View Full Version : Do I need sensorless vector vfd?

will gilmore
04-15-2010, 08:58 AM
I'm building a small CNC lathe (thread: http://cnczone.com/forums/showthread.php?t=92337) 4" swing over cross slide, 8" swing in front of slide (probably wont turn anything close to that big). Headstock is Dunham 5C max rpm 3600
Motor is Marathon Inverter/Vector duty, 230/460V, 6/3A, 1760RPM, 60Hz, 2HP, SF1.0, Max speed 3600, 3HP 30min duty.

I'd like to have single phase input power.

This is what I'm looking at:
Hitachi X200 volts hertz $202

Hitachi SJ200 sensorless vector $283

Right now I'm planning on a 1:1 belt drive which gives me the max headstock speed of 3600rpm. I'd like to be able to reduce the speed though the VFD and not have to swap pulleys and have enough torque.

Do I need the sensorless vector VFD?

04-15-2010, 10:18 AM
In order to handle the torque requirements you will want at low speed, YES, I would get the sensorless vector drive! With a standard drive, the torque falls off drastically at low rpm. On my 1725 rpm motor, I can't run it below about ~300 rpm and still have a usable product.


04-15-2010, 10:21 AM
SVD is most beneficial at drive frequencies below about 20hz. If your applications do not require slow speeds, the V/F drive may prove adequate.
If purchasing a VFD rated motor, and a new drive, go with the more sophisticated SVD, You won't be disapointed.

04-15-2010, 11:25 AM
SVD's rock.

But to make that motor live if it isn't rated for it, you may need an additional fan to cool it at low rpms. It's not hard to rig one up and most SVD's can even close a relay to turn it on at low speed.

The problem is without the fan the motor doesn't spin fast enough to move any air with its shaft fan.



04-15-2010, 11:56 AM
A real vector drive (with encoder feedback) is what rocks. It will run exactly the commanded rpm, (no guessing) or it will fault on overload trying. That is what I would recommend on a lathe which generally might be running with a much higher torque requirement than a mill.

Real cnc lathes typically use a monster motor to develop the needed torque at their lower rpms. In other words, your 3 hp motor will do 3hp of work in any gear when running at full rpm through a gearbox, but it will only do 1.5hp of work if running at half nameplate speed, because the current limiting smarts of the drive will prevent it from overdrawing (which the motor is quite willing to do).

Best would be to bump up the motor size. Failing that, it might pay to at least buy an over-rated VFD so that you can get that ocassional 200% out of your motor for a few seconds without nuisance tripping of the drive.

I put a VFD on a cold saw that was theoretically matched to the motor on the saw. It was always nuisance tripping. I didn't save any money selling off the small VFD to get the next size up :D

04-15-2010, 12:29 PM
Hu, you'd be surprised at what an SVD can do versus a non-vector VFD, though you're right, there are some advantages to the flux vector (encoder). But the flux vectors are so much more money than a vector, and the vector in this case is only a little bit more.

Here's a good discussion of it:


The remarks in that link about the relative horsepower are also good to consider.

The SJ200 will deliver full torque down to 15 rpm. However, the bigger issue for a "real" lathe is positioning accuracy. You want to be able to index the spindle and a sensorless is just not very good at that relative to a full position sensored (flux vector) drive. They're also weak at holding position, relative to a flux vector drive.

This matters a lot less for mills, which just need positioning close enough for a tool change and can use an encoder feeding back to the CNC control (rather than the VFD) for rigid tapping. So you see a lot of SVD's on mills.

Many do seem to feel that a VFD with higher HP rating is helpful as you suggest, Hu.



will gilmore
04-15-2010, 02:39 PM
Thanks for the brain power everyone.

I bought a 4 pole motor to get more torque at low speed. I figure that if you run out of torque on a lathe its going to be turning a large diameter piece which would happen at low rpm. If you are turning high rpm you are probably cutting close to the axis of rotation and probably cant take deep cuts anyway because of deflection.

Eventually VFD will have feedback via spindle index pulse --> G540 --> Mach3 --> G540 --> VFD

I guess I'll pony up the $81. Would hate to regret it later. Is Hitachi the way to go?

04-15-2010, 03:14 PM
If anyone can supply a graph of the REAL performance advantages (Torque,Frequency) of SVD over V/f
I would be VERY interested. I have only seen one, and the "improvement" was not significant above about 20hz. I'm looking for the reference now.
Speed regulation capabilities of the SVD are recognized and NOT part of this question.


04-15-2010, 03:20 PM
I have a 3/4 HP 2 pole motor with VFD on my minilathe (not SVD) and it runs nicely form 20 Hz through to 100 Hz without any additional cooling. I geared it about 1:2.5 so the spindle is about 2200 when the motor is at 100hz.

This set up works very well, only hitting it's limits with an M40x1.5 thread, as I couldn't run Z fast enough to give a fast enough spindle speed, with a decent cut the wouldn't rub.

Given my data point, I am guessing your 3hp motor with a SVD VFD running 1:1 and 10-100hz ( with additional cooling on the motor at low speed), will chomp though anything you can swing on that lathe you are building.

Nice build you have there.

04-15-2010, 03:41 PM
I have to go with Hu on this one, practically all CNC spindles have a feedback to the drive, of some kind, This is besides whether a spindle encoder is fitted or not.
When I started doing retrofits with using P.V. instead of closed loop vector I ran into a few problems with speed variation between load/no load.
The option often is not that much more.
The local WEG salesman told me they now have a vector motor/VFD combination that will retain 100% torque down to 1rpm and no fan on the motor.

will gilmore
04-15-2010, 03:48 PM
CalG--Its going to depend a lot on the application. 20hz on my motor is going to be ~600rpm. If the sensorless vector gives improvements 0-600rpm I think that is probably going to be worth it for me

RotartySMP--Its actually a 2HP motor. The data plate says its rated for 3HP for 30min duty. I originally was looking for a 1 or 1.5 HP motor but I found this one locally.

Al the man + Hu -- What are you guys recommending? I don't want to spend the money to get flux vector drive plus encoder.

04-15-2010, 05:15 PM
Will The graph data I am interested in is not dependent on any particular application.
Frequency/Motor RPM vs Rated torque overlayed with % of actual torque.
Rated torque is a straight horizontal line, frequency is the ordinate, developed torque is also on the y axis.

All such curves have torque fall to zero when the motor obtains synchronous speed ;-)

And in the search I did read a comment that V/f drives are generally good to 10% of rated freq. That would be only 6hz.!

Knowledge is power! Searching still

Ahh! Here it is...Comments on the plot seen on the last page Very welcome!


04-15-2010, 07:29 PM
All such curves have torque fall to zero when the motor obtains synchronous speed ;-)

That makes sense, an normal induction motor can never obtain synchronism.
Slip has to occur for it to work.

04-16-2010, 01:37 AM
I missed the part about this being an 8" swing lathe. Swing means max diameter, so with 2hp you should have lots of omph. Probably you would not be turning 4" shafting down in a machine that size :D

02-16-2012, 04:18 PM
Just remember if you want to use single phase input to de-rate the drive by 50% if the drive isn't already specified for single phase. So to be safe make sure your drive is rated for single phase or go with greater than a 12 A drive.

I know Mitsubishi makes a great vector drive that comes with a 5 year manufacturer's warranty. You can find it on VFDs.com (http://www.vfds.com). Here's that link:
Variable Frequency Drive (http://www.vfds.com). The model number you'd probably be looking at is an FR-E720-175-NA. That's a 17.5 A drive. Just search the product number on there site and it will come up.

02-16-2012, 08:55 PM
If anyone can supply a graph of the REAL performance advantages (Torque,Frequency) of SVD over V/f
I would be VERY interested. I have only seen one, and the "improvement" was not significant above about 20hz. I'm looking for the reference now.
Speed regulation capabilities of the SVD are recognized and NOT part of this question.


calg, really perceptive question!

I will look for docs to post showing that above 20hz there is NO difference in available torque output from a v/hz drive vs a svd, that the only difference will be 0-50rpm variation on v/hz vs 1-2rpm on svd from no load to rated torque. there will also be a time associated with the speed correction and that may make folks feel it is a lot better....

02-16-2012, 09:13 PM
Sometimes we fall into the trap of claiming universal improvement due to the perception of some detailed improvement.

I like graphs! They tell me much, and I have learned to read between the lines.

Love to see more.


06-06-2016, 10:03 PM
Hi, in order to maintain the large torque even at low running speed, a sensorless vector motor drive would help. Different from common drives, most of the vector motor drives (http://www.veichi.org/product/electric-drive/ac70-sensorless-vector-control-inverter.html) have large torque when the motor runs with a low rpm.