[ger21]

I haven't used it on a machine yet, only for testing. I have a UC300ETH, and will be using it on my new router, whenever I find time to build it.

I haven't done a Vectric post yet, as nobody's has really asked. It's pretty much the same as the Mach3 post, except for the Park position M code.

[Bad Wolf]

Just a quick update since I posted this a few years back, I have lost all interest in other software and have stuck with UCCNC. In fact I have just bought a UC300ETH further committing myself to the UCCNC platform. I haven't had chance to fit it yet and there really is no hurry as the UC100 hasn't missed a beat.

[ger21]

Wait until they get the S-Curve planner done

[ger21]

I have read mention of the UCCNC not supporting the 4th axis

It's not that it doesn't support a 4th axis. It doesn't have support for mixing rotary and linear feedrates, and does not display a rotary toolpath properly. But it will work with a rotary 4th axis.

Is that different than slaving the A to the X motors so you can have 2 motors drive the X axis?

Yes, it's different. UCCNC supports slaved axis just fine.

[Bad Wolf]

It doesn't have support for mixing rotary and linear feed rates,

Please correct me if I am wrong but what I think you are saying that it can control X,Y,Z,A at different speeds, and it can control all four axis at the same time, but it lacks the intelligence (for the want of a better word) to factor in how fast or slow the A axis needs to go to create a smooth cut?

[battwell]

i mainly make cabinet parts with uccnc- where everything must clip and fit together- so accuracy is paramount.
not once have i had to switch to g61 exact stop mode :-)
i upgraded from usb 300 to 300eth (as id nagged balacz to make it in the first place. -and so glad he came through with such a great product)
i leave my pc on all the time, unlike mach3 it doesnt require restarts after a few days when things stop working etc. uccnc only gets restarted when i add a new plugin etc now
here is a video of it running a nested job- its doing this day in day out - couldnt be happier!
im running uccnc with uc300eth, 3 cheapy £5 chinese break out boards, mitsubishi servos in position mode.
ive since added a toolchanger rack and have tool length probing for fast start work with any new tool.
https://youtu.be/veb6rAmkS9Y

[victorofga]

by the way if you can adjust in mach3 the segment lengths that producing the corners then mach3 can be very smooth

[ger21]

Sort of.
Say you command a feedrate of 100ipm. When doing moves that include the rotary axis, the feedrate can be higher or lower, depending on where the Z axis is.
Basically, it can't maintain the programmed feedrate when using a rotary axis.

This is on their "To Do" list, and hopefully they'll get to it later this year. They've added a lot of features in the last year, and have spent the last few months adding G41/G42 cutter comp.

[jfong]

What a DSP/FPGA or microcontrolller does well is the fine grained timing requirements that is necessary for low jitter step pulse output.

There are some step generation hardware that is good to several MHz. I use a uc300eth that can do couple hundred kHz step pulse. A super fast PC cannot do that without using some kind of external hardware or pci card.

[Bad Wolf]

What a DSP/FPGA or microcontrolller does well is the fine grained timing requirements that is necessary for low jitter step pulse output.

There are some step generation hardware that is good to several MHz. I use a uc300eth that can do couple hundred kHz step pulse. A super fast PC cannot do that without using some kind of external hardware or pci card.

Sort of see where your coming from, but this chip isn't processing data, it is timing data, unless "processing" is being used in a generic sense as a given amount of throughput.
I know a lot of the people who understand audio far better than me talk a lot about timing, so I could see how an external chip to perform timing could be beneficial for this purpose.

[ger21]

How the trajectory planner is coded has a lot more to do with how the machine runs than the processor.

[Beefy]

As a calculated GUESS I would also tend to think it's all about timing, and not data throughput. Yes an ethernet connection is super fast with data throughput, but I bet it can't time that throughput precisely in a way that's needed for step pulse generation.

Windows PCs are also multi-tasking machines where they jump from one process to another. And don't ethernet connections send data in "packets". If a PC could deal perfectly with timing I'm sure we wouldn't have any external motion controller boards.

[A_Camera]

As a calculated GUESS I would also tend to think it's all about timing, and not data throughput. Yes an ethernet connection is super fast with data throughput, but I bet it can't time that throughput precisely in a way that's needed for step pulse generation.

Windows PCs are also multi-tasking machines where they jump from one process to another. And don't ethernet connections send data in "packets". If a PC could deal perfectly with timing I'm sure we wouldn't have any external motion controller boards.

Not only that, but Ethernet packets collide all the time, so there is a collision detection procedure which makes sure that packets are sent and received without error, which means some packets are resent. Packets are also not always sent in sequence, so it is pretty complicated. The motion controller board receives the packets, assembles in the right order and sends the timed signals to the steppers. There is probably no other processing, processing takes place in the UCCNC software.