CNCdrive UC300ETH Mach3 plugin is now available

[Clive S]

Now that looks good,

And support the relevant Gcode like G76 (linuxcnc) or what ever flavour is used.

[A_Camera]

Well I do and I'm telling you in my expereince they are cause of most troubles on CNC machine.(After broken/lose wires)

Fine, but still, the thing is that if you want to use a super expensive BOB you can do that with the UC300ETH, nothing prevents you from duing that. The cheap ones I have are good enough for me at the moment, and until I really see problems which I can't fix I'll continue using them.

It's like the USB argument I've had so many times with people like you.! . . . People who have insisted there's nothing wrong with USB for CNC it works fine, they never have any troubles with it.!! . . . . Total Bollocks. Compared to Ethernet it sucks.

I don't think we ever had a fight about USB. Of course Ethernet is better, MUCH better, no question about that. Also, I never made any secret about my USB experience, you can also read about them on my blog, and it has been there since over a year now:

http://adapting-camera.blogspot.se/2...ry-part-5.html

http://adapting-camera.blogspot.se/2...ry-part-6.html

Both describing problems, the second is a serious warning. Since that second post you can no longer buy that USB motion controller at all, it was simply removed from the market. It was after that I bought the UC300USB, which I was very happy with and which never gave me similar EXTREMELY serious issues. Yes, it lost contact with Mach3 2-3 times, but that's all, and when it happened it just stopped, unlike the other, which literally run away and continued sending pulses to the stepper, there was no stopping it except through power off emergency stop.

I have also made a video a while ago:



I don't know if that stop and connection loss was caused by the fact that I was recording the screen and at the same time I also had a backup task running, probably the backup caused it because screen recording I have done many times before, but I have not kept that a secret.

Never the less, USB works if it is a good motion controller, like the ones CNC Drive is making, but "never" is a long time. I can tell you that I have not had any issues with my UC300USB more than 2-3 times, and one of those is the video and NONE were serious problems. Sure, Ethernet is definitely better I will be the last to deny that, but until recently it was pretty expensive to venture into Ethernet motion control, and for an amateur if would have meant a significant cost, so yes, USB is a very good alternative, and if I'd have money issues I'd still be happily using the UC300USB and we would not have this discussion because I KNOW it worked well for me.

Anyone who works and maintains CNC equipment will tell you it's pain in the arse and can't be trusted not to lockup. Well the Same goes with Cheap Bob's they can't be trusted and will fail or cause issues at some point, often sooner rather than later.

If they are working for you then great but that doesn't change the fact they are potential trouble causers best avoided.

Of course, if there is a better alternative why not use that? If I would disagree and would claim USB is as good as Ethernet I would have not bothered buying the UC300ETH at all. I also know the benefits of Ethernet vs. USB, and if the UC300ETH plus the Mach3 plugin was available a year and a half ago I would have definitely NOT bought a USB controller at all because Ethernet is definitely better. But at that time I found no better alternative which I could motivate the costs for.

Also, like I said before, it is different if you need a machine to be running H24 reliably because your daily bread is depending on it, but in my case, and I believe in many others case as well, this is just a hobby, so in that perspective, USB is just fine. Of course, the rules of the game have changed with the arrival of UC400 and UC300ETH, since with those two available there is no reason to buy a USB controller at all, which probably will show up on CNC Drives sales figures also, since I believe that these two Ethernet controllers will seriously knock on the sales of UC100 and UC300USB.

[m_c]

BOB's and features are always a compromise.
.
The question came up on the Dynomotion Yahoo Group recently, as Tom was thinking about a more universal kind of BOB, and was wondering what type of features users would like.
Things like Step/Dir outputs to screw terminals, analogue outputs/inputs, voltage of general inputs/outputs. Off course opinions varied.
.
One thing that surprises me, is that there are not more people using the KFlop+KStep combination. 4 reasonable stepper drivers, 16 12-24V inputs, 2 relay drivers, along with a fully isolated analogue output. I used one for a probing machine, and to me it seems like an almost ideal setup for smaller stepper driven machines. I suspect the problems are you need a small C Program to make use of them (A basic C program is included to get up and running with no alterations), and the fact you have to use above 5V to make use of the inputs.
The combination costs $450, yet people are happy to spend $155-180 on a SS, $300 on G540, and then bodge it all together with questionable wiring or a cnc4pc BOB (Been there had one, biggest heap of CNC related junk I've ever owned!), and finally buy a Mach license.
.
I can understand hobbyists wanting to keep costs down, but there seems to be plenty who'll spend far more just to go with the flock.

[Ger21]

One thing that surprises me, is that there are not more people using the KFlop+KStep combination

I can tell you exactly why.

1) The lack of a good, easy to connect breakout board.

2) Having to write C programs.

It's too complex for the average person to use, when there are simpler options that work just as well.

I have a friend that was fed up with Mach3's bugs, which constantly caused ruined parts for him.

He bought a KFlop, but didn't have time to install it.

I had him buy a UC100, which he could switch to by just plugging in the parallel cable, and he started using UCCNC until he had time to switch.
He also bought a second machine for his business, which already had a KFlop installed.

With the updates that UCCNC have made to their software over the last year, he now sees no reason to switch to the KFlop, and will be selling bioth of them and moving the second machine to UCCNC.

[A_Camera]

I watched the video and frankly I wasn't impressed. You quote 10,000mm /min speed but on a machine such as yours you don't have the room and I'll bet it never gets over 2500mm / min
I'm certain that a standard mill running a standard copy of M3 could do this file the same.
Perhaps UCCNC is capable of more but it will take it to be installed on a larger machine than that. I feel the file doesn't do justice to the software

I didn't post that video to impress you or anyone. Never the less, I don't think there are many DIY machines which run at that speed with only direct stepper motor driven single 1605 ball screws on each axis. Most people when they want speed in this range use belt and gearbox, dual screws, higher pitch, servo drive and so on. I seriously doubt that there are many Chinese moving gantry machines which reach more than half of this speed. Never the less, this is not a pissing contest, so I don't care if you are impressed or not.

As for the maximum REAL speed of my machine, again I think you are wrong, but I still have to measure that using real instruments, which I don't think I will ever do because as I said, it is not a contest and is not my goal to impress anyone.

Anyway, just for the sake of discussion and for answering your claim, UCCNC and Mach3 both display the speeds, in this video at least once, at the 40 second mark you can see it reaches 10,000mm/min. If my calculations are right then with 700mm/s/s acceleration I am up at 10,000mm/min after 19.8mm. Watching that video frame by frame shows 33mm for the Y to move from 170mm down towards Y0 before it reaches 10,000 (you can see it in the FACT message box) and deceleration starts at the Y 19.18 mark, which is almost like my calculation.


Remember that the frame rate of the video is not high enough for accurate measurements, but I am pretty sure it is correct as far as the top speed is indeed reached at some point, probably long before the 33mm travel of Y. Also, the settings used was constant speed, which probably affects the distance needed as well, but never the less, I am sure that I actually reach 10,000mm/min. Also remember that this drawing is only 170mm Y movement, and my table is almost twice that size if I need it to be.

The 2500mm/min is the feed rate for the pen when the pen is down, not the rapids, which is obvious from the video as well.

I am aware that my machine is tiny by your standards but again, it does not matter. It is large enough for me, weights over 80kg and it takes up more than enough space as well. Lastly, it definitely beats most, if not all the kits which can be bought with similar table size, in terms of speed and real accuracy, but regardless, it is NOT a competition against anyone else.

[Boyan Silyavski]

I believe that almost all board makers are ignorant people that live in their dreams and refuse to make and sell what we really need. I may write it down somewhere or open a new thread for the people to say what is really needed in a board.

How hard could be to design a FPGA chip like most of the new boards use and put that chip on a tiny board . Then make that tiny board fit in the middle of and connect directly to a bigger board which has high speed opto isolation, differential inputs and outputs, 24VDC or 5VDC by our choice. Come on guys, go to Galil web site, download a manual of their boards and see how its done. Swap of cheap resistor block changes board from 5VDC to 24VDC, swap of IC direction, or even better from software changes board active high or low and so on.

What year we live in? i find strange that a new board will not output 4mhz at least signal and still talking about khz. How much more expensive could be for manufacturer, 10$? I am sure its not more than 20-30$ to make a perfect board. Now that we have ARM chips so low priced.


And how hard it's to put that board in some kind of shield box?


I agree with Dean though. in my limited experience problem was either the board or the cables. And the problems that board can give , there is no other part in the machine that can frustrate me as the BOB can


People don't use Dinomotion for the same reason i don't use Linuxcnc with Mesa boards. For that same reason Apple became a big player in the PC game. Its called "ease of use"

[m_c]

I can tell you exactly why.
1) The lack of a good, easy to connect breakout board.

With a KStep, all you need is a basic 26 pin IDC breakout board. You could say the same about some of the UCNC headers.

2) Having to write C programs.

Only if you want to do something non-standard.
As I said in my post, if all you want to use is the basic features, there is a C file you just need to link to, either via the Mach plugin, or via the KMotionCNC config page. In Mach3, all the inputs/outputs are configured like any other controller. KMotionCNC it depends on what you're wanting to do, but the new screen editor lets you assign inputs/outputs to buttons/leds, along with little C snippets to do what you want.

It's too complex for the average person to use, when there are simpler options that work just as well.

I have a friend that was fed up with Mach3's bugs, which constantly caused ruined parts for him.

He bought a KFlop, but didn't have time to install it.

I had him buy a UC100, which he could switch to by just plugging in the parallel cable, and he started using UCCNC until he had time to switch.
He also bought a second machine for his business, which already had a KFlop installed.

With the updates that UCCNC have made to their software over the last year, he now sees no reason to switch to the KFlop, and will be selling bioth of them and moving the second machine to UCCNC.

I'll agree that the KFlop can be complex, as ultimately it is a pretty high end controller with lots of capability that lots of users will never use, but it's only as complex as you want to make it.
.
I do find the programming thing a bit of strange argument, considering plenty people doing anything non-standard with Mach and even UCNC, are usually happy enough to learn enough of the required language to write Macros, but mention C, and they break out in a cold sweat!

[Ger21]

And if you don't want to use the Kstep?

Only if you want to do something non-standard.

Like Homing?

I'll agree that the KFlop can be complex, as ultimately it is a pretty high end controller with lots of capability that lots of users will never use, but it's only as complex as you want to make it.

And my perception is that to do the same things that I can do in Mach3 and UCCNC, that the KFlop is much more complex.

Dynomotion needs to change that perception to get more people to use it.

I've had this discussion with my friend with the 2 KFlops. We both feel that it would be much more popular if it were easier to use.
It's not unlike Mach4. Mach4 is incredibly powerful, and at some point may well be the best software control available. But it's incredibly complex, if you want to do anything "non-standard".

My take is that all of the available software controls are designed for a wide range of application.
For serious users, they all require a fair amount of customization for specific applications. This is where the complexity comes in.

[m_c]

Ger, I'm not going to disagree with you, but it has proven a point.
Users complain when a controller can't do something how they want it, yet when they're given the option to do it however they want, they complain it's too complex.
Controller designers just can't win.

[JAZZCNC]

I'll back Jazz on this and throw a bit of history into the ring.

Jazz is unique in that he works on routers, I work on lathes and mills, believe it or not, totally different animals. A mill can work fine on a USB card as the speeds are very much slower than the kernel speed the computer is able to reach. On a mill or lathe the controller is always waiting for the machine unlike a router which can easily outrun the controller, especially if servo's are used.

Thanks John I was hoping you would jump in because know you have had same BOB issues, As do many Cnc Engineers.

You also hit the nail on the Head with routers being different animals. Routers are always pushing the edge on Torque and Pulse rates because of the speeds they work at. In my experience the only way to avoid issues is by building using quality components and with performance to spare so they are running well below there maximum.

Even then It's fine line between working correctly and not. Even with experience this can catch you out as I found today. (by the way this as nothing to do with Bob's or Electrics just high lighting the differences in machines.)

Deliverd machine to Kent few week ago, 500mile round trip for me. On Friday Customer informed me was having issues with loss of position.

This is twin screw machine connected with belts so after talking with customer I suspected we had slipping pulley or some mechaincal issue, which to me is totally unacceptable and very hard to swallow or believe.!

So in the car this morning at 5am on the way to Kent to find issue. Thankfully and to relief of my pride all mechanical was fine and was just case of motors being overtuned. Bummer for me because was long trip just for bit of motor tuning, but still needed to happen for customers happiness and my own piece of mind.

But this highlights that the line between working fine and not on router with steppers is so very close. If this was Mill/lathe then wouldn't have happened (without being ridiculously overtuned which it wasn't) because motors are running no where near there maximum.