PDA

View Full Version : Recommended stepper motor drivers - G202 (or G203V?)


Gerald_D
Fri 27 October 2006, 23:53
This is probably the part of the MechMate that needs the least discussion - get Geckodrives (http://www.geckodrive.com/product.cfm?pid=10)! http://www.mechmate.com/Forum/clipart/happy.gif
The G202 (http://www.geckodrive.com/product.cfm?pid=10) is the right one for now. It is 15 year old proven technology, but still up there with the best.

You need 4 of these G202's for a MechMate

(March 2007 - G203 (http://www.geckodrive.com/product.cfm?pid=38) drives are becoming a reliable option - see post lower down)

(If you are using an already packaged control box, then you don't need to buy drivers...)

Gerald_D
Sat 28 October 2006, 04:18
GeckoDrive will soon launch a newer version of the G202, which will be known as the G203 "Vampire". These are the features:

1. Short-circuit protection
2. Reverse polarity protection
3. Over-temperature protection
4. Over-voltage protection
5. DISABLE input is optoisolated
6. STEP, DIRECTION and DISABLE are COMMON ground
7. STEP, DIRECTION and DISABLE work with 2.5V, 3.3V and 5V logic
8. STEP, DIRECTION and DISABLE need only 2.5mA
9. STEP pulse frequency to 350 kHz
10. No funny DIRECTION to STEP pulse timing. 200nS setup/hold times
11. 0 to 7A phase current
12. 18 to 80VDC supply voltage
13. Recirculate mode while stopped for reduced motor heating
14. No user jumpers or settings inside
15. Top settable ADJUST trimpot
16. Green POWER LED, red ERROR LED
17. Internal socketed power fuse
18. Crystal-controlled 19.531 kHz switching frequency
19. No external 470uF/100VDC capacitor needed
20. Power-on reset

Snag is, the G203 is a totally clean sheet design with zero exposure to the real world yet. The G202 has evolved and is fully debugged. The people living close to GeckoDrive can "experiment" with the G203, but I suggest to stick to the G202 if you live more than a 1000 miles from California. Geckodrive will not be withdrawing the G202 from the market until there is a natural swing away from it.

Gerald_D
Tue 31 October 2006, 22:30
The press release by Mariss last night:....

Good evening! The G203V drive gets introduced for sale tomorrow.

The G203V is our first attempt at a truly bullet-proof drive. Short
it, run it without a heatsink, overvoltage it, miss-wire motors,
reverse power supply polarity; throw everything at it save dunk it in
saltwater or fill it with metal chips. The worst consequence it will
suffer is to replace an internal socketed 20-cent fuse.

The "V" in G203V stands for "Vampire" as in cannot be killed. It
incorporates everything we have learned in the last 6 years about how
good drives wind up dead.

New features have been added and many annoyances have been removed.
The G203V does'nt need external capacitors, the ADJUST tripot is now
top-adjustable, the motor stays very cool when stopped, there are no
jumper settings at all, The DISABLE input is optoisolated, the STEP
and DIRECTION inputs are common-ground, it works with 3.3V logic (no
breakout board needed), no "funny" G2xx timing restrictions, power LED
and error LED indicators and more.

All the good things have been kept. 33A at 100VDC MOSFETs and 100VDC
tested power section, same package size (G201 package) and the same
terminal pin-out, the same mounting-hole pattern.

What this means to you if you are a reseller is drive-related service
calls, support and drive replacement should drop to nil. The worst
case customer service response will be "sort out the problem and
(maybe) replace the fuse". The G203V is targeted at never having to
replacing the drive.

Caution: The G203V is an entirely new clean-sheet design. Circuitwise
the G203V shares absolutely nothing with its predecessors. As such it
has absolutely no history. The older G2xx drives have a 20-year design
history where every vice has been bred out of them. They simply have
no uncovered bad behaviors or surprises left.

The G203V is completely new. I will not feel comfortable with this new
design until there are at least 6 months and 5,000 user's experience
under our belts.

Pricing is $147 singles, $114 at 1K quantity. See the files section
for the preliminary user's manual.

Mariss

Dan Mauch
Fri 17 November 2006, 07:13
I have been testing some of the first batch of the G203. They are awesome. There is only one problem with them. That is he uses a new disable function. I use this for manually position my machine. The issue that Mariss said he will address is that once you disable the outputs it take about 1-2 seconds for this to occur. At the speeds you use for routers this can be quite a problem. The other issue is that if a motor does cause a reset then only that axis is affected. I have suggested that he change the design so that if one driver faults they all fault.

Gerald_D
Fri 17 November 2006, 08:11
Welcome to this forum Dan. So, my opening line at the top of this thread "The G202 is the right one for now" still holds true http://www.mechmate.com/Forum/clipart/happy.gif

I don't envisage a router user wanting to disable the drives (to make motors freewheel) while anything is moving - so the fact that a router may move faster than other CNC machines doesn't seem to apply.

Dirk Hazeleger
Fri 17 November 2006, 17:38
Hey Dan
What do you see as the performance difference between a 203 and a 202. It sounds like you can get higher rpm out of the steppers. Is this true?
Any idea as to why?
Dirk

Gerald_D
Tue 05 December 2006, 01:23
Sales of the G203V have been suspended for a while because there are bugs to iron out. The next update is due out on 22 December.

The tried and tested option is still the G202. It is suggested that you only consider the G203V if you are near to the GeckoDrive offices and that returns for updating are viable.

Gerald_D
Sun 14 January 2007, 23:00
Something mentioned by Gecko on their forum last night:

"........all our step motor drives morph from a
sine/cosine reference to a full-step (quadrature) reference at higher
speeds.

Microstepping has a huge benefit below 2 revs/sec. It makes the motor
move smoothly without resonances and other bad bad motor behavior.
Microstepping loses all benefit above that speed and becomes a net
deficit above 4 revs/sec. It's an "area under the curve" thing; the
sine function has pi/4 (78%) the area a square-wave has. That means
only 78% of the potential torque a full-step drive has.

Secondly, persisting with a sine reference generates aliasing
frequencies between the drive's 20kHz switching rate and the step pule
frequency that pump the motor into parametric resonance (mid-band
resonance).

So what do we do? We morph the reference from sine/cosine to a
full-step quadrature square-wave reference beginning at 4 and
finishing at 6 revs/sec. The result is the motor delivers every last
Watt of its potential performance to the load.

The motor 'thinks' it's being driven by a full-step drive above 4
revs/sec but it still takes 10 step pulses from your controller to
generate a full-step (1.8-degree motion) from the motor.

The anti-resonance (mid-band or parametric resonance)damping circuit
insures the the motor will not waste its available torque resonating
above that speed (4 revs / sec). It effectively adds about 80 degrees
of phase lead to the system loop.

End result? Your motor delivers all its potential to the load. Nothing
is wasted; stratigies that add performance at low speeds
(microstepping) get switched-out when they become a hurt instead of a
help. Other circuits prevent the motor from doing wastefull things
(mid-band compensation). Your motor delivers all it has to give.

Other circuits manage motor heating when it is stopped and resting.

More than any other motor type, step motor performance depends on the
drive running it. A crap drive gives crap results on a step motor. The
same motor driven with a good drive tames all the motor's vices and
makes it sing. It is a motor on steroids when driven with a good drive."

Gerald_D
Mon 05 March 2007, 03:44
Revision 5 of the G203V has just been launched. Apparently the teething problems are ironed out now.

From the Geckodrive company, one could use either the G201, G202 or G203V drives. With the door of the control box closed, I don't think that anyone will be able to tell which drives are in the box when the MechMate is cutting away.....

The G201 needs an external capacitor and a fuse. $114
The G202 needs an external fuse only, but it is wider. $134
The G203 needs nil external components and is back to the narrow footprint. $147 (needs a different supply to the Common pin)

The internal fuse of the G203V is unknown out here. You might also have difficulties in locating an exact replacement. Explore the option of buying spare fuses with your G203.

James Webster
Mon 05 March 2007, 10:36
The internal fuse of the G203V is a 5A axial fuse (it looks like a green resistor, not glass). It is made by LittelFuse part #0251005.MXL . $.43 USD each:

http://www.mouser.com/search/ProductDetail.aspx?R=0251005.MXLvirtualkey57610000 v irtualkey576-0251005.MXL (http://www.mouser.com/search/ProductDetail.aspx?R=0251005.MXLvirtualkey57610000 virtualkey576-0251005.MXL)

Gerald_D
Mon 05 March 2007, 10:50
Mariss on his Yahoo forum this morning said he could supply a spare fuse with a G203V at cost ($.22).

I can find similar looking fuses here in S.Africa, but not the Littelfuse version. Don't know if these will fit the socket when crunch time comes.

Gerald D
Wed 13 June 2007, 02:28
Rev 5 of G203 has a third yellow LED added. From the manual, under troubleshooting:

MY YELLOW LED NEVER LIGHTS: You are not going fast enough to get full power from your motor. If you don’t need to go any faster, use a lower power supply voltage. This indicator is a good application diagnostic for motor gearing and power supply voltage choice. Using it correctly will help you to optimize your system.


. . . . . . interesting...

Gerald D
Sat 08 September 2007, 00:44
There are reports that the G203V and the Oriental Motor PK296A1A-SG3.6 might be a problematic combination. (Roughness, resonance). Note that G202 has always been the recommended drive for the MechMate. Apparently the G202 is fine with the PK296A1A-SG3.6 motor.

If you do have PK296A1A-SG3.6 motors on G203V's then you should be going for the half-coil wiring - series winding is what causes the roughness (too much inductance).

Doug_Ford
Sat 08 September 2007, 07:50
How about the 7.2 motor? Have you heard if it works okay with the G203V?

Richards
Sat 08 September 2007, 08:43
The G202 has a jumper setting that allows the use of a low current stepper motor. Although I wouldn't normally put the PK296A1A-SGxxx motor in that class, it does allow you to use that particular series of motors using full-coil connection (series).

The G203v does not have that jumper. However, if you wire the PK296A1A-SGxxx motor using half-coil connections, the motor has 7mH of inductance rather than 31mH of inductance.

In my extensive testing with the PK296B2A-SG3.6 motors, which have 6mH inductance full-coil and 1.5 mH inductance half-coil, they work perfectly with the G203v when they are wired either half-coil or full-coil.

If you have not yet purchased motors and Gecko stepper drivers, I would recommend that you try the PK296A2A-SG3.6 or the PK296A2A-SG7.2 motor with the Gecko G203v stepper driver and a power supply using a toroidal transformer with output windings in the 18 to 25VAC range. If you wire the motors half-coil, the 25VDC to 35VDC from that size toroidal transformer (after rectification and capacitor filtering) works perfectly with the G203 and those motors.

The choice of whether to buy the 3.6:1 model or the 7.2:1 model depends on how you're going to use your machine.. With 3.6:1 gearing and a spur gear with a 1.5 inch pitch diameter (30 tooth when used with the 20-pitch rack), you will move 0.000654 inches per step. With 7.2:1 gearing you will move half that far, or .000327 inches per step. (Don't expect to be able to cut at either of those resolutions. Due to all of the factors that work with and against each other, a cut resolution on a CNC router in the 0.001 to 0.003 inch range is considered excellent. Remember that a piece of copy paper is about 0.003 inches thick. Expecting a CNC router to give 5X to 10X better resolution than that is just wishful thinking.) The 3.6:1 motor will jog 2X faster than the 7.2:1 motor, but the 7.2:1 motor will be working closer to it's optimum speed (yellow light flickering on/off with the G203v stepper driver). EITHER motor will work fine. Don't waste a lot of time worrying about which motor is the best. Depending on a lot of unknown factors, either motor may work better than the other on any particular day but either motor will do the job.

If you have already purchased the PK296A1A-SGxxx motors and the G203v stepper drivers, do as Gerald suggested and wire the motors half-coil. Use a 50VDC to 70VDC power supply. That motor is rated 3.3V, so those voltages give you between 15X and 21X rated voltage. That range works very well with other motors that I've tested. That combination should work fine. However, please report back to the forum so that we can all learn from each other.

Doug_Ford
Sat 08 September 2007, 08:48
Thanks Mike. I've already purchased four G203Vs but not the motors. I'll stick with my plan.

Richards
Sat 08 September 2007, 10:26
The unique feature of trying to size a motor to a CNC machine is the fact that (usually) the feed rate and distance of each motor's move is unpredictable. Take, for instance, the need to cut a taper along one side of a 100-inch sheet. Assuming that both X-axis motors have to move 100-inches and that the Y-axis motor only has to move 1-inch during that move, it becomes very clear that a very small Y-axis motor (with a high gear ratio) could be used but that we would want the largest and fastest X-axis motors that we could use. On the other hand, if the cuts were reversed, and the Y-axis motor made the long move and the X-axis motors hardly moved at all, we would want a large and fast Y-axis motor and small X-axis motors. That's the main reason that I don't worry too much about getting a perfect motor and a perfect gear ratio. With the kinds of cuts that I make, finding something that would always be ideal would be impossible. So, I try to find a good compromise - a motor that can move its axis at a comfortable speed and still have little or no resonance at slow speeds.

On the other hand, when I'm wearing my process-control hat, I can often find a motor and a gear ratio that works perfectly for the process. Right now, I'm designing a 6-station turn-table that requires the table to stop every 60-degrees so that a pneumatic action can happen when the table stops. Since the motor will always be moving an exact number of steps, and since the slowest part of the process that I'm controlling is the time that it takes for the turn-table to rotate from station to station, I need to use the largest practical motor to spin that turn-table. The Oriental Motor PK299-F4.5 motor driven by a Gecko G203v stepper driver at about 35VDC is the best fit that I've found so far. With this process an occasional missed step is not a problem because the table re-homes each time it passes the #1 position.

Several years ago I had to design a paper advance system to be used in Kodak S-printers. The limiting factor in that process was the back-printer that printed a copyright on the back of each photo as the paper advanced. A little PK268 motor with a 5:1 gear ratio was an ideal match. With that gear ratio, the motor never lost a step and the available speed matched the back-printer perfectly.

Gerald D
Sat 08 September 2007, 12:47
I have just been reliably informed that the G202 can handle 1mH to 40mH inductance range while the G203V can handle 0.5mH to 20mH.

gmessler
Thu 18 October 2007, 09:14
Can anyone help?
There is a picture that shows up at the top once in a while that shows a 4 fuse power distribution block. I can't seem to find one anywhere. Does anybody have any suggestions?

Thanks

smreish
Thu 18 October 2007, 09:34
Greg,
If your referring to the block that Gerald uses to fuse his power supply to the Gecko's, then I have an answer for you!
Here in the US, I have found the only easy match, yet it's a 6 blade block.
- Advance AutoParts, ATV (Bussman part #15600 - 8 dollars
- Newark electronics charges 28 dollars for the same unit.

Good luck.
Sean

Gerald D
Thu 18 October 2007, 10:11
If you are using Gecko 203's you can drop the external fuses - they have internal fuses. (Fuse type is not available in Cape Town - one of the reasons I am staying with the G202)

domino11
Sat 20 October 2007, 19:12
Gerald,
What kind of fuses do the 203s use?

Gerald D
Sat 20 October 2007, 21:57
I don't remember now, but it is mentioned in the 203 manual that no substitutes will be accepted. It is a tiny, pluggable thing made by Littelfuse. Geckodrive will sell fuses with the drives, but how many should one buy?

driller
Sat 20 October 2007, 22:17
LittleFuse 0251005.MXL

digi-Key and Mouser sells them, 10 for less than $5.50, so 55 cents each. I would buy a pack of 10.

Dave

Gerald D
Sun 21 October 2007, 00:28
Another reason I am staying with the G202 is because we have 2x control boxes running 8x G201/2's off 2x PMDX-122 BOBs. (Have another 4x G202's on the shelf - Sean has 1 more). Should a G202 need replacing, the G203 cannot slot straight in because of the different voltage on the common pin. Most BOBs allow you to jumper either for 0V common (G203) or 5V common (G201/2) but this jumper is for ALL the drives connected to the board. Drive types cannot be mixed. So, we are locked into the G202 for the forseeable future.

However, if I was starting from scratch today, I would probably go for the G203 and a packet of fuses. But then again, the extra cost doesn't bring any more performance . . . . . . Snag is, the G201/2 may be declared obsolete one of these days - already Mariss says he is battling to find the older style components needed to produce it. I have a feeling I have bought my last G201/2's . . . . .

Any suggestions on how to mix drive types on a PMDX-122?

driller
Sun 21 October 2007, 07:30
mixing drives would be pretty simple if you have an interface board.

all you need is a board that takes the signals and isolates them. the isolated signal could be either 0 ref or 5v ref on each channel, simply by a jumper.

But, I prefer to run the Geckos directly from the parallel port and not get involved with a BoB that relays the signal. all BoB's that take the parallel port signal and repeat it are adding some time delay.

and since the Gecko's are optically isolated in the first place, there is no real benefit to the BoB repeating the signal.

If you ask Mariss, he will lay out a great circuit that will work with his unit and any BoB. He does this often and without charge and puts it in the open source. - public domain.

Dave

Richards
Sun 21 October 2007, 09:56
Gerald,
At least with a PMDX-122, it looks like you can interchange G201/G202/G203v very easily. I just plugged the headers from a G202 into a G203v drive. Other than disconnecting the COMMON (Term 10) from its PMDX-122 J1 COM connection and then connecting the COMMON (Term 10) to the J5 GROUND pin on the PMDX-122 nothing else was modified.

My test setup consisted of three G202 stepper drivers, one G203v stepper driver, a PMDX-122 break out board, Oriental Motors PK268 sized motors, 35V power supply, Mach 3 software. The Mach 3 test file that I ran seemed to run the same with mixed Gecko stepper drives as it had when all the stepper drivers were G202s.

EDITED: Because the signals for the G202 are inverted when compared to the G203v signals, you may lose or gain one step pulse when the stepper drivers are first turned on. That is not a problem to me, but others may need to adjust their cut files.

Gerald D
Sun 21 October 2007, 10:13
Thanks for that test Mike! A single pulse (1/2000th motor rev) is not going to bother us.

Hugo Carradini
Mon 22 October 2007, 11:06
Hello Mike.
I was reading your post and need to understand something. ¿Do you change the position in JP1 instead of "+5volts pins 2-9 " to "pins 2-9 GND " for the G302V? I ask this because I did like the PMDX-122 shows in the first diagram for setting the Card to the Geckos. Now reading the Geckos instructions I read they set to DIR; STEP and GND:confused:
¿What could happened if they run with +5v in the COMON instead of GND?:(

Richards
Mon 22 October 2007, 12:01
Hugo,
The instructions to Gerald on Oct. 21, 2007 were for mixing a G203v device with G202 devices when the PMDX-122 was set up for the G202 devices.

If you're using G202 devices only, set the JP1 jumper to +5V. If you're using G203v devices only, set the JP1 jumper to GND. Connect the G20x to the PMDX-122 using the J1 through J4 terminals to connect individual Step/Direction/Common signals.

I have not tested the G203v while using +5V as COMMON. It should not work. That configuration would mean that you would need to SINK current through the opto-couplers on the G203v. The G203v is designed for interfaces that SOURCE current through the opto-couplers. To SOURCE current through a device, you need to use a COMMON GROUND.

Gerald D
Mon 22 October 2007, 12:04
Mike, while it won't work, it shouldn't cause any damage? I think that Hugo is worried that he may have cooked something.

Gerald D
Mon 22 October 2007, 13:13
I asked Steve Stallings of PMDX about mixing drives on the PMDX-122. Here is his reply:

Because the PMDX-122 does not have opto-isolation (to the drives)
and thus separate grounds, you can use any GND terminal
or any +5 AUX terminal to feed the common of a Gecko
drive connected to the PMDX-122. The COM terminals
on J1 - J4 and the jumper JP1 are just a convenience,
there is nothing magic about the COM terminal on these
connectors.

Richards
Mon 22 October 2007, 13:33
Gerald, Hugo,
There should not be any damage to the circuit. What happens when the circuit is working properly is that there are a few milliamps of current going through an LED. In order to work, the Anode end of the LED is connected to a higher voltage and the Cathode end of the LED is connected to a lower voltage. In the case of the G202, the Anode end of the LED is connected to +5 volts and the cathode end of the diode is connected to a signal line. When the signal line goes LOW (near ground), current passes through the LED and the LED lights up. In the case of the G203v, the Cathode end of the LED is connected to GROUND and the Anode is connected to the signal line. When the signal line goes HIGH, current flows through the LED and the LED lights up. In either case, the amount of current going through the LED is limited to a safe level by a resistor. The JP1 jumper on the PMDX-122 determines whether the Anode is connected to +5V or the Cathode is connected to GROUND.

One of the functions of a diode (like the diodes in a bridge rectifier that we use to convert AC to DC) is to block current. So, in most cases, when an LED is connected backwards, no current can flow through the device and the LED will never light up. The exception would be if we connected an LED to higher voltage than it was designed to handle. In that case, the LED could be destroyed; however, a Break Out Board only uses 5VDC on its signal lines, so no damage will occur.

Because I have both G202 and G203v stepper drivers, I have often forgotten to move the JP1 jumper when I've connected different devices to the PMDX-122 board. In every case, as soon as I moved the jumper to the correct position, everything worked properly.

Hugo Carradini
Tue 23 October 2007, 09:00
Thanks Gerald for your interest in helping .:)
Mike, is a pleasure reading your comments. I am really learning more of my electronics (the heart of the baby beast)then if I had not go through these problem so I am glad with the situation. I heard of so many cases that is just a simple thing what is wrong that I am sure this is one of those situations. These baby beast show me how strong and precise she is so I have to learn more, how to take complete care of her.:)

Pilar
Thu 03 January 2008, 14:22
What is the best drive to start a machine G203V or g202 or others?
I pretend to use G-Rex G100 / drive g202 / and motor Phase Stepping Motor
- PK296A1A-SG7.2 2-Phase Stepping Motor . Any sugestion, idea etc? sorry of my English.
Thanks,
Marcos Pilar
Brazil

Richards
Thu 03 January 2008, 15:36
Marcos,
The stepper motor determines the stepper drive that is best for you. If you already have PK296A1A-SG7.2 stepper motors, then the "best" stepper driver for that motor would be the G202. That is because that motor has high inductance and the G202 is better suited for that motor. If you have not already bought your motors, I would suggest the PK296A2A-SG7.2 motors, the G203v stepper drives and the PMDX-122 breakout board.

The PMDX-122 is a very good breakout board if you're using a parallel port. There are many other good breakout boards that others prefer. I've only used the PMDX in my testing. The G100 is being discontinued. I have a G100 and also a G101/G102. They are excellent devices, but Art and Steve (at Artisoft) have had serious difficulties making the G100 work as well as they had hoped. However, they are working with other devices that should give us most of the good features of the G100 without the expense or the difficulty.

The power supply that you use will depend on the motors that you select. The PK296A1A-SG7.2 wired either bipolar series or half-coil can use a 70VDC power supply. The PK296A2A-SG7.2 motor would use a 30V to 35V power supply if you wire it half-coil or a 50V to 70V power supply if you wire it bipolar series. (I prefer half-coil, but either will work.)