seem to be as reliable in this particular model - the PW-SE), we decided to replace with a new spindle controller. The engineer there helped us source a replacement drive package.

Looks like we'll get to integrate that, but at least we avoided a lot of trouble going off half-cocked.

Thanks to all again!

SRDC said:

Sorry - I didn't even think to post the part #. It is a FR-SE-2-5.5K-A.

That said, based on all the feedback, and after a call to our local CNC service company who recommended a new power supply rather than repair (he said the repaired ones don't seem to be as reliable in this particular model - the PW-SE), we decided to replace with a new spindle controller. The engineer there helped us source a replacement drive package.

Looks like we'll get to integrate that, but at least we avoided a lot of trouble going off half-cocked.

Thanks to all again!

Click to expand...

quick search found one on ebay for under $1500 with a warranty!!!!!!!!!!!! Did you search online??

Mitsubishi FR-SE AC Spindle Drive Controller, FR-SE-2-7.5K- A-C, Used, WARRANTY | eBay

Agreed on all the comments RE using a VFD drive for spindle orient and spindle performance - you will be extremely frustrated trying to make an open loop V/Hz drive work in this manner which will ultimately culminate in a ton of wasted effort at best and a cooked spindle / damaged tool changer at worst.

IF . . . you wanted to replace this with a brand new drive, I have helped customers to replace this series drive with a UniDrive M702 series drive. The spindle orient feature is easily programmed in the drive and the V/Hz profile of your motor is easily accommodated by the drive as well. It comes with a broad range of feedback / encoder options standard and has a built in position / motion controller that has a nice velocity control interface.

Comes with Ethernet as well for easy commissioning with the free software - Unidrive M7 - AC and Servo Drive from Control Techniques

We recently did a project with a DirectLogic PLC / HMI / M702 drive for a Hollywood camera positioning system that you can operate wirelessly from an iPad/iPhone or Android phone . . . an amazing amount of performance bundled in that drive.

mike_kilroy said:

quick search found one on ebay for under $1500 with a warranty!!!!!!!!!!!! Did you search online??

Mitsubishi FR-SE AC Spindle Drive Controller, FR-SE-2-7.5K- A-C, Used, WARRANTY | eBay

Click to expand...

Actually we did - but our boss said to avoid used parts from 3rd party sellers, as we don't know their history and don't want to increase the risk of having to address this again in the middle of a job... so we went with a new drive.

Thanks!

motion guru said:

...extremely frustrated trying to make an open loop V/Hz drive

Click to expand...

I think I confused everybody here. I was never meaning to operate it in open loop. My only question was about closing the loop in the drive via an encoder feedback board (or, optionally, closing it in the controller).

Based on everybody's feedback, we purchased a drive setup that will close it directly in the drive, and pass the encoder input on to the controller so it can be aware of actual spindle speed.

motion guru said:

an amazing amount of performance bundled in that drive.

Click to expand...

Thank you for the heads up! I will definitely be looking into that drive for some of our other machines.

Thank you again to all.

mike_kilroy said:

your base speed setting will NOT be 50hz but rather 150hz, with 267hz as top speed.

Click to expand...

Mike,

I got to thinking about your reply again today and was wondering if you would explain why the base speed would be 150Hz?
How was that calculated?

Here is a very detailed video about repairing the power supply on a Frequol drive. This guy is pretty sharp.

Thanks! That will definitely come in handy because we are planning on trying to repair the power supply and recommission that drive in another machine.

However, I'm still really curious how Mike Kilroy got that base frequency. I don't understand it - and I figure the more I can learn, the better it will be in the future as other things pop up.

Got it! (?)

So I think I finally got it ...

I've been studying the V/F relationships and how it all works - a great explanation in particular by Mike on the big Z with a nice little drawing. In that scenario, 110Hz @ 240V was chosen as the base frequency because the rated base was 50Hz @ 110V ... so upping the bases accordingly kept the same V/F relationship while maximizing the CT portion of operation. That makes a lot of sense, helped as well by a discussion on eng-tips with Mike and JRaef helping someone out.

With all that, I still didn't understand how Mike determined where the max voltage was reached for our motor. Obviously it wasn't at 50Hz ... but I couldn't figure out where.(Mike's statement " In the MT spindle world it is not uncommon for us to rate motors where the VOLTAGE reaches max available from a supply such as 230v, & 460v" made lots of sense, but I was missing that voltage max point)

Then I just was scrutinizing the motor specs again, and the little statement "reduced output is obtained for speeds of 4500RPM and above..." caught my attention. Looking back at the motor nameplate information I put in my first post, I realized that it was all right there 1500RPM, 4500RPM & 8000RPM - 'standard' base speed @ 50Hz, 'real' base speed @ max voltage, & max 'safe' (balanced?) speed. Like I originally stated, I had assumed it was low/medium/high ranges. That wasn't making sense anymore, but I wasn't sure what to make of it until the various pieces fell together.

SO ... then it was straightforward:
frequency = #poles * RPM / 120

For the point where this motor was designed to reach max voltage:
frequency = 4poles * 4500 RPM / 120 = 150Hz

Now ... I'm sure Mike & others will point out some things I've got wrong, but thanks for all the help!

(btw Mike ... have you considered being a professor? I wish you'd been teaching my electrical engineering 101!)

More important to us is the torque curve. Most likely the peak torque is at the base speed. Any RPM gained above the base will come at reduced torque.