[A_Camera]

I marked this with the prefix "Newbie" because I am new to UCCNC and the UC300ETH...

I received my UC300ETH a few days ago and first yesterday I had some time to test it a bit more seriously together with the UCCNC. I was very pleased with some functions, which I could not test with the demo version before, but what impressed me most is the huge speed I can get out of my CNC with this combo. Previously, the top speed for rapids was 5500mm/min, which I found very fast, so I never thought I could to that but never the less, I increased to see what happens, when will my motors start stalling. So I increased the speed first to 6000mm/min, then to 7000mm/min and played around for a while. Nothing bad happened, so I increased to 8000mm/min and... still nothing, no stalling. I just could help myself, played with the X and Y watching the whole table and the Z flying across at this, for me incredibly high speed. Not a single glitch or a sign of struggle... Didn't really reach top speed yet, no stalling, so I will continue playing tonight.

I changed nothing else except the motion controller and the software. Maybe the same speed would be possible with the same setting even with Mach3, but for now, there is no plugin for Mach3, so I can not test that. I can also not test how UCCNC with the UC300USB because my UCCNC license works only with the ETH, so I can't really compare apples with apples, but it certainly proves that my machine is not all that bad, it can run very fast without any additional hardware upgrades. Considering the fairly small table I have, 8000mm/min is pretty scary fast and before I keep this, or even higher speed, I need to test the reaction speed of hitting the limit switches. I don't want inertia to destroy those switches or smash the table or my Z into the sides, so that is definitely needed to be tested as well, especially at these speeds.

Another speedy reaction I noticed is at probing. Perhaps that is just my imagination, but I have a feeling that the UC300ETH and the UCCNC are much faster in detecting the probe contact and stopping than Mach3 was. I still need to get this tested out, but playing with it certainly gave me that impression, and if I can confirm this, it would be great. OK, maybe no surprise, since UCCNC is a much more modern software compared with Mach3, and is not written for parallel port at all, so it should be faster, but anyway, it is good to know.

Other things I like about probing is the possibility of probing with jog. It is a very simple, but genius idea because it makes probing so much easier in many situations. No need to type anything, just jog and the machine stops when the probe gets touched.

Things I don't like is the discovery that it does not support radius compensation. I initially thought that this is no big deal, I can wait until they fix it. Unfortunately, yesterday I thought I might as well start using the new setup for real, but I hit this problem immediately because the code I thought will be the first to use needs the radius compensation. This would have been fine, if that was the only file, but then I loaded another, and another and another... I guess you get the picture. Most of my files use it, and regenerating the code for all of those is not going to be easy, so I probably have to wait for the Mach3 plugin and continue using Mach3 until G40, G41 and G42 are implemented. I knew that this was not supported in UCCNC but never really realized how dependent on them I was.

Never the less, the UC300ETH with the UCCNC is a real speed monster, and I hope that it is also with the Mach3 and that the Mach3 plugin becomes available soon.

[jfong]

Yes with the low pulse jitter of these motion adapters, you are able to spin the motors much faster. I still haven't plugged in my uc300eth to bench test it yet.

I probably have several dozen different stepper drivers purchased over the years. I test them out using a Tektronix pulse generator. With a very good pulse stream, I've spun motors over 9000rpm. That 300khz pulse with a geckodrive. Using a good motion adapter will make the motor spin much more smoothly.

[Beefy]

Yes with the low pulse jitter of these motion adapters, you are able to spin the motors much faster. I still haven't plugged in my uc300eth to bench test it yet.

I probably have several dozen different stepper drivers purchased over the years. I test them out using a Tektronix pulse generator. With a very good pulse stream, I've spun motors over 9000rpm. That 300khz pulse with a geckodrive. Using a good motion adapter will make the motor spin much more smoothly.

Got to 2nd that.

When I first got into this game and didn't know anything I bought some really crappy Chinese stepper motors ("Smooth" brand) off Ebay. Ha ha, yeah right, "smooth".

My first stepper driver was an electronics kit with some stepper driver ICs which did full or half step. I got up to about 2 or 3 revolutions per second before the motor stalled.

Then I went to one of the many double the price Australian suppliers and got a Leadshine stepper driver unit. In Mach I had to set the step pulse length to 8 to get the Leadshine drive to accept the pulses. I could only get a few rpm before the motors would squeal and stall, that's on the bench with no load.

Then I heard about Gecko drives and bought a G203V. The price was about half what I paid for the Lead****e (spelling mistake ??) and it ran my crappy "Smooth" stepper very well. Also the step pulse length is set at 2 now. The rapids on my plasma table are set at 10,000 mm/min. I tested this a long time ago with a crazy zig zaggy test run that went all over the area of the table. Was running for about 20 mins. Afterwards did a "verify" to confirm no lost steps and everything was good.

So yes I have to say my experience has been to get good quality stepper drivers.

Keith.

[OlfCNC]

I changed nothing else except the motion controller and the software. Maybe the same speed would be possible with the same setting even with Mach3, but for now, there is no plugin for Mach3, so I can not test that. I can also not test how UCCNC with the UC300USB because my UCCNC license works only with the ETH, so I can't really compare apples with apples, but it certainly proves that my machine is not all that bad, it can run very fast without any additional hardware upgrades. Considering the fairly small table I have, 8000mm/min is pretty scary fast and before I keep this, or even higher speed, I need to test the reaction speed of hitting the limit switches. I don't want inertia to destroy those switches or smash the table or my Z into the sides, so that is definitely needed to be tested as well, especially at these speeds.

I experienced the same and what I already figured out about this is that UCCNC has a far better motion planner than Mach3. While Mach3 often struggles to keep speed on the paths and jerks a lot in corners UCCNC runs still nice and smooth there.
Maybe the better pulse timing also plays a rule in this but I beleive the overal better performance is mostly due to the better path planning.

[A_Camera]

Yes with the low pulse jitter of these motion adapters, you are able to spin the motors much faster. I still haven't plugged in my uc300eth to bench test it yet.

I probably have several dozen different stepper drivers purchased over the years. I test them out using a Tektronix pulse generator. With a very good pulse stream, I've spun motors over 9000rpm. That 300khz pulse with a geckodrive. Using a good motion adapter will make the motor spin much more smoothly.

9000rpm would give me 45.000mm/min... Now, I don't think I can ever reach that on my fairly small machine with my steppers, but I will find the maximum speed as soon as I have time. Yesterday I had no time to play with speed, but maybe during the weekend. Never the less, 9000rpm is I think simply too much for me. The slightest mistake and there is no chance of stopping the motion since my table is only 450 x 300mm, so with that speed it would take only about a half second to cross the table, and even if acceleration and deceleration takes some time, it is simply too fast. ...but it would be fun to try and see if possible. so I'll definitely try and if I can reach that, I will document it.

BTW, my stepper drivers are just simple and cheap DQ542MA, so not as fancy as yours. Pulsing on the UC300ETH was set to 400kHz. The specifications of my machine is documented here:

Adapting Camera: CNC upgrade progress (mechanical parts)

Skip to the end if you are not interested in reading, just want to see the specs.

[daniellyall]

I am waiting to see what Ger21 and terry have to say before I move over, but if you can in-cress you speed that much with not to many problems, it's making it's case over Mach4 more inviting

[A_Camera]

Got to 2nd that.

When I first got into this game and didn't know anything I bought some really crappy Chinese stepper motors ("Smooth" brand) off Ebay. Ha ha, yeah right, "smooth".

My first stepper driver was an electronics kit with some stepper driver ICs which did full or half step. I got up to about 2 or 3 revolutions per second before the motor stalled.

Then I went to one of the many double the price Australian suppliers and got a Leadshine stepper driver unit. In Mach I had to set the step pulse length to 8 to get the Leadshine drive to accept the pulses. I could only get a few rpm before the motors would squeal and stall, that's on the bench with no load.

Must have been something wrong there. The drivers I have are Leadshine fakes and they work excellently, have been working for a year now without a single issue.

Then I heard about Gecko drives and bought a G203V. The price was about half what I paid for the Lead****e (spelling mistake ??) and it ran my crappy "Smooth" stepper very well. Also the step pulse length is set at 2 now. The rapids on my plasma table are set at 10,000 mm/min. I tested this a long time ago with a crazy zig zaggy test run that went all over the area of the table. Was running for about 20 mins. Afterwards did a "verify" to confirm no lost steps and everything was good.

So yes I have to say my experience has been to get good quality stepper drivers.

Keith.

I am using DQ542MA drivers and I am very happy with those. Have 5, but only use three in my CNC, of course, I tested the other two as well, so they are just as good. Where could you get the G203V cheaper than the Leadshine? Either you paid not only twice the price of the Leadshine, but many times more, or you found some real crazy person who sold you the Geckos and missed a zero at the end of the price tag or something similar. Noramlly the Gecko is sold for about $350 while the DQ542MA for $35 which is about half of the original Leadshine price, so the seller must have lost his mind if he sold you Geckos for that price.

I will make the tests like you did as soon as I have time, but the few cross runs on my table definitely not indicated any lost steps at 8000mm/min, so if I can keep that I am pretty happy. Have no real motivation to spend over $1k when I can get the same for about $100, but sure, is I could get the G200V for less than the Leadshine original costs, I'd be willing to buy them just too how they really are...

Perhaps the problem you had was actually with the Smooth stepper, not the drivers... Perhaps the Geckos are more tolerant to noise or whatever, which makes them more usable. I would not dismiss Leadshine just because it is made in China...

[A_Camera]

I am waiting to see what Ger21 and terry have to say before I move over, but if you can in-cress you speed that much with not to many problems, it's making it's case over Mach4 more inviting

It's OK if you trust them more than you trust me, I really don't mind, but why post it here? What's the relevance and the connection to this thread? I am not trying to convince you or anybody about anything, just describe what I found. Also, even if they would decide that this is not for them or they can't reach the same speed or totally disagree with my conclusions it is totally irrelevant for me, isn't it? What I found is about my setup, not theirs.

Never the less, if you consider the UCCNC you should also consider that it is currently free of charge if you buy the UC300ETH, so basically it is a deal difficult to turn down. But... beware of the still missing G40, G41 and G42 support, and currently no one at CNC Drive has an idea about when it will be ready because apparently it is pretty complicated, so if you are dependent on those codes, like I am, then you are stuck with Mach, unless you want to regenerate all the G-code without cutter compensation. I don't know if the UC300ETH will have a Mach4 plugin, they are working on a Mach3 plugin, which supposed to be released soon.

[A_Camera]

I experienced the same and what I already figured out about this is that UCCNC has a far better motion planner than Mach3. While Mach3 often struggles to keep speed on the paths and jerks a lot in corners UCCNC runs still nice and smooth there.
Maybe the better pulse timing also plays a rule in this but I beleive the overal better performance is mostly due to the better path planning.

It can be so, but as soon as there is a Mach3 plugin I'll redo the tests. Perhaps it will not be able to drive as fast, but I suspect that even the UC300ETH plays a role and suspect that it is in many ways better than the USB version of UC300, which I had before. Mach3 parallel port drive is a dead end, I never really used that because it was behaving like you describe, struggling a lot, but have not seen that with the UC300USB. Never the less, I expect that the UCCNC is in many ways better than Mach3, though it is not fair to compare the two, if compared it should be compared with Mach4, which I have no idea about how that is.

[jfong]

A_camera,

9000rpm was just a bench speed test of the motor. Nothing connected to motor shaft. Just wanted to see how fast a very low inductance stepper can run with a geckodrive. At that rpm, there is no torque available anyway.

With the same motor on the gantry router, I get 300ipm rapids with the stepper spinning at 1500rpm. That's about the best that motor can do with ballscrew attached.

I switched to servo motors for even faster speed.

[dontpanic]

Beefy you must know there are lots of types of STEPPER MOTOR brands, not all perform equal with the same drive.

A stepper + a drive is a marriage, you have to find the right match. One of the hardest part is to get impedance
of the windings match to the driver electronics, and the other part is the motor inertia and finally the inertia of the
complete axis (if it has mech.reduction etc) which will affect the resonance of the system. Resonance is a BALANCE
that will provide that your system will not vibrate too much and will not run smooth.

The best choice of buying a stepper system is ALWAYS pay attention to the manufacturer. If you can buy it in
combo motor+drive you will get the most out of the system.
Leadshine is an excellent brand , but it will probably not perform well with XX types of motors.

Gecko 203V has done a remarkable work with its mix of micro and full stepping in an intelligent approach.
But even G203V have a best match motor, that will run better than none. I will put pacific scientific and annaheim
steppers as one of the best in quality and performance.
(I had the chance to experiment and use for 15 years ago Swiss and German steppers, that were called Hybrid steppers
with 5 phases windings and 3 phase high voltage drives , this was the cream of the cream in stepper techn. but they were
10 times the price of a modern gecko drive !).

[dontpanic]

Jfong
there is a limit of axial rigidity in a leadscrew system that can not be braken even with high speed servos.
Because of its long length with respect to other components, the ball screw is typically the greatest source of deflection.
And it is not possible to use ballscrews at high speeds more than 2000 rpm is difficult to see. Depending on the mass (diameter) and the length and the
assembly of your screw if it is a double support with fixed endings f.e.), you will have a critical speed of the screw that can
not be overcome without entering in resonance and deflection of the screw.

[ger21]

I am waiting to see what Ger21 and terry have to say before I move over

While it may be a while before I get a machine running with UCCNC, I'm all in right now. I have my UC300eth, and am currently working on a screenset for UCCNC.

It's highly unlikely that my Mach4 license will ever get used on a machine. They've made far too many decisions that I disagree with.

I don't know if the UC300ETH will have a Mach4 plugin

I think that it's highly unlikely that CNC Drive will write any Mach4 plugins for any of their motion controllers.

Perhaps it will not be able to drive as fast, but I suspect that even the UC300ETH plays a role and suspect that it is in many ways better than the USB version of UC300

I would expect their performance to be similar, except for the higher pulse rate in the ethernet version. I think that the trajectory planner can make a big difference with the correct settings.
A friend of mine experienced a 30% speed increase for UCCNC over Mach3 with a UC100.

[jfong]

Jfong
there is a limit of axial rigidity in a leadscrew system that can not be braken even with high speed servos.
Because of its long length with respect to other components, the ball screw is typically the greatest source of deflection.
And it is not possible to use ballscrews at high speeds more than 2000 rpm is difficult to see. Depending on the mass (diameter) and the length and the
assembly of your screw if it is a double support with fixed endings f.e.), you will have a critical speed of the screw that can
not be overcome without entering in resonance and deflection of the screw.

Yes there is a limit. They are 20mm NSK ground ballscrews with bearing supports on both ends. X&Y axis riding on THK linear rails very smoothly. Moving to servo was more for the slightly higher acceleration than rapid speed and I got the motors/drivers really cheap on eBay anyway. My biggest problem was the poor Mach3 CV. The machine was never really able to get fast continuous motion when doing complicated 3D profiling. I think the fastest I've seen Mach3 move simultaneous 3 axis was about 170ipm. It just does poorly when there is a bunch of short line segments.
I still haven't hooked up the uc300eth yet but I'm sure it will do much better than MAch3 with its newer CV implementation. I would be happy to get 250ipm continuous 3 axis fluid motion reliably with my machine. That will be 1250rpm for the ballscrew. I really don't care to about single axis rapid speed as much.

Edit Mach3 is capable of moving a CNC faster than 250ipm. Just my particular machine with its heavy gantry (I guess about 100lbs-125lbs), had issues with getting good results at higher cutting speeds. Sloppy corner rounding and not particularly accurate path following. Everything I see points to Mach3 CV issues. Atleast that is what I have been told and have seen on various YouTube videos. I have to slow down the Machine to get good results with Mach3. I'm hoping that UCCNC will be able to do better.

Edit2
A couple of years ago I had borrowed/tested a UC100 USB adapter. The axis did run smoother due to the better pulsing but it did not alleviate any of the Mach3 CV problems I was seeing. I did not have a chance to test UCCNC back then.

[daniellyall]

It's OK if you trust them more than you trust me, I really don't mind, but why post it here? What's the relevance and the connection to this thread? I am not trying to convince you or anybody about anything, just describe what I found. Also, even if they would decide that this is not for them or they can't reach the same speed or totally disagree with my conclusions it is totally irrelevant for me, isn't it? What I found is about my setup, not theirs.

Never the less, if you consider the UCCNC you should also consider that it is currently free of charge if you buy the UC300ETH, so basically it is a deal difficult to turn down. But... beware of the still missing G40, G41 and G42 support, and currently no one at CNC Drive has an idea about when it will be ready because apparently it is pretty complicated, so if you are dependent on those codes, like I am, then you are stuck with Mach, unless you want to regenerate all the G-code without cutter compensation. I don't know if the UC300ETH will have a Mach4 plugin, they are working on a Mach3 plugin, which supposed to be released soon.

Dude I did not mean I don't trust your word or any disrespect at all, I was just meaning you are seeing A big improvement, if ger and terry say the same thing that is 3 people haveing the same outcome, I trust ger and terrys advice like most people who know who they are, I have seen you around here long enough to know you are not full of **** at all and give good advices as well. the relevances is if you plus many more have the same thing it's gold. and my setup is similar to yours I have 213V

[robertspark]

daniellyall

Uccnc is much smoother than mach3 (even with an ess) [personal experience]

There is a YouTube clip showing a motion comparison. Mach4 was fairly smooth too... but just no ess support for my plasma functions, and every time an update landed the screenset was different (couldn't run a custom screenset)... plus something stopped working in the ess maybe

note, to be unbiased, there is no indication of what motion controller he is using (it may be PP or more likely the Mach3 UC plugin [s its easier to compare directly]
https://www.youtube.com/watch?v=UJ8NM1MVFdM

As you know it's not about speed, the name of the game is acceleration and accuracy (repeatability). Doesn't miss a beat, and I've not had a brain fart, sure I don't cut hundreds of metres of plate, but I'd keep an eye on this, the official forum, the latest release, manuals and documentation to smooth over any transition you may think about. It's not my forte but there is also a good YouTube clip on spindle synchronisation.... get the m3/m4+ ess to do that now and I will be impressed! But it's all down to whatever suits your needs

EDIT: I WAS WRONG, SORRY, this was Mach3 + UC400eth.... note the hand turning of the spindle... still be impressed to see someone to do that with the uSS/eSS + the same software
https://www.youtube.com/watch?v=25FtHEo4Ujo

The rigid tapping with uccnc is impressive:
https://www.youtube.com/watch?v=QqNfKvWuOBI

Rob

edit: Yes, there is a special promotion on for the product launch of the uc300eth (only) where is you buy on between now and 15/10 (dd/mm), [or current initial stocks last] you'll get a licence for uccnc free (saving ~$60 / 55euro + taxes if relevant, ~ 44% saving against UC300eth+uccnc). the licence is for the serial number of the uccnc device, ie 10 devices, you need 10 licences to correspond to the serial numbers... you can run multiple devices off one pc (you add a switch to the end of the shortcut to call for the licensed device you want to start with the profile.... but you cannot run multiple instances of uccnc from one pc at the same time). The wise will realise that the promotion has little cost to cncdrives, given the software is already produced, and I do not believe that there is a uc300 mach3 plugin available at this time, but given all other devices have mach3 capability, it will probably be issued (probably= i don't know).... so you may have a fallback option. However no uc device has a mach4 plugin... so they probably won't be developing one for any uc device on current precedent.

[A_Camera]

As you know it's not about speed, the name of the game is acceleration and accuracy.

It´s not that simple... Speed is just as important as the other parameters. High speed is necessary, especially if you are doing a lot of small cuttings and must move the spindle often (like I am doing), or if you have a large table and work on large items a lot with long movements. Of course, high speed without acceleration which let you reach that speed is pointless, as well as if accuracy is bad and you can´t stop where you need to.

Doesn't miss a beat

So far I have not seen any misses, but on the other hand, I had no such problems with Mach3, even that was very good in my case.

[A_Camera]

Dude I did not mean I don't trust your word or any disrespect at all, I was just meaning you are seeing A big improvement, if ger and terry say the same thing that is 3 people haveing the same outcome, I trust ger and terrys advice like most people who know who they are, I have seen you around here long enough to know you are not full of **** at all and give good advices as well. the relevances is if you plus many more have the same thing it's gold. and my setup is similar to yours I have 213V

Dude, it really doesn't matter.

[A_Camera]

I think that it's highly unlikely that CNC Drive will write any Mach4 plugins for any of their motion controllers.

I agree. They have never done it before, so now with the UCCNC and the slowdown of Mach4, there is no reason.

Perhaps it will not be able to drive as fast, but I suspect that even the UC300ETH plays a role and suspect that it is in many ways better than the USB version of UC300

I would expect their performance to be similar, except for the higher pulse rate in the ethernet version. I think that the trajectory planner can make a big difference with the correct settings.
A friend of mine experienced a 30% speed increase for UCCNC over Mach3 with a UC100.

The UC300ETH base plate is the same as the UC300USB, zero differences. However, the fact that it uses Ethernet interface makes it more reliable and probably faster and smoother. However, unless I buy a license for the UC300USB I will not be able to compare fairly. I will however be able to compare UCCNC with Mach3 together with the UC300ETH and will see if it makes a difference.

[ger21]

However, the fact that it uses Ethernet interface makes it more reliable and probably faster and smoother.

All the movements are buffered, so connection speed should have zero effect on movement.
Did you have reliability issues with the UC300?

note, to be unbiased, there is no indication of what motion controller he is using (it may be PP or more likely the Mach3 UC plugin [s its easier to compare directly]

A UC100 was used for both Mach3 and UCCNC in this video. I saw the same comparison in person at his house.