View Full Version : bbs stepper drive design

robert phillips
06-06-2003, 03:28 PM
seems there are more than a few people who hangout here that have hardware/software design experience. with the apparent drive board problems and seemingly high cost of commercial boards, we may want to consider tapping our experience and creating our own "public domain" driver boards. also, some recommended procedures for quickly characterizing your motor drives on your equipment to determine maximum slew rate, calibration routines, etc. we seem to have a pretty sharp crew here.


06-06-2003, 03:43 PM
Here is a lot of good info on the subject.


and another one:


And this is interesting:

"A program that reads a HPGL drawing and controls steppermotors connected to parallel pc port
For stepper motor driven XY(Z) engraving / milling / router tables
For Hobby use"


[This message has been edited by Evan (edited 06-06-2003).]

[This message has been edited by Evan (edited 06-06-2003).]

robert phillips
06-06-2003, 04:48 PM

i was suggesting the bbs pool it's talents and spec a board that competes/beats commercial products hopefully at a fraction (25~33%?) of the cost. Like some others here, i have had design experience in positioning systems, hardware and software and metrology. i've been doing strictly programming for quite some time now, but the hardware end can be refurbished as needs be. someone (you?) used a 6502, i used a 6802 on my first stepper controller project way back when it was a new processor, with discrete components and SSI for a bipolar drive design. the assy program generated the quad-phase drive signals through a PIA with a few xor gates. current sensing and positional and angular feedback. very low L/R and large cemf (winding field collapse) capability for max response out of the motors.

i think we could give some of these companies who believe they have a captive audience a real run for their money, at least within the local circle here....


John Stevenson
06-06-2003, 04:51 PM
Some more stuff here:-


John S.

06-06-2003, 05:13 PM
Yeah, I posted something about a plotter I built a long time ago. I put up the links just to get people thinking. I don't see why it should be a big problem to make a reliable stepper driver, especially using VMOS power fets. You can parallel them as needed and they don't thermal runaway. The real secret is cooling. BIG heat sinks, active cooling with fans. There are lots of nifty step/direction controllers available. I just bought three as I am putting together a stepper controller in the next few weeks for my telescope. That will be a low power, slow rate, constant speed unit but the only real difference is the output drivers. I won't need any for this application since the chips I have will sink up to 1.5 amps. The manufacturer spec sheets often have example circuts using the part and it shouldn't be very difficult to come up with a PCB design. I'm afraid I don't have the time for the next few months to get involved with such a project. I have to finish up my telescope project in time for the Mt. Kobau Star Party at the beginning of August and then the grandkids will be visiting during August.

[This message has been edited by Evan (edited 06-06-2003).]

[This message has been edited by Evan (edited 06-06-2003).]

robert phillips
06-07-2003, 03:22 AM

you must be geared way down and/or be belt coupled to integrate the impulses without jitter in the field of view. a bellows coupling would smooth this out too. i assume the steppers are used for tracking too. is this smooth enough for long term photographic exposure? sounds like a satisfying project. John S, i see the stepper driver used a few parts like the uln2003 array for switches same as i used way back when the 8085 first came out. it's ubiquitous and long lived, just like the neon bulb.


06-07-2003, 11:15 AM
Spur gear and then a worm drive. The controller chip I am using (BIMOS UNIPOLAR 5804) has a microstepping mode. The 'scope has a best resolution of .7 arc seconds (calculate how far the earth turns in a second). To make the steps un-noticable I need at least three steps per second (not difficult with the gear train). The chip with the microstepping mode will give me 24 steps per second, plenty good enough. The stepper is used for making the telescope track the apparent sky motion.

06-07-2003, 01:43 PM

I have some (5 or 6) 3 phase motors and gearboxes. I think the gearboxes are about 4" x 3.. That with a real servo is what you need. I only have one tiny servo left. Smooth.. they were fabric feed for a automated sewing machine.. If you had 3 phase, well you could just wire the motor on it. I know all your equip must run by battery thou. Problem with gear box is that the spur gear is made onto the shaft on the motor, you'd have to turn it off and drill it to match a small motor.


As you can see, a working drive over a year old has the proper amount of heat sink grease and was applied properly to the heat sink. This connection is still determined by the pin springing this transistor to the heat sink. Now I have disturbed it, I plan on opening the other one, physically bolting a aluminum strip across the transistors and mounting them in front of a fan.
If it will work a hour a week for a year, but not 8 hours a day, well it is a heat dissipation problem. I have had the geckos bolted to the door heat sink of the bridgeport which is 3 feet by 2 feet and finned. Tranfererrance to the outer heat sink is the problem thou.

[This message has been edited by ibewgypsie (edited 06-07-2003).]

robert phillips
06-07-2003, 04:09 PM

the tabs are isolated or at the same electrical point yes?


robert phillips
06-07-2003, 04:11 PM

the tabs are isolated or at the same electrical point yes? a t/c would give you a quick way to check tab temperature and you'd know for sure then about the heat.


06-07-2003, 06:24 PM

My meter says they are all on the same supply. They are the same mosfet, a IRF530. The Drain is the tab on a N channel Mosfet. Lots of small parts on this 201 drive. I got a letter into Mariss at Gecko asking him.

I will do one like that and play with it.
I have a zener regulated supply in the works too. It uses a 6.25 volt zener turning on power transistor across a voltage divider for base voltage.
Mariss just emailed me back... According to him all the tabs are at different potentials. If I want excellent thermal contact he recommends shimming the mosfet case against the circuit board so tightening it will put transistor pressure on the case. The inside of the case is non-conductive, but thermal conduitive.

[This message has been edited by ibewgypsie (edited 06-07-2003).]

06-07-2003, 11:24 PM
http://www.epanorama.net/links/motorcontrol.html#stepper]http://www.epanorama.net/links/motorcontrol.html#stepper]http ://www.epanorama.net/links/motorcontrol.html#stepper (http://www.epanorama.net/links/motorcontrol.html#stepper)

I have a sketch of what I want, A modular one with plug in boards, except for the power transistors, they have to be remote and fan cooled.

[This message has been edited by ibewgypsie (edited 06-08-2003).]

robert phillips
06-09-2003, 10:59 AM

we use to use SIL-pads instead of grease and a nylon spacer (ferrule) in the hole in the heatsink to isolate the screw from the heatsink. nylon washer between the nut and the heatsink. i think it was #6 hardware. no electrical reason you can't do this.


ps: how's new graphical setup?

06-09-2003, 11:56 AM
I dont like sil pads. The are not nearly as good as a well fitted and properly greased heatsink/device setup. The only reason sil pads exist is to speed up assembly and reduce grease cleanup problems. When I build a computer that is likely to see hard use, say for instance it will be overclocked, I always scrape off the stupid little pad on the heatsink, lap the heat sink on some fine paper on a piece of glass and use grease.

Ibe: A servo motor could be used to drive the telescope but is not necessary. The stepper has way more than enough resolution to avoid any jitter being noticable. In this particular application the accuracy of the tracking is not an issue since this telescope will not be used for long duration time exposures. This instrument is meant for primarily visual observing and webcam photography with exposure times no longer than 1/4 second. The truly cool thing that has come about with the use of webcams is taking 20, 30 or more images and then using software such as the really excellent program "Astrostack" and automatically combining all the images together. Astrostack will align the images even if they are displaced a lot and the combination of 20 or more images greatly increases the amount of information used to make the final image and so results in a much higher resolution image as the result.

robert phillips
06-09-2003, 12:30 PM

yes but SIL pads are also the insulator for the device. i've used both methods many times and each has its application. multiple packages on a shared heat sink IS easier to assemble with SIL pads and has been adequate for all applications i've ever used them in. (including commercial stepper and dc servo drives..). higher power linear regulators, triacs, scr's etc were usually greased but they also had individual isolated heat sinks.

i wish i had the $$$ and especially the time (5 kids...) to do the telescope thing. i used to make a lot of use my father's homemade 6" reflector. he machined all the parts on a lathe in the basement. he also machined all the parts for a homemade 1/2" reel to reel and built all the electronics for it (tube) and an amp and tuner (early 1950's). all i'd see here in cleveland now is city light and clouds anyway...


06-09-2003, 01:07 PM
I wonder? Which is better as an insulator in terms of lesser thermal resistance, a sil pad or mica? I worked for Xerox for 23 years and have worked on many a power supply with fried devices on sil pads. The best practice when space is not an issue is separate isolated heatsinks.

You really should get a scope together. It isn't that expensive any more with the advent of diamond tipped automatic mirror milling. They can now turn out mirrors that only need a couple of hours of hand figuring as a final step. I was recently reading about a milling machine that can make mirrors from solid billets of aluminum with finish accuracy 1/4 wavelength of light. They use these in laser weapons since it is much easier to cool. If you want to scare yourself a bit just do a Google on "metal mirrors".

Here is a photo of my scope, finally at the usable stage. I had it out last Friday and was able to see the ring nebula, not bad for a six inch scope.


robert phillips
06-09-2003, 02:00 PM

that scope's a beauty! you must have been working on it for a while with all those machined parts. about the mica thing, i was forever cracking and chipping those little buggers. i don't know which had better thermal characteristics. i think the sil was better than the mica insulator alone but then the grease sandwich was better than the sil for really high heat applications. it's just simplicity vs mess. i still have sil pads laying around in a 20 yr old flood marked box somewhere in the basement. all my 5% R's went south probably 15yrs ago....
eternal remodeling in the house and a strong willed irish wife sort of put hobbies on hold. i've had a new saw table since xmas, still in the box, unassembled in the garage. i can open it and put it together when i start the addition.....


06-09-2003, 03:26 PM
Just a note about heat-sinks. To get rid on heat from a power transistor, or any other tabbed device like the irf530, it's best to hold the device against a copper(best), or aluminum piece directly, then isolate that piece from the heatsink proper with an insulator. Grease being used to fill all gaps as usual. Any insulator will reduce the rate at which heat can escape from the device, so if the device is going to dissipate lots of heat in the application, it's best to start by spreading the heat from the device into the first metal without an insulator. This first piece need only be a sq in, or so, and as that is about five times larger than the tab of the device, the method works to keep the actual die temperature down, while allowing the device to dissipate more power. As the die is centered within the plastic block of the device, that is where the heat is created, and where it should be removed from. In that light, it's best to put mounting pressure against the case of the device. This gives better performance than using the hole in the tab to mount the device. As in some car amplifiers, a spring clip can be used, or an aluminum bar can be bolted across multiple devices, taking care that all parts are clamped evenly. The bar will also remove some heat from the device, not much, but if the devices are used near the maximum ratings, every improvement in heat conduction helps.
One thing I have always been intrigued by, but haven't tried,is the possibility of mounting the devices directly on an anodized surface. This is essentially an insulator, and for low voltages might be totally fine. If anyone has any experience with this, I'd be happy to hear the feedback.

[This message has been edited by darryl (edited 06-09-2003).]

06-09-2003, 03:30 PM
By the way, the isolated tab devices don't dissipate heat quite as well as non-isolated tabs. Heatsink grease is still required, more so with an isolated device. I've seen many products where no grease was applied at all. Hmmm maybe that's why they came in for servicing.
The last point I'd like to make, while it's still in mind, is that, for higher current levels, it's very important to make low resistance connections to the device leads, so there's no excess heat developed at those points. A poor connection here will get out of control, and can 'blow' the device by overheating the leads. Make good-sized contact areas on the pcb, if you have the option to do so, or enhance the current-carrying capacity of the traces with some small diameter copper wire.

[This message has been edited by darryl (edited 06-09-2003).]

06-09-2003, 03:58 PM

I am guessing I have spent about 300 or 400 hours working on the scope. At last count it has about 120 drilled and hand tapped holes for fasteners. The design is unique and is not based on any thing else that I have seen.


No way would I trust a regular anodic coating to insulate. Aluminum oxide is an excellent insulator but a standard anodised finish is about .3 thou thick. It might be a different story with "hard" anodising as that can be built up to 75 microns (about .003"). But then I still wouldn't trust it.

[This message has been edited by Evan (edited 06-09-2003).]

robert phillips
06-09-2003, 05:25 PM

the two stage method works well with adequate room. i'm talking about 4 to 8 devices tied to the same sink due to physical constraints. unless the collectors or drains were tied to gnd or vcc/vdd, they would have to be kept electrically isolated so an insulator is necessary either under the body or under the first stage dissipator. black surface works better than shiny surface for radiation too if i recall correctly. forced air a big plus.

i think there is too much risk of scratching / breaking through the coating in any direct mount attempt without an insulator. why risk it. besides, the grease doesn't fill in the 'bumps' in the coating well enough. that's why the device outline is usually masked off from the coating process and left smooth in industrial power/drive units i've seen.


06-09-2003, 07:36 PM

I put silicone under the mosfet, let it almost harden then put on the heat sink. It FIRMLY adheres to the heat sink, looking from all directions it they are flat against it, silicone on top next to circuit board, heat sink grease properly applied to metal tab. THAT will work I hope. I then mounted them on a aluminum plate and put a fan over them. I also left the neat gecko-drive cover off for more airflow.

Mariss (of gecko) suggested putting business cards between the circuit board and the mosfet, I used silicone. (firestarter?)

Now I am making, (so far smoked) a regulated supply for them incorporating a regen dump. (the part that smoked)

OUT of frustration I have created two PCB drawings on pcbexpress software to make a 4 axis stepper with individual cards for decode, and chopper PWM. My plan is to make them plug into a bus, except for the mosfets, they will be housed seperate for heat.
I been really agitated nothing is available that pumps 7-12 amps at 68 vdc to stepper motors and is reliable as is.
I will email the circuit boards to any that want to look at them, check my work. Right now they are in the design mini-board stage @ $62 for 3 delivered boards. NO edge connectors yet, they are free wire.. Lots of places to make mistakes, lots of things probably missing.

[This message has been edited by ibewgypsie (edited 06-09-2003).]

06-10-2003, 11:20 AM

I would like to have a look at your pcb design. Please mail to:

"v i k i n g at v t s.b c.c a"

Just lose the spaces and quotes.

robert phillips
06-10-2003, 11:44 AM

i'd like to look at the schematics and pcb layout too. mainly the schematics to check out power capability....once upon a time did circuit design, board layout and silkscreen, and schem. dwgs for a job, on a board because cad wasn't really around yet.


06-10-2003, 05:59 PM
Well... jump on the wagon...

I have the ExpressPCB software. You will need to download it to view it.
AS I said they are on the mini-boards. NOT a productiond design yet but a test bed. I need all the help I can get.

06-10-2003, 06:14 PM
I just installed ExpressPCB and am playing with it. I don't do much in the way of PCB design these days so I won't have much use for it but it looks pretty cool. I sure wish I had it back when I was doing serious computer PCB design. I did all that on a drafting table and laid out prototypes with tape and Letraset. This would have made life a lot easier!

06-10-2003, 07:36 PM

Had a look at your boards. If you are doing the mini board it doesn't include silk screening. I would strongly suggest that you orient all ICs with pin one in the same direction and then indicate with copper text where pin one is. Much less chance of a goof. Another good practice is to surround the board edge with a ground trace although space is a bit of a problem. More important, you shouldn't leave gate inputs floating. They should be grounded. It is all right to leave outputs open. The golden rule for logic devices is "All input states should be well defined, either LOW or HIGH".

You have a lot of unused gates on the driver PCB. With some extra work you should be able to cut the chip count and gain some space.

06-10-2003, 08:06 PM
Well.. Nice ideal.. I am so stunned I didn't see it. Using a board for each axis drive, then one for each heat sinks and opto isolators, shunts.

By using a driver-pwm board for each axis you could swap them to try them out. yep.. That might fit on a mini board too..

Minimum order, 3 boards, Min needed, 3 axis.. sounds like a plan to me. I will cut and paste the boards together tonight.

I got the decoders, the PWM and a place for a single channel op amp and room for terminal strips.. That mini board looks empty now. Geeze, thanks for suggestions. Room for ? the needed isolators maybe, I got to go to the catalog and the data sheets. 4 channel fired by the decode, PWM via the driver to a common circuit to transistors?

[This message has been edited by ibewgypsie (edited 06-10-2003).]

06-11-2003, 06:47 AM
I need a circuit that will pulse single shot each time the step moves. I want to throw a higher current for a adjustable millesecond each move then idle back. I have extra op-amps and two logic gates left.

Also, I need some info on FET drivers. I want a bipolar drive I think.

06-11-2003, 11:50 AM

When you say Bipolar I presume you refer to the transistor configuration, not the transistor type? In this case, I assume you are using Bipolar steppers, not Unipolar steppers. Assuming you are using a single supply voltage then the usual driver circuit is an H bridge.

This is an example of an H Bridge with FETs. The part marked motor would be one of the windings of the stepper and the device marked TZ is a tranzorb, a special type of avalanche diode that soaks up the back emf of the winding. The FETs are energized in opposite upper/lower pairs and it is imperative that your logic circuit does not energise two FETS on the same side at the same time or you have a short and will fry something. You need an H bridge circuit for each motor winding.


For best reliability there should also be a Schottky diode placed across the source/drain of each FET to provide a path for back emf past the FET (obviously polarized against the supply). For gate protection against static, especially if these devices will be on a separate removable board, I would place a resistor on each gate input of perhaps 100K to ground.

I'm still thinking about the current booster.


[This message has been edited by Evan (edited 06-11-2003).]

06-11-2003, 12:12 PM
thanks evan..
I have never saw some of the components on the drawing. They don't come in 500,000 volt switchyards. ha ha.. I will research.
I just found a single shot timer, made with simple 555 timer and two capacitors, three resistors. I got it on the board, now have to figure out how to fire it with the step. It requires a negative pulse, I have to figure port state at switching. 7414's are in the port inverter. Possibly? put a meter in it and check state?
A matched pair, 3904-390? to switch potimeters bases balanced to ground? fire it runs on high curr, off it idles at lower?
Bipolar, yeah.. should add some complexity to the circuitry on the switching board. But I think a desired evil.

Current booster? how about a darlington array opto-isolator?

[This message has been edited by ibewgypsie (edited 06-11-2003).]

06-11-2003, 01:05 PM
I suspect you are talking about the tranzorb when saying you haven't heard of it. It is a special type of zener diode, actually a pair of them in a package. It is bidirectional. Known as an avalanche diode, it is the cadillac of transient suppression. It has a response time of about 4 nanoseconds and as long as its maximum dissipation ratings are not exceeded it does not degrade with use. The breakdown voltage is very precisely defined. If it is rated to switch at 67 volts then it does, completely. It has virtually no difference between the on point and the off point and can handle large surges.

I don't think you need the added complexity of boosting the current at the start of each step. The capacitor across the H bridge will serve that function. Put your drivers as close as is practicle to the motors and use the heaviest wire possible to wire them.

To check the validity of your logic design you need to draw a state diagram for the design. Do a google on "Karnaugh map" and state diagram.

What I would use the 555 for is as a missing pulse detector. After five seconds with no pulses to the stepper circuit use it to turn off a hefty FET with big sink that throws a resistor into the power supply lead to the motors and reduces the current by half. Cut down on driver heating when things are not moving. Combine the phase outputs with diodes (or an or gate) and use them to trigger the threshold pin on the 555 in monostable mode. When the pulses stop the 555 will time out and the output will go low. Use that to switch the FET. Put the resistor across the FET, a good wirewound will do nicely and there will be no interuption of current. The moment the steps start the FET will conduct giving full power. Since at all times a good portion of the current will pass through the resistor the FET doesn't have to handle the full supply load.

[This message has been edited by Evan (edited 06-11-2003).]

06-11-2003, 01:49 PM
Pulses are at variable rate thou, evan. Steppers run at different speeds. Missing pulse detector, isn't that a set freq pulse detector? Perhaps the reset could come earlier than asked for?

What the diagram above shows is a servo by the way, not a stepper. My unit would have 8 mosfets controlled via npn and pnp transistors for the logic. I am looking for fet drivers at the moment. To see what they actually do.

Not sure yet. still playing with the board and other parts, Learning as much as anything else.
There are a lot of other people on this web that are a lot better at this then me, but getting them to design a stepper board when I need one? they would want to be paid.. ha ha.. yeah..

My motors run at 7amps at 68 volts =476 watts of power, x 2 phases = 952 watts of asorbed energy. 746 watts is a horsepower in normal thinking, not sure about a stepper.

Not many drives out there in that market in the price range I can afford.

New gecko just got here....

[This message has been edited by ibewgypsie (edited 06-11-2003).]

06-11-2003, 02:17 PM
The H bridge is not specific to a servo although it can be used to reverse any DC motor. A bipolar stepper requires the polarity of the drive impulse to reverse each step and the H bridge does just that. Just think stepper coil instead of "motor" in the above circuit. You need to arrange the logic outputs so the correct FETs energize in sequence. You can connect top left and bottom right gates together and top right and bottom left gates together. Just make sure that you cannot energise both at once. I assume your NEMA 42 style motors can be configured as Unipolar or Bipolar. When configured as Bipolar they are two phase instead of four phase and effectively have only two coils.

The 555 as a missing pulse detector won't care about the step rate. What I am suggesting is that if no pulses at all are seen for 5 seconds then it reduces current to all the drivers to keep them cooler, motors too. The 555 would be set with a 5 second time out and if ANY pulse arrives in that time then it is set for another 5 seconds by that pulse.

About your power calculation. The two phases are not energised at the same time, only one at a time so the gross power is 476 watts. I'm curious, how have you measured the current draw?

[This message has been edited by Evan (edited 06-11-2003).]

06-11-2003, 02:29 PM
Something else I have thought of. Your description of the failures in the Geckos makes me think they are failing for the exact reason I suggested the 555 missing pulse detector to reduce current. If the steppers are operating the effective duty cycle in a four phase stepper for each phase is 25% assuming wave step mode (single step mode). In a bipolar stepper it is 50%. BUT, if you stop stepping one of the drivers and coils is operating at 100% duty unless you remove or reduce power. 476 watts is rather a lot. From the pictures you have posted I can't see those devices lasting more than a few seconds at that level. Just how does your present controller deal with this? Is it leaving power on when nothing is moving?

06-11-2003, 02:33 PM
The current with the geckos steps back milliseconds after stepping is started and load accellerated. nice drive, this allows the capacitor to pump back up to allow full charge for next step. I have that jumper on. I also have the nema42 hysterisis jumper on. NO clue what that means thou. The nema steppers run about 1/3 the speed of a smaller motor. But the large diametere has more torque so the gearing is upped. Meaning less resolution but same power.
I wish I had started with servos.
All the manuals are online at www.geckodrive.com (http://www.geckodrive.com)
I was wanting to copy the operational features of the gecko into a more rugged design. After it works with my LARGE motors, it will be bulletproof with smaller ones.
The gecko is all surface mount technology thou, hard to decipher. Especially for a instrument tech. (hammer, pipewrench and tubing cutters have no effect or use)

06-11-2003, 03:06 PM
Have you turned on Auto Current Reduction on the Geckos? Jumper pins 2 and 6 on the options block. This will reduce current to 33% of normal 1 second after stepping stops. If you have not done this it would likely explain the failures. Also, have you calculated the current setting resistor correctly?

06-11-2003, 05:03 PM
Yes current reduction is on, No resistor needed, wide open is 7 amps.. WOT, got me a ticket on a motorcycle once..

06-11-2003, 06:41 PM

On your Telescope image, it looks like there is a cable or two running to your diagonal mirror. What’s that? Did you machine your own Focuser.

As to heat sinking, I use, whenever possible, a beryllium pad to insulate the device electrically, but, has very nice thermal specs without grease. Can be costly though.

Someone mentioned Aluminum Oxide for heat sink. The Geckos use a "Heavy Anodize" layer for just that. (Makes me nervous) http://bbs.homeshopmachinist.net//eek.gif

Tom M.

06-11-2003, 06:48 PM
Those are the ends of the string that secures the velvet lined bag that protects the secondary mirror.

Yes, I made a crayford focuser:


[This message has been edited by Evan (edited 06-11-2003).]

06-12-2003, 10:38 AM

Looking over the 297, I need to erase a lot on my mini-board. Pretty good ideal, thank you.. I would have to use it without the 298 of course. Translate the logic to opto-isolators directly, current feedback will need some looking into. Does away with 4 chips on the present design... and half the board is freed up.

Kiss principle, keep it simple (for) stupids..
(I was a engineer once) but I regained my common sense.

Got to go to the shop, I promised, I sold a item on ebay I had not made yet. (shame on me)


06-12-2003, 10:39 AM
With the room gained, how about a encoder to binary output? Encoder to store-serial output?
I want feedback.. thank you very much..

I built a binary encoder to input into a plc once. each section was a 2x ration drive, was the size of a cigar box but outputted a 16 bit position into the plc ttl logic board.. Could not lose it's position either.. each power up it knew exactly where it was.. This was on a edge guide printer.. yep.. top followed bottom exactly.. and I got a raise there..
I'd like to have one on my mill too, each axis.. all that takes is money you know..

[This message has been edited by ibewgypsie (edited 06-12-2003).]

06-12-2003, 12:24 PM
Looked up the transistors that were in the textron conversion of the bridgeport mill.. ECG53.. Remember that these had a double coil reactor inline with them. The outer coil looked to be shorted across a diode and resistor. It was UNIPOLAR by the way.

The reactors are gone. One of my drunk buddies saw the copper laying around and it worried him till I gave them to him to stop him from asking. (promoted kidney failure?)


Bummer... 297's are not available at Digikey or Hosfelt or allelectronics.. I thought that was a common chip.. Like the japanese motorcycles,, here today.. gone tamale.. (japanese motorcycles just a few years old you can't get parts for)

[This message has been edited by ibewgypsie (edited 06-12-2003).]

06-13-2003, 11:16 AM
L297 in stock:




[This message has been edited by Evan (edited 06-13-2003).]

06-13-2003, 12:40 PM
Thanks Evan.. did I send you a copy of the pcb so far? email me and I will..

It looks pretty simple.. socket everything so I can replace the 297's when I blow them working on the pwm circuit... huh? (reckon I need a zif socket on the first one?)

http://www.storm.ca/~larken/steppers.htm check this stepper board out, he says he can fix it.. the problem is a low impedance motor problem..

[This message has been edited by ibewgypsie (edited 06-13-2003).]

06-13-2003, 02:15 PM
Looks like decent heat sinking. I think you can build for less though. I sent you a copy of your PCB design that I fiddled with.

06-13-2003, 11:36 PM
Pretty cool.. a IR2110 Mosfet driver.. Replaces a lot of smoky parts.. Used in everything from small battery racers to ??

Anyone familiar with it?

robert phillips
06-14-2003, 12:21 AM

been tied up for a while. just downloaded expresspcb so i can see where you're at with your boards. will have to wait until monday since the emails are at work, not here at home. BTW, did you ever swap motors around to see if the problem moved with it? how about an itermittent short in the cables (flexing) to the motors? it's kind of funny that this suddenly happened and kept on happening....

about tool path layout form (graphical) -- added undo feature for gcode generation events and a diameter trace feature to show the relative width of the cutter along the tool path. next would like to work at configurable xy work area and then 3rd axis display for cut depth.


06-14-2003, 10:12 AM

A real neat 1/2 H-bridge driver.. Really neat.
Push pull circuit.. check this out..

Looking around.. for better..

Better... Ir3221... Sip design, space saver H bridge control..

[This message has been edited by ibewgypsie (edited 06-14-2003).]

06-14-2003, 01:12 PM
That is the same as a unipolar driver.

06-14-2003, 02:50 PM
Evan, only if you wired it like it shows. Two chips, four coils, it would work as a bipolar.

The full bridge I posted, well it has a ramp up, plenty of extras. More complications.. I am window shopping right now for the best ideal..

06-14-2003, 03:27 PM
Ir3220 IS the 6 amp solution for Servo drive conversions using auto windshield wiper motors on shop machines. I am thinking of gantry tables, motorized harbor fright machines>... here we go.. And, them at Autozone have a lifetime warranty on thier wiper motors... yep.. lifetime (forever).. a gilmer tooth belt, sprocket made to fit output shaft..

Pretty cool.. one chip does it all.. Makes me want to trash the steppers..

ONE chip solution, geeze where are the bipolar stepper 70vdc - 6 amp solutions?

[This message has been edited by ibewgypsie (edited 06-14-2003).]

06-14-2003, 03:47 PM

I will have to look closer at what you suggest with the two half bridges. Monday. I have about two acres of lawn to mow.

06-14-2003, 04:44 PM
I also was looking for stepper motor controllers and I found a device with three NEMA 23 sized motors. It was an ATM printer built in the 80's. It was based on the Z80 processor and has TIP122 output transistors. It was programmed in firmware to be a printer and I had zero chances of getting it to do what I wanted. Lucky for me I found a great person who thought the project was challenging and who knows firmware machine code like the back of his hand. Now I have a machine that can accept commands and it will do much of what I wanted it to, but I realized later that I really needed to go all the way to the milling machine to do what I wanted instead of playing around with a computer controlled router. Designing circuitry and logic controls is what Grahame does and he does it well. I've invited him into this discussion and if he does join in, I'm sure his contribution will be very very valuable. He has many years of experience and he is very willing to share his knowledge of electronics and of business with others. That means he would make a great Home Shop Machinist like all of us.

Would it be logical to develope some parameters or standards for the HSM SMC?(Home Shop Machinist Stepper Motor Controller) Some kind of wish list at least? I would like to know the dimensions so I can plan a box and fan. It would be fine if it mounts on stand-offs. It should have ample capacity so we don't smoke em' when we are making deep cuts. Would a 4 stepper board fit most needs or would separate modules added in a serial fashion make more sense? Why not skip the parallel port and serial port device and go directly to network device level? It seems to me that we need two types of devices. An I/O controller with some memory and separate inputs for safety switches, indexers or feedback loops, and maybe a joystick, and also separate output modules for the steppers from 1 to X number steppers and I mean BIG steppers like on my mill.

06-14-2003, 09:09 PM
One of the sites sent to me, A electronic supply has a PCB service. You can purchase 3 6inch by 4 inch PCB's cheaper than the expresspcb mini-board service...

I wonder....

06-16-2003, 03:31 PM
Servo motor driver .,.. National semiconductor
http://www.national.com/ds/LM/LMD18201.pdf A full Hbridge driver, with 3amp output, 6 amp peak.. This is most of the Servo drives, they use a resolver to add encoder input. Usually a pic chip..

06-16-2003, 03:44 PM
Hmmm... I still have several servo motors from an old scrapped Xerox machine. They are really nice motors with 22 pole spiral commutators. They can't get into a position where they won't start and are smooth as silk when running. I need a power crossfeed for my antique South Bend. I smell a project. That chip looks just right. I just have to figure out a mount for the motor on the far end of the crossfeed and some way to attach a gilmer pully to the crossfeed screw. Then all I need is a chopper circuit that can be controlled by the tach output on the servo to maintain speed as set when the load changes. That shouldn't be too hard.