[radiantheat]

Please excuse the different idea post , and also the fact that my ideas are against the flow and my knowledge is limited..

I think open source is a great idea sometimes , but not for a case like mine. I just need MORE POWER than a Gecko 320 can deliver. The Gecko's or there copycats are very highly developed , cheap , and much is known about there behavior. I have been watching the UHU project for a while. The UHU project is going no were (that I want to be).

THERE MUST BE A WAY TO TAKE A INEXPENSIVE GECKO controller , which is just like a commodity at this point , and add a high output power stage. That way we do not have to re-invent the wheel , and can focus on getting more power.

I envision a stand alone card which is connected on the input side to DC power , plus the output wires from a gecko or ANY OTHER SMALL DRIVE. The output side would be higher amperage current and would 'follow' the gecko. Perhaps a 'dummy' load would be needed on at the inputs so the 'gecko' or other drive thinks it is driving a normal load.

Consider how much simpler such a board would be. I guess it would be a lot like the output stage of a UHU.

No doubt a few people will say . Cant be done , impossible , etc., those folks might have believed in a flat earth in another time.

If a dc drive expert would care to discuss how this could be done and perhaps sketch its layout and parts basics, I would be glad to spend my time and money trying to make it work.

Or if someone could explain in simple terms why it is fully impossible , I will accept that.

[Al_The_Man]

Would the Larken Viper drives suffice?
I believe they were aimed at the 180VDC @ 25amps.
What size are your motors?
Al.

[H.O]

Hello,
First I got the feeling you wasn't looking for something DIY then you say you want to take apart a Gecko, use its logic circuitry and then come up with your own outputstage that matches the drive-signals of the Gecko - talk about DIY.... ;-)

The reason the G320 is so popular is because it's cheap and it (usually) works really well. The reason it works really well is because Mariss and crew have put a LOT of work into designing the drive and laying out the board. Fitting higher voltage/higher current MOSFET to the board has been discussed on the Gecko group and Mariss says it simply can't be done, I have an E-mail where i states why, I'll be happy to quote it if you like.

Just tapping off the PWM or MOSFET drive signals and running it to an external bridge circuit MIGHT work but without knowing much more about the G320 it's hard to say. For someone knowlegable enough to pull THAT off and do it PROPERLY (I'm NOT that guy) I think it would be easier to start from scratch.

The first thing that springs to mind that proably will be a problem is the current limit circuitry. Should that be handled by the present G320 circuitry or be "outsourced" to the new output stage? If "outsourced" how will it affect the G320 circuitry? If not outsourced you'd need to match signal levels to the G320 circuitry etc.

I'm not saying it can't be done I'm saying that IMHO it'll be a much bigger "DIY mess" than the HP-UHU is.


There's been problems with the HP-UHU, no questions about it! I've had quite a few of them myslef and worked thru them all. There's still the possible problem with the false transitions on the encoder input that hasn't quite been verified and/or solved yet. And not that IT ain't a big issue but other than that I think it's been proved to work quite good.

Let's say we figure this encoder issue out once and for all, what else is there that holds you back from trying it?

Again, I'm not trying to poke holes in your ideas I'm just trying to understand.

/Henrik.

[radiantheat]

First thanks for your thoughful reply. The UHU project is pretty scary to me , its not like I can go online to a single website and buy a complete UHU-HD kit. And its not like I know enough to build what I am talking about.

I always thought the reason that no-one makes a 'hobby' class drive for large motors is because of product liability. After all the end result of MOST experimental hook-ups is smoke and a violent crash.And yet there are so many large CNC machines from the 80's and beyond that are just waiting for a second chance.

My current project is a 1979 Ikegai 20 HP Turning center with 12 position turret. Everything has been stripped except the motors. I have fried 3 Gecko's 320's, 2 Rutex 2020's , and a few other drives. The X and Z axis motors are Fanuc-Getty 10M and 5M, but I also want to be able to drive a DC motor fitted with encoders for the spindle drive. That should allow threading and other tricky operations.

The whole point behind my idea is to NOT TOUCH the Gecko's (or other drive) but to amplify the voltage and current coming off the motor outputs. What I wish to do is to make a drive system that would create a 'false' load on the gecko (perhaps with some inductors) , so it thinks it was driving a motor , and then amplify the output from the gecko's to drive the large motors with some power transistors. no need for perfection for my intended use.

Problem is I am a experienced digital designer and programmer , but a newbie with 'analog' or whatever you call dc power circuits. Still if there is a simple circuit that could serve as a starter , I would be ready to try. And when I can go online and order a complete UHU-HD kit , with english instructions, I would try that also.

Thanks again for your insights

[Al_The_Man]

The Fanuc DC 5 motor is rated at stall torque 55Kg-cm and a torque constant of 4.87Kg-cm/Amp which would equate to 11.3 amps at this torque.
The max speed is 2000rpm and the BEMF constant is 50v/Krpm which means they are rated at ~100VDC.
You can allow approximately twice the torque and current values for the size 10.
Al.

[radiantheat]

Thanks, I am looking into your Viper Drive Suggestion. It seems there will be a new HIGH OUTPUT version soon which might drive my Servos and Spindle Motors, problem solved ! But I still have a soldering gun and some wire so watch out for smoke !

[H.O]

I'm not sure doing what you suggest with the G320 is even possible, and IF it is it won't be easy or someone would've done it already - me thinks.....

The reason Gecko doesn't build a high voltage drive is, acording to Mariss, that the market is MUCH smaller and he fears US lawyers will come after him after some stupid idiot have electricuted him/her self by not following the docs, rules and regulations.

Then we have the Rutex products which you apparently have hands on experience with, we have the Mammut from CNCDrive which I have hands on experience with (not good so far), we have the Viper drives that by judging purely from a photo seems to be using a 32V rated automotive fuse on a 170V rated drive. Designing high voltage, high current drives is NOT easy - that's probably why there aren't any good ones around....that you can afford....

What new version of the Viper is this you mentioned? Any additional info or links?

I'm not trying to push the HP-UHU here but I do feel it MIGHT be an option for you. Like I said, there's still this problem with the encoder input that needs to be figured out once and for all but then I think it might be a good candidate for your axis motors atleast. The 20HP spindle motor is whole different animal though....

I'm not sure about the current status of kits but I know Paul and Irfan mentioned they should put a few more together. By getting the kit you get everything you need except the UHU-chip and X-tal which you have to get from Uli Huber in Germany.

And, as far as I know no one has had a problem with the powersection on the HP-UHU yet, (if you don't count me shorting the output that is) so that part of the circuit seems to be working really good. And, since you are a programmer you might be interseted in developing the next generation controller chip for the HP-UHU ?

Yes, it's a kit, yes you'll have to some soldering but if you're conciddering the Gecko/external H-bridge thing it seems that part doesn't scary you ;-)

Which ever way you go I wish you the best of luck and be sure to let us know of your results! Good luck!

/Henrik.