Honest Appraisal of UCCNC

[Bad Wolf]

I was actually referring to how the UCCNC speaks to the UC100 in the case of the software talking to the motion controller.

[Bad Wolf]

Just casting my mind back a decade to when I wrote some software that auto tuned an after market ECU (via a USB to Serial adaptor). After much begging the manufacturers gave me some protocols to interface with the ECU.

I am figuring this is probably similar in that the parallel port break out board has fixed pins on which it can receive 0 and 1s to trigger a particular movement, so regardless of software or usb / parallel motion controller in front of it, everything after the parallel port plug must be the same (or otherwise it would be doing something different)

But what the different software output through the USB could be vastly different, hence why the firmware needs to be changed between uses of the different software, to interpret the command correctly.

Am I close?

[BanduraMaker]

Looking for perceived or measurable differences between the two, and if one is favoured over the other based on performance rather than bias.

I know I'm late to the party here but I thought I'd throw in my two cents.

I've run Mach 3 for about the last 4 years and recently switched over the UCCNC and have directly compared the two. In fact, I made a video smackdown here:

To summarize, the UCCNC trajectory planner is far superior to the Mach 3 planner, especially when using CV mode for complex surfacing. The Mach 3 TP has a bug in it where it may violate acceleration settings by up to +100% which can cause the machine to stall if you don't have enough safety margin on your acceleration settings.

Additionally, Mach 3's lookahead settings are rather cryptic and you don't really know how tight a tolerance you're cutting to when using Mach 3. e.g., to get the accuracy I need for my parts, I had to drop the lookahead all the way down to 10 lines which resulted in very jerky motion of the machine with visible marks on the parts. This also increases the machining time significantly compared to smoother mothion. Motion can be smoothed by increasing the lookahead but then the parts are not machined accurately.

UCCNC has settings to define how closely to the toolpath it will actually be following and it in fact does. Parts machined using UCCNC are coming out more accurate and more quickly than when using Mach 3.

The bad news about UCCNC: It's young software and v1 still has bugs in it. I've been using UCCNC for a couple months now and I've found some strange behavior. The software can sometimes start hanging when calling a lot of macros and I've had some very strange behavior where the machine seems to forget for a single line whether it's in machine coordinates or work coordinates. This particular bug hasn't caused any problems as it seems to only affect the Z axis at the start of programs.

There are some other minor bugs e.g. soft limits have some issues when using offsets...sometimes...

At the end of the day, I'm sticking with UCCNC over Mach 3 because of the cutting performance. UCCNC has a better trajectory planner period. What I've also learned is the first symptoms of the UCCNC software thinking about going into the weeds is that the M-codes have a bit of a delay before getting started. If I detect this at the start of a job, I'll re-start my control computer and try again. If it happens in the middle of a job, I let it finish and then re-start.

I'm really hoping that the next version of the software fixes these bugs as they did cost me a couple of broken tools early in the process.

I've also recently purchased another machine that's running Kflop and KmotionCNC software so as soon as I can get off my lazy rear end and get that machine running, I'll be able to compare all 3 packages.

[Bad Wolf]

Hi Andy, thanks for chiming in. To be honest I actually saw your video before I purchased the software, and promptly dismissed it. I thought you were working for the company and were just giving it a big rap. Because you have the comments disabled, I figured you might know people would dispute what you were saying, and as such really couldn't put any stock in your recommendation without hearing both sides.

But I have seen other posts of yours here and come to realise that you are just an enthusiast like the rest of us. Mach 3 is known to have issues with trajectory and that's is one of the huge improvements in Mach 4, pity you couldn't do a similar test against Mach 4, but if nothing else you are willing to say its better than Mach 3.

[ger21]

Fwiw, I've seen Andy's machine running in person with both Mach3 and UCCNC, and it's more apparent in person than in the video how much smoother UCCNC actually is.

[victorofga]

Gerry

I might don't see it right.. but...

mach3 can be used as using the main processor (of computer) for clock and timing..
versus uccnc as I saw on thei r website, that can be used only with their ""plug""
and if so, then the two cannot be caompared because an independent chip don't have interrupts from windows..

I believe the result could be same, under same circumstances.. if mach 3 used in conjuction with driverboard..

but again, I might wrong..

just a sidenote, after using mach3 without driverboard, now im using ncstudio..
im quite satisfied with..... due my signals coming from the pci card..

with mach 3 , I experienced very bad things, due windows can not be turned off.. sudden movements, were looking some blocks were skipped..
what coldnt be occur with an independent driver board..

[dubble]

The difference is not in the timing, but the trajectory planner works different.
By the way the UC100 motion controller (which you've mentioned as "plug") works with both Mach3 and the UCCNC,
if using the UC100 with Mach3 then the timing is the same precise because then the timing is done on the DSP of the UC100 and not on Windows.

Don't get me wrong, but our company exchanged lots of NCstudios for customers on machines imported from china, that software has lots and lots of problems...

[dubble]

Just got this video, rigid tapping with the UCCNC on a customer's machine.
The spindle is a 4kW AC motor, it has a 100 line encoder feedback to the UC300-5LPT motion controller.
The tapping is done with G33.1 / G33.2 codes.
The motion controller keeps the syncron between the spindle rotation and controls the Z position.

[victorofga]

my concern is trying to void using control directly under windows..

I don't want to enter in discussion what program better..

approaching this question from differently, artsoft or your company cncdrive has reasons why they made trajectory planner as it made.. I believe in expert hands both can be set same effectively..

many ncstudio user not capable to setup properly ncstudio, either way

because they have a Chinese version

or simply set wrong start and connection accelerations..

or simply they cant start to work with any way..

many user changed ncstudio by a fashion manner.. and it doesn't shows mach3 or cncdrive would be better..

my point is, again, with ncstudio I have pci card, a lot safer than using mach or any controlprogram simply under windows trough on db25 or usb..

another point about ncstudio, most folks identify ncstudio with the clone, what can be purchased for 50 dollar..
it contain a card, breakoutboard, cable and ncstudio 4.5 or 5.5..
but this is the clone version..
made by former weihong employees

if you go to weihong website you can see there are more than simple 3 or 6 axis control..

[dubble]

You don't understand how the system works, again: The pulse generation does not happen on Windows, either with the UC100 + UCCNC and with UC100 + Mach3, all pulses are generated on the external DSP of the motion controller (you refered that as "plug", but it is not just a plug, but a high power DSP board inside the tiny connector enclosure), so the timing does not rely on Windows at all and the timing is very precise in both cases, the jitter is in the 100nsec range.
As ncstudio using a PCI card the pulse generation happens on the PCI card which probably has an FPGA as the processing unit which handles the timing ... so there is no difference, in all systems the timing is handled by an OS independent processing unit.
Ofcourse I was not talking about Mach3 + LPT port where the timing relies on Windows ...

The big difference can be is the software itself (means GUI, functions, flexibility, etc.) and how the motion path is precalculated by the different softwares, that can make a huge difference.

[dertsap]

its great that its fast and has all the gadgets but how long do you predict that it will be before z arc's are going to be implemented (g18-19) , d comp and other typical g codes ?

[dubble]

The G18/G19 is on the todo list with high priority.
The tool radius compensation is on the end of the same list, as today's CAM softwares and post processors can do the job and because only very few customers asked about it and also because implementing this function is hard.
It can be only important on mass production especially on lathes, but currently the UCCNC's maintarget is CNC mills, for lathes we still need to do a few other things (like lathe screenset) to give full support...

Other typical codes? Please give me a list of codes you are interested in, I do not know what else missing.

[victorofga]

I call it simply plug.. yet im clear about it is not a simple usb-db25 converter..
in th evideo is not detailed if mach3 were used in conjuction with your dongle, or dsp

yes im very clear mach or your program calculate a """virtual toolpath""" called trajectory where each point is precalculated..

its all about settings..

acceleration deccelration

also the corners precision.. make it tight permitting only a very small diff will result jerking
and its beyond the cv settings by angle..

great effect on the smoothness the virtual steps by the motor 1/16 steps wont result so smooth movement than 40K microstep..

the trajectory planner even it would be """better""" wouldn't ensure a smoother tool movement
there are too many factor has effect on how smooth a movement on a cnc

and im closing here.. because theres no way you can tell us why yours or any program trajectory planner better..
you cant just post a part of the programs and analyzing why your algorithm faster or better...

unless fenerty set up a mach on same machine and you... we wont see it right..
we haven't any clue how we could program something like this..

in fact the cnczone 99 percent user have no clue.. how and where we could start to make a program like mach or yours..
that's why its a dead end street to talking about what program better..

its indeed about users... what you can use, might others cant..
what you like program might others don't..

but beyond everything its a great performance and hat off you all made another controlprogram..

on other hand... I can make a youtube video, install on a win 98 turbocnc and it works pretty neatly..
then I install cncdrive and it wont work at all.. no way..

then I can say turbocnc better because it works?
and yours don't?

there are too many the variable around.. just simply comparing programs...

[ger21]

If I were you, I would target router users, as they seem to be the largest market by far.
Especially the inexpensive chinese imports that are all over Ebay.

As far as required features. If you want this to be successful on a large scale, then you need to support all standard features, including G17-18-19 and G41/G42.
As a router user, I use G41/G42 all the time to compensate for resharpened tool diameters. But I understand that many don't use it, so it's not a priority.

If you're not going to have a complete package that can compete with Mach4 and others, then you'll be limited to users with less demanding needs.

You're in a unique situation, as your both in a sort of partnership with Artsoft as well as competing with them.
Unless you're software is competitive with Mach4, you'd probably be better off working on Mach4 plugins, which will sell more hardware.
Mach4 seems to be starting to get a lot more use in recent weeks, but it's still a little bit away.
It looks to me like whoever get's their software finished first will have a huge advantage.

[mactec54]

Fwiw, I've seen Andy's machine running in person with both Mach3 and UCCNC, and it's more apparent in person than in the video how much smoother UCCNC actually is.

There is a big difference using the UCCNC control, But when you use the UC100 with Mach3 then you get the same performance from Mach3, as you do with the UCCNC control, the UCCNC control is new & is getting better, add the smoothstepper to Mach3 & you are way ahead of the UCCNC for performance

I have tested all on the same machine, UCCNC, Mach3, & Eding Plus some others, all licensed copies, Eding is hands down the best control at this time, the Eding control still needs some minor tweaks also, but it is superior in performance, operation & setup

I will be redoing/testing Mach4 soon, I played with Mach4 about 31/2 years ago, it was & looked very good then, I know they have rebuilt a lot of it since then, what I have seen though, it still looks like the same, layout,as when I tried it out, it should be interesting to see how Mach4 stacks up against these other controls

[Bad Wolf]

Gerry I hypothesised that scenario in a previous post and deleted it as ultimately we all have to take the path we think is best for us. I would love to see a M4 / UC100 plugin even if I don't ever use it. You are right about the amount of people buying cheap Chinese CNCs, and the DIY market isn't that small either.

[Bad Wolf]

There is a big difference using the UCCNC control, But when you use the UC100 with Mach3 then you get the same performance from Mach3, as you do with the UCCNC control, the UCCNC control is new & is getting better, add the smoothstepper to Mach3 & you are way ahead of the UCCNC for performance

I have tested all on the same machine, UCCNC, Mach3, & Eding Plus some others, all licensed copies, Eding is hands down the best control at this time, the Eding control still needs some minor tweaks also, but it is superior in performance, operation & setup

I will be redoing/testing Mach4 soon, I played with Mach4 about 31/2 years ago, it was & looked very good then, I know they have rebuilt a lot of it since then, what I have seen though, it still looks like the same, layout,as when I tried it out, it should be interesting to see how Mach4 stacks up against these other controls

Look forward to seeing what you think (about new version of M4), after all a lot of better or worse is perceived by the person using it, and not always entirely derived from performance based testing.

[victorofga]

its probably confusing for many
older controls were working withouth continuous velocity
they slowed down at each node, line or arc endpoint, to zero, then started the new segment

modern programs with CW, continuous velocity didn't stopped at nodes, but pending on the included angle they continued with same speed

it can result 10 times or more speed up within a work

but..
with continuous velocity the corners will be rounded, pending on the included angle and ""corner precision""

the picture shows it..

and by this achieving a given velocity corners will be rounded..



how much? it has affect first the motor resolution, how much it can step per revolution..
also effect the corner precision, what you can set tight and machine wont ""get"" the corner but will ""jerk""

and by this, I truly believe all trajectory planner will result same..

because no matter what calculated the relity will decide what speed a machine can achive..

on the picture I tried to show two type corner rounding..

I called ""dynamic"" due it changing by settings.. this is what you can observe actually when speed set too high compared other settings..

and again in the video nothing shows mach3 were used conjuction with any driverboard
and if so, then windows interrupts will mess up the best calculated toolpaths and it wont be smooth but noisy..

while a driverboard, or call ""dsp processor"" will work ""outside"" of windows interrupts..


edit
I just saw mactec posted..... exactly what I saying.. make same conditions, circumstances and we will see it..

no one questioning cncdrive is an outstanding performance to make the program itself.... just the qualification somehow wasn't right..

[dubble]

Hi Gerry,

Yes, I understand that the G17-G18 can be an important feature... again especially for lathes it can be really important, because they work on the XZ plane.
However we have a bunch of serious machines here, most of them are in all day use, all running the UCCNC now and nobody asked me about the G41/G42 yet,
also the number of guys who asked about these codes is very minor, let's say 1 out of 500 people, so this is why we put the G41/G42 to the end of the todo list.
And yes, I agree with you that the G41/G42 can be a great tool to have in some cases, but only if it is implemented correctly, to be honest with you I don't know any hobby software on the market today in which this function works perfectly in all circumstances.

And you are right, we are in a unique situation, but we are not the only one, there are other motion controller manufacturers who has plugins for Mach and also their own softwares ... I don't want to say out names here, but I'm sure you know who I'm talking about.

Also we are in a really great friendship with the Artsoft guys and we love Mach3 becaue when it appeared (it was called something like Master-something, uhh, can't recall the first name of Mach now, but I'm sure it had a different name than Mach.) it was a unique software on the hobbycnc market with lots of features which nobody thought about before.
I'm also sure that Mach4 will be a great software.

So, we don't see this as a competition, we do the developments for the lots of fun and challenge.

[dubble]

Nice explanation Victorofga.
Just to extend your explanation a bit:
The corner rounding is one part of the trajectory planning, there are several other things a trajectory planner does.
Even very complex algorithms can be implemented to do better and better CV path blending, but ofcourse there is a tradeoff, because the processing unit which calculates the path (being it the PC or the DSP or FPGA) has resource and processing capability limits,so the calculations cannot be more complicated in any circumstances which the device can calculate in a reasonable amount of time.
So, ofcourse the CV planner not only rounds corners, but a good planner optimises the paths at many more points using many different algorithms ...

Therefor it cannot be said that all CV planners work the same, I venture that on this hobby cnc software market there are no 2 totally identical.