What the geck are "Geckos" - Mach they drive a ShopBot?

[gerald_d]

From time to time the terms Gecko and Mach are mentioned here. Let's get to grips with these beasts.....

I have a little experience on this front, having built a machine that uses PRT style motors, but does not use a SB controller. Why did I do this? Because:
- it was fun
- it uses readily available components that I can repair/swop out here in Africa
- it uses the industry standard G-Code, for which I can find trained operators off the street.
- it performs at least as well as the SB controller, if not better.
- setting up "ramping" is a no-brainer.
- and (touch wood) so far have had no mystery z-dives and other erratic moves.

This thread is for those who want to understand the concepts and constraints.

Right now I am buiding my second iteration of a Gecko/Mach controller to replace the SB controller on our first machine. The power supply got knocked together on the weekend - you can see it here (http://www.mechmate.com/Forum/messages/17/670.html?1145902622).

I hope that the others with Gecko/Mach experience also chip in here to answer questions that may arise.

[btk]

Gerald,

With all do respect and you can certainly do as you like (and I am certainly not looking to start a long drawn out issue), however do you really think that the Shopbot sponsored forum is a good place to post this.
I find that Shopbot markets their products in a very honest and open way as very good performing machines that are both affordable and easy to use. Perhaps that post is better suited for a different forum that is not owned by shopbot.

Brian

[mikejohn]

If you start with a Shopbot and improve it, how does this harm Shopbots marketing?
One of the biggest attractions of the Shopbot over similar priced machines is the ease of adaption for your particular needs or wallet.

I am against any idea of censoring posts here which might not be Shopbot approved. Does it mean we can only discuss the software that Shopbots sell, or not talk about 'knitting needle' probing?
I feel this is Shopbots position also. They never seem to delete or intervene in technical matters such as this, only when things get personal!

..........Mike

[gerald_d]

(Edited to say I cross-posted with Mike - my reply below is only in response to Brian.)

Brian, various people have raised the Gecko/Mach option on this forum for a couple of years already, and it is often spoken of as a variation or modification of a ShopBot - the title of this topic on the forum. Folk have also discussed Milwaukee vs the Porter-Cable router, where SB supplies the one and not the other.

I am under no illusion that a discussion on replacing the controller cuts to the very bone of the ShopBot. But it is a whispered discussion, with some exaggerated claims at times. This thread is an attempt to bring the discussion into the open. Somehow I think that ShopBot would prefer to have such a discussion over here, where they can pull the plug if it becomes harmful to their business.

[mrdovey]

I'm interested! I like my 'Bot; and I've been using it to build a more specialized machine for horizontal spindle operations - and have plans to use it to build parts for two more (metalworking) specialized machines: one to cut to length and form aluminum ribbon for solar panel heat exchanger production, and one to drill aluminum tubing with program control of position and angle.

If anything, what I'm doing highlights the wide range of applications and materials that can be addressed with a ShopBot. It's not limited to kitchen cabinets and signs!

So, Gerald, what is a Gecko (beside a lizard)? I'm aware that Mach-n is software that runs under Windows to drive a parallel port step/direction protocol. Is Gecko a stepper motor power/control box, or a brand/type of stepper motor, or both?

[gerald_d]

Hi Morris, a Gecko is only a "driver" of a motor. When I sensed that this discussion might not be welcome here, I wrote up this description (http://www.mechmate.com/Forum/messages/16/675.html?1145947627) for a Gecko.

It is little black box, one for each motor. The important things that connect to a Gecko are:
- 4 wires to the 2 coils of the stepper motor
- 2 wires from the DC power supply
- A couple of thinner wires (normally indirectly)from your PC's parallel port (LPT1 etc).

Mach2 or Mach3 are programs you load on your PC that read DXF or G-Code files and pushes out signals to the Gecko's via the parallel (LPT...) port.

[gerald_d]

This (http://www.machsupport.com/artsoft/index/index.htm) is the home of the Mach software. Note that Mach is only 4 years old.

Connecting little wires to the pins of your parallel port are not fun. So some guys are building "parallel breakout boards" to make that part easy, and to give outputs for relays and inputs for proxy switches. Here is an example of a breakout boards: PMDX-122 (http://www.pmdx.com/PMDX-122/index.html). Your printer cable plugs into this board.

[btk]

Mike,

How did I know that you were going to be the next post on this thread :-)

My post was not againt discussing upgrades and modifications to the shopbot (which is obviously greatly encouraged on this forum), however to me it appears as the spirit of the initial post of this thread (particularily a few of the bullet points) were worded in an unfair way (a bit of a "dig" if you will).

For Example:
I think if someone came here looking for a machine to buy, after reading that post, they might feel that shopbot has a less than adequate Controller and that someone who uses G-Code would not be able to use a shopbot.

What I am trying to convey, is that
If I hosted a forum for my company and worked as hard on creating such a customer support system as well as a open forum for users to experss ideas (and upgrades), I would hope that it was not used against me.

However you (and shopbot for that matter) may see differently, however as you said, I am against sensoring and just wanted to write what I felt.

Brian

[mrdovey]

Gerald...

Thanks! I'm working with a 4-axis control board from hobbycnc.com to drive smallish (200 in-oz) steppers in a 3-1/2 axis machine (photo at the bottom of this page (http://www.iedu.com/DeSoto/JBot.html)). I'll read up on the Gecko drivers as I suspect they'll allow driving larger steppers.

BTK...

I think it's pretty much inevitable that at least some 'Botters will either already have or will develop a serious interest in robotics. I'd like to suggest that it reflects particularly well on ShopBot when such people speak well of ShopBot and its products. The message is strengthened when presented in a candid forum.

I also have a sneaking suspicion that there may be some direct benefit to the ShopBot staff from our wider discussions of new technologies, techniques, and solutions on the forum. If we use the forum only to sing the praises of our 'Bots, then we make improvements and advancements much more difficult...

[mikejohn]

BTK
I understand your point, and I support everything that Shopbot do.
My "Shopbot" experience has been good from my first reading of this Forum.
I also agree a superficial reading of some threads might lead a 'newbee' to feel that the Shopbot is somehow inadequate, but I do read it differently.
There is a very telling statement in Gerald's original post which applies to me also it uses readily available components that I can repair/swap out here in Africa (not Africa for me, of course.)
I am sure I am not the only non-USA based Shopbotter who feels frequently 'out of the loop' when you guys seem to be able to pop down to the corner store to get almost anything.
We need to have alternatives.
Before I bought my Shopbot somebody (probably Gerald
) wrote about drilling holes in the Shopbot "Who would dare do that to a Thermwood".
It is this aspect which will be appealing to many.
Don't take my post as a slur on Shopbot, in all its aspects. I don't always agree with them, but I do support what they stand for. Accessible and good quality CNC for the masses.

.........Mike

[gerald_d]

I think I've said it many times..... The ShopBot is perfectly fine as a tool for someone who wants a ready-made solution. Then there are the few oddballs like myself who are always tinkering. Note that I didn't start this thread saying that SB is junk and that if you paid me a few $'s I would tell you how to fix it!

[earld]

When I bought the 'Bot, the Alpha was hitting the streets and I became a 'tweener- new reinforced frame and PRT motors. My concerns were that there will be spare parts available in case of breakdown, and, who's pink hat was was that in the crate?
Of course, I would like to upgrade someday to the alpha and a spindle, and afford gasolene.
But, the Gecko system is also a possibility depending on costs.

Earl

[richards]

When it comes to modifying electronic controls there needs to be a big bright red sign that says, "PROCEED AT YOUR OWN RISK". Shopbot has done an excellent job on my Alpha of protecting me against myself. They've designed the controls so that I would have to tamper with the wiring before I could hurt myself. On the other hand, people like me who like to play with electronics, often leave out the safety circuits until we get shocked once or twice.

When I looked at the Photos that Gerald posted, I was impressed with his power supply. So impressed, in fact, that I went down to the local auto supply store to buy some automotive type fuses for my experimental Gecko setup. Unfortunately, here in Utah, I can't buy 80V automotive type fuses at the local auto supply shop. If I didn't know the danger of using low voltage fuses on a high voltage circuit, I might have tried using 12V fuses. I also noticed that he hasn't yet installed the bleeder resistor across the capacitor's terminals. This post is not a flame against Gerald or of his designs. His designs and excellent photos have often helped me solve problems here in my shop. In fact, one of his posts last year led me to experiment with Gecko drivers, toroid transformers, PMDX breakout boards and G-code.

What I am concerned about is the fact that messing with electronics powerful enough to drive high performance steppers can be lethal. Not having a dump circuit to immediately discharge a power supply can be lethal. Not having an E-STOP switch properly connected to immediately kill the power to the router/spindle can be lethal. In short, before enhancing the drive system on our machines, we'd better be sure that we have all the safety circuits fuctioning properly.

[gerald_d]

80V automotive blade fuses (http://www.littlefuse.com/cgi-bin/r.cgi/en/prod_series.html?SeriesID=3184&LFSESSION=wQw7oxdcn Z) / Fuseholder example (http://www.littlefuse.com/cgi-bin/r.cgi/en/prod_series.html?SeriesID=3199&LFSESSION=wQw7oxdcn Z)

The capacitor is adequately discharged by the four Gecko's across it. A bleeder resistor that will discharge the cap any faster will be too inefficient.

Of course....PROCEED AT OWN RISK!...this stuff is not for the timid, careless or stupid.

[mikejohn]

That rules me out of making those alterations!

[gerald_d]

There is a discussion on bleed resistors going on at the GeckoDrive forum as we speak. I am sure that triggered Mike_R's comment above. The jury is still out on whether it is essential or not.

Guys, we are getting bogged down in minutae, legalese and politics. I say again, this thread is for those wanting to find out more about the Gecko/Mach alternative that has often been mentioned here.

[richards]

Gerald,
Thanks for the Littelfuse reference. My local auto supply store said that they can order them in for me.

You're correct about the four Geckos being able to discharge the capacitor as quickly as a bleeder resister if the Geckos are functioning. In my paranoid world, where disaster is waiting just around the corner, I use a bleeder resistor just in case all four fuses between the power supply and the Gecko drives blow. (I'm so paranoid that I also have a fuse between the Gecko and the motor -figuring that I need a fuse to protect the motor as well as a fuse to protect the drive.) Is it likely that all four fuses would blow, thus isolating the Geckos from the power supply? No, it's not likely. Could all four fuses blow at once? Yes, they could.

[gus]

Gerald,
Keep at it, those who can do and those who can't need a big sticker on the box that says "No User Servicable Parts Inside".

[gerald_d]

Ted, when I switch on a power supply like that for the first time, I put it out on the driveway with a 30 ft cable back into the garage...and I close the garage door while I hide inside to flick the switch!

[watswood]

Gerald,
What have you noticed as far as performance increase over SB control(specifically over non-alpha prt)?
Is it faster?
More powerful?
Smoother/ cleaner cutting(diagonals/arcs)?

What is the approx. cost of:
The Geckos?
The mach software?
The power supply?
Misc.?

I'm always in search of improving my bot as long as the cost:performance ratio is not too great.
Thanks

[richards]

Gerald,
I should check the Gecko forum more frequently. Typically I check in about every two weeks to see how the G100 project is moving along. You're right there is a discussion going on about bleeder resisters. When I finish today's set of cabinets, I'll have to take the time to read through them.

Since this is your thread, what would you like to see posted here? Questions about interfacing Geckos to the existing Shopbot controller? Replacing all Shopbot control components with other manufacturers electronics? Programming with G-Code? Switching to Mach software? Or, just general hints on how to use locally available parts as substitutes for parts only available from Shopbot?

(My biggest concern as a reader of this forum is how help could ever be offered when someone has a little of this and a little of that mixed in with a shopbot controller and shopbot software. I see a lot of CNC operators getting hair as white as mine as they realize that they don't have the experience to fix a problem and no standard FAQ to turn to.)

[mikejohn]

I bet you hide in the garage and make Agnes switch it on

[stevem]

“I see a lot of CNC operators getting hair as white as mine as they realize that they don't have the experience to fix a problem and no standard FAQ to turn to.”

That’s the beauty of the ShopBot. Not only are most of the mechanical components easily replaced by anyone who is capable of assembling the machine, but the machine is easily modified by those who like to tinker (I define “tinkering” as fixing something that doesn’t need fixing).

All the electrical components, as well as the software, on my Bot were replaced with Geckos and Mach2 before I produced a single part on the Bot. This was done because my home built router was operating with Geckos and Mach2 and I didn’t want to relearn a new software. There are many advantages to those Bot owners who want to upgrade their older machines without the Alpha cost.

A typical Gecko/Mach conversion should cost approximately USD $800.00 although other less expensive software is available, such as TurboCNC.

[drodda]

To those of us who have a PRT Alpha is there any advantage to using the gecko and Mach setups? Or is this mainly for the people who live outside the US and own older shopbots? I am a little lost as to where this thread is heading? Please enlighten me. I am still trying to understand the timid, careless, and stupid comment?

[stevem]

Dave, the Alpha system is a huge improvement over open loop steppers and will operate with step and direction signals that are put out by the Mach control software. If I was going to invest in additional upgrades to my Bot, I would replace the Gecko drives and original steppers with Alpha components, but keep the Mach control.

[richards]

Dave,
I have an Alpha 120-60 and, on my test bench, I have Geckos, both stepper (G212-4 each) and servo (G340-1 each). I have a licensed copy of Mach3 (free evaluation copies are also available), driving a PMDX-122 breakout board. I have a 50V toroid transformer connected to a PMDX-135-8020 power supply giving 70VDC to the steppers and to the servo. I have a Gecko G101/102 high speed controller running with MachIV. On my test bench I have a dual channel 40mhz oscilloscope and 2 Fluke digital meters. Before the Gecko stuff, I used Oriental Motor stepper drivers and Oriental Motor stepper motors. Before that I designed from scratch stepper controllers for process control applications. In short, I believe I have enough "stuff" and enough experience to make a statement comparing the Geckos to the Oriental Alpha series stepper drivers and stepper motors. At the present time, the hands down winner is the Oriental Alpha series stepper drivers and stepper motors because it is a closed loop system with feedback. In the future, when the Gecko G100 controller is finished, and the encoder input and feedback module is completed, the situation may be different.

If I had an older model PRT with worn-out motors or drivers, I would consider replacing the old stuff with Gecko drivers, Oriental Motor PK299-F4.5B stepper motors, and 500 CPR encoders. Even then, I would still have to build and program my own interrupt driven pulse comparator to verify that pulses sent to the stepper motors actually resulted in verified steps. And, if I decided to switch to Mach3 software, I would have to reprogram every one of my cut files - which is not difficult, if you have had 30 years experience programming computers, as I have, but still a task that would take a substantial amount of time. BUT, if I already had an Alpha, which I do, I would consider myself lucky and blessed to have such a fine machine direct from the factory.

[gerald_d]

If you have Alpha ShopBots, Gecko/Mach is not for you. Please stop reading the thread.

If you have PRT standard ShopBots, running without glitches, then Gecko/Mach is also not for you. (From my limited experience, there is only a marginal improvement in cut quality).

However, if the ShopBot controller has ever frustrated you, then consider changing to Gecko/ Mach. We learnt to live with our SB's unusual coding, the occasional blown driver, the odd diving/creeping z-axis, susceptibility to electrical noise, having to stay in DOS because metric windows was odd and forever setting ramps - but felt that we couldn't grow the business with 2 or 3 routers that behaved like that.

Your first reaction is probably that I should have gone from standard to Alpha, but I did not have the confidence to totally lock myself into the ShopBot software and hardware. With the pre-Alpha, I could at least look after the hardware out here, and we run an old trusty DOS version that we've grown to understand - all the old system needs is a new driver (+$100) every couple of months or so. Being out here, this experience did not swing me to the Alpha, rather it swung me to what other folk out here are using; Gecko/Mach.

As I said at the beginning of this post, if you have an Alpha, or a standard SB running without glitches, then you shouldn't have read this far. :-)

[richards]

Gerald,
I immediately switched off the breaker to my Alpha and covered the machine with a tarp (out of sight, out of mind) so that I could read your last post.

Assuming that I had a shopbot, either PRT frame or PRT/Alpha frame, without electronics (no controller, no motors, nothing at all), what parts, including stepper motors, and software would you recommend to bring the machine to life? (Please respond as quickly as possible. I can't leave my alpha covered up all day :-)

[gerald_d]

Please excuse the delay in the reply, I read your post three times to make sure there is no catch....

Stepper motors with facilities for encoders, but with no encoders yet. Gecko 202's. PMDX-122 (because of charge pump). Mach3 software.

[richards]

Excellent choices. I completely agree with your selections. I would suggest the Oriental PK299-F4.5B motor which produces 800 oz-in of torque at 400 RPM when wired parallel and driven at 60V or higher. It also has a double shaft, making attachment of an encoder easy. Price for the motor, here in the USA, is $213.

(Thanks for the quick reply. The City Fathers let me start work at 7:00 a.m. - it's currently 6:45 a.m. ;-)

[gerald_d]

I am using PK299-01AA motors because that is what was fitted (direct drive) on our early PRT, and because I found a couple of slightly (ab)used ones from folk upgrading to Alphas. I have a way of installing encoders on their rears when the time may arise (when GREX/Mach4 is industrialised). The 1/2" shaft extensions look too big for elegant encoder installations. (I am considering similar motors with 1/4" rear shafts at half the price from Taiwan.......)

Am I hung up on Gecko/Mach? No, there are viable options to them - and this makes them even more attractive. If a bus runs over Mariss of Gecko tomorrow, then the alternative Pro-Drive 2000 will fit straight in. The Mach software is also carried by a large group of "volunteers" (many metric) who could take over if Art Fenerty bows out - or I could even go LinuxCNC if necessary.

Another thing that makes Mach/Gecko particularly attractive is the accessibility of Art & Mariss for support. Mariss actually convinced me to not buy his G202 last year. He said that the product was then too new to travel out here - and I admire/respect that frankness. He has an unusual repair policy - if a customer blows up a drive by doing something silly it will be repaired/replaced for free (once) provided the customer comes clean and explains how he blew it.

[stevem]

Gerald, what kind of performance are you getting with your Gecko/Mach setup?
Rapid speeds?
Max. cutting speeds without lost steps?

[gerald_d]

Steve, those are loaded questions. Sorry for the lengthy reply that is going to follow.

The current limit resistors for the Gecko's were chosen to limit the motor currents to the motor specs. (I don't have the figures to hand but I seem to remember an 18k resistor for a 2 Amp limit?). The original SB supplied drivers (IMS481?) were also limited at about 2 Amps. So, it was no surprise that rapid speeds and cut speeds are similar to the original SB system if the current limits are about the same.

Rapid speeds were set by holding the gantry back with a fish scale to about 10kg (22lb). I think that speed, without losing steps is around 100mm (4") per second. It could have been faster if I didn't want the reserve of 10kg.

Max "cutting speed" doesn't mean much to me because it depends on the "effort" needed to make the cut. If I want to cut lightly (below 10kg) then I can have a cutting speed higher than 4 inches/sec.

A couple of people have told me that I should be changing that current limit resistor to get much better performance. Apparently this is very common. They say to run the motors on half windings at the higher current - if that half burns out, drop the current and use the remaining good half...??!!

The impression is that a boost in voltage from 48 to 75 makes only a slight difference - the real difference comes from turbo-charging the current and the Gecko is very comfortable to do that for you. I saw the effect of max current, when (by error) I dropped the output a bit, and the max jog/rapid speed dropped substantially. (See message 7824 on Gecko yahoo forum)

Out of curiosity, what motors do you have and what resistors are you running? (If you have a power supply similar to mine, what is the outside diameter of the transformer, or the capacitor size in MF (mF))

[stevem]

Gerald, I'm using this power supply from Acopian, U45Y500, with a 10,000 MF capacitor (cost was $29.00 NOS). The resistor on the drives is 18K. The motors are stock, with gear boxes.

My machine operates comfortably at 10" per sec. rapid speed (measured with a stop watch), although I did not measure cutting force to calculate the speed limit. I increased the speed until the motors lost steps, then backed off the setting by 20%. Most of the cutting I do is at 2.5" per sec., The small amount of wood I have cut indicates that the cutting speed is limited by the power of the PC router, although 4" per sec. is easily reachable in shallow cuts of .375".

Since I've never cut anything using the ShopBot control, I can't compare cut quality. My home built router produces much better cut quality than my Bot, but uses THK linear slides and ball screws and operates at a max. rapid speed of 2.5" per sec.

[gerald_d]

Steve, that is a 50V supply (third one down this page (http://www.acopian.com/single-u-goldbox-44to180vt.html))

I'm off to bed now, but Mike is good at looking up specs
- it looks like your 18K resistor is giving about 2 Amps to a motor that is rated about 1 Amp? A bit tricky to compare your geared drive to my direct drive.....

Message 7828 on the Gecko list is interesting...

[dirk]

Gerald
If the motor you have is eight wires why don't you wire parallel? You would take a small hit on low speed torque but the torque curve and your top speed would increase dramatically, 1/4 the inductance over series bipolar (faster speeds) but double the amps.

[richards]

Some of these posts make interesting reading. Turbo charging a motor by forcing more current through its windings than specified by the manufacturer will (should) increase its speed; but at what cost? An automobile engine, when fed nitro, will turn faster for a brief period of time. However, when the life of the engine becomes important, running the engine within factory specs would seem to be the prudent thing to do.

Now, looking at four motors from Oriental Motor that are all the same size, but differ in the amount of current that they are rated to handle, shows some interesting facts as far as speed is concerned. In order to have some reference point that makes comparisons possible, let's say that the motor must produce 600 oz-in of torque in order to make the desired cuts. The PK299-01 will turn about 80 RPM @ 48 volts while drawing about 1.4 amps and still produce 600 oz-in of torque. The PK299-02 will turn about 120 RPM @ 48 volts while drawing about 2.1 amps. The PK299-03 will turn about 165 RPM @ 48 volts while drawing about 3.18 amps. The PK299-R4.5 will turn about 550 RPM @ 60 volts while drawing about 6.3 amps. My conclusion is that, if you were going to use a G202 drive and if you had about 7.0 amps available at about 70 VDC and if you were going to select a motor that moved as fast as possible with that drive and still produced 600 oz-in of torque, THEN the PK299-F4.5 motor would be the most likely candidate.

Let's translate RPM into move/jog speeds. My Alpha jogs at 30 ips, or 1,800 inches per minute. The pinion gear on my Alpha has about a one-inch pitch diameter, which means, as far as I understand, that the motor needs to turn 1,800 / 3.14 RPM per minute to move the various axes at 30 ips. So, 1,800 / 3.14 = 573 RPM, which is very close to the ~550 RPM available for the PK299-F4.5 motor. The PK299-01 could be expected to jog at about 4.18 ips. The PK299-02 could be expected to jog at about 6.28 ips. The PK299-03 could be expected to jog at about 8.63 ips. Of course, the one fact that is missing in these jog speeds is that the torque required to jog should be substantially less than the torque required to cut. So, based on whatever factor you wish to use, the jog speed with each motor could be increased by 50%, 100% or 200%. Again, the weight of the router/spindle, the tightness of the rollers, the amount of gunk resisting the gantries movement would have to be factored in.

What does all of this really mean? Simply, if you already have motors, and if you want the motors to last for a while, limit the current to the current specified by the manufactorer (whether you use RMS or PEAK voltage when calculating current is up to you because that is another topic entirely). If you've inherited a chassis without electronics and want the most bang for the buck, consider the PK299-F4.5 as being a really good match to a Gecko G202 and a 70V power supply. One last word of caution: Unless you really know what you're doing, plan on spending some substantial time bench-marking the Geckos and the motors. Plan on frying a motor and/or Gecko driver if you're really pushing the limits. Above all, look at retro-fitting a Shopbot with anything other than factory equipment as being a time-consuming experiment that may or may not be interesting, depending on whether you need your machine to produce cuts immediately or whether you have time and money to play for a while.

Getting the most speed out of your shopbot, at the expense of quality, might be counter-productive. This morning I cut 64 raised-panel MDF drawer doors in less than two-hours using moderate speeds. Sanding the edges to have the panels ready took almost four-hours. Although the Alpha is excellent in most aspects, it does flex more that I would have imagined. To get the most productivity out of the machine, it probably would have been more prudent to tighten up the axes by using some linear rails and possibly a ball-screw rather than trying to spin the motors faster than quality dictated.

[gerald_d]

Mike, just a quick question; If we have two stepper motors of similar physical size (PK299), one set up for 2 Amps and the other for 6 Amps, won't the one run much hotter that the other at creep speeds?

[gerald_d]

Further to the question above, I wonder if the high current version of the PK299 is meant for high-speed running only, due to cooling considerations? Anyway I have PK299-F4.5A's on order and I will found out in a couple of weeks or so.....

Will get back to some of the other issues tonight - have to go for a walk after mowing the lawn on this public holiday.

[richards]

Gerald,
I can't answer the heat question except to say that in a conversation I had with Marris, he said, as best as I remember, that if a stepper motor is run in the knee area of its torque curve, which is the speed before torque starts to fall off rapidily, that the heat build-up is minimal compared to running a motor so fast that it is running past the knee on the torque curve. He also said that a size-34 motor is meant to run hot, too hot to comforably hold in your hand. The specs allow a temperature rise of 80-degrees C with a maximum ambient temperature of 50-degrees C. The insulation is Class-B, rated at 130-degrees C.

In my personal tests, running PH299-03 (old round style motors) PK299-02 and some high current Superior brand motors, Marris's statement about heat build-up seems to be accurate. When the motors were run at speeds that allowed maximum torque, they ran warm but not hot. When I increased the speed to the point that they produced minimum torque, the heat build-up was significant.

As a point of comparison, the AS911 motor used on the Alpha seems to be a very close cousin to the PK299-F4.5 motor. The motors on the Alpha get warm, but at the 6-8 ips range where I normally run them, they never get hot.

(I'm playing with a 34-size brushed servo motor right now. It must be geared down at least 4:1 to produce adequate torque at Alpha-like speeds. It was certainly harder to tame than stepper motors; but, since it requires an encoder to work, it seemed easier, at least on paper, to build a servo system that would fault if 'steps' were missed than it would have been to build a similar system with steppers and encoders added to detect missing steps. That kind of begs the question, because, if a stepper is run properly, it won't miss steps. At any rate, that's a long-winded explanation of why I don't have test data on the PK299-F4.5 motor. As soon as I finish testing the servo, the next item on the list is the PK299-F4.5, which should arrive here at about the same time you receive your motors.)

[gerald_d]

After a rough start, this thread is now going well - thanks guys.

To summarise the key points for me.....

- The Mach part of the equation is only software, I believe very good software, and there is very little to debate around that.

- The Gecko part is the electric driver of the stepper motor, capable of accepting a wide range of voltages and driving a wide range of currents. Again, the Gecko's are good, and there is very little to debate on them per se.

- The power supply feeding the geckos must be looked at from a safety angle because mains power is involved.

- When first hearing of Gecko/Mach on this forum, the overwhelming anecdotal opinion was that the higher voltage of around 80V made it all go lightning fast. That seems to be a myth now. Steve's fast machine runs at 50V. Other fast machines run at 50V. ShopBots standard supply voltage is 48V. Also, from other forums, increasing the voltage from around 50V to just under 80V will not make a major difference. (The Gecko pops/burns at 80V - that is its max limit). Sure, when ShopBot increased from 12V to 48V (400%), that made a huge difference. The Alpha going up to 160V (330% up from 48V) also experiences a major difference. Going from 48V to 75V (50%) pales in comparison.

- "Speed" whether Jog, Rapid, Move or Cutting is a fuzzy concept for comparing Mach/Gecko systems to ShopBot controlled systems. A description of speed always needs to be qualified with load/force at that speed, considering always that a twin-motor carriage should produce double the force/load of a single motor carriage. I set max speeds conservatively because I don't want sawdust/grease on a rack to trip up a jog - nor must the 5" dust hose, nor the drag of the bundle of shielded cables.

- The max driving currents from the Gecko is adjustable (by installing different resistor values). Very little has been said about how this affects "speed". I thought that the max current only applies at crawl speed, but I discovered by mistake it also affects the max speed. Steve's machine appears to be over-driven, I've been told that I should over-drive my motors as well, but I am too timid, lest I burn the motors. This brings me to.......

- the motors. If you have a particular stepper motor in your hand, there could be as many as 8 wires coming out of it. The Gecko can only supply 4 wires. Look at a standard ShopBot and see that some wires are simply cut off near to the motor (sometimes too near to salvage). Nothing wrong with chopping the wires that is conventional practice. But, with a Gecko we can drive a lot more amps than the ShopBot drivers ever could, should we not try and pump current down all the wires now? If we do use all the wires and max current, won't we cook the motors?
To cut a long story short, Gecko/Mach are non-issues for me. They give the freedom to get the best out of given motors. The overiding question in my mind now is "What is the best way to connect stepper motors and what current limits should be set for them?" Time for a new thread on that?

[gerald_d]

Step Motor Basics (http://www.geckodrive.com/photos/Step_motor_basics.pdf) by Mariss is my choice of reading for tonight.....

My feeling that the current setting affects the max speed would only be true if the speeds were all below the knee-point. Now to figure out where the knees are.....

Steve, that 18K resistor gives you about 2.2 Amps. The SB controller for driving the same motor as yours was limited to 1.5 Amp. Allegro 3955 (http://www.allegromicro.com/sf/3955/)

[richards]

Voltage can play a large factor in a motor's speed. One concept that must be understood is that most stepper motors are low voltage devices, usually in the range of 2 - 5 volts (look at the face plate on your motors to see its voltage rating). Basically, windings on a stepper motor resist change in current, following the old law that things in motion like to stay in motion and things at rest like to stay at rest.

So, if you have a motor rated at 5 volts and tell it to move one step, it will - eventually. The windings in the motor take a little while to saturate, meaning that instead of "feeling" the 5V pulse immediately, the motor sees a 0V signal that eventually grows to 5V. Seen on a scope, the signal looks like a linear ramp. Okay, with that in mind, we can force the motor to react faster by sending a brief 70V pulse to the motor and then quickly dropping the voltage to 5V. That, in effect is what a 'chopper drive' does. It overcomes the inductance of the motor by using high voltage until the winding is at the rated voltage and then drops the voltage to normal levels to not burn up the motor.

That is one reaon the alpha drive works so well at 160V. The alpha drive is forcing a reluctant motor to change states more quickly than a lower voltage drive would be able to. Again, looking at the waveform on a 'scope shows what happens. If the drive's timing is setup properly, the scope will show an almost vertical pulse which never goes above 5V. The drive trottles back before the motor reaches its rated voltage, so the scope shows that the voltage stays within the motor's rated voltage. The only thing that changes is the amount of time required for the pulse to reach 5V.

[gerald_d]

Dirk, an interesting comparison:

621
from the Oriental Motor page on Unipolar/Bipolar Connections for 2-Phase Stepping Motors (http://www.orientalmotor.com/support/Unipolar_connection.htm).

They also have a Frequently Asked Questions (http://www.orientalmotor.com/automated/faq/Stepping%20Motor%20Systems.html) section.

[gerald_d]

Mike, there ia an odd marking on Oriental Motor's graphs that I don't catch. Along the zero torque (horizontal baseline) line of the graphs, there is a small "fs" mark at fairly low speed. Any idea what it means?

[gus]

Gerald,
Did a Google search and found........;>)

fs is max starting freq. of the motor

http://www.orientalmotor.de/de/uploads/documents/21122005144800.pdf

Load this page and do a Ctrl-f and look for fs. It is down in the document a ways.

[richards]

Gerald,
Ted found the answer while I was away from all computers today. Quoting from the document cited by Ted:

"Maximum Starting Frequency (ƒS) This is the maximum pulse speed at which the motor can start or stop instantly (without an acceleration or deceleration period) when the frictional load and inertial load of the stepping motor are 0. Driving the motor at greater than this pulse speed requires gradual acceleration or deceleration. This frequency drops when there is an inertial load on the motor."

When I visited Oriental Motors several years ago, to receive training on basic fundamentals of stepper motors, they advised us to start the motor at as high a pulse rate as possible (close to the fs mark) to avoid unnecessary vibration (resonance) that is inherent in steppers at certain speeds.

The resonance problem can be a problem with CNC machines that use stepper motors. As an example, if the X-axis were moved 96-inches while the Y-axis moved only 5-inches (MV, 96, 5), most likely the Y-axis would resonate (shake) because it would be moving so slowly. The good thing is that a motor under heavy load resonates less than a lightly loaded motor and resonance usually only occurs over a fairly small range of frequencies. The problem, of course, is that if you're making a cut that requires movement in that range of frequencies, the motor is going to sound like something is wrong, and the cut may show more than normal 'jitter'.

[gerald_d]

Why search if it is much easier to ask the question here and make someone else search?
Thanks Ted and Mike.

Interesting how the various Oriental Motor sites around the world carry different docs and part numbers. The F4.5 becomes E4.5 in another country (the UK I think) and there they don't have the 01AA.

[richards]

I just spent a few pleasurable hours testing an Oriental Motors PK299-F4.5B stepper motor being driven by a Gecko G212. One of the more surprising things that I learned was that the motor performs very well at 35VDC. At that voltage, it easily spins as fast as the Mach3 software can supply pulses. The voltage test started at 25VDC (Geckos require at least 24VDC to operate) and ended at 45VDC (Geckos should be voltage limited to 25 times the rated voltage of the motor). I configured the motor as parallel with current set at 6.15 amps (assuming that the 330K resistor is actually 330K). The motor sounded too loud at voltages above 35VDC, so I cut the voltage back to that level.

My torque test showed that the motor matched the torque sheet printout published by Oriental Motor. My very unscientific approach to measuring torque was to hang a spring scale from the ceiling, attach a pulley to the motor and loop a nylon cord around the pulley. Then I gently pulled on the cord until the motor stalled. (An 'exciting' lesson learned was just how fast friction builds up when a nylon cord starts to melt!)

My tests led me to the conclusion that the PK299-F4.5B motor can produce Alpha-like speeds. However, instead of seeing how fast the machine can run, it might be more worthwhile to gear the motor down 1:3, or so, and see if the resulting quality is better. (I'll never be satisfied until the quality of the cut matches the quality of the Shopbot's rails - getting higher quality would probably be impossible, but getting lower quality means that something isn't working properly - yet.) Gearing down the PK299-F4.5B 1:3 would limit the jog speed because the Mach3 software just can't send pulses fast enough - but the MachIV software and the G10x Gecko boards can - when they finally get to market.

Just one final note to those of you who think that your're going to have to throw out your controllers and motors in order to get a real machine. DON"T DO IT! I'm not planning on changing anything on my Alpha until the controller and motors die a natural death. This Gecko 'hobby' is something that I do to feed my need to play with process control computers without ruining my Alpha.

[dirk]

I also had some fun today. I tested the top speed on the megabot.
Here are my results:

Gantry weight 300#s
1303 oz stepper motors Belt gearing 2:1, 20 tooth pinion
Gecko 212 Drivers
60 volt power supply
Alpha controller (SB software)

Top speed:

Motors wired bipolar series (4.5 amp):
16 inches/sec before lost steps

Motors wired bipolar parallel (9 amp, Geckos only supplying 7amp)

20 inches/second (top step frequency reached)
90#s of force required to stall motor

40+ inches/second (using step multiplier in geckos, and raising speed limit in software)
5 to 6 seconds to travel 24 foot (including ramp)
3-foot travel from 0 to full speed
15 minute constant jogging to each end , no lost steps
15 lbs of force required stall motor at 40 inches/sec.

Notes:
Scary at top speed, 20 inches a second would be plenty.

[richards]

Dirk,
I wish that you weren't so far away. Watching your machine perform would be worth nearly anything - except a cross-country trip.

Do you have photos of your motor mounting method, or can you describe your pullies, belt, bearings, etc.? Since testing my motors, I've been thinking of using a belt-driven reduction unit, but my preliminary designs are cumbersome and heavy.

[dirk]

Hey Mike
I've posted some new pics on the blog.
http://cncshare.blogspot.com/
I cut the aluminum brackets on the PR. I went with L size belts and pulleys because of the large motor and I also had enough hub that I could bore it out and use a 5/8 shaft. I felt I needed as much support as possible because the location of the rack caused me to have more overhang. I went with flanged bearings mainly because it was easier. I was tempted to just use bronze bearings because it would have reduced the size of the reducer quite a bit. I used bronze bearings on the PR and it worked out quite well.
Dirk

[richards]

Dirk,
That is impressive! Your motor mounts/gearing showed me just how much I need a tutor. After looking at your design, it's obvious that my design wouldn't work. Thanks for taking the time to post everything. I'll be spending some serious time carefully going through your blog later tonight when I finish up some computer work for a customer 50 miles to the south.

One thing I did notice when doing a very quick scan of your blog, is that you've listed 15mm Hiwin rails and blocks. I've considered using 25mm rails and blocks in a project much smaller than yours. Do you think that 25mm is overkill? (I'm not much of a mechanical engineer, so I often tend to specify parts that are larger than necessary.)

[dirk]

That was a mistake on the drawing, it should have been 20 mm. Actually 20 mm was speced out on Bob Campbells design on t-slot rail, and that was the original plan to build entirely with T-slot Aluminum. Their isn't much difference in price between 20 and 25 mm, just kind of depends on what you need. I've had my own tutors, I was inspired by a fellow shopbotter Randy Arendt who emailed a pic of his reducer. Very nice, he used a cam to tighten his timing belt. I was tempted to have him build mine. I went the easy way.

[gerald_d]

Mike & Dirk, thanks for posting in the last couple of days. All very impressive stuff! It will take a while to digest. Dirk, what is the pitch of your rack/pinion?

[richards]

Torque tests for a 'standard' Oriental Motor's PK299-02AA motor matched the PK299-F4.5B motor; however, the -02AA motor was connected as a UNIPOLOR device, current was limited to 3amps, and voltage was 70VDC. (The PK299-F4.5B motor was connected using the parallel mode, 6.3amps and voltage was 35VDC.) Either motor would work very well.

Hooking up a motor using unipolor connections requires a six wire motor. You use the center-tap wire and one of the end wires for each of the two windings in the motor. On an Oriental Motor's motor, I use the black conductor and the yellow conductor for the 'A' coil. The red conductor and the white conductor for the 'B' coil. If the center-tap wires have been clipped, you'll have to use series wiring, which severely limits the motor's speed. Although I am extremely partial to Oriental Motors, other size 34 motors are available. I would look for a motor having the holding-torque rating at ~800 oz*in and a unipolor voltage rating of at least 3V and a unipolor current rating of at least 3amps. In my opinion, a square style motor with those ratings would be very similar to the PK299-02AA motor that I've tested.

[dirk]

Gerald, the rack was 20 pitch.

Mike, I was wondering if you ran accross any specs on your motors as to inductance rating. I couldn't find it on Orientals site. I was told to look for a real low rating to get high speed performance. I was wondering how it compared to the motors I have on the bot. Here's a spec sheet on my motor.
http://www.mcg-net.com/electric_motors/downloads/step_motors/IH34_step_motor.pdf
Mine is the last one in the column IH34114

Dirk

[richards]

Dirk,
The PK299-F4.5B is rated 2.5 parallel, 10 serial, and 2.5 unipolar. The PK299-02AA is rated 24 serial and 6 unipolar.

(When you get to the motor's main data sheet, click on the 'More Specifications' button to get the data on all motors with that frame size.)

The specs for the motor that you're using look really good to me.

[stevem]

Dirk, where did you buy the motors?

[dirk]

I bought it from Camtronics
Dirk