Trade in 4G Controller for New Model Controller?

[rb99]

Is anyone with a 4G controller interested in upgrading/ replacing it to the new one?

Thanks,

RIB

[chipster]

A new one? I just bought the 4g.

[gerryv]

I'm confused by what you're saying here richard, please clarify.

[rb99]

There is a new controller using Oriental Motor Drivers instead of Geckos.

RIB

[erik_f]

I talked to the guys at SB about it. They stated there is no major advantage to the new box, other than its construction (the new box is sort of the like an alpha box). Speed and power they said were a tiny bit better, but not enough to spend the money. It would pretty much be a sidegrade not an upgrade. At least thats what I was told.

[gerryv]

Companies do not always change suppliers because they can offer better output, better quality or, for that matter, better price to their customers. Quite often, it may be a situation where they're able to negotiate a little better margin by combining volume or save a little on labour costs to offset other increases as a means of holding prices. Sometimes it's part of a strategy to reduce the number of suppliers so that you have more clout with those that remain.

Personally, I doubt that there is any issue whatsoever with Gecko quality or dependability and, although I grumble about wanting more from Shopbot and Vectric in terms of 3D/4-axis ease of use due to my own weaknesses and interests (no harm in trying eh!) I'm inclined to believe that what Eric said above based on a chat with SB is accurate and noone'ss getting shortchanged or missing out on something big here.

[richards]

A change of controller is not always a better choice. The 4G uses Gecko stepper drivers, which are about the best that you can buy for an aftermarket upgrade. The new controller uses a motor/driver combination from Oriental Motor that gives similar torque and speed and the Gecko and the PK296A1A-SG3.6 motors wired bipolar series that are similar to the proprietary motors/gearbox that Shopbot buys from Oriental Motors.

If you read the forum carefully, you'll probably notice that many posts deal with erratic behavior, i.e. ELECTRICAL NOISE. I'm not aware that Shopbot has addressed that problem with the new controller. Typical methods used by others in the industrial controller arena includes the use of opto-isolators on each input and on each output. I don't think that the new controller uses opto-isolators on each and every I/O line.

For those who want/need the highest possible performance from a 4G machine, I would suggest that you spend about $1,000 and buy four PK296A2A-SG3.6 or PK296A2A-SG7.2 motors, a 35VDC 12A power supply and four 33K 1/4-watt resistors. Replace your Shopbot supplied proprietary motors with the PK296A2A-SGxx motors, replace the power supply, replace the current limit resistors on the board with the 33K current limit resistors, wire the motors half-coil (Black/Yellow, Red/White) and watch you machine take on a new life.

The 3.6:1 motors have about 1/2 the available torque of the 7.2:1 motors, but they also have 2X the jog speed. Many people that I've received emails from use the 7.2:1 motors. I, personally, have 3.6:1 motors on my testbench. If I were to do it again, I would probably buy the 7.2:1 motors and sacrifice a little top end jog speed for higher torque.

Solving noise issues is complex. Designing and building a "noise-free" controller is difficult, especially when the manufacturer has a relatively noise-free facility.

I guess what I'm trying to say is that if you already have a 4G that is working well, the newer controller probably won't buy you additional speed and will probably not solve any electrical noise problems that you might be having.

[Gary Campbell]

If I read his post correctly Richard is looking to trade backwards? Mike may have the reason?
Gary

[ted]

Hi Guys,

I will just chime in with a little data here.

As part of our general process of trying to improve our products, we have been working at upgrading the Control Box for our PRSstandard ShopBots. About 2 months ago, we started shipping the new system. It is fully described and illustrated at: www.shopbottools.com/prsstandard_motors_&_drives.htm (http://www.shopbottools.com/prsstandard_motors_&_drives.htm) for those who would like to have a look.

It is a nice, sealed, industrial control box with an integrated disconnect and router/spindle control as well as a key-activated, contactor disconnect for disabling power during tool changes. As stated in the description, ALL 12 inputs and 12 outputs are FULLY optically isolated (please note that Mike is wrong about this, above).

The Control Box also uses new OM drivers. These are very robust drives that have similar performance to the gecko 203. They have a few additional features that make them attractive to us including more step rate options and a nice case for mounting and heatsinking.

We have this Control Box available for upgrades for older PRTs, but do not feel it represents a performance improvement for those who already have a Version4g. It is probably also obvious that it is a more expensive box than the old PC case, and it includes a number of new features. So it is more expensive than the Version4g upgrade and did force the price of the PRSstandard up a bit.

The current tool in my barn is one of those old PRTs, and I've just added one of the new Control Boxes to it (I'm going to put a Retro-Z on in the next few days). I hope to be reporting about my MAKEcation and using this tool over the next few weeks. FYI, my MAKEcation finally got started just last week, about 1 month later than planned. See: www.shopbottools.com/teds_report.htm (http://www.shopbottools.com/teds_report.htm)

Ted Hall, ShopBot Tools

[richards]

Ted,

I apologize for saying that the Inputs and Outputs were not opto-isolated. My information was wrong.

Isolating the external signals from the internal signals is something that *should* eliminate many of the problems that 'botters have posted here on the forum for years.

We emailed back and forth about opto-isolation years ago, including schematics. It's good to know that the new design incorporates some of those changes.

-Mike

[rb99]

Does the 4G have this optically isolated thing?

RIB

[richards]

I'll let someone from Shopbot answer that question; however, opto-isolation can easily be added to any Input/Output line on any controller. All you need are a few inexpensive parts. The actual opto chip, which you can buy from www.jameco.com (http://www.jameco.com), costs about $0.25 each. Part number 4N25 is popular. You'll need a resistor on the input side that limits current to about 10mA, so you could use a 1K or 1.2K resistor if your external power supply is 12VDC, or 2K or 2.2K if your external power supply is 24VDC. If you use 5V for the external power supply, you would use a 330 ohm resistor.

On the output side of the opto-isolator, you would need a 4.7K pull-up resistor.

You can easily hand wire opto-couplers to a Radio Shack proto board. So, the total cost to add home-brew opto-isolation to all inputs and all outputs on a machine is only about $10.

The opto-isolator separates all outside signals from all inside signals through the use of an L.E.D. (Light Emitting Diode) and a phototransistor. When a signal is active, the LED turns on. When the signal is inactive, the LED turns off. A phototransistor detects the state of the LED. If the collector of the phototransistor is connected to the Input Signal, the input goes LOW (active low) when the LED is ON and the controller knows that the input line is active.

Because there is no direct connection from devices outside the controller to the circuit board inside the controller, electrical noise is greatly reduced.

Because an opto device is slow (a millisecond or so, instead of a nano-second), false triggering is greatly reduced.

Because an opto device can easily be used with 12VDC and 24VDC, it is easy to interface ordinary proximity sensors. In fact, when one of the Shopbot supplied proximity sensors died unexpectedly, I added a home-made interface that allowed me to use a 12VDC proximity sensor. Since adding that opto interface, I haven't had a single false trigger from either proximity sensor or from the E-Stop.

[gordon]

The above description will work well for the isolating the external signals, but this is only half the battle of eliminating noise that can interfere with the controls and communications. The new RBK controller board that ships with all standard tools now also isolates the internal controls from the other power in the box with and isolated 5V power supply and has truely seperate isolated grounds that prevent noise from the drivers and other devices from injecting noise back through the grounds into the controller and USB circuits. Unfortunately, this isn't the case on the V4G boards and won't be an easy job to change, in fact, this would be almost impossible on the V4G boards. Though, there are advantages to optoisolating the I/O other than noise reduction as Mike points out.

To address the second subject brought up, there is an advantage in speed by changing the gear ratio on a standard(V4G) that has 7.2:1 gearboxes, as well as the lose of torque. So, you need to weigh the benefits before making this change. I wouldn't recommend purchasing the stock 3.6SG motors unless positional accuracy isn't a concern. The off the shelf 3.6SG motors use spur gears as opposed the tapered hob gears that the special motors that we ship with the tools use. Compare the rated backlash of the 3.6SG gearbox of 90 arcmin to the 25 arcmin of the 3.6TH gearbox. This is a difference of 65 arcmin or approx 0.012" more backlash when using a 25 tooth pinion and a 3.6SG gearbox.

[richards]

Here is a schematic showing how an opto-coupler can be used. NOTE: In the schematic, the GROUNDS are NOT tied together, which is a point of contention between designers. For years, I followed the same theory as Gordon and isolated the ground on the internal power supply from the ground on the external supply, but Mariss Freimanis, the owner and designer of Gecko products, pointed out the advisability of connecting all grounds. After checking grounded and non-grounded signals with my Oscilloscope, it was obvious that Mariss was correct. So, totally isolating the power supplies, as Gordon wrote, is entirely acceptable but tying the grounds together is also acceptable is is the method normally used when connecting printers, etc. to your computer. With the grounds tied together, a static charge will not cause the potential to rise anywhere near an opto-coupler's 1,500V isolation range. With the grounds separate, it is fairly easy for a static charge to exceed 1,500V.


7572

[richards]

Gordon,
You've brought up a good point about the differences in gearboxes. The TH gearbox has an internal adjustment that allows the gears to be meshed tightly together to about 1/2-degree.

The SG gearboxes do not have that adjustment.

Yesterday, I spent a few hours searching the Oriental Motor website trying to get definitive data on those gearboxes. All that I could find was that the SG gearbox has between 1-degree and 2-degrees of backlash. Your figure of 90-arcminutes is right between those two figures, and is 60-arcminutes, or 1-degree, greater than the published data for the TH gearbox.

I have four of the SG gearboxes (PK296B2A-SG3.6 motors) and three of the TH gearboxes (Alpha motors supplied by Shopbot as part of the Alpha upgrade). I measured two of the SG motors and found that I have about 1/3rd-degree of backlash on those two motors. That is expected because those motors have only been used on the test bench and not on a machine, so the gearboxes have not had much of a load. I dropped one of the Alpha 7.2:1 gearboxes and measured its backlash. It had about 1/2-degree. That is also expected because that is the advertised backlash and that motor has been in regular production since it was installed.

Geared motors have backlash. The TH motors have an adjustment to reduce backlash. The SG motors do no. But without a minor amount of slop, a geared motor would freeze up.

My preference is a belt-driven transmission. Before upgrading to the 7.2:1 geared Alpha motors, I built several belt-drive transmissions for the original non-geared Alpha motors. A belt-drive has no internal backlash. (There would still be the backlash between the pinion gear and the rack, which would be equal no matter what motor/gearbox/belt-drive were used.)

The belt-drive is big and bulky and somewhat expensive. It will probably always be a "home-made" solution, but it works very well and is worth the time and cost for those who have motors without gearboxes. It also has the advantage that it can fully multiply the motor's torque. The SG-3.6:1 gearbox is limited to 20-lb*in and the SG-7.2:1 gearbox is limited to about 40-lb*in. The Alpha 7.2:1 gearbox is rated at about 80-lb*in. The 3.6:1 belt-drive, with a 450 oz*in motor (PK299 wired half-coil or PK296 wired bipolar parallel) produces just over 100-lb*in.

What I'm trying to say is that the SG gearboxes that I have (and those of other users who actually have them on their machines) seem to give excellent results. As they wear, they will probably develop more backlash, but at $257 each, they can be replaced easily. The TH gearboxes have an adjustment, but I have no idea whether any of us can perform that adjustment, or whether the motor and gearbox would have to be returned to the factory for maintenance. For those who have the desire, time and patience, they can build their own belt-drive and not worry about backlash.

Does any of this matter for the "regular" 'botter? No really. Some of us like to tinker, but most 'botters would probably rather let you and Ted select the proper components for their machines. I agree with that idea. My PRT-Alpha is "stock". It has Shopbot supplied motors. In fact, everything, except some replacement proximity sensors (and an opto-isolated I/O board), are all Shopbot supplied. So, I have several test benches loaded up with stepper motors, stepper drivers, power supplies and various odds and ends, but the CNC machine is left alone.

[bleeth]

Will the new TH Alpha motors function properly with a 4g type controller?

[Brady Watson]

Yes. Any Step & Direction drive will work on a 4G or Alpha.

I'm not going to tell you it's plug & play, because you'll have to make up a few adapters and account for a few of the Alpha driver's features when you do wire it up. Definitely do-able. I've done it.

-B

[bleeth]

Thanks Brady: I'll get back to you if/when I decide to do this. SB wants too much for my budget for the whole "official upgrade package" but I haven't checked out pricing for motors only or if they can be done from Oriental motor. Might pick your brain on the way to making a decision too!!

[br928]

Will the new TH Alpha motors function properly with a 4g type controller?

I don't think an alpha motor can be driven with a 4G controller. Someone correct me if I am wrong. Without the closed loop feedback, what would be the point? Increased torque? And I am not sure you would get increased torque w/o the associated alpha drive.

[Brady Watson]

No problem, Dave.

I doubt you will find a cheaper vendor for the motors, drivers and often overlooked (and expensive) cables to go between them, than ShopBot. You may find it more economical in the long run to buy a used Alpha & then sell your machine. Something to consider anyway.


I don't think an alpha motor can be driven with a 4G controller. Someone correct me if I am wrong. Without the closed loop feedback, what would be the point? Increased torque? And I am not sure you would get increased torque w/o the associated alpha drive.

Stan,
The 4G controller processes information from the SB3 program and sends this data (How many Steps & What Direction to each respective Alpha or Gecko drive. The positional feedback of the Alpha system is fully contained in each Alpha drive. The control board really doesn't do anything dealing with positional feedback.

I think you may have just had a brain fart, as many of us do from time to time, and thought that the conversation was about driving an Alpha motor from a Gecko drive. While this can be done, you would lose feedback functions as well as about 2/3 of the motor's performance. Alpha motors run at about 160-170v, which is one reason they can jog so fast...Without electronic governing, an Alpha can run well over 60 IPS. :eek:

-B

[br928]

That's exactly what I thought Dave was asking. Driving an alpha motor with a (4G) gecko drive. I was pretty sure you could not do it.

[Gary Campbell]

Brady....
Dont tell that to Dave! Then he would have to remind you I told him that 2 years ago. At least there were no problems or downtime! :D

[Brady Watson]

Brady....
Dont tell that to Dave! Then he would have to remind you I told him that 2 years ago. At least there were no problems or downtime! :D
:rolleyes:


Gary,
"Oldtimer's Disease" is on the rise...watch out! :D

-B

[bleeth]

Brady-You are a long time from oldtimers disease!

Stan-Thanks for weighing in and getting more deeply into my initial question.

To be clear-The object of the overall upgrade has always been to increase speed and accuracy as well as allowing the addition of a seperate drill.

The "4g type" controller I have is, of course, Gecko driven. The object of going to the next step (alpha geared feedback motors) is to increase speed, torque and accuracy further. Obviously there is a point of no return for value in which so much has been put in incrementally that one may have been better off changing machines. But the increments are cash purchased and a new machine means a new note. I have been given what I consider pretty reliable information that even with the up to date gantry the lower torque/power of the PRT motors will not handle the mass of an added drill hung off the z without further compromising speed and loss of steps. At this point, practically, my actual top performance without losing steps in cutting applications is 3ips cutting and 8ips jogging. Although it does better in "air" tests, when running parts that is the limit. So it seems that that information would be accurate.

So the natural thought occured to upgrade to the feedback alphas to both increase the torque and receive the benefits of the feedback which (at least conceptually to me) would eliminate the loss of steps I have now when pushing the machine further as well as allow the safe installation of a drill head and keep consistant operation as well as increasing speed further.

This complete discussion is not my area of expertise, at least at this point, so the initial question may well have not been complete enough or even a good one. If loss of feedback and inability to run at improved performance speeds is not possible with a gecko board, then the next obvious question to me is "would the SG motors be a better option?" If so, what other issues might occur that I don't know about?

Gary: Your memory is better than mine-It will probably be tomorrow that I go "oh, yeah, Now I remember that!!"

Buying a recent alpha and selling mine is not an option right now. Just as buying a new toolchanging SB or someone elses isn't. Regardless of the money, the current economics just don't allow it.

Once again-Thanks for the educated feedback guys.

[Brady Watson]

Dave,
SG (Straight Geared) motors are not desirable due to high backlash. You would do better to use a 1:1 motor with a 3:1 or greater belt reduction setup like Mike Richards did.

All this re-engineering of the SB stuff opens up a whole can of worms...there is always some caveat that gets you, especially if you are trying to do it on a shoestring. I hope that you are able to get satisfaction with your machine soon, because that's not an area you need to 2nd guess when you've got work to do.

-B

[bleeth]

It hasn't really been a shoestring. The overall costs in parts and time is pretty considerable. The main differences between simply buying a new machine and upgrading mine in stages has been that I have no new major bills at once, I have and still am learning more new things regularly, and this is currently more interesting to me than directing a plant full of men in building boxes and desks. Gotta have some fun!
As of now, even if there is no practical advantage to go further with this machine than I have, it is running overall somewhat faster, much smoother, and more accurately than it ever has.
Before I rebuilt the frame I was running 10ips jog but could only cut at 1.7. I have decreased cutting time for a board full of cab parts by about a third. Not too shabby, and I could simply jack up the link and let it lie. There is just this little voice inside me saying "you can put the drill on and run a little faster without yet breaking the bank". That's the bottom line I'm after. Reading back through this thread again more carefully Mike put in the following paragraph:

"For those who want/need the highest possible performance from a 4G machine, I would suggest that you spend about $1,000 and buy four PK296A2A-SG3.6 or PK296A2A-SG7.2 motors, a 35VDC 12A power supply and four 33K 1/4-watt resistors. Replace your Shopbot supplied proprietary motors with the PK296A2A-SGxx motors, replace the power supply, replace the current limit resistors on the board with the 33K current limit resistors, wire the motors half-coil (Black/Yellow, Red/White) and watch you machine take on a new life."

That, along with a couple of other things he said, seems a pretty good case to me for the Alpha 3.6 's on my machine. Am I correct that he is saying I would get more torque than I have now, maybe a little jog speed, but probably more cutting speed, and no problem in the added mass and moments of the drill head cantilevered off the spindle?

[ken_rychlik]

Dave,

I found a guy selling an alpha machine and having a hard time getting his price. I paid him to swap systems and then he sold his as a std easily. It was a win-win deal. It may be worth looking into.

Anyway if you go alpha your gonna love it. Then you will need a better vac hold down system. I am getting by with the twin feins, but I have to surface more often. Jogging at 20 ips is kind of scary though, after running a std for several years.

Kenneth

[richards]

Dave,

The black-art of stepper motors makes giving an absolute answer inpossible, BUT, using a low inductance motor gives better performance than using a high inductance motor. Inductance is the AC equivilent of DC resistance. The higher the inductance, the more voltage that is required to force current through the motors.

The PK296A2A-SGxx motors have 1.6mH of inductance with wired half-coil.

The PK296A1A-SGxx motors have 7.7mH of inductance when wired half-coil and 30.8 mH when wired bipolar series.

The geared motor that was often shipped with the Shopbot was electrically equivalent to the PK296A1A motor, but it was attached to the TH gearbox. The TH gearbox is a tapered hob design, meaning that the gears can be adjusted for lower backlash than the SG gearbox.

The problem with the A1A motors is that most of them were supplied with only four wires exiting the motor housing, meaning that they could only be connected using bipolar series wiring. The maximum voltage required to run a motor at peak performance is: 32 X SQRT( Inductance). For a 30.8mH motor, that means that you would have to use a power supply that generated 177VDC. The Geckodrives can only handle 80VDC. So, the A1A motor cannot be driven hard enough with a reasonable power supply to get the full potential from the motor.

The A2A motors, when wired half-coil can easily work with power supplies ranging from 28VDC to 40VDC. I use a 35VDC power supply with excellent results. With a 35VDC power supply, I can spin the shaft at 1,500 RPM, which means that the gear-boxes shaft is turning at just over 400 RPM. Since the gear box on a 3.6:1 motor is limiited to 500 RPM, that is a good match. That shows that the jog speed would be just over 26-ips if you used a 1.25 diameter spur gear (25-tooth), Ramping would have to be adjusted to allow sufficient time for the motors to attain that speed. Personally, on my Alpha, I limit the jog speed to 12-ips. That's plenty fast - especially when I get careless and the code sends the gantry into the end-stops.

The limiting factor with the PK296A2A-SGxx is the gearbox. Oriental Motor derates that motors from 4.5A down to 3A because of the gearbox. So, wiring the motor half-coil does not cost you any torque, it just lets the motor run at top performance - with that gearbox.

The SG7.2 motor, with 2X the torque of the SG3.6 motor, is even a better match IF your controller can generate pulses fast enough for a 7.2:1 gearbox. Remember that a Geckodrive requires 2,000 pulses to turn the motor's shaft one revolution. With a 7.2:1 gearbox, you need to send 2,000 X 7.2 = 14,400 pulses to turn the spur gear one revolution. A twenty-five tooth spur gear would move an axis 3.927 inches per revolution. A pulse train of 30,000 steps per second would move that axis about 8.18 ips.

A lot of CNC machines, worldwide, use the PK296A2A-SG7.2 motors with Geckodrive stepper drivers. A few users have complained about backlash, but most have found the SG motors to be sufficiently "tight" so that backlash is not a problem. However, many of those machines have a controller that can generate 45,000 steps per second, so they can jog at 12-ips.

A 3:1 belt-drive transmission would allow you to use a PK296-F4.5 or PK296-03AA motor at 4.5A for a holding torque of about 900 oz*in which is about 3X higher than the PK296A2A-SG3.6 motor and about 1.5X higher than the PK296A2A-SG7.2 motor - with the added advantage that a belt drive has no backlash.

[ken_rychlik]

Mike, You should sell dave your alpha system and then sell me the colombo. :eek:

Then you could sell your basic machine for less and still come out ok.

Kenneth

[bleeth]

Mike:
Being electronically challenged as I am I will have to read your response a few times over the weekend to begin to get it but I am grateful for your input for more than one reason. I have pretty much decided to stick with the controller I have and see if I can come up with the right motors if possible. "Building" a gearbox is not something I would have anywhere near the time to learn properly and do. I agree with you that if I ever saw my gantry flying down the ways at 26ips I would probably have a heart attack!!
Ken: Just keep thos good ides comin!!

PS: I guess this conversation is about ready to move over to my "upgrading" thread but I am loathe to lose quick scrolling back here.

[bleeth]

How does one determine how many steps per second their controller can put out?

EDIT BELOW:
OK-At current transmission efficiency I am putting out 37K steps per second max so it sounds like the 7.2 would be limited by that. Power supply is OK for 2A's already, so I get more jog speed out of the 3.6's.

It would then get down to how much more in smoothness and torque and cut speed. We get plenty of smoothness out of the PRT motors now. 8 jog is OK but not great. I would like to be at 12 consistantly. I also could not get the "Alpha chatter issue that was endemic in earlier Alphas. I am assuming until corrected that the 3.6's would not have that issue. Oriental is selling them direct for $257 each. Is there any other "special" considerations I am missing like SB customized shaft or anything?

I don't get the definition of "backlash". Does Mike's suggestion to use these motors wired half-coil with the replaced resistors overcome that issue? Or keeping jog to 12ips?

[br928]

Backlash is the amount of 'slop' in the gearbox. When the motor comes to a stop in one direction it is how much the motor has to move in the opposite direction before there is actual movement in the machine. Backlash is purely mechanical. Like steering in an old car when there is play in the wheel between turning left and right.

The TH or tapered hob gearboxes have less backlash than the straight gearbox you buy from Oriental Motors. They are able to adjust the gearboxes of the TH to minimize the backlash. That is also why they cost more. The belt drive reducers are desirable because they can have zero backlash.

[Brady Watson]

It would then get down to how much more in smoothness and torque and cut speed. We get plenty of smoothness out of the PRT motors now. 8 jog is OK but not great. I would like to be at 12 consistantly. I also could not get the "Alpha chatter issue that was endemic in earlier Alphas. I am assuming until corrected that the 3.6's would not have that issue.

Dave,
The 3.6:1s coupled to microstepping (1/10) Gecko drivers will give you 2.5 times finer resolution when cutting than the original PRT 1/4 stepping drivers. Assuming that your pinions are in good condition, you should not have 'chatter' issues. The newer Alphas with 7.2:1 motors have about 30% better resolution than a comparable Gecko equipped setup (@2400 UV compared to 4G 1800 UV) You should have enough resolution with your setup to get the job done - cutter deflection being the common denominator with chatter...

7.2:1s non-Alpha motors are OK on a Gecko equipped machine. They have a boatload of torque, and lost steps are really only an issue when you try to go too fast - E.G. - Jogging. You also gain twice the resolution over a 3.6:1 setup.

The jog speed on your machine is completely governed by your communication speed.

-B

[bleeth]

Just had a fascinating chat with Oriental Motor. Seems like there is a third option that will at least bear a bit of investigation. That would be to take the PK296-F4.5A motors and fit a planetary gear box to them. They gave me the names of two companies that do that. I'll fill you in when I have more info, but basically it means that the torque would be much like the belt drive add on (Huge).

OK-Just had a chat with one of the companies he suggested. Price is out of hand. By the time I got through I would spend as much to buy new motor's and controller from Ted-Not something I'm against except for not being able to afford it (Not that they are overpriced from Ted either)!!

[Brady Watson]

Geez Dave...I could've saved you a load of time. The planetary gearboxes are more like outer planetary with the prices they ask for them.

The bottom line is, a Gecko/RBK or other non-Alpha solution is always going to have a compromise to it. If you are out for performance, then the Alpha is very hard to beat. ShopBot spent a LOT of time in R&D on the Alpha machines. Lots of different drive systems were tried out - with the Alphas winning. They are the best of both worlds in terms of servo performance and stepper low speed torque. They are faster than comparable servos as well because they can resolve small moves quickly with all of that torque, but also rapid position when needed with their wide power band.

Hate to say it - nothing wrong with a PRT/4G/RBK setup...but they will never live up to the performance of an Alpha. I don't care how many different systems, motors, drives, servos etc you try out...you'll be disappointed. I've been there...done that, as they say...Why torture yourself?

There is little that you can do, aside from upgrading to an Alpha, with your current setup that would be worth doing. You are limited by what a regular stepper can do with a given gear ratio, top speed is limited by pulse rate and power supply voltage, and it's still open loop. Alphas have positional feedback and internal 'electronic gearing' which will give you higher top speed per given pulse rate.

Roll with what you've got...Save your cash...buy what you can, when you can to upgrade.

-B

[bleeth]

Brady:

If we weren't the type of people to check things out ourselves we wouldn't be who we are. You know you wouldn't have half the knowledge you do if you simply accepted all as related. This entire program has taught me more about my machine than I ever knew and that can't be anything except good.
Never expected to get current Alpha performance out of my machine without going all the way, but will get vast improvement over previous constraints. Your and everyones advice has all been well taken and mucho appreciated. Time to digest for a while and move on in order.

Thanks again;

Dave

[Brady Watson]

You are oh so right, Dave. Sometimes we forget about how valuable the process of finding out for ourselves can really be...

-B

[richards]

I have to disagree in part with what Brady wrote about the Alpha vs standard steppers. This morning I checked the Oriental Motor web site and compared the AS98AAE motor to the PK296-F4.5A motor. Both motors have almost identical torque curves out to 1,500 RPM. Beyond 1,500 RPM, the Alpha motor has a slight advantage; however, at 1,5000 RPM the Alpha requires 25,000 steps per second, so the ability of the controller to produce step pulses becomes a factor.

The PK296-F4.5A is a non-geared motor, so it would require a belt drive to boost its torque to a usable figure. A 3.6:1 belt-drive and 30-tooth pinions would give a per-step resolution of 0.0006545. (1.5" x PI / 3.6 / 2,000 pulses per revolution).

The AS98 motor attached to a 7.2:1 geared motor would require the same number of pulses to move the same distance because it uses 1,000 pulses per revolution.

The AS98 motor is rated at 280 oz*in holding torque. 7.2 X 280 = 126 lb*in of torque. The gearbox is rated at about 80 lb*in, so I would use 80 lb*in as the actual expected holding torque figure.

The PK296-F4.5A motor is rated at 310 oz*in half-coil or 440 oz*in bipolar parallel. That gives 69 lb*in or 99 lb*in, depending on how you wire the motor and how much heat you're willing to tolerate. (Bipolar parallel connections produce A LOT of heat.)

The fact that the Alpha motor can recover from missed steps has caused me more grief than I ever expected. When I first got my PRT-Alpha, I had a lot of unexplained "divots" in some of the parts. When I started paying closer attention, I realized that I was pushing the Alpha motor (AS911) too hard and that the motors were entering "Alpha mode". Most of the time, the Alpha function worked and the motor was able to recover, but at the expense of a divot in the part. When I slowed things down, the motors stopped missing steps, the Alpha mode never kicked in, and the parts turned out as expected.

The PK2xx motors don't have an Alpha mode. If they're driven too hard, they miss steps and usually stop. The solution is to run them at a speed that is less than their "missed steps" speed.

So, why choose the PK296-F4.5A motor, a Geckodrive G203v or G201x stepper driver and a linear power supply over an AS98 Alpha motor/driver? Cost. The PK296-F4.5A costs $139. The stepper driver costs $147. A power supply to run four motors costs $125. So, list price for four motors, four stepper drivers and one power supply is $1,239. The list price for ONE AS98 motor/driver geared 7.2:1 is $1,326.

What you end up with is the same resolution between the two motors, the same pulse train requirements, the same torque (up to 1,500 RPM) and a huge difference in price.

An added advantage is that you could substitute the PK299-F4.5A motor, which costs $208 instead of $139 and you would have 2X the torque of the PK296-F4.5A motor.

The big question that everyone needs to ask themselves is whether saving a few thousand dollars is worth the aggravation of building belt-drive transmissions and maintaining a non-factory electronics package. That route is not for everyone.

[Brady Watson]

Mike,
We'll have to agree to disagree. If Alpha performance could be achieved at a much cheaper price, then I am sure ShopBot would be one of the first to implement it, if for nothing else than to save money. (Otherwise us Alpha owners are all just a bunch of suckers...)

Alphas have internal 'electronic shifting' that switches from microstepping to full stepping to allow a wider speed range than a standard driver. Microstepping is not needed at higher speeds. The newer Gecko drives will do this to a lesser degree (morphing) - but not as efficiently as an Alpha will. The unique aspects of an Alpha driver allows for high resolution and speed, while not requiring an insane communication speed or pulse rate. This is key. This is the real advantage of an Alpha over anything else! Positional feedback is icing!

At the end of the day, does the tool do what you need it to do, at the speed and reliability that you need to do it? If yes, then you have the right tool, regardless of drive system. If not, there are solutions out there.

-B

[curtiss]

Quite a bit of information above, but could it be sorted out in a

"good, better, best format" for motors & drivers for the

prts
prt alphas
prs
& prs alphas

not that everyone will agree but ?

[richards]

Brady,

Stepper motors belong in the "black magic rhelm", but unless I'm mistaken, if I want any stepper motor to move 100,000 steps, I have to send 100,000 pulses. Yes, the Alpha has two modes, one for jogging where the internal multiplier is manipulated and one for cutting, but the high-speed comes at a resolution cost which would make that mode undesirable for cutting.

The Gecko's do morph from 1/10 step to full step internally, but they are always given a pulse train corresponding to 1/10 steps. The internal morphing keeps the torque at its highest level.

Positive feedback can be added to any stepper motor for about $35 or less. An encoder from Digi-key costs around $25 and a microprocessor, oscilator and glue chips add another $10. That minimal cost circuit will not compensate for missed steps, but it would show how many steps were missed before the micro-processor sent a fault signal to the Shopbot controller.

As an experiment, I temporarily replaced the X and Y axis AS98 7.2:1 geared motors on my PRT-Alpha with some PK296B2A-SG3.6 motors, some G203v stepper drivers and a 35VDC power supply. After adjusting ramping and the unit values, the inexpensive PK296B2A-SG3.6 motors cut exactly the same as the much more expensive AS98 geared motors. The test was run on particle board and MDF, because those are the materials that I cut most. I limited jog speed to 10" per second with the PK motors, although they could have run at 15" per second without pushing the Shopbot controller at more than 25,000 steps per second. That experiment showed that when the Alpha motors die (if they ever do) that I can safely use the must less expensive PK motors as replacements without giving up speed or quality; however, I will lose the missing step detection/correction feature of the Alpha motors, which can be offset with the encoders mentioned above.

Comparing the PK296A1A motors (Shopbot uses a different part number, different gear box, same electrical specs) that were wired bipolar series to a PK296A2A motor wired half-coil with a properly sized power supply is comparing apples to oranges. The A1A motor is an excellent motor, but its inductance is much too high for a Gecko stepper driver. It would require a 170VDC power supply to get the same performance as the A2A motor gets with a 35VDC power supply. Matching the components to get the most bang for the buck turns a Gecko driven standard machine into a very high performance machine.

The SG gearboxes CAN have more backlash than the TH gearboxes supplied to Shopbot by Oriental Motor, but the SG motors that I have exhibit no measurable backlash. In time, they may loosen up. That's the main reason that I prefer a belt-drive transmission. It's more bulky, but it is adjustable and it transfers the full torque of the motor multiplied by the gear reduction to the axis.

The point that I'm trying to make is that Shopbot provides two excellent choices, a Standard machine and an Alpha machine. For the do-it-yourselfers, there are two more choices, a PK296A2A-SG3.6 geared motor or a PK29x-F4.5A motor with a belt-drive transmission. Few people would order a Ford pickup without an engine, which is essentially what I'm suggesting. Those "hot-rodders" who know their engines might want to use something other than the standard Ford engine to get the most out of their Ford pickups. A few Shopbot users who like to play with electronics might find a do-it-yourself approach the way to greatly improve performance at an affordable price.