Experimenter board +g540 instead of 4g upgrade?

[jseiler]

I really like the shopbot drive system mostly for the shopbot language. What I don't like is the limiting of the output current on the 4g upgrade. I was thinking about driving G540 geckodrive with an experimenter board to have the best of both worlds...shopbot language and 3.5A/50V drive. To meet the speed requirements I have, I would need 60KHz which would be easily achievable with Mach3, but would a shopbot experimenter board do this? I'm also having trouble deciphering the shopbot developer page for the cost of the board needed. Would I need the 4g upgrade minus drives, or could I use something less expensive?

thanks,
John

[ted]

Hi John,

Let me ramble a bit about some of your questions.

The output on drivers for the Version 4g Interface Board is set to ~2A (and we power it in new boxes with 63V) because this provides a good match to motors on older ShopBots and motors on new tools. Because we also wanted a drop-in replacement that did not require people fussing with wiring-up drivers, we came up with a snap-in board mount for the Geckodrives on the board. This puts all the wiring on the board and keeps things simple -- especially for ShopBot owners who are just swapping in a new 4g board.

The disadvantage of this convenient driver mounting technique is that there is no heat-sinking of the Geckodrives (they do run warm) and the wiring is limited in current capacity by the traces on the board. Geckodrive says to heatsink their driver over 3 amps. And, as it is, one needs to be careful to keep the Control Box cool with 4 Gecko's in them. So the heating issue and the power wiring of the Gecko's on the board are the reasons we recommend not running the drives at the higher currents on the 4g board. While the Geckos are rated to 80v, we've found over the years that the 60-70v range for the power supply is a little less likely to create problems than pushing the limit. (You could modify the 4g board and mount the Geckodrives off-board on heat sinks, but this would be more work than my suggestion below.)

Rather than a 4g board, what I think you want for your project is one of our Control Cards. These are the same cards that run all our current PRS tools and are the cards that current Version 4g's are fitted with. You can either wire this card up directly to your drives and other I/O, or you can connect it through one of several Interface Boards. The Card essentially provides the connection from the ShopBot software running on the PC to your own CNC tool or robotic device, via a USB interface.

Our idea is to have a selection of Interface Boards for developers from the most basic (which only provides connectors and optical isolation of I/O lines), to one that has full CNC safetly relay interlocks for router/spindle control and accessory relays (this would be the same interface board as on a PRSalpha).

We've been working on pulling together these offerings for most of year but don't have it all finalized or available yet. I'll try and get some more specific info and pricing posted here in the next day or two. [We actually showed several variations of developer kits at Maker Faire in San Mateo this year; running everything from a baby CNC to a light-show device (thanks to Mike Richards) -- then we got distracted by getting Buddies and PowerSticks out].

With respect to speeds, the Control Cards (V201, V202, and V203) are currently running at speeds that support 30-40Khz stepping (actual top speeds depend on the practical bandwith available to the USB controller on the PC). These Control Cards are all capable of communicating at twice the speed we are running at with the current firmware. We have been planning to switch to the higher speed with a firmware upgrade later this year (firmware upgrades to the controller are included in the ShopBot software install). But we have not thoroughly tested the controllers at higher speeds. We are expecting to achieve 60-70Hz pulse rates and there will certainly be an increase over the current rates. However, we don't know yet what the maximum solid rate will be.

I'll see if I can add to this as I pull some more specific info and specs together over the next few days.

Ted Hall, ShopBot Tools

[jseiler]

Wonderful. Thank you for the indepth answer and I look forward to the follow-up. Posters like Mike Richards have gotten some of us experimenter types fired up to try out some things, like different motors and drive combinations.

thanks again,
John

[richards]

John,

I can't see the need for such a high step rate. Here are some figures that I've used:

Spur gear = 30 tooth (1.5 inch pitch diameter)
Gear ratio = 7.2:1 (Oriental Motor PK296A2A-SG7.2)
Linear travel per (motor) shaft rotation = 0.65 inches

That means that it would take 2,000 pulses to move 0.65 inches. (Each pulse would move the axis 0.000327 inches.)

So, 30,000 steps per second would move that axis 9.82 inches, which would probably be as fast as those motors could reliably move the axis. Even using a 25-tooth spur gear would let the axis move 8.18 inches per second.

The G540 is limited to 50VDC, so those motors would have to be wired half-coil (black/yellow wires, red/white wires) to stay under 50VDC. Wiring the motors full-coil would require a 70VDC power supply (or something close). In other words, the torque loss from running the motors wired full-coil at 50VDC would be greater than the torque loss running the motors wired half-coil at 35VDC. To me, that is pretty good performance from a motor running on a 35VDC power supply.

Remember, that is with a 7.2:1 gear ratio and a 2,000 pulse per shaft rotation pulse requirement. Using a 3.6:1 motor (i.e. PK296A2A-SG3.6) would give you 2X the speed and the same resolution as the PRS-Alpha (7.2:1 gear ratio, but only 1,000 pulses per shaft rotation). However, the gearbox on the PK296A2A-SGxx series motor is not as robust as the gearbox on the Alpha motor, so I would be hesitant to run those PK296A2A-SGxx motors continually at high speed under heavy loads.

In any case, the limiting factor is probably the torque of the motors, not the pulse rate of the G20x board.

There are two other concerns that I have with the G540 stepper controller. The first concern is an obscure notation in the instructions. Mariss wrote: "Choose a motor that has a rated winding inductance of 2.5mH to 3mH if maximum power output (>100W mechanical) is a requirement."

Oriental Motor does not sell a geared PK296 motor that falls within the 2.5mH to 3mH range. The PK296A2A-SGxx motor is rated 1.5mH half-coil and 6mH full-coil. The PK296A1A-SGxx motor is rated 7.7mH half-coil and 30.8mH full-coil. I've posted a note on the Geckodrive forum about the 2.5mH - 3mH statement, but, so far, there has been no response.

The second concern is the use of DB25 connectors to connect the motors to the G540. At one time, I built a process control computer to replace the original electronics in the Kodak S-series photo printers. A competitor used DB-25, DB-37 and DB-50 connectors. I used much heaver duty connectors. I was often called in to trouble-shoot the competitor's electronics. Often the cause of the problem was an intermittent connection on one of the DB-xx contacts. Many DB-xx female contacts use a split (two-piece) slanted 'finger' assembly. That means that 100% of the current passing though that DB-xx contact may very easily be passing through a single contact point that is smaller than the head of a pin. That's not very much area to carry the current required by a stepper motor. If a good quality DB-25 connector is used, and if the contacts on both sides of the connector firmly mate with the contacts on the G540, then things should work properly. However, to me, that is a problem just waiting to happen. Sometime down the road, after the contact point has arced several times and carbon has built up between the two conductors, that contact point may easily become intermittent.

Even with those two concerns, if I had an extra $350 or so laying around, I would order a G540 to play with. It looks like an interesting product that could easily be connected to the G201 that Shopbot was kind enough to let me test. On the surface, it looks like the G20x from Shopbot, the G540 from Geckodrive and the PK296A2A-SGxx motors from Oriental Motor would be a good match if the motors were wired half-coil and used with a 35VDC power supply.

[jseiler]

From my perspective, I have a pr32. Its ancient, but it runs really well and its wickedly accurate. It runs small motors (PK266? I think) on 5tpi screws. It originally came with a gen 1 control box that ran on 12V. The first step was changing out the current control resistors and upgrading the power supply which made it run stably at .6 in/sec. I next bought a PRT control box + new motors and that got me up to a nice .9 in/sec. Now I'm looking at wanting 3+ ips rapids. At 200 steps per inch, 10 microsteps per step, and 5 rev per inch, we get to 10000 steps/in. To get 3 ips, I need at least 30K steps/sec from the controller. I'd like the ability to have a bit more to play with. At this point screw whip becomes an issue (but not much of one...the screw span is rather short and I've got the screw secured on both ends.)

For people with racks and gears, you'd not ever really need 60K steps/sec. This is also for future expansion. I have another machine with 10tpi screws. 60000 steps/sec = 3ips.

John

[richards]

John,

I apologize. I made the assumption that you were using a standard rack and pinion machine. Sorry for jumping to conclusions.

[jseiler]

No apology necessary. I appreciate the long answer because the math is essentially the same.

As far as the "obscure notation", Mariss does that a lot with heuristics that border on the wierd. Personally, I like to see derivations so I can tell if the math fits my situation. Torque x speed=power. Great. Torque is proportional to current. Speed is proportional to voltage. Current is proportional to inductance at a given pulse rate. I think somewhere in all of that, his drive might not be able to provide full current (torque) at higher speeds if the inductance is too high. Might have something to do with his pwm frequency (don't know for sure).

Got a suggestion for a stepper motor that will pump out a reasonable amount of torque at 1200 rpm? Most of them look like they want to stall fall off the torque curve there. I'm looking for over 100 in-oz at 1200 rpm in a Nema23 package with under 3.5A of drive current in the 2-3mH range.

I'm currently using Keling 276's and they aren't too bad, but if there was something significantly better recommended by someone with more automation control experience than I, I'd wouldn't mind a head's up.

John

[ted]

Hey John,

Here's my first installment. This is mostly the work of Mike Richards and Bill Young. Bill is trying to get some of our developer material organized.

http://shopbotdev.pbwiki.com/

I'll post some specs on the Controller Cards shortly. And, hopefully, a little pricing info. The interface board described in the developer wiki won't be ready for another month or two.

Ted Hall, ShopBot Tools

[ted]

Here is a link to a diagram of the connections for ShopBot's Control Cards, both the Ver201 that is our current controller and the Ver204 which should be available in a few months.

http://www.shopbottools.com/files/ShopBotControlCardConnections.pdf

Both Cards will work with the latest version of the ShopBot software. These cards require a 5v power supply and are generally described at the wiki address in the post above.

The Cards are the core of our tools. You'll get a pretty good understanding of their features by just looking at the designations of the various lines. You'll note that there are a number of new features that will be implemented on the V204 Cards. Be aware that things might change a little on the 204 and that some features will also require getting the application side working in the Control System Software.

Both Cards will step at a little over 30KHz at the moment, and we expect both to be running at ~60KHz with new firmware later in the year, pending some serious testing.

I'm linking this file to the development wiki as well, and we will keep working on updating offerings there.

Ted Hall, ShopBot Tools

[dana_swift]

Ted- thanks for the info.

It looks like exciting things coming out, I'm sure a lot of us are looking forward to the future capabilities in our next bots purchases, and upgrades!

A suggestion: it would be really "nice" if there was a header that just had a set of ground terminals. Then wiring an input or output circuit would be much easier. The existing G terminals on my bot have two wires on them already, and putting 3-4 on then becomes a royal pain. With the additional I's and O's it would be even more helpful. I'm guessing my suggestion is too late as the new controller is probably already laid out and the design is frozen.. oh well. Maybe next time.

D

[bcammack]

Is there anywhere on the ShopBot Tools website that lists the individual prices for all of the controller boards under discussion presently?

[bcammack]

{insert sound of crickets chirping here}

[dana_swift]

Brent- that is funny! You got a chuckle out of me-

D

[Ryan Patterson]

http://www.shopbottools.com/developers.htm

[bcammack]

Thanks, Ryan! They seem to be adept at keeping their light under a bushel sometimes.


You're welcome, Dana.