Low cost / high speed / high torque stepper solution

[Brady Watson]

It is not a plug and play kind of thing though. It would take someone with some electrical knowldedge and some research on their part as well.

VERY good point, and readers should understand this.

All theories aside, and not to minimize Mike's research, but it is very hard to beat the performance and value of a PRS Alpha drive system. Yes, it is not "cheap"...but what is "cheap" at the end of the day, on a tool that is used to make a living?

Forget that Alphas have positional feedback and can go into 'Alpha Mode' or self-correct. The tool will run all day long (and day & night) without missing a beat, if you aren't running a beat computer, and cut with 150 pounds of cutting force, reliably, up to 12 inches per second. It will rapid position at 30 inches per second as well, although the gearboxes are really only rated for around 22 IPS. It was carefully researched and engineered to run all day, every day - production duty.

300 pounds of cutting force???? 2X that of an Alpha? Who wants that? I've only put my tool into 'Alpha mode' exactly two times since 2006. Both times were induced by parts moving and wedging the machine on one side of the gantry, forcing the tool to adjust for this & eventually stop with a safety alarm. 300# of force on a slow moving metal mill is fine. On a fast moving router - no way, Jose. Think about safety. 300#s on ANY part of your anatomy ain't pretty.

If you are tinkering around, no doubt you can get impressive numbers with off the shelf parts...but what kind of warranty and support are you going to get?

-B

[Gary Campbell]

Kenneth...
Thanks for the clarification. Mike is a dear friend of mine and we have spoken about this issue. I am sure after those conversations he would be the first to agree that he is NOT talking about the current PRSalpha motor/driver combo.

Like Brady says, there may be hundreds of lurkers out there, there is no reason to have them think that a few $200 motors, $69 drivers and some homemade gearboxes will double the performance of the alphas currently being built.

Mike's posts here, and on other forums, has always been spot on. He has offered up a great solution for those wanting a DIY budget upgrade for one of the earlier SB models or a very cost effective solution for a DIY machine build.

There has to be some common sense applied here. All the motors are coming from the same place. When could we expect to get, from the same vendor an under $300 package that outperforms an over $1000 package? Even nitrous cant do that!

The current alphas are very expensive. And worth every penny if your living depends on them.

[bleeth]

If any of you "lurkers" haven't noticed there is a current batch of SB owners in the process of adding the PK299 F4.5A's to their machines with gearboxes. See the subject (Gear Boxes) under modifications to Shopbots. Some have the early Alpha's with 1:1 motors that need the gearboxes to get improved resolution, and some have earlier PRT's or PRS's Standards and need to step up resolution and/or speed with motors and gearboxes. The program is designed for those who already have Gecko drive controllers (Either independantly built or SB 4G types). No one who has checked into it thinks that doing so will equal the current PRS Alpha. Or if they do they are fooling themselves.
Not all of us can afford to or wish to possibly overextend ourselves and upgrade to the latest and greatest. Therefore the above conversion is a way to vastly improve our current setup without as high an investment. I can tell you, since I am in the middle of it, that the complete process of identifying ALL of the specific parts and engineering is not a "123-I'm done" deal and takes a definite time and research investment. I have also pointed out in previous posts that if I had my druthers I would just buy a new Alpha.
The defense of the new Alpha against "hotrodding" an older PRT or PRS standard is akin to an owner of a new Northstar Cadillac braggin that his car is technically superior to a Studillac. (If you don't know what it is look it up). By the same token claiming that it is the same or superior to do those conversions to an older machine and have the equivalent of a new one is also patently absurd.
Not pointing any fingers here or flaming anyone-we ARE all friend here-just wishing to keep the conversation to the point.

[fozzyber]

Thanks to all here. I am watching this post as well as the gear box thread,
I have a early prt circa 2001, with non g4 , I don't need an alpha but would like to increase speed and res, I am watching these postings with great interest. I think my upgrade path will be controller, gear box, motors.
in that order. thank you to all the creative minds out there.
Now hopefully after I give Obama his cut for last year (#*&%@*&%$) I can toss some jack at the shopbot,,,,


Jerry

[robtown]

Must we... really? Don't we get enough "political discourse" in our faces on a daily basis from EVERYWHERE else?

[richards]

Talking about stepper motors and drivers is a lot like talking about religion or politics, i.e. without a great deal of study and experience everything is superficial.

There's no magic in one type of stepper motor over another or one kind of driver over another. A stepper motor that is properly selected for the job will do the job properly all day long. A stepper motor that is too small for the job or mismatched to the driver will cause problems.

Rule Number 1: Determine how much torque you need to do the job.

Rule Number 2: Determine how fast you need to move and axis.

Most people would think that the answer to Rule Number 1 is "how much torque is available?" That's not a good answer. The material you're cutting might limit the torque required. The cutter you're using might limit the torque required. The ability of the CNC's frame to push a cutter without having the frame flex might limit the torque required.

What's a safe answer? 80 lb*in. Why is that a safe answer? It's safe because that's the torque provided by the Alpha 7.2:1 motor/gearbox, so we can assume that the PRS machine can handle 80 lb*in of torque without too much flexing.

Rule Number 2 is dependent on Rule Number 1. As soon as you know how much torque you need, the next step is to open a spreadsheet and run the numbers, using factory specs from good companies. (I use Oriental Motor. Shopbot uses Oriental Motor. Oriental Motor is a great company.)

As you 'run the numbers', you'll find a lot of motors can produce the torque required when they're used with a belt-drive gearbox, so which motor do you choose? My answer is to choose a size 34 motor that has eight wires or leads that has the lowest inductance. A size 34 motor has more surface area to dissipate heat. (Stepper motors can and do get hot - really, really hot.) Oriental Motor has two popular motors that meet those requirements: The PK296-F4.5 and the PK299-F4.5. Both of those motors can be wired bipolar series, half-coil or bipolar parallel. Bipolar parallel gives the most torque at high speeds, but at the cost of high temperatures. Bipolar series gives the most torque at low speed, but but at the cost of VERY poor high speed performance.

Now look at the numbers again. 80 lb*in of torque is the same as 1,280 oz*in of torque. BUT 80*in at what speed? Holding torque (motor stopped) or jogging torque (motor moving at maximum velocity)? Stepper motors trade torque for speed. Look at the manufacturer's torque curves and you'll see what I mean. The faster the motor spins, the lower the torque the motor produces. If you look carefully at the torque curve, you've probably see a 'shoulder' or 'knee' point where torque quickly drops off. The secret is to keep the speed on the high torque side of that 'shoulder' or 'knee' point.

Look at the AS98AA motor and compare it to the PK9x motors. It's very similar to the PK296-F4.5 motor. For all three motors, that point is about 1,250 RPM. The AS98AA motor has about 225 oz*in of torque at that speed, the PK296 motor has about 200 oz*in of torque and the PK299 motor still has about 300 oz*in of torque. So, we're still comparing apples to apples.

At this point, we know that with the proper power supply and the proper stepper driver all three motors will cut at the same speed. If we give the AS98 motor a 3.6:1 belt-drive, it would have the same usable torque as the PK296 motor with a 3.6:1 belt-drive.

BUT, we have a problem. The maximum belt-drive ratio that we can get without going to a multi-stage design is 4:1. AT JOG SPEED, with a 4:1 ratio, the PK296-F4.5 motor will give us about 800 oz*in of torque or about 50 lb*in which is less than the 80 lb*in that we're looking for; however, the PK299-F4.5 motor can handle the problem. It has just over 300 oz*in X 4:1 = just over 1,200 oz*in which is just over 75 lb*in or just under 80 lb*in, depending on how you look at the chart.

Now that we know the PK299-F4.5 motor can match the AS98 motor for torque and that a 4:1 belt-drive can give matching high-speed torque, it's time to see if the pulse rate to drive that motor/belt-drive at 22 IPS (remember, that's the recommended maximum speed for the Alpha 7.2:1 motor).

Using a 1.5" diameter pinion gear (30-tooth) would move an axis 4.7124 inches per revolution. So, 22 IPS would require 22 / 4.7.124 X 4 = 18.6741 revolutions per second (remember, we have to compensate for the 4:1 belt-drive). 18.6741 revolutions per second equates to 1,120 RPM. We're still within the good torque range of the motor.

The other fly in the ointment is the pulses per second. A Geckdrive requires 2,000 pulses per revolution. The Alpha driver normally requires 1,000 pulses per revolution. So, 18.6741 revolutions per second with a Geckodrive requires 2000 X 18.6741 pulses per second. That's pretty fast. The Mach controller can easily handle that speed. Some of the Shopbot controllers might have a hard time keeping up.

So, depending on the controller that you have, jog speed might suffer when comparing an Alpha motor to a PK motor.

Cutting speed would not be affected, at least speeds up to 12 IPS.

Now it's time to address the 600 lb gorilla that's been sitting in the corner. What about 'Alpha Mode'? The AS98 motor can recover when it momentarily slows down. What I'm going to say will not make me popular around the water cooler. In my opinion and from my experience, Alpha Mode is not what it's cracked up to be. In my early days with my PRT-Alpha when I tried to push the envelope as far as possible, I often forced the machine into Alpha Mode. What it bought me was a divot in the cut that usually ruined the piece. Think what happens when the machine enters Alpha Mode: an axis can try to recover for up to 1/2 second before the driver sends a fault signal. (The actual time depends on a 'pot' setting on the stepper controller.) If you're cutting at even a lowly 5-IPS, that 1/2 second delay means that one axis is up to 2-1/2 inches behind its expected position. The three axes are not coordinated to all slow down proportionately. Result = ruined part.

I quickly learned to slow things down so that the machine NEVER entered Alpha Mode. Result = NO ruined parts.

Stay within the capabilities of the motor and the driver and the power supply and the transmission and the cutter and the flex of the machine and you'll have a good experience. Push any one of those things too far and you'll be pulling out your hair.

------

If you're really into electronics, you can always add an encoder to a stepper motor that would display any missed steps and, with the proper programming, send a fault signal to the controller to stop the cut if a motor missed too many steps during a pre-determined time period. (One missed step would be 0.00059 inches. Depending on the type of cut, a few missed steps might be acceptable with no visual indications that a problem had occurred.) The encoder and microprocessor (not counting design time, programming time, or the cost of the initial prototype circuit board) is about $50 per motor.

[richards]

The fact that a Geckodrive might require a faster pulse rate than the Shopbot controller can provide when fast jogging speeds are desired drove me to the Geckodrive site where I noticed that their NEW step multiplier board board for the G201x and the G203v has been released. (This is a new design that purportedly eliminates the 'jitter' that Brady discovered on the older design that was used with the G202 and G201 stepper drivers.)

A step multiplier board lets you use a slower pulse stream to drive the stepper motors. The documentation on the new board is a little fuzzy, but it looks like you have the choice of 1:1, 2:1, 5:1 and 10:1, so, using the 2:1 setting, the Shopbot controller would only have to furnish half as many pulses per second. The drawback is that resolution is also cut in half; however, at 2:1, the resolution would be 0.0012" per step. I doubt that you could get repeatable movements any better than that on the PRS doing production work. (That's less than 1/2 the thickness of a piece of copy paper.)

Now for the fun part.

If someone were doing step and repeat work where he needed the fastest possible jog speeds between parts but still needed the highest possible resolution when cutting the parts and he wanted to use the PK299-F4.5 motor and a Geckodrive stepper driver, it looks like he could use the new pulse multiplier board and an electronic switch to switch between 2,000 steps per rotation for the highest resolution and 200 steps per rotation for the fastest possible speeds. (A similar 'trick' is used by Shopbot to drive the Alpha motors. One resolution is used for cutting and another resolution is used when jogging. The Alpha stepper drive has a provision to allow on-the-fly speed/pulse changes.)

Ramping would be critical, but it is conceivable to also alter ramping withing the program so that the jogging routine used one set of VR parameters and cutting used another set of VR parameters.

As long as I'm just writing about potential and not about what I've been able to do at the test bench, let me point out the advantages of using a standard stepper motor with a Geckodrive stepper driver and a pulse multiplier board with electronic switching:

1. Better resolution with a 4:1 belt-drive than the Alpha motor gets with a 7.2:1 gearbox. (Remember that the Gecko uses 2,000 steps per rotation and the Alpha uses 1,000 steps per rotation. The Alpha would have to use an 8:1 gearbox to equal the resolution of the Gecko driven 4:1 belt-drive.)

2. Faster jog speeds. The Alpha is limited to 22 ips due to its gearbox. The belt-driven standard stepper is limited only by the ramping times and torque of the motor used.

3. Cost. The basic motors, drivers and power supply using a standard motor and Gecko for the entire system cost about as much as one Alpha motor/driver; however, to be perfectly fair, if you hire someone to design and build a system using standard motors and Gecko stepper drivers, it may actually cost more for the standard motors and Gecko stepper drivers.

And now back to that 600 lb. gorilla that I mentioned in my previous post. The standard motor and gecko system does NOT have a feedback/Alpha Mode.

[steve_g]

Mike:
I see your point regarding ruined parts even though the Alpha mode was entered… In defense of Alpha mode, I often cut small step and repeat parts, these jobs are likely to be unattended in that I’m in another part of the shop. In the few times that alpha mode was engaged, it resulted in one bad part while the balance were still indexed properly with previous cuts.

Thanks for the time you put into intelligent, researched postings!

Steve

[richards]

Steve,

You have a perfect application for Alpha Mode.

To be perfectly fair, almost always, my problems came when I was cutting panel parts from plywood, i.e., long multi-axis cuts at high speed. Plywood is notorious for being inconsistent. My practice cuts, where I determined feed speed, where likely on pieces that were more consistent than the production sheets.

Until I went to the first Maker Faire in San Mateo, I didn't know that the Alpha stepper controller's fault time was selectable. Ted Hall took the time to show me how it worked. His eyes lit up when he explained how it all worked. That's a sure sign that he really enjoys what he does. That's also when I gained a lot of respect for him and his products. After all, my ignorance on the proper procedures to use and the proper settings could have been cured by picking up the telephone years earlier and simply talking to someone about my problem. Had I set the delay time to a very short setting, I would probably have learned in a hurry that I was pushing the machine too hard for the kind of cut that I was making.

I never regretted having purchased the Alpha model.

If my health ever returns and if I ever buy another machine, I'll probably order a Shopbot without electronics and install my own controller - not because my design would be better but simply because I enjoy the electronic side more than the production side. To me, breathing sawdust is a necessary evil to get to see electronics at work. (In my former life as a process control computer designer, my last major project had options to use from six to twelve stepper motors to control a Kodak 5-S professional photo printer. Instead of using air cylinders and rotary solenoids, everything was controlled by stepper motors, filters flags, paper advance, mech plate, paper curtains, everything.)