Motors Minimum RPM ( wrt rigid tapping )

[spumco]

Some thoughts from somebody who is not remotely competent to fine-tune a PID control...

The Acorn's encoder will need to be enabled for rigid tapping. Otherwise it can't synchronize the Z-axis. Don't forget to enable it and see if starts fighting the GS3 when you do some 'air' tapping. If the Acorn doesn't have a spindle PID settings page or screen, then it's probably not doing anything but sending out a static 0-10V signal to the GS3 (for a given programmed RPM) and it shouldn't be fighting the VFD.

Once you've got the initial tune sorted out, I suggest taking a fairly heavy cut with a fly cutter on a chunk of scrap. The intermittent load should help you see if the GS3 is tuned appropriately or if it's hunting for an appropriate speed as the tool does the usual thump-thump-thump deal.

Then do a heavy(ish) profiling cut with an end mill to generate a constant load and see if that's good to go or if you need to fiddle with the PID settings some more.

An interesting feature (bug?) I found in my controller is that if I program a rigid tapping cycle but leave the S word (spindle speed) out, the controller shifts to synchronized Z-axis mode and I can turn the spindle by hand and it will advance by the programmed thread pitch distance. Put an indicator on the head and rotate the spindle a couple of times and it's cake to check if the programmed pitch matches your tap TPI.

Once you've got the VFD sorted out - and you will - you might see if the Acorn has a similar feature. It's pretty handy.

-S

 

[cut2cut]

Some thoughts from somebody who is not remotely competent to fine-tune a PID control...

The Acorn's encoder will need to be enabled for rigid tapping. Otherwise it can't synchronize the Z-axis. Don't forget to enable it and see if starts fighting the GS3 when you do some 'air' tapping. If the Acorn doesn't have a spindle PID settings page or screen, then it's probably not doing anything but sending out a static 0-10V signal to the GS3 (for a given programmed RPM) and it shouldn't be fighting the VFD.

Once you've got the initial tune sorted out, I suggest taking a fairly heavy cut with a fly cutter on a chunk of scrap. The intermittent load should help you see if the GS3 is tuned appropriately or if it's hunting for an appropriate speed as the tool does the usual thump-thump-thump deal.

Then do a heavy(ish) profiling cut with an end mill to generate a constant load and see if that's good to go or if you need to fiddle with the PID settings some more.

An interesting feature (bug?) I found in my controller is that if I program a rigid tapping cycle but leave the S word (spindle speed) out, the controller shifts to synchronized Z-axis mode and I can turn the spindle by hand and it will advance by the programmed thread pitch distance. Put an indicator on the head and rotate the spindle a couple of times and it's cake to check if the programmed pitch matches your tap TPI.

Once you've got the VFD sorted out - and you will - you might see if the Acorn has a similar feature. It's pretty handy.

-S

I got it a little bit more tuned in tonight. Thanks for the thoughts about putting a load on it. I was actually seeing more constant rpm when i turned off the **motors** encoder initially but realize the encoder’s value comes into play when a load is applied. I ended up with very low integral and proportional settings and now the rpm stays close to constant when I use the vfd’s controls. So some trial and error is getting me closer. I’ll try cutting something next to see when it starts balking again.

I ran across this video last night of the lathe version of the same Centroid software and Acorn hardware I have. He used an m code to allow him to manually turn the lathe chuck so the “carriage” servo would move accordingly, just as you are suggesting. I don’t know if the command he used will work on the mill version of the software but I intend to try it or an appropriate variant.
He used G1W1F.5

Here is the link :

Best Regards,

Jake

 

[spumco]

Check the GS3 manual to see if it has a positioning and/or homing mode. My Hitachi does, and I could pretty easily connect a pin from my BOB to one of the digital inputs on the VFD. Program Mxxx to send the pin high, and the VFD responds to that input. Program the VFD so that that input switches it to positioning mode and it'll go to the index signal or rotate xx number of degrees or whatever.

If the motor speed is still hunting a bit when the motor encoder is off, check the 0-10vdc signal from the Acorn BOB for stability. If it's fluctuating, that could be source of RPM hunting. It could just be the Acorn, or you could be getting noise that's resulting in the VFD trying to match the wonky speed input signal. There may be a tolerance setting in the VFD that smooths things out - but at a cost of less fine resolution in motor speeds. If you dont' care whether it's 500RPM or 500.1 RPM, then adjusting the VFD to respond to larger voltage steps might help.

 

[spumco]

Oh, yea - and another thing if fiddlign the PID settings doesn't get you exactly where you want...

If/when you check the 0-10VDC output from the Acorn to the GS3, do it first without the motor running. i.e. disconnect the signal from the VFD and see if it's a stable reading. Command an M3 Sxxx that will result in a low, middle and high voltage and check at each speed variant.

Then do it again with everything connected and motor turning and check at the Acorn output terminal and also at the VFD input.

If there's a difference in stability between motor on & off, you've probably got RFI/EMI from the VFD cables getting in to the signal wire. One method for dampening this is to adjust the GS3's PWM frequency a bit lower. Another is to shield the cables like I mentioned earlier and make sure the 0-10VDC input wires are routed well away from the VFD motor cables.

P.S. - just checked and the GS3 doesn't have a parameter to adjust analog input voltage fluctuation tolerance. Boo hoo.

 

[cut2cut]

Oh, yea - and another thing if fiddlign the PID settings doesn't get you exactly where you want...

If/when you check the 0-10VDC output from the Acorn to the GS3, do it first without the motor running. i.e. disconnect the signal from the VFD and see if it's a stable reading. Command an M3 Sxxx that will result in a low, middle and high voltage and check at each speed variant.

Then do it again with everything connected and motor turning and check at the Acorn output terminal and also at the VFD input.

If there's a difference in stability between motor on & off, you've probably got RFI/EMI from the VFD cables getting in to the signal wire. One method for dampening this is to adjust the GS3's PWM frequency a bit lower. Another is to shield the cables like I mentioned earlier and make sure the 0-10VDC input wires are routed well away from the VFD motor cables.

P.S. - just checked and the GS3 doesn't have a parameter to adjust analog input voltage fluctuation tolerance. Boo hoo.

Yesertday I meant to reply, but it was a bitter sweet. I was able to rigid tap a couple times successfully but then dinner gave me food poisoning ! so its been a battle the last 24 hours....

I was looking around in the manual today for the adjustment you suggested ( analog voltage tolerance control ) and only found P6.37 which is for "Hunting Gain" which "helps control frequency oscillation". Not sure it applies but its the closest I could find.

I tested the VFD using only the control on the VFD to rule out any ground loops or other interference and wandering voltage by removing control via the acorn. The rpm would flicker just a couple rpm after I optimized the integral and proportional encoder settings. Then I compared the fluctuation when using 0-10v control coming from the acorn. I am seeing some wandering of frequency and accordingly rpm too but also an increase in fluctuations. Its not bad, but for rigid tapping I think it needs to be better. So I definitely need to figure this out before I get too confident it'll work reliably. I was able to rigid tap but I also had it automatically abort because the rpm had drifted enough that it didn't have time to sync up. I'm able to manually get it close using different analog input offset settings in parameter P4.02. I'm pretty confident with all your suggestions I can get the kinks out. I thought maybe temperature might have something to do with the change in rpm but I haven't had time yet to go through with a volt meter to check output voltage from the acorn. Probably Sunday I'll get to that and make sure the wires are as you suggest.

I'm also considering a proper though bore $450 encoder good for 8000 rpm instead of my rigged up $25 dollar encoder using RC car belts and pulleys ( see picture ) Its working so far without too much fluctuation from the skinny timing belt but over-all its max rpm is not going to allow me to leave it connected all the time.

Also, I am considering going 1:1 with my motor pulley and try using solely frequency to get the max RPM of 5400 from the motor. It is nice, however to have the possible 7000 rpm with the 2:1 pulley configuration I have now at 120 hz max instead of using 1:1 and 180 hz. Tough to really know whats best but surely low speed torque will benefit and probably some stability in rpm when tapping and slow rpm fly cutting ?. Do you have an opinion as to what would be best ? Is 7k rpm worth it, or should I try to run 1:1 and therefore probably have better torque but lose 2k rpm. I do tend to mill small items requiring small end mills ( 1/4 inch and smaller down to 3/32, for instance ), so rpm is important, especially for carbide end mills.

Best Regards,

Jake

 

[cut2cut]

P.s. spumco i’ll Need to look for the index related function on the VFD. Not sure it has anything similar to your hitachi however if I get the bore through encoder it’ll provide an index directly to the acorn.

 

[TomS]

Yesertday I meant to reply, but it was a bitter sweet. I was able to rigid tap a couple times successfully but then dinner gave me food poisoning ! so its been a battle the last 24 hours....

I was looking around in the manual today for the adjustment you suggested ( analog voltage tolerance control ) and only found P6.37 which is for "Hunting Gain" which "helps control frequency oscillation". Not sure it applies but its the closest I could find.

I tested the VFD using only the control on the VFD to rule out any ground loops or other interference and wandering voltage by removing control via the acorn. The rpm would flicker just a couple rpm after I optimized the integral and proportional encoder settings. Then I compared the fluctuation when using 0-10v control coming from the acorn. I am seeing some wandering of frequency and accordingly rpm too but also an increase in fluctuations. Its not bad, but for rigid tapping I think it needs to be better. So I definitely need to figure this out before I get too confident it'll work reliably. I was able to rigid tap but I also had it automatically abort because the rpm had drifted enough that it didn't have time to sync up. I'm able to manually get it close using different analog input offset settings in parameter P4.02. I'm pretty confident with all your suggestions I can get the kinks out. I thought maybe temperature might have something to do with the change in rpm but I haven't had time yet to go through with a volt meter to check output voltage from the acorn. Probably Sunday I'll get to that and make sure the wires are as you suggest.

I'm also considering a proper though bore $450 encoder good for 8000 rpm instead of my rigged up $25 dollar encoder using RC car belts and pulleys ( see picture ) Its working so far without too much fluctuation from the skinny timing belt but over-all its max rpm is not going to allow me to leave it connected all the time.

Also, I am considering going 1:1 with my motor pulley and try using solely frequency to get the max RPM of 5400 from the motor. It is nice, however to have the possible 7000 rpm with the 2:1 pulley configuration I have now at 120 hz max instead of using 1:1 and 180 hz. Tough to really know whats best but surely low speed torque will benefit and probably some stability in rpm when tapping and slow rpm fly cutting ?. Do you have an opinion as to what would be best ? Is 7k rpm worth it, or should I try to run 1:1 and therefore probably have better torque but lose 2k rpm. I do tend to mill small items requiring small end mills ( 1/4 inch and smaller down to 3/32, for instance ), so rpm is important, especially for carbide end mills.

Best Regards,

Jake


View attachment 259295

I'll jump in here and say that more RPM is better. My mill has a 2.5:1 pulley setup that gives me 8800 max RPM (120 hz) and 345 RPM (30 hz) at the low end using 120 hz. Could crank up the Hz but the spindle bearings may not take it.

I too typically use small end mills. Normally I don't use anything over 1/2" except a flycutter. Seeing your mill run at higher feed rates and getting good surface finishes is very satisfying, at least for me.

My 2 cents.

 

[spumco]

I'll jump in here and say that more RPM is better.

I agree, except if his spindle bearings are dodgy. If that's the case, I'd stick with the 1:1 or 1:1.25-5(ish) and live with lower feed rates on really small end mills until you're ready to upgrade the bearings to something expensive. Or if you want an 'opportunity' to change the bearings sooner rather than later, go with the higher top end. Run the mill until the bearings grenade and bite the bullet sooner.

The one reason I (sort-of wish) I had more low-end grunt is for making large holes in thicker steel. Drills are so much more efficient than helical interpolation, and I can get a nice long drill bit whereas a 4" long 1/2" end mill is going to be a noodle. What I want is a 10k spindle with tons of grunt down low, like a Fadal with an auto gear changer. Not happening, so I compromised and limit myself to 3/4" drills and a 7kRPM top end.

I read the Acorn manual again (see how nice I am?) and it's got a pretty sweet feature where you can tell it the encoder is on the motor or spindle. I think it does the math, based on the pulley/gear ratio you input and maybe knows where the index should start as well as the actual spindle RPM. You may be able to get away with just a motor encoder and not have to have a spindle encoder at all.

Look at page 14-20 in the Acorn manual. Parameter 36, Bit 4. That's your huckleberry.

If the Acorn can do the pulley math but not do the index signal math (index on motor), then you're in the same situation I ran in to setting my machine up.

It should still rigid tap with the index signal on the motor going to the Acorn, but unless you have a 1:1 pulley you will not be able to peck tap since the index signal is what starts the synchronized axis movement. An index signal that varies in relation to the spindle will cause the tap to start at a different rotational location each time it tries to re-start. Much excitement and buggered threads result.

It may be possible to hook up a proximity sensor on the mill head and trigger it with a plastic pulley that has a steel set screw sticking out. Wire the proxy to the BOB and tell Acorn that pin is your spindle index signal. This is what I did instead of adding a separate spindle encoder.

The problem you may have with this hack is that the Acorn is looking for a differential signal. If it were me, I'd try sending a 5v signal to the Acorn Z+ input and see if it accepts a non-differential signal as an index trigger. If it does, you're in business and can run the encoder wires to a terminal block and then to both the VFD and Acorn. Index (proxy) is connected to Acorn Z+. Acorn Parameter 36 is set to Bit 4, and away you go.

If it doesn't accept a non-differential signal, there are open-ended to differential converter boards for your index signal. Check out Mesa Electronics (of Linux CNC fame) - they have a couple. I happen to have a spare I'm not using, and they're cheap. In this case the proxy goes to the converter and then to the Acorn as a Z+/Z- spindle index signal.

Just read the GS3 manual again, and unless I'm mistaken the FB card doesn't even have inputs for an index signal. Nor is there any provisioning for motor positioning or homing. No joy on rotating the spindle with the VFD alone like with my Hitachi.

If you're dead-set on having a separate spindle encoder AND motor encoder, it just so happens I have a through-bore encoder, essentially the same things that's already on your motor (Dynapar). 5/8" bore, 2048PPR. PM me if you want a deal on a grown-up brand-name encoder. I took it off my Black Max 'cause I can't use it; my VFD can't take encoder input frequency higher than 2kHz and I had to drop down to a 100PPR encoder.

I know, I know TLDR. Last thing... The minor spindle speed variation you're getting shouldn't really matter much. The Acorn is moving the Z axis according to the encoder, and it's constantly adjusting the axis speed to match the RPM. Otherwise when the spindle slows down due to cutting load (big tap) it'd wreck the threads. Even if you've got noise on the 0-10VDC line to the VFD and the spindle is hunting, the Acorn knows what's going on and tweaks the Z-axis advance because the encoder is reporting the actual position (RPM).

-S

 

[JimDawson]

Just read the GS3 manual again, and unless I'm mistaken the FB card doesn't even have inputs for an index signal. Nor is there any provisioning for motor positioning or homing. No joy on rotating the spindle with the VFD alone like with my Hitachi.

I agree. That is the reason I am using the external spindle encoder on our lathe. I spent quite a bit of time researching this. I'm closing the loop at the controller rather than at the GS3. I didn't even order the encoder card, I couldn't see any advantage to having it, and a couple of downsides, like the motor encoder and the spindle encoder battling for position. We'll see if I can set up the system tight enough to beat the original 1 degree indexing capability. Bet I can

 

[TomS]

Agreed, good quality bearings are key to high spindle speeds.