Geckodrive user input requested

[Mariss Freimanis]

Hi,

I'm cross-posting this everywhere because I need a lot of replies fast.

I'm nearly done with a G201/G210/G202/G212 replacement design called the G203 for right now. It is a complete, clean-sheet CPLD based new circuit. Same size and pinout as the G201 drive but a whole lot of new features.

Here's where I need input. The G203 uses 3.3V logic which makes the step/direction opto-isolator interface work with much lower current. The quandry:

1) Be 100% backwards compatible with the existing drives. That means supplying +5VDC to the COMMON terminal and require 5VDC step/direction signal levels. The G203 would keep an unplug and replace compatibility with our current drives.

2) Change the COMMON terminal spec from +5VDC to GND. The G203 requires only 2mA max to operate step/direction optos instead of 16mA. The G203 has a max step pulse frequency of 350kHz (>10,000 RPM on a 1.8-degree motor) without any funny G201 step/direction timing restrictions.

The result is supply step, direction and PC ground to the G203 and it will run. It will work now with 3.3V or 5V logic; no level-shifting breakout boards are required, no +5VDC supplies.

Please let me know: Option (1) or option (2). The first 1,000 boards must go to the PCB fabrification house Monday.


Mariss

[turmite]

Originally Posted by Mariss Freimanis 1) Be 100% backwards compatible with the existing drives. That means supplying +5VDC to the COMMON terminal and require 5VDC step/direction signal levels. The G203 would keep an unplug and replace compatibility with our current drives. 2) Change the COMMON terminal spec from +5VDC to GND. The G203 requires only 2mA max to operate step/direction optos instead of 16mA. The G203 has a max step pulse frequency of 350kHz (>10,000 RPM on a 1.8-degree motor) without any funny G201 step/direction timing restrictions. The result is supply step, direction and PC ground to the G203 and it will run. It will work now with 3.3V or 5V logic; no level-shifting breakout boards are required, no +5VDC supplies. Mariss

Mariss is the last paragraph attached to the 2nd option? If so, my vote is #2

if fact my vote is #2 anyway!

Mike

[Mariss Freimanis]

Yep; it is.

I have been torture testing a hand-built SMT prototype for the last week. It works so good that only minor changes were required on the PCB gerber files. Blows the G201 clean away.:-)

Mariss

[pminmo]

I'd also suggest #2, backwards compatability isn't that important in this minor of a wiring change. You'll have less support issues, more satisfied customers...

[Mariss Freimanis]

The 3.3V thing with new PCs has been a real headache. Some of our drives "sort of work" with them, others flat-out do not. Both kinds are being operated completely out of spec.

The typical support exchange for this begins with "I got 3 of your drives. One works, the other 2 don't. What's wrong with them?". Well, what's wrong is they are not designed to work with 3.3V signals. Some may, most won't.

Normally I stick with 100% backwards compatibility with any new design but this 3.3V/5V thing is a real and growing headache. It's very tempting to completely make it all go away using option 2.

Mariss

[2muchstuff]

There is something to be said for keeping things the same for the ease of "plug and play". Since your headaches stem from the use of 3.3v logic in todays computers and there will be more of them to come. It's not hard to move a few wires around if you need to replace an older drive. I vote for option #2.

You already provide a quality drive for a good price with an excellent reputation. It would be hard to improve on a good thing except for trying to make it more idiot proof. Heck, I thought it was idiot proof already.

[phoneman]

Mariss: I hope to soon be a new customer. How do the two options relate to using the drive with the Grex? Planning Mach3/Grex plugin, Grex, 5 stepper drives, Compaq laptop computer is that doable with the new drive?

Thanks

Dave

[Mariss Freimanis]

ROHS compliance and the growing unavailabilty of 4000-series CMOS logic has promted this design. I'm a lazy person who is loathe to fix something that isn't broken.:-) I have read the handwriting on the wall however and it says "Design a new drive series if you still want to be here in 5 years".

The G201 traces its ancestry back to 1985 so it's been a good run but it's time to move on. The G201 will be in stock and available for at least the next 3 to 5 years. We intend to choke G201 demand by pricing the replacement at parity while offering significant performance enhancements once the production teething problems are overcome.


Mariss

[mikie]

Mariss

Option 2 is the only way to go. At the moment people running G201's already understand how the unit works. People who don't have one are finding it hard to wire up (as i did when I got one).

Make a product as idiot proof as possible in my opinion

/Michael (an idiot)

[Mariss Freimanis]

It has no impact on the G-Rex since both GND and +5VDC is available in each axis connector group:

+5VDC
GND
STEP
DIR
CH_A
CH_B
CH_I
LIM


Mariss

[Mariss Freimanis]

They always make better idiots though.:-)

1) The G203 uses a lossless (no current shunt resistor) short-circuit detect/protect circuit that senses top MOSFET drain-to-source "on" voltage, level-shifts it to ground reference and trips a protection latch if it is excessive (>10A).

2) No internal jumper settings at all:

2A) The G203 covers a 0A to 7A range instead of 1A to 7A.

2B) The G203 goes into a motor heat reducing recirculating switch mode when the motor is stopped while maintaining full holding torque.

2C) Auto-gain adjust for NEMA-42 anti-resonace.

2D) No external capacitor needed at currents above 3A. A ripple current canceling design takes most of the stress off of the internal 100uF/100V cap.

2E) No weird G201 step/direction timing constraints. Simply have the direction input true before the active step pulse edge. It's clocked in now.

2F) Opto-isolated DISABLE input. Run it from your PC if you wish just like step/direction. DISABLE free-wheels the motor and clears the short-circuit protect latch.

2G) A green "POWER" LED and a red "FAULT" LED. This eliminates the G202/G212 quandry of "Why isn't it running? Is it short-circuit or is it something else?".

2H) Over-temperature protect. You just can't burn it down if you forgot to put it on a heatsink while you are running your 7A motor at 80VDC. It shuts down into FAULT when the board temp reaches 105C.

2I) Still trying to decide on this: An internal fuse that blows if you:

a) Put more than 114VDC on the G203 power supply input terminals.
b) Apply reversed power supply polarity (Get '+' and '-' all mixed up).
c) Something inside goes very badly south.

Problem is the fuse would be a thru-hole soldered-in PicoFuse. You blow it and the drive has to come back for replacement unless you are a really good electronics type person with really good multi-layer board rework skills in which case you wouldn't have blown the fuse in the first place.:-)

2J) Top-side adjustable ADJUST trimpot. You won't believe the number of requests for this.
---------------------------

Again, a completely new clean-sheet 2006 design, not another evolution of the G201. The G203 has only 2/3 the number of SMT parts so we can turn out half again as many a day as G201s. That means 450 a day instead of 300.

Downside is a CPLD is a programmable device and each must be JTAGed which takes 15 seconds. Doesn't seem like much? That's 6,750 seconds or nearly 2 hours; a quarter of a work day.

We need to see how the "upside" and the "downside" balance out. I'm hoping for a draw in 3 months. That will affect the price.


Mariss

[Greolt]

Mariss is it possible to have protection built in so that if someone swaps in a
new drive without changing the necessary wiring it won't blow up

Be good if it just doesn't work rather than have a disaster