[ger21]

What pushed me to buy UCCNC was real simple, you can run it under Windows XXXXX 32bit or 64bit it doesn't care.. and what really sealed the deal for me was watching a dude run his cnc from his Tablet.

You do know that if using a UC100, you can do the exact same thing with Mach3?

[dharmic]

Hi Steve

> having been in the IT/T industry for most of my life
I started on an IBM 7044 with punch cards in the mid 60s.
Since then I have done everything from writing drivers in assembly, scientific crunching, creating DB systems (the DB kernel), debugging compilers on OSs, building robotics and image analysis systems ...
I have forgotten more languages than I can remember. Seriously. And I still say OO is just plain inefficient.
Why do old programmers drift into CNC????

Cheers
Roger

Bloody OO.

Writing a system for work in C++ on Linux and the whole OO thing makes it so much easier to read, test, maintain, everything. But it comes at a cost.

Tried to code up an embedded system in C++ as well - again, it was a thing of beauty, but it didn't work. And the bug kept shifting around. Stack running out of space from all the auto variables. Rewrote it all in dirty C with function pointers and globals and it worked like a charm

[InMesh]

I read this thread with interest I like the way my cheap Dell works at the moment it has never missed a beat I have a moldy back up PC but there will come a day ...

[curiosity22]

Hi Steve

I really really doubt that the PP had anything to do with the max size of job you can run. That would make zero sense. Just ain't so.
Running Mach3 on a 32 bit system is also fine, and won't be the limit. Most of us ARE running on a 32 bit system - WXP SP3 in many cases!

But Windows is really sloppy bloated code, and if the box did not have enough processor grunt or memory or graphics speed, THAT could kill it. The system could be page-swapping when it should be generating tracks, and then the PP driver runs out of data and hangs. I think those limitations might go with the age of the box.

Mind you, if you use exactly the same box and use an ESS instead of the PP, you might find the job cruising through.

Cheers
Roger

Hey Roger g'day,

You are probably right mate, page swappping would kill it, the processor was under powered not enough ram and yes windows is bloated with all sorts of things I don't need. You know I forget about page swapping, I switched page swapping off on my current system a long long time ago,.


Cheers,

Steve

[curiosity22]

Ger g'day,

Thank you for your post, yes I know you can, the point for my purchase is I'm happy at the price to support a team that is at least prepared to make and develop something current. I'm sure there are differences between the two what they are for me the end user is moot point - I can only supply feed back and ask for features. So long as x moves in the direction its supposed to go and I don't have to sit watching the machine like a hawk rdy to hit the Red Button I'm a happy chappy.

Cheers,
Steve

[curiosity22]

I read this thread with interest I like the way my cheap Dell works at the moment it has never missed a beat I have a moldy back up PC but there will come a day ...

LOL you know you will jinx it now that you've said that.

Its funny sometimes I feel like a worm that's just been pulled out of the dirt. Just wriggling around cause some bugger decided to change the rules about how things run.

I'm waiting on the postman to deliver this parcel, even Aussie post has changed, no more quick deliveries, gone are the days if you post before 6pm that it was a good chance your parcel would arrive the following day.

Cheers,
Steve

[Beefy]

I'm into plasma and have my eye on the UC300 & UCcnc motion controller software for the future. I'm hearing very good things about it and having downloaded the demo, it seems to have a lot of similarities to Mach3. One thing I hear repeatedly is that it's trajectory planner is great, something important for plasma or wherever CV is needed. It seems to have the general plasma features built in but for some reason does not have corner lock integrated. Maybe they'll add that too later on. I've also heard it's quite stable.

The reason I'm holding off for now is they only have a USB version, and from what I've read, USB motion controllers are prone to noise when used with plasma. I have heard though they are working on an ethernet version. I've also heard rumours they are working on a screen designer program. As soon as those arrive, I'll be hitting the buy now button.

So for now I'm sticking with the parallel port and designing my own torch height controller and communicating with it from Mach3. The computers I use have parallel ports built in but I don't even use them. I just buy a dual parallel port PCI card and use that. So far this is the only way I can make my own custom designed plasma system (screenset, communicating with external devices, etc). I'm hoping UCcnc will end up being the Mach3 replacement and have the same customisation, scripting & communication facilities as Mach3, time will tell.

Another option for parallel ports is buying the brand new Mini-ITX motherboards. Many of these have one parallel port built in and one PCI slot.

Something else to consider is I think Windows 8.1 is the last version to support parallel port cards. I think Mach3 works with Windows up to 8.1 but only 32 bit.

[OlfCNC]

I am running Mach3 with Windows 8.1 and 64bit with UC100 and it works without any issues. I recently upgraded to UCCNC and it also works on the same OS without issues.
And I had no problems with Mach3 + UC100 on my plasma machine eventhough it is USB.

[RCaffin]

Yes, my understanding is that the UCcnc group is getting an ethernet version ready. Dunno when though.

And yes, there have been problems with the USB connection to the Smooth Stepper. The signal voltages are low, there is no galvanic isolation between PC and peripheral, and the USB protocol per se has no provision for error detection and correction (EDC), unlike the Ethernet which is designed to handle EDC.

However, I suspect the guts of the problem lies with the FTDI handler for the USB interface used on the USS. It's the handler which hiccups on some machines. If that handler had good EDC all might be well - but it doesn't. Which leads to the thought that perhaps the UC SW does include EDC. Certainly, the speed of the USB connection is fast enough to be able to handle EDC. Once you have a reliable connection, life does improve.

Mind you, I still like the galvanic isolation of the ethernet interface.

cheers
Roger

[OlfCNC]

What about using a USB isolator?
Adi ADUM4160 USB Isolator Protective Board STM32 Arm J Link Infineon TI DSP Pic | eBay

[RCaffin]

The USB isolator might help, but I (and others) did a lot of testing with really good ground links between the PC and the I/O device, and still had noise problems. Different ground connections were also tried, without success. I tried a USB booster cable as well, and still had crashes. Crashes in a CNC machine are slightly less than desirable.

I suspect that the lack of EDC in the SW FTDI device handler could still allow the odd crash to happen, despite the isolation, because the signal levels in the USB are very low, around 3.3 V I think. The Ethernet signal levels are more like 15 V. That matters.

I was getting about one crash a day with the USB interface at one stage. I don't know exactly why. Since changing from a USB SS to an ESS, I have not had a single crash (well over 1 year). But I know others have not had this problem.

Cheers
Roger

[curiosity22]

Funny how some days are good days and others are well, left to be desired.

My mate just had a cry over the phone, can't get his roughing to finish. Stopped 3 times so far, 3rd time z is doing a fresh air cut. He has got some real gremlins today. Normally he runs ok, but I think today has pushed him over the edge to look for alternative solutions.

Rodger what type of ESS are you using if you don't mind my asking.

Dam postman still hasn't delivered my package, don't know about you guys but I think Aussie post has gone to the dogs, my experience in getting parcels quickly has gone out the door it seems.

Cheers,
Steve

[RCaffin]

Hi Steve

No problem at all. I am using an Ethernet Smooth Stepper from Warp9:
Products

It needs an external +5 V supply, but so does the rest of the interface logic. No problems there.

Off this I have two Homann Designs Breakout Boards MB-02V6:
MB-02-V6 Bidirectional Breakout Board [MB-02] - US$39.99 : Homann Designs!, The preferred CNC Component Supplier
Actually, I power the BoBs from the +5 V supply and the ESS supply comes from the BoBs via the connecting ribbon cables, but that is a minor detail. You don't have to do it that way.

Each BoB gives the equivalent of a full PP, with extras like a safety keep-alive heartbeat. The pins on the second PP can be configured more flexibly than the first. I can hang a third BoB off the ESS if I ever need more I/O pins, but so far I have an excess of pins. Mind you, I have also added optical isolation almost everywhere, including on the encoder feedbacks from each axis. (Not relevant if you are using steppers.)

What I have found is that the Z axis is especially susceptible to drift in many machines. The main reason for this is that the Z axis cabling is longer than for the X or Y axes, plus it often has the spindle cabling running bundled up with it. If there is any noise on the spindle drive, the Z axis encoder can pick it up. Sigh! As it is almost impossible to find a spindle drive which does not radiate LOTS of RF noise, this happens.

I have moved the spindle drive away from everything else - into a different metal box in fact, and I have rerouted the spindle drive cable away from everything else as well. I put a good copper braid screen over the spindle drive wires, plus I have thrown in several suppression ferrites back at the spindle drive connection. This has turned a horrible static noise on a small nearby transistor radio into quiet enjoyment of the ABC or 2MBS stations while machining.

Yes, the ESS and BoBs cost money. So do broken cutters, wrecked jobs, and hair torn out. Feel free to ask any questions.

Cheers
Roger (BSc Hon, MSc, PhD - I knew they would be useful for something...)

[curiosity22]

Hi Steve

No problem at all. I am using an Ethernet Smooth Stepper from Warp9:
Products

Roger (BSc Hon, MSc, PhD - I knew they would be useful for something...)

Hey Roger G'day,

I will pass on the info re : Z axis to my mate, from memory I'm pretty sure his Z axis is routed a different path to everything else I will find out soon enough. He is a digital and electronics engineer and loves to speak to me in shorthand, I need to remind him every so often that I'm only human and need to breath between heartbeats. ...

Thanks for the info on your controllers - I've fallen over warp 9 before but never stooped long enough to have a good read and see what they are offering, seeing as your using their product I'll have a closer look.

You know I don't mind throwing money at some things - provided they work. The greater cost is spending the penny's and cents which add up to pounds in the long run.
Ok you write as though you are using more than one BoB on your machine or am I misreading that? Also what type of cabling are you using where you can use an optical isolation diode (I'm assuming its a diode). I take it your using servos.

What was your paper on "PhD" if you don't mind my asking?


I'd throw some letters at you as well but I won't :P..
My lovely wife says that I should - she even gave me permission to do so lol...
According to her I'm the smartest man she knows, after all I did marry her.....

[OlfCNC]

Roger, thanks for info about the USB.
I did not have a single issue like this with my UC100 in about 1 year i'm using it on my plasma machine.
May be the issue is with the usb smoothstepper itself?! I did hear about the same issue you telling from a few other people too.

[RCaffin]

need to breath between heartbeats.

Then you will be breathing awful fast ...

Warp9 and ESS: to some extent you could call the ESS a reference standard. I am sure some of the other vendors, such as UCcnc, will disagree, but if you look at volumes and at presence on the Forums, that's how it looks to me. That does not mean that the UCcnc stuff won't work; I believe it does, but someone running Mach3 through a UC100 seems to be the exception rather than the rule. (In my experience.)

The greater cost is spending the penny's and cents which add up to pounds in the long run.

I don't even mind the pounds bit. It's when I spend the dollars and the many hours and it still fails - that's when I get most upset.

Yes, TWO BoBs on my machine. The first BoB is full, but the second BoB is only half-full.

Optical isolation - I make up my own PCBs using the CNC for this. A Catch-22 situation to be sure.
Attachment 290106
Optically Isolated outputs: optical MOSFETs actually. Like a relay, but only ~4 mA input to the IR diode.
Attachment 290108
On the left, optically isolated encoder inputs - they go between the HEDS encoders and the electronics. Two layers: top layer is actually the spindle Index pulses going into the BoBs; the bottom layer goes to the X, Y & Z Gecko320X servo drivers.
On the right: the eStop sub-system. It monitors all the Geckos and the eStop button. It can disable the Geckos and the spindle driver and alert Mach3.
At the top right, the MOSFET outputs.

Caution: what you see in the photos is not exactly how it looks now. Yet another rebuild (to move the noisy spindle drive away) is just finishing ...

BSc and MSc were in Physics at Melbourne Uni; PhD was in Systems and Control in London. I am 'retired' these days. And the research world is not nearly as nice today as it used to be. Sigh.

Cheers
Roger

[RCaffin]

Hi Olf

May be the issue is with the usb smoothstepper itself?!

Well, yes and no.
Choosing to use the USB in the first place was a mistake imho, but the FTDI driver for the USB interface chip on the Warp9 USS is where the hiccups occur. I can only suggest that UCcnc have a different chip and SW for their system.
Kudos to UCcnc for the better interface design.

Cheers
Roger

[curiosity22]

Hey Roger,

Thank you for the explanation, I know and can follow what it is you've done and are doing. Clever!!
I hope your enjoying your retirement, I know I'm not.
You must have one hell of a noisy spindle to do so much work around it. Or are you just being pedantic to make sure?
Dam it Jim your making me want to pull out cables lol. I don't need to but looks like I want to reroute as well.
Got email from Aussie post, dear john your package is delayed we will let you know when its not. Which is generally a knock on the door from the postman.. @#$!@%#!@%!#@

My mate gremlin has disappeared in the same way it appeared, I have a suspecting theory that his newly installed Solar Power System is to blame. Or its just coincidence that the sun goes down and his system is running right again.

Cheers,
Steve
I've cut and pasted in a knowledge folder I keep the information you have supplied - I hope you don't mind? If you have any objection to that let me know.

Cheers,
Steve

[RCaffin]

Hi Steve

Cut and paste as much as you want! You're welcome.

I mentioned my academic background because it often strikes me just how hard it must be for novices with little or no knowledge of electronics when they want to 'build their own CNC'. I struggle at times to debug the damn thing, and I have 40+ years experience in electronics.

You must have one hell of a noisy spindle to do so much work around it. Or are you just being pedantic to make sure?

Chuckle. Perfectionism is certainly part of it, but ...
You see, my CNC is more of a VMC, and I routinely set the zero to within 5 microns. It can hold that accuracy over a few hours, and needs to. But then, after replacing a dead spindle driver recently I noticed that my Z axis was drifting by >40 microns over a 15 minute machining cycle. Noise. Not good enough!

I have a 180 V 0.5 kW industrial DC brush motor on the spindle. MUCH grunt. Glorified commercial triac lamp dimmer circuit direct off the mains for the spindle power supply. Electrically fairly quiet, but the rectified 100 hz into the motor made it sound rough. Then a triac died, mit spitzen und sparkzen. I 'upgraded' to a more modern MOSFET chopper drive, which was certainly far, far nicer to the motor, but the RF from the chopped MOSFETs was incredible. And the Z axis went wandering. So, all the electrical system is being upgraded, thoroughly, to ensure reliability. Isolation, separation, screening, earthing ...

The solar power system may also be using a chopped MOSFET drive to pump 240 VAC back into the mains. IF the problem seems sync'ed to the sun, that is probably the source. A very large 'pi' filter on the mains inlet to his machine might help.

Cheers
Roger

[Beefy]

Another good thing about Uccnc and the UC300/100 is they are made by the same company. This means you don't have one company passing the blame to another company when something isn't working right. Example, the Mach people could blame the UC300 when something isn't going right, or the UC300 people could blame Mach3. I read a forum post where this was happening and who are we to figure out which case is true.

Candcnc uses the Smoothstepper on one of its systems. It came to light the Smoothstepper could not do the M10/11 commands fast like they should (M10/11 are fast laser commands) and on the Candcnc forum it was declared there's no easy fix for this problem. Don't know if that has changed but last time I checked the thread it wasn't. I asked Balaaz if the UC300 can do the laser commands as they are intended and he assured me they can. I've also found the communication better with the UC300 company. Last time I emailed Smoothstepper I was ignored so I quickly forgot about using that board.