food for thought (s) about getting away from the pololu

I discovered this during my searches on stepper motor drivers thing is that the stepper motor drivers we are using stick us with only two sources (one being open source, but still we are stuck using an smd. Which makes home fabrication quite difficult. However the above link shows how to use a pic processor to drive a stepper motor with some external circuitry (being that a pic can't muster the power but can drive pwm's) anyway there implementation and source code may give some of you cracks at programming an idea, on how to use later generations of processors to drive all of our stepper motors even in microstepping mode without the aid of stepper drivers but instead using a next generation (maybe a 32 bit one arm or so) to steer the motors directly through pwm's regulated by a chip that will enable us to go much higher when it comes to printing rates. Maybe just maybe we can hit the natural limitation of the technology this way, you can only print so fast till what your printing just becomes a melted glob ( however if your printing multiple parts at once that limitation is higher of course. Imagine printing a raprap (4th or 5th gen in a matter of an hour or less) and I mean a generation that uses much less vitamins and more plastic.
Imagine no need for a heated bed because the abs/pla part is finished before the bottom can cool down too much.
Imagine 6000mm/min print speeds the electronics are part of what is holding us back 20mhz at 1/16 microstepping is just not gonna be able to do it , well we'll need more powerful stepper motors as well and a better drive system :p
Just some food for thought one can dream can't we :p

terramir

Open loop PWM does not work as well as a chopper drive. If you look at fig 14 you see the current rises slowly during the first few pulses. A chopper would keep the power solid on until it hits the target and then starts chopping. Also it gives constant current irrespective of supply voltage and coil resistance.

If you use open loop PWM the motor will drift as the coil with most current heats up relative to the other. So you need a sense resistor and a micro that can do very fast A to D conversions as on times can be a few microseconds.

---

[www.hydraraptor.blogspot.com]

I think gen7T will fill that niche fairly well if it works out. The drivers are TB6560 through hole chips.

[forums.reprap.org]

yes it would but traumflug already stated there may be some resistance problems, that could cause troubles on that end
making it a software related solution could have many advantages, if we can replicate what the Toshiba and pololu as well as stepstick drivers do by using chip Ports possibly over could bring the cost's down tremendously, however this would require faster processors, but then again there are chips that can run easily on 200 mhz that don't cost an arm (pardon the pun) or a leg LOL.
If we can figure out how to use the newer generation of processors (well I wish you could find a more powerful one of the atmega with a dip package but much faster speed would keep things simpler. but that doesn't seem to be in the cards
terramir
PS: However I saw cortex arm development boards while they might only run at 8mhz or so aren't arm chips like 32 bit that should be a heck of a lot more processing power and those development boards can be like less than 25 bucks only problems there the coding as well as the circuits would have to be redesigned.

Edited 1 time(s). Last edit at 01/05/2012 10:08PM by terramir.

It's not so much processing time as fast A/D sampling and comparison with a threshold. With four motors and 8 coils and switching times down to about 1us you need a very fast A/D just to emulate 8 comparators.

A stepper driver contains a translator to convert step /dir to sine and cosine, comparators, level translators and two H bridges with shoot through protection, etc.They only cost a few $ each. It is only the first part you can do in software, so it doesn't save much unless you go to unipolar drive and find a micro with 8 comparator inputs.

Edited 2 time(s). Last edit at 01/06/2012 04:18AM by nophead.

---

[www.hydraraptor.blogspot.com]

All the ARM CPUs I've seen so far are SMD, so with your current plans you get a through hole stepper driver, but require an SMD CPU instead. That said, a nice thing with a dedicated PIC for running the stepper might be to put the whole ramping algorithm into this chip, so the ATmega gets rid of the extremely time consuming step interrupt.

One-chip stepper drivers with built in ramping algorithms exist already, but so far I'm not sure how they can be driven to run in concert for more than one axis. A custom PIC would be open to adjust for the required synchronous communications.

---

Generation 7 Electronics	Teacup Firmware	RepRap DIY

I see nophead's point that it would be difficult to get away from stepper driver IC's and while I was looking into economic feasibility while you can get dual comparators for 33 cents it's be difficult to emulate the other stuff. The toshiba chip is the only real alternative when it comes to economic feasibility and it has a 2.5A continuous rating 3.5A peak.

As for switching to arm it might be our only option in the long run as the printers natural development progresses, however we'd have to pick one that is readily available as a prototype board, which will need to fulfill three requirements.
#1 it needs to give us greater speed and capability then we have now, we are about to hit the processing barrier on those old ardino chips and the faster chips are not availible in a form we can socket it seems
#2 it needs to be fairly cheap and simple to hook into our project
Question there how would a 20 MHz 8 bit chip compare to a 32bit arm chip @ whatever MHz in terms of processing/ I/O power.
#3 it would need to be a chip Branch we could all agree on so we don't fork the heck out of the source. And later as it get's more mature, we might just be able to convince one of those protoype and pcb manufacturers to offer something on eBay with the next faster but still cheap version chip. you never know.
terramir

Quote

we might just be able to convince one of those protoype and pcb manufacturers

The RepRap market is easily big enough to do make such designs on it's own and "convincing" a PCB manufacturer is typically done by sending them a set of Gerber files.

---

Generation 7 Electronics	Teacup Firmware	RepRap DIY