Has Anyone Gone Ascension after 4G?

[flyguy]

I am looking at getting the 4G board but I can truly not afford to upgrade my machine twice. I don't care about speed even a bit only in quality of the cuts. So has anyone gone to the 4G board and then traded up to the Ascension? In other words is the Ascension enough of an upgrade from the 4G board to be worth the extra money??

Thanks
Mike

[richards]

Mike,
If we strip away the hype about the 'advantages' of non-Shopbot controllers and look at the actual electronics involved, there would be no differnece in either speed or cut quality between a 4g and any other upgrade that uses Gecko G202 stepper drivers with a stepper pulse train of about 30,000 steps per second. The G202 divides each rotation of a stepper motor into 2,000 steps regardless of whether the G202 is housed inside a 4g or in another box. With equal power supplies and adequate current, the speed would also be equal. (The highly touted 'constant velocity' function available in the other controller should be renamed as the 'constant corner rounding' function. There are parameters that can reduce or eliminate the corner rounding produced by constant velocity. I think that advertising might have eclipsed reality in that particular case. Constant velocity on 'big iron' machines that have on-the-fly spindle speed control is a horse of a different color and not to be confused with the 'constant velocity' that I'm writing about here.)

If you're concerned about the 4g, why not call Shopbot and ask them about their return policy? I'm sure that they stand behind their product and would return your money if you had a problem. They've always done everything possible to help me. In fact, I've never before dealt with a company as customer conscious as Shopbot.

[jeffreymcgrew]

We upgraded to the 4G board, and it's been great. Once we worked out our USB communication issues it's been smooth sailing. And I do mean smooth. Our cut quality has noticeably improved, as has our speeds. I think if we did upgrade to anything at this point it would be into other add ons, like a second Z axis or big spindle. Even a friend of mine in the City with an Alpha commented on how quiet and smoother our machine is than his sense we did the upgrade. But his is still faster, mind you...

[jhicks]

Michael, I have no personal experience with G4 but Mikes comment about constant Velocity is really not the big deal. It does however add quite a nice set of features that allow one to make adjustments and move through 3D quite quickly or switch to squared corner parts for exacting corner work. Depending on many variables one can really run most common files without discriminating but 3D and circles, curved work, letters,and the like are definately efficient with the CV operation. And you can change move speed on the fly. Maybe 4G will do the same but thats primarily software.
The key thing I noticed with the ascension and Mach G code software seems to be a more efficient cut code,or fewer lines to accomplish the same cut.
I am no technocrat like Mike so all I can say is it appears to be more efficient in G code and that seems to translate to smoother cuts and more rapid move speeds.
Even with higher move speeds other mechanical limits come into play. The motors, spindle/router, gantry, and other mechanical limits can be over run by these faster speeds and feeds.
That being said, I assume the new 4G which was not available at the time I purchased the ascension, may have the same Gecko drives and perhaps IF G code is more efficient, the 4G would be relatively equal when comparing apples to apples on CAM processor and actual operation.
It would be very nice to somehow compare side by side but it looks like both add a lot to the older options so it probably depends a lot on what you intend to do with it.
Good luck and have fun

[conceptmachine]

Michael,
I'm tossing my opinion in here as far as a programer/operator.If you want to stay with the shopbot software then by all means get the 4g upgrade.If you are familiar with g-code and want to have more control over your sb then go with the ascension/mach setup.I'm not a technical wiz on the driver's so i can't say which is better but i've been a tool maker for the last 15 years and program cnc's daily at work and can tell you that you will get more bell's and whistles with the mach software because you can customize it to fit your need's.I downloaded the trial version and i must say that i was impressed.I'm not putting down the sb software but for me the mach software makes me feel more at home.Just my .02 worth.
shawn

[richards]

I'm going to throw another wrench in the works here and cause this thread to drift just a little. Since I'm a programmer and have designed, built and written software for process control computers since 1977, I have a feel for what it takes to get the very most out of electrical/mechanical devices. Most of you know that I have an Alpha and that I refuse to modify the Alpha to use anything except Shopbot software. You probably also know that I've been very actively researching and testing Gecko products for the last 18 months. (Currently I have four G202 modules, four G203 modules, one G320 module, one G101/G102 GRex and one G100 GRex along with a dozen different stepper motors and one servo motor.)

If you are a programmer and really want to get the ultimate out of your machine, get a G100 and use it in native mode. It is incredible what you can make it do. For those who like constant velocity, you can turn on constant velocity on-the-fly for any move that would best use that feature and turn it off when you want square corners. You can set a different ramp value on-the-fly for any axis and use either S-curve or standard ramping. You can select from pulse speeds from 32,000 steps per second to 4,000,000 steps per second on-the-fly.

For those who think that I'm really up in the night on this one, you could easily write a file processor program to read a standard SBP file and output native GRex code. It's almost exactly the type of thing that I did with Bruce Clark when we wrote the FreeDoors/doors.exe programs.

I just received the G100 yesterday and loaded up Mach3 to test it out. Today I found a minor error in the Mach3 slave axis feature. In trying to work around that bug, I decided to test the G100 in its native mode. Suddenly, the light came on.

If anyone wants to discuss writing an open-source interpreter for SBP files, send me an email at miker@xmission.com (mailto:miker@xmission.com).

[conceptmachine]

Mike,
I'm assuming that you would need the g202's to go along with the g100?

[richards]

Shawn,
Yes, you'll need a stepper driver, such as the G201, G202 or G203 to drive the stepper motor. However, if you have a spare AS911 alpha motor and stepper driver, the G100 can also be used with them.

The G100 is basically an elegant pulse generator that can provide pulses for almost any stepper driver that accepts step and direction signals.

[searayx350]

Mike,

I hate to jump around boards but I find your posts fascinating.

I have a quick question for you;

My tests – with oscilloscope, hardware port monitors and so on, conclusively have shown the following;

1.

[richards]

Josh,
I'm thinking in terms of a minimalist system that has adequate power, adequate speed and is cost effective. By cost effective, I mean that the entire machine, including electronics, but excluding the spindle/router and computer could be sold for $8,000 to $10,000. In keeping with that idea, four Oriental Motor PK299 size stepper motors, and the Gecko G100 pulse generator with four G203 stepper drivers and power supply components could be bought/built for $2,000. That would give you a electronics package that would match the Oriental Motors Alpha package for torque and speed. The Gecko products, without encoders, would NOT have feedback like the Alpha products - but the feedback on the Alpha is not synchronized, so I don't consider NOT having it to be a liability. The G100 can generate pulses at 4,000,000 pulses per second, or 20X faster than the G202 stepper driver can handle, and more than 100X faster than the 4g controller board can produce pulses - so speed is not an issue. The G20x series of stepper drivers produce 2,000 steps per revolution compared to 1,000 steps for the Alpha motors (as typically configured), so resolution is 2X better than a factory configured Alpha. In fact, I'm willing to believe that the components that I've described would cost no more than $200 more than the electronics used in the PRS. The extra $200 buys a tremendous amount of extra features - 16 digital input lines, 16 digital output lines, 6-axes with encoder and limit switch inputs on each axis, 4 analog channels and TCP/IP communications. The beauty of TCP/IP communications is that major accessories - such as an automatic tool changer - can be added without rewriting all of the code. It basically only requires one digital output line to signal the tool changer to do its thing and one digital input line to receive a status signal from the tool changer that it has done its thing. The master program only has to have a tool changer function added and another TCP/IP controller to interface to the tool changer. I'm sure some are already saying, "Yeah, but that would cost a fortune." Well, the Z-World Rabbit module that would control the tool changer is expensive, if you consider $70 to be expensive. Not to mock the efforts of those who have actually built a tool changer - because I haven't - but an old fastioned dot matrix printer has much tighter timing restraints than a tool changer.

My main concern, when I designed process control computers, was to give added value and the possibility for expansion without adding to the cost of the equipment. I believe the parts and pieces that I've written about today would do that very well.

[searayx350]

Mike,

Good answer and makes sense.

By the way, check your e-mail please.
The 2nd I sent you is what you should look at.

JM

[richards]

Let me clarify something that several of your have questioned in private emails: I am not suggesting that anyone compete with Shopbot in building a new machine. As far as I can see, and from the posts of Danny L. Ray II and others, the PRS is a great machine, even better than the Alpha that I own - and I consider the Alpha to be an excellent entry level CNC router.

The heading that we're posting under is Variations and Modifications of Shopbots. To me, that means that we are discussing what we would do to our machine to vary or modify it from the normal Shopbot.

I hope that all of you are totally aware of the fact that I have not modified my Alpha electronically - nor do I plan to until or unless the electronics die a natural death. I have made some minor modifications mechanically. The 'bow' finally broke one too many times so it was replaced with a real Rube Goldberg contraption. I also added belt-driven gearboxes to the X and Y motors, which convinced me that I could get the quality that I wanted from an almost stock Alpha.

Unless I've misread the specs on the PRS models, both the 4G model and the Alpha model have gear boxes which means that Ted has addressed the ONLY thing that I took exception too to the point that I did something about in on my own machine. Personally, if I were going to build my own machine, it would look almost like the PRS - and since it would always cost more to design and build an original machine than it would to buy the PRS and then modify it, I'm sure that I would buy the PRS and then add my variations and modifications to it.

It's kind of like hot rodding a car. Everybody that I know who has done that has started with a car and then hacked away until they either got what they wanted or ran out of money. I see the Shopbot in the same way. Compared to a car, it is a darned good machine right from the factory. Some people like to tinker - people like me for instance. Some people have long experience in messing around with electronics and software - me for instance. Some people would spend a year of their life fixing something that isn't broken, just for the challenge of doing it - me for instance.

That's kind of the way the Gecko G100 is to me. Because it exists, I have to mess with it. In fact, last week it was important enough to me to get my hands on the G100 and the G203 modules that I drove 1,500 miles to get my hands on them. (For those who think that only a crazy man would do that, you might be right; but, driving 1,500 miles through the desert also gives me a lot of private time to think and contemplate and mull over designs and ideas without being pestered with the telephone every few minutes. It's a good thing that I needed the time to think and comtemplate because only the G203 were available the day that I stopped by. The G100 was shipped and arrived the day after I got home.) Because it can be hacked, I'm hacking it. Hacking code is just as real to me as modifying mechanical stuff. When you take the ramping routines that Shopbot has - which are very good and work great just as they are - and find that by using other electronics (the G100 for instance) that you can do things that you only dreamed of being able to do, then it makes playing with the new electronics interesting. Or take the constant velocity function that others really like in their non-shopbot electronics. To me, if I had to choose one or the other, I would choose to NOT use constant velocity because it messes around too much with the corners; however, if I could turn constant velocity on or off at will from within the program, I would certainly use it. Creating a simple routine to cut an arc or a circle is soooo much easier when constant velocity is turned on. Instead of fiddling with thousands of line segments, you only have to compute a few - maybe as few as thirty-two - to get a smooth and accurate arc. Cutting an arc using any two axes is another feature that I sometimes wish I had. It's just another thing that can be hacked given the right electronics and a little time.

Hey, lets not forget that there's a hobby component to owning a Shopbot. Although it is a workhorse machine, it can also be a very interesting hobby. When you get to be old and white haired like me, it's really nice to not have to hunt and fish - especially when your legs can't hold you up any more and your eyes can't tell the difference between a horse and an elephant. Fiddling and twiddling with electronics and software fills that void in my life. I hope that some of your get the same satisfaction from electronics and programming as I do.

[cjohnson]

Mike, you mean to tell me you are old AND you have hair? I'm jealous.



It appears to me that the discussions around modification and variations tickles my intelect. If we are drifting the discussion from the orginial question about 4g and ascension, maybe another thread should be started around the Gecko or something. Would that satisfy those who have concerns?

cj

[jhicks]

The thing I now find as a limit with the ascension / G code is my 1 amp motors. Not being a techno guy, can anyone tell me if i can simply add 2hp motors to my PRT and get the extra UMPHH, I need to handle the new speeds and feeds without stalling my motors?
Beside considering new THK rails, i think I could really gain with more powerful motors.
Thanks

[richards]

Jerry,
The Gecko stepper drivers can handle motors up to 7A. I don't know what Gecko model you have (G201, G202 or G203) but they are all very similar in how they are connected.

First, you'll need to make certain that your power supply can handle the increased CURRENT. Mariss suggests that you add the current requirements for all motors and then multiply by 0.667 to get the minimum necessary current flow. For instance, four two amp motors = 8A. 8A X 0.667 = about 5.5A. So, at a minimum, your power supply would have to be rated at 5.5A. It's also a very good idea to use a large capacitor with your power supply - even when you use a switching power supply. Stepper motors like to pull hard on a power supply and a large capacitor acts like a temporary battery to store the necessary current between sine waves on the power supply (yes, even a switcher has sine waves; they're just closer together than the sine wave on a linear power supply).

You'll also need to change the current set/limit resistor on each Gecko stepper driver. With a 1A motor you probably have an 8.2K resistor (or something within about 10% of that value). For a 2A motor, you'll need to use an 18K resistor (or the nearest match from Radio Shack). 1/4 watt is adequate. The formula to determine the resistor (with the auto current reduction jumper turned OFF) is 47 X Amps / (7 - Amps) = the resistor's K value. That means that an answer of 8.2 is an 8.2K ohm resistor. If the auto current reduction is jumpered ON, then you are limited to 2A motors and you would use the formula 47 * Amps / (2 - Amps) = K. (There's a lot of controversy about using or not using auto current reduction. I've made my view known, but I respect the opposite view from others. They make a very valid point for leaving auto current reduction OFF. The motors that I use are 3A minimum so the question is not valid with my equipment. I also use gearing to keep the torque 3X higher than a non-geared motor.)

Anyway, that should get you going.

Edited: Brady has made a very good case for using Alpha motors with their excellent gearboxes instead of using the PK motors and Gecko stepper drivers. The Alpha motors are smooth and quiet and expensive, but they are the ultimate in stepper motors. You wouldn't need the Gecko drivers because the Alpha motors come packaged with their own drivers. You might drop him an email.

[jhicks]

Thanks, guess I have more research to do

[richards]

Jerry,

You might have a simple option to get a little more power from your existing motors. You're probably using a PK296A1A-SG3.6 motor (it may have a different number, but if it is a geared motor from Oriental Motors, it will be either the PK296A1A or the PK296A2A if it is a single shaft or B1A or B2A if it has a double shaft. If the resistance between the black and the green wires is close to 4.4 ohms, measured at the wago connector with the connector disconnected, you have the PK296A1A motors. If the resistance is about 0.96 ohm, you have the PK296A2A motors. If you have the PK296A2A motors, send me an email, and I'll walk you through the conversion.). You're probably using the motor with a series connection where the black wire is connected to Phase A, the green wire is connected to Phase /A, the red wire is connected to Phase B and the blue wire is connected to Phase /B on the Gecko (the wires may be connected differently to the four Phase connections, but if those colors are used, your motors are connected in series).

If you use either the black and yellow OR the yellow and green wires for one of the Phases and the red and white OR the white and blue for the other Phase, your motor will be converted to unipolor. Unipolar will give you 1-1/2 amps instead of 1A.

You will also need to change the current set resistor from the 8.2K (or similar) that you're now using to the closest standard 1/4 watt resistor to 12.8K that you can find. It is important that you find a resistor or two resistors in series that add up to 12.5K to 13.5K or you won't get the last drop of power out of the motors (A 10K resistor and a 3K resistor wired in series will probably be the closest match from Radio Shack.)

Your power supply should work fine with the motors converted to unipolor as long at it is rated at 4A or higher - preferably higher. Converting the motors will change their voltage rating from 4.4 volts to 3.3 volts; however, the Gecko is designed to work at 4X to 25X the motor's rated voltage (it's black magic). I usually run mine at about 15X the voltage to keep the heat down a little, but steppers are made to run hot and 25X the voltage is fine. The usable voltage range for a G202 is 24V to 80V, so if you're within that range, the motor will be happy. (If you're interested the voltage rating change keeps the wattage constant. 4.4V X 1A = 4.4 watts. 3.3V X 1.5A = 4.95 watt. Oriental Motor must think that 4.95W is close enough to 4.4W to not cause problems with the motor.)

Be sure to drop the motors down from the machine when you first try them out. If a motor runs backwards you'll need to reverse the wires on one of the phases - with the power off of course. For instance, if you have black connected to Phase A and yellow connected to Phase /A, and the motor runs backwards, you would connect yellow to Phase A and black to Phase /A - without changing anything on Phase B or Phase /B. Depending on how the two X motors are configured, the controller might invert the direction signal from one motor, or the wiring might be adjusted, as described above because one X motor has to turn CW while the other motor turns CCW or you'll really rack the gantry in a hurry.

If you have questions, send me an email miker@xmission.com (mailto:miker@xmission.com).

[bruce_clark]

Mike,

I know I am being totally nitpicky here, but actually, it is not unipolar--Geckos cannot drive stepper motors in unipolar mode.

What you are describing is bipolar HALF-COIL mode. This gives you the same performace as running the stepper in parallel (a 4-phase 8-wire motor) but 70% of the series torque.

So, yes, you will get a faster turning motor, but it will only have 70% of the series rated torque.

To get the full rated torque AND increase the performance, you need to run them in parallel, which is not possible with 6-wire motors.

Sorry to be so nit picky, but Jerry needs to know that he will lose 30% of his current motor's torque by rewiring them.

Bruce

[richards]

Bruce,
Normally I totally agree with you. You are right about the correct name being half-coil. I normally use that name myself; however, it has led to confusion in the past when I've said half-coil and then received emails saying that Oriental Motor's charts don't have a listing for half-coil. Since the important values - current rating and voltage rating - are the same for half-coil or unipolar, I've gotten in the habit of specifying unipolor. (A true unipolar stepper driver would have twice as many transistors as a half-coil driver and would use the center tap wire (yellow)as the power connection and the other two wires (black & green) as the transistor connections. Long ago, before low-cost stepper drivers were available, and we had to build our own L/R stepper drivers circuits, that's the way we hooked things up. Thank goodness that there is a better way now.)

The output power is a little more tricky to answer. It's related to Ohm's law. If we have a voltage, say 1V and a resistance of 1-ohm, the resistance will allow 1A to flow through the circuit. Then, we multiply 1V times 1A to get the power rating, which would be 1W. Now if we reduce the resistance to 1/2-ohm (since we're only using 1/2 of the coil), we would get 2A flowing through the circuit and we would also get 2W of power. Or twice as much power from a half-coil connection compared to a full power connection. Since motor makers don't normally build motors with the ability to handle 2X the wattage in half-coil usage compared to full-coil usage, they do us a favor by multiplying the voltage by 0.7 and multiplying the current by 1.41 in their specifications, which basically gives the same wattage whether the motor is wired series or half-coil.

And now for the magic answer why Jerry's motors can produce more power (torque) when wired half-coil compared to being wired full-coil: The Gecko G202/G201/G203 stepper drivers are designed to feed the stepper motor 25X the motor's rated voltage - without damaging the motor; however, the Gecko is limited to a maximum of 80V. In the case of the PK296A1A-SG3.6, the full coil voltage rating is 4.4V and the unipolar (half-coil) voltage rating is 3.3V. Those figures show that although the stepper motor could be driven at 110V if the Gecko stepper driver could handle 110V, the Gecko limits the highest power supply voltage to 80V. When we do the math, 25 X 3.3V is 82.5 volts, much closer to the maximum rating for the half-coil winding.

So, in very simple terms, as long as the power supply voltage is less than 25X the unipolar (half-coil) rated voltage, a motor used in half-coil mode will produce higher watts (power) than the same motor used in full-coil mode when used at the same power supply voltage.

The real advantage of using a motor in half-coil mode is its ability to ramp up faster and to achieve much higher speeds without losing torque. (The torque curves published by OM will make believers of all of us.)

Now, just so that everyone doesn't rush out to their machines and rewire the motors to get extra power, doing what I have suggested only works with certain motors. The PK296A2A-SG3.6 motors that I use have a much lower voltage rating, meaning that I have to lower the power supply voltage when I wire the motors in half-coil mode, which also means that the wattage produced by my motors is basically equal when they are properly wired as half-coil (at reduced voltage) or as full-coil (at higher voltage). I wire them as half-coil to get higher torque at higher speeds. And, I only talking about using the stepper motors with Gecko G202 or G203 stepper drivers. Other drivers will have other requirements/limitations when compared to the Geckos. (If you are geeky enough and like cheap thrills, hook up a PK299-02AA motor in half-coil mode and then run it at 3,000 RPM using a G100 connected to a G203 stepper driver. It will make you start thinking of the endless possibilities that high pulse rates and high stepper speeds offer - with gearing of course.)

Please, anyone, if you have questions about doing this, don't bother Shopbot with questions about hot-rodding the motors. What Shopbot offers works very well and will keep you out of trouble. If you have questions, send me an email, and I'll try to help. miker@xmission.com

[jhicks]

Mike, that's a Mouthfull there between you and Bruce. BUT I will print this out, take a closer look and once I am able to digest what I have and what you are both saying here, I'm sure I'll be in touch.
Thanks

[bruce_clark]

Mike,

Now, my momma always said, "Better let them think your are an idiot than to open your mouth and prove it." But here is goes....

You have a wire that is say 2 feet long. Roll that up into a coil and feed it 1 amp. Say that coil makes 2 imaginary "flux" units (2 feet * 1 amp = 2 FLUX). Now, cut that wire in half and give it 1.5 amps. You get 1.5 FLUX (1 foot * 1.5 amps = 1.5 FLUX).

From my understanding of steppers, it is AMP/Turns in a stepper that determine its power. Increase the turns of wire OR increase the amperage is what determines the power output.

Now, I agree that running a motor in UNIPOLAR (or bipolar half coil) will keep the torque curve more level as speed increases, but it will be shifted DOWN compared to a series wired stepper. Now, this assumes that you are using the same voltage for either wiring method.

If you up the voltage, I would contend that any rating you show me in bipolar half-coil (or unipolar), I would show you an equal improvement in series wired mode as well.

I would like to see the OM pull-out torque curves for both series and unipolar wired motors. If you can scan them or point me to a link to view them, I would like to check it out.

Remember, this is all in good fun!

Bruce

[richards]

Bruce,

You've brought up some interesting points. I'm fairly certain that your reference material differs from my reference material. I'll send you an email with two of the documents attached that show the basis of how I arrive at the figures that I use. One of the documents is Mariss's White Paper that can be found at www.GeckoDriver.com (http://www.GeckoDriver.com) in the support section under the title "Step_Motor_Basics.PDF". The other is from Oriental Motor. It is titled "StPk29.pdf". It lists data for all of the PK29x sized motors with torque curves showing both bipolar series and unipolar curves.

I was hoping to be able to copy a page from Mariss's paper, but the graphs didn't copy (perhaps you'll have a way to post them from the documents that I'll email you); however here is the text from two important paragraphs:

"The motor power output (speed times torque) is determined by the power supply voltage and the motor’s inductance. The motor’s output power is proportional to the power supply voltage divided by the square root of the motor inductance. If one changes the power supply voltage, then a new family of speed-torque curves result. As an example, if the power supply voltage is doubled then a new curve is generated; the curve now has twice the torque at any given speed in region 2. Since power equals torque times speed, the motor now generates twice as much power as well."

... remember the graph is missing ...

"This shows a half winding connected motor delivers twice as much power as a full winding connection at a given power supply voltage. This is because full-winding inductance is four times higher than the half-winding inductance."

Hopefully, you'll agree that I've posted pertinent and fair information when you see the entire document.

The torque curves published by Oriental Motor show that a unipolar connection gives sustained torque at much higher speeds than a serial (full-coil) connection. The chart shows both 24V and 48V for series, but only 24V for unipolar, so I'll use the 24V figures and assume that Mariss is correct when he states that "if the power supply voltage is doubled then a new curve is generated; the curve now has twice the torque at any given speed in region 2. Since power equals torque times speed, the motor now generates twice as much power as well." The chart shows that for a series connection on a PK299-02AA motor (a motor that I have and prefer for low power applications), the motor produces about 800 oz*in of torque at rest, 600 oz*in at 50 RPM, 200 oz*in at 100 RPM and ~50 oz*in at 150 RPM. The torque drops below 50 oz*in as the speed increases.

The same motor connected as unipolar has about 600 oz*in at 50 RPM, 400 oz*in at 100 RPM, 300 oz*in at 200 RPM, 200 oz*in at 300 RPM, 100 oz*in at 600 RPM, and 75 oz*in at 1000 RPM.

In the case of this motor, if used without gearing, and assuming a feed speed of about 4-ips, which equates to about 100 RPM, depending on the spur gear used, the unipolar connection would have 2X the torque of the bipolar series connection. At 8-ips, the unipolar connection would have more than 6X the torque of the bipolar series connection. Other motors would give different results.

My favorite motor of the PK29x series is the PK299-F4.5B, since it can be wired as parallel and give 400 oz*in at 800 RPM.

Anyway, Mariss's paper gives good background for those using Gecko products and the Oriental Motor's chart shows how voltage levels and resistance affect speed. For people like me who like to read data sheets, they are treasures.

[bruce_clark]

Mike,

On the surface, it looks like the data says your are right...but I do have to nit pick again.

First off, Mariss's white paper says exactly what you said above BUT his graph compares a Bipolar series to a Bipolar Parallel wired motor. If I remember correctly, a few years ago this was pointed out and there was a clarification, but I cannot remember what it was--getting old does wonders for your memory...
Second, I looked at the pullout torque graphs and yes, a unipolar motor has more torque at the same RPM than a series wired motor at 24V. I would contend that 24V on a series wired motor is a) too low for this low a current motor and b) barely 4X the rated voltage for a constant current driver (which they used). Also, the 48V graph on the series wired motor looks totally different than the 24V.

I am trying to find an old PacSci catalog which has graphs for unipolar, series and parallel wired motors. If I remember correctly, there were some hi voltage drivers used on their test and I think it will prove my point. So, on the surface, for the PK299-2XX motor, it appears that running it at 24V bipolar half-coil gives more pullout torque than 24V bipolar series. I will conceed to you for this motor/voltage combination. Pending further data...

Bruce

[richards]

Bruce,

Did I misread the paper? This is what I thought I read:

"The following graph shows the effect of rewiring the motor from full-winding to half-winding while keeping the same power supply voltage." I think that clearly states that we're comparing full-coil windings to half-coil windings. Perhaps there was a retraction in some other paper, but I don't have any other documentation.

I also totally agree that running a motor at 24V is neither practical nor efficient; that's why I quoted Mariss when he said, "If one changes the power supply voltage, then a new family of speed-torque curves result. As an
example, if the power supply voltage is doubled then a new curve is generated; the curve now
has twice the torque at any given speed in region 2. Since power equals torque times speed, the
motor now generates twice as much power as well." , so that it would be evident that if we used twice the voltage on a full-coil winding, that we would get similar results with twice the voltage on a half-coil winding. My own tests with a fish-scale show that what Mariss wrote is provable. That's why I've taken more than a year to test and prove some of these theories. Those theories totally contradict what I thought I knew from past experience. In fact, when Gerald Dorington suggested the use of Geckos and of toroidial transformers, I resisted him here in the forum because it went against what I had been doing for almost twenty-years. However, because he had been right on so many other subjects, I bit the bullet, and ordered a G212 stepper module and a toroidial transformer. The results were totally amazing. Then, I called Geckodrive support and talked directly with Mariss, who was totally amazing with his extensive knowledge of how a stepper works and what a stepper driver should do.

In fact, late last night, I ran some more tests on the G203 stepper drivers, but this time I used a knocked-together power supply using a 24V 12A TRIAD F-226U transformer that I had that was left over from some other project. Mariss suggested that the proper way to determine the size of a power supply was to sum together the amps required for all the steppers (4 X 3A) and then multiply the sum by 0.667 (8A) and then use that number as the minimum current requirement for the power supply. I thought, "A 12A power supply is larger than an 8A power supply, so I'll give it a try." After about an hour of testing, I knew that the TRIAD transformer was really getting hot, so I measured its temperature with an infared thermometer. It read 146-degrees F., which is much hotter than I like. I thought that the test showed that Mariss was wrong and that I should have used a larger transformer. However, when I used a Greenlee CM-600 Clamp-on Meter to measure the transformer's current, it said that the secondary side was drawing 3.5A and that the primary side was drawing 1.5A. That did not make sense to me, since 3.5A should not have caused the TRIAD to heat up so much. So, this morning, I bought a Fluke T5-600 meter for $100 to cross-check the Greenlee meter. Guess what? All that I proved was how quickly a fool can be separated from $100. The Fluke meter gave a reading of 3.3A on the secondary side of the transformer and 1.2A on the primary side of the transformer. The figures were slightly different, but well within the tolerance specs of the individual meters. The end of this long story is that when Mariss says something, it most likely is correct.

I hope that we're not getting into a contest to see who is right; rather, I hope that we're all involved in a quest to find new and better ways to use existing technology to improve our machines. I totally complement Ted and his crew for looking at and then adopting the Gecko product for the low end machines. It took time and money and effort and revisions to get everything working, but the Shopbot community is better off with a stepper driver that gives 4X the resolution and 2X the speed (depending on how deep one cuts and how hard one pushes his machine). Brady Watson and I have exchanged several emails where we discussed various stepper motors. (Brady, I hope you don't mind my mentioning this on the forum.) He is a stalwart advocate for the Oriental Motor's Alpha stepper and alpha stepper driver - particularly the newer motors with attached gearboxes. I totally agree with Brady that a machine using Alpha steppers and stepper drivers will out-perform a machine using PK steppers and Gecko stepper drivers. The Alpha product is in a class of its own, just like a BMW is superior (in engineering) to a Chevy. But, just like a BMW costs much more than a Chevy, so does an Alpha stepper/driver cost much more than a PK/Gecko. In fact, for the price of one Alpha stepper/driver, I can purchase several PK/Gecko units and get functionally the same results.

I think that we're in the same situation in this thread. There are lots of other options available that might give better performance. There are other options that are less expensive. But when we consider the entire package, especially if we can re-use our existing motors, I don't think that there are many options that we'll find that are superior to using a Gecko G20X driver and a PK motor. Luckily, we also have a choice of whether to connect the PK motor using full-coil or half-coil connections. By carefully studying the charts, we can determine which would probably work better for each of us. If we make the wrong choice, we'll have to spend another dollar to buy four more 1/4-watt resistors and ten minutes of time to change the wiring to change the motor to the other wiring configuration.

[conceptmachine]

I took the plunge and ordered the ascension today, i think that this route will suit me best.Now i need to get a spindle and add hardened rails and i'll be all set!just curious to how many of you fellow botter's are using the ascension and what other mod's you have done?
shawn

[richards]

Shawn,
I'm curious why you would choose that particular upgrade. Four Gecko G203 stepper drivers would cost $600. (The G202 costs a little less). A PMDX-122 breakout board costs $85. A Mach3 license costs $160. If you own a PRT you already have a case and power supply. So, for $850 you have all the parts and pieces to upgrade from a standard PRT to a G-code PRT with the latest and greatest Gecko stepper drivers. Why spend an extra $800-900 to have someone do thirty minutes of wiring for you? (Lest you think that I'm minimizing the time to wire up a machine, I've done it many times with the various G202, G203, G101 and G100 modules that I have. Everything is straight-forward. A step wire, a direction wire and a common wire from the PMDX breakout board for each G20x module, a 1/4-watt current limiting resistor connected to the G20x, and four motor leads connected to the G20x are the 'tricky' connections. In addition to that, you have to connect the + and - conductors from your power supply to the G20x modules. There you have it - $800 - $900 of wiring.)

For those who cry 'foul' and insist that you should also change the power supply. That's a $95 'upgrade' - $75 for a 500VA Avel toroidal transformer, $15 for a 10,000uF capacitor, and $5 for a bridge rectifier.

So now what is left that costs $700 - $800. Is it the case? Well, my personal opinion is that, even though the black case that is part you the package that you specified looks pretty good, it's not a NEMA-4 case that will protect the components from dust and debris any better than your current PRT controller case. A NEMA-4 case from FactoryMation will cost $160.

In the process control world that I lived in for thirty years, there was a concept called "Value Added" that allowed us to make a living building "Add On" components for various electro-mechanical machines. Typically I would take a Kodak printer than had tube-type analog electronics and upgrade that machine to a digital computerized modern machine with networking, advanced multi-point slope control, adjustable filter-action software for $6,500. The "Value Added" portion of that package was the software. The electronics were certainly an enhancement, but the software increased production from 250 prints per hour to 1,500 prints per hour. Since one machine could do the work of six, the payback for the customer was the labor savings.

I see the "Value Added" component that Shopbot adds to their electronics - that, of course is the software. The package that you bought seems to just be a gathering together of parts and pieces that others have designed and then adding $750 - $1000 for packaging everything together.

[rustnrot]

Out of curiousity, if one has a 4G board (has Gecko drivers, correct?), and one wanted to run Mach3, what would be needed to convert? Humor me here, as there is probably no logical reason to ever do this, just curious.

[srwtlc]

Mike, ya gotta remember that not all come from or have such a background and knowledge in electronics as you. All the testing and trouble shooting has been done for you and the system is known to work.

Just as I used to build over-the-road semis from a "Kit" (just the cab and frame, nothing else), not every driver would want to or could do it themselves. That's were those of us that can come in. ;-)

[jhicks]

There is a difference between tinkering, hot rodding, and buying a well designed, street ready, Shelby Cobra Mustang.
Many try to duplicate and have great fun in the process. Others prefer performance out of the show room.
Some like being mechanics, others like to get in and race. Either way its all about doing what you have a passion for and what works best for you.

[richards]

Tom,
The 4G board (with control box) has all of the physical parts needed to run Mach3 EXCEPT a breakout board and Mach3 software. I use the PMDX-122, but there are several other boards that would work just as well. Essentially, you would need to remove the G202 stepper drivers from the 4G board, mount them on to a piece of aluminum or directly to the controller case, and then connect a few wires from breakout board to the G202 modules. Before doing that, however, I would suggest that you download the trial version of Mach3 and play with it. If and when you find features in that software that are not available in Shopbot's software (and there are a few features exclusive to each software program that make it hard to choose one over the other), then you would need to determine how you're going to generate the necessary tool paths because PartWizard won't do G-code. There are many programs that will generate G-code, some of which are both powerful and inexpensive, but they are something that you'll have to learn to use. Personally, I have a licensed version of Mach3 and several versions of tool-path generating software. I also have all of the electronics necessary to build control boxes for several CNC routers, but I'm still using my Alpha with Shopbot software. As long as the factory electronics keep working on that Alpha, I'll leave it alone. If I do venture out and decide to use G-code, it will probably be on a small CNC-router either built from scratch, or more probably built from a new PRS machine sans electronics.

[conceptmachine]

Mike,
Simple answer; plug and play,i just got a job to more than cover the cost of the upgrade and maybe even get a spindle out of it!I could have built my own,but knowing my luck with electronics i probably would have fried it on the first try.

[conceptmachine]

Tom,
If you would like see the difference in the sb software and mach you're welcome to come to my shop i'm only a few hours from augusta.It will be a couple of weeks before it's ready though.

[richards]

Scott and Jerry,
I agree with you 100% about building or buying. However, I think that there is a mystic surrounding simple electronics that keep many from trying do-it-yourself projects. To me, if you made the simple wiring connections necessary to get your Shopbot running, or if you can replace a thermostat on your furnace, you can easily wire up the electronics that I posted about further up in this thread. On the other hand, if you hired an electrician to wire up your Shopbot (as suggested in the Shopbot documentation) or if you would rather call a repairman to replace a thermostat, I would not argue with your decision. Everyone has his own comfort level, and, I suspect, knows what will be easy and what will become a hair-puller. (Everytime I tackle a plumbing problem my wife suggests that I call a plumber. After doing and re-doing the repairs at least three times, I know that she knows, and has always known my limitations when it comes to plumbing repairs.)

Edited: Shawn, you posted while I was composing. I agree with you also. If you have work waiting, it's nearly always best to buy ready-to-run upgrades.

[conceptmachine]

scott,
I see that you sold your bot,did you get something else?I like the mod you did to your y car,i'm either doing something similar or making a new one altogether.
shawn

[srwtlc]

Shawn, I'm running a PRS now. The plate and the hold down setup worked out pretty nice. The plate did a good job of reinforcing the carriage and not having to mess with the four hold downs ever again, priceless.

[tommyr]

Scott,

This is off subject but I value your opinion, in short what is your opinion of your new PRS machine.

Tom

[Forum Admin]

Thread closed to prevent major drift. Start a part2 if more discussion on original topic. Thanks.