10EE Restoration and VFD Closed Loop Conversion

[tailstock4]

How does the encoder wire into the VFD? I’m very fuzzy on what the encoder actually does within this sort of set up/use.

Will


Sent from my iPhone using Tapatalk

The encoder attaches to the back of the Black Max. The feedback cables into the VFD. This supplies the drive with exact parameters inside the motor. A sensorless vector uses feedback based on an algorithm which is an approximation of motor parameters. You can find more information on the internet that explains it far better than I can.

 

[will.mcray]

The encoder attaches to the back of the Black Max. The feedback cables into the VFD. This supplies the drive with exact parameters inside the motor. A sensorless vector uses feedback based on an algorithm which is an approximation of motor parameters. You can find more information on the internet that explains it far better than I can.

Thanks!

 

[mksj]

Many VFD's do not support the use of a motor encoder or require an additional plug in circuit board. The encoder needs to have positional control and the appropriate resolution/voltage, this is usually specified by the particular VFD. Typically the the use of encoders is relegated to the premium VFD models, so at the end of the day you are probably looking at 3-4X cost increase vs. w/o and the parameters need to be dialed in. Newer VFD's use flux vector control which is the next step forward in sensorless vector control, the speed accuracy is around 0.1%, moving to a motor encoder you are getting in the range of 0.01% or better with finer resolution of motor stability and zero speed control (locked motor as used in cranes). Used in this build, in addition to the belt drive isolation in the 10EE may have an effect if one is looking at 50-100 microns tolerances. The VFD with an encoder from my read is an equivalent (as good as) to the original DC drive, and would be expected to have more precise dynamic speed control as well as acceleration/braking repeatability. Whether you need '0' speed control and/or that degree of resolution, probably 99.99+% of lathes you would never now the difference. Most lathes you would never be able to detect a surface finish difference, given the gear drive and the tolerances/vibration in the machine. Other than the CNC world where you may use BLDC motors with encoders for positional control, I am not aware of any current lathe manufacture (other than the 10EE) that use an encoder in a manual lathe. Similar to CNC there are also tuning issues to get the most out of an encoder system.

There have been several lathe installs that I did in the past where I had the option to use an encoder (came with the vector motor), but at the end of the day the lathes would hold +/-1 RPM regardless of load using sensorless vector and mid quality VFD's. The newer flux vector my understanding is that the control algorithm builds an increment mapping of a motors properties and more dynamic motor control through the 360 degree rotation. They are complex calculations that take in many factors of the motor properties. Running the VFD autotune is important so it maps the specific motor parameters you are using. For us mere mortals, a good VFD with a decent 3 phase motor is about as good as it is going to get. I am not an engineer, just a user of the the technology. There are many online white papers and discussions on methods of VFD motor control.

 

[tailstock4]

I handled the wiring and control cables as follows:

• Used shielded power cable between the motor and VFD, between the VFD and the power switch, and also to the braking resistors.
• Used shielded control cable in the motor compartment and the VFD compartment and tried to avoid running these cables parallel with the power cables.
• Used a magnetic contactor for the power up of the VFD so that in case of power loss the drive could not restart.
• I also used a separate transformer for all secondary power requirements except the encoder which uses its own DC power supply rather than the drive’s power supply. I also mounted and wired in the old interlock switch for the spindle lock.

For the back gear adapter plate, I used a 1” ground 1018 steel plate. This was mounted on an 18” face plate on my American Pacemaker. The outer recess for the motor and the initial hole for the motor shaft (which is 1 1/8”) were done in one setup to ensure alignment. I then made a bushing and alignment pin to locate the back gear box on the opposite side of the adapter plate.

Many conversions use the original dowel holes, but I found that this caused the shifting to be a little rougher because of slight misalignment. So, I mounted the motor and gear box on a large angle plate in a radial drill where I trammed all this in then adjusted the back gear housing until shifting was completely smooth. I then clamped this housing and removed the outer cover and drilled and reamed new dowel holes. I then plugged and sealed the old holes. This produced a better result in shifting. The four mounting bolt holes were drilled and tapped in this plate.

The other modification I had to make was that the 7.5 hp Black Max motor shaft had to be turned down from 1 3/8” to 1 1/8”. In addition, this also required a 5/8” motor shaft extension to maintain the correct length. To do this, a temporary brass bushing for the motor shaft was needed in order to mount this in a steady rest. A hole was drilled and tapped. This then received a threaded extension that was screwed in, Loctite’d and taper pinned in place. The final thing I did before removing bushing and steady rest was to put a new center in the end of this extension. This allowed the motor shaft to be mounted on the new center and turned to final dimension. After this the existing keyway had to be slightly deepened. The sliding gear then had to be bored to a close sliding fit over the motor shaft and a keyway broached into this gear housing.

The final steps were to make a new elongated key which I surface ground to size and mounted to the shaft keyway with two counterbored cap screws. The last two steps were 1) to enlarge the drive shaft hole in the adapter plate to accept an oil seal, and 2) to provide a shallow counterbore for the top jack shaft and bearing.

One other change I made to the mechanical drive was to make a new top pulley to make the final drive ratio 1:1. I sped the drive up to retain the original 4,000 rpm. I encountered two problems when I did this. The first was increased belt vibration which was ultimately solved by a pair of matched Browning belts. The second was a small increase in vibration in the sliding gear and spline coupler in the motor shaft. This was solved by disassembling and individually balancing each of these parts. After this I came to the conclusion that I would not increase the reduction of the final drive greater than 1:1 and maintain the 4,000 rpm. I believe if a person wants greater reduction or greater pulley size on the spindle, then the resulting slower spindle speed should be accepted. In other words, 4,000 rpm I thought was about max for the belts and gear box.

 

[tailstock4]

Rabler said:

---

I'd be interested in a little more description of the closed loop encoder used with the drive. I understand the concept, would like to hear more about what your process was for coming to this particular design. I have a DC motor that I could try to resurrect using some form of modern controller, but a VFD approach has merit. I would be inclined to retain the back/reduction gear but would need to cut a few gears to replace worn out ones in my current reduction gearbox.

To understand my choices in drives it helps to remember that my original goal was to get to the performance of the original DC drives. One of the things that people really like about the DC drives is their abundance of torque even at low rpm’s. Closed loop systems do supply a high degree of control over speed and position. This coupled with the braking does amazing things for the ESLR. But these are not the only advantages.

Closed loop systems such as mine have the ability to supply a 200% boost in torque throughout the range including all the way down to 0 rpm or stall. The reason they can do this is the encoder takes information from the motor in real time and relays it to the drive. This information is supplied regardless of speed.

Sensorless vector drives can boost torque up to 100% and down in rpm to about 10% of the motor’s rated speed which in the case of these four-pole motors would be around 170 to 180 rpm. The reason torque falls off so much as the rpm’s go down is because with a reduction in motor speed there is less feedback for the drive to interpret and make adjustments.

An example of this I believe can be found in a test that the late Harry Bloom performed with his 5 hp sensorless vector drive and published on PM. His machine was a 4,000 rpm machine. The following test was performed in direct drive. For the test he used a 1 ¼” piece of 1018 steel. His spindle speed was between 150 and 200 rpm. Depth of cut was .180. The feed was around .015. Harry said that upon attempting this cut it stalled the drive. He also said that he would not repeat this test as he believed it was close to the maximum ability of the machine.

I however did repeat this same test with the mg machine which is an 4,000 rpm machine that was reduced to 3,000 by an increase of the top pulley size. The result was a slowdown approaching a stall, but torque was quickly applied and then continued to advance the cut. The result from the close loop VFD machine (which was a 4,000 rpm machine with 1:1 open belt drive pulleys) was a much smaller initial slowdown and a much quicker response back to the original rpm’s. Another example where something like this might be encountered might be the engagement of a large drill bit where the need for torque rises quickly and speed drops.

I believe the close loop machine more closely imitates the performance of the original and perhaps even an improvement in terms of a more seamless transfer of power from an unloaded condition to a heavily loaded condition. One area I believe the DC motors are superior is if the DC motor and the VFD close loop motor were coupled together, I have no doubt that DC motor would win this tug-of-war. But at this level of power, the 10EE would lose. There would be more power than needed.

The cost of the drive, interface module, encoder, and cable along with the Black Max 7.5 hp motor and brake was $4,000 in late 2019. There is more need for additional tuning beyond the initial run in and set up of parameters. This drive was hooked up to an oscilloscope twice – once for set up and once for some changes I wanted. DP Brown supplied all the consultation and technical support including tuning for no increase to the initial costs. One side note, this same system with 5 hp would have been $3,500. For the extra $500 I decided to go with the 7.5 but this also comes with a little more complication with the back gear box.

My reason for this post was not to promote this drive or even suggest it is the best method as there are other methods such as larger motors without back gear and even servo drives. But I cannot comment on them as I have no firsthand experience with these in conjunction with a 10EE. In the final analysis I believe I accomplished my goal and am pleased with the results.

 

[Dabbler]

@tailstock4 A great description of your tradeoffs and your mods. Thank you for all this detail.

 

[Beckerkumm]

I appreciate the write up. The 10EE has a cult like status among those who have used a good condition machine with the original drive system. The technology involved for the 1940s and 1950s is pretty impressive. The design seemed to combine pretty good rates of metal removal with high precision in an industrial type setting. Monarch evidently wanted the highest precision but didn't want to sacrifice efficiency. The D1-3 choice kind of confounds me though. The early 10ee with the clutch and oil bath bearings probably had some inherent limits in the machine itself, as with the flat belts. The evolution of the machine is interesting in its own right. I know Monarch was constantly improving the design. CVA on the other hand licensed the early 10ee and made changes very sparingly.

I understand the fun of designing a system to replicate the original and whether the economics work out becomes secondary to the experiment. We are all better for your efforts, whether we use them or not.

Dave