Newbie (ageing rapidly!)

Hello all,

I'm new to CNC but I have a fair amount of machining experience over the last 40 years. I am a professional electronics and electrical engineer and former systems programmer (a long time back).

I have a mill that I am converting to CNC if I can ever get a motion controller to work. I'm on the fourth one now (from different manufacturers) and have tried each on both a Windows 10 (freshly installed) laptop and an old indestructible Toughbook XP machine (that ideally I want to use for the CNC).

I'm ageing rapidly as nothing I have tried so far does what the suppliers claim. The instructions are just plain wrong in most cases and the software seems hopeless. I'm four months and several hundred hours into this project now and losing the will to live!

I would be most grateful if anyone can recommend a USB motion controller for Mach3 that actually works. Or should I give up on Mach3 and use something else?

Your thoughts, words of encouragement or sympathy gratefully received

Which motion controllers have you tried so far then ?

When you say systems programmer are you referring to PDP-11 type equipment ?

Professionally, mostly IBM 370s, IBM Series 1 some PDP and VAX-11s, Various Micros: SC/MP, 6502, 6809, 68000
My own first computer had 256 bytes of RAM and controlled my model railway.

Ah for the days when computers did as they were told!


So far I've tried:
A breakout board that the supplier claimed would work on either USB or parallel but they actually meant it could be powered from USB (I know rookie mistake).

A UC-100 added to the breakout board, no life signs. No activity on the USB.

A Chinese USB motion controller that they said would work with Mach3 but later confessed only worked with their own proprietary software. Not that I could get that to work either.

A UK supplied (European made) SODIAL USB motion controller. Mach3 didn't even recognise the plugin provided.

A UK supplied (Far east made) NVUM USB motion controller that was supplied with a plugin driver that didn't match what the instructions claimed. They swear it's the correct plugin. Mach3 recognises the plugin but will not communicate with it.

I've downloaded Mach3 multiple times and tried copies provided by the controller supplier. All claim to be the same version, 3.043.066 . I've reinstalled Windows 10 on a formatted laptop. I've tried both the USB 2.0 and USB 3.0 ports and there is data activity on the USB. I've also tried my old Panasonic Toughbook XP machine (ideal for the workshop).

I am close to binning the whole project and sticking to manual machining.

Starting at the top.
Mach parallel port drivers won't run on any 64bit windows, or any 32bit windows after 7.

Was it a genuine UC100 board? Amazon and eBay are full of counterfeit copies, none of which are supported by CNC Drive.

Cheap Chinese is always a gamble.

The only SODIAL board I've found says Mach 3 substitute, so I would doubt it's meant to work with Mach 3.

NVUM do seem very plentiful, so I would expect them to work.


Are the motion controllers showing up correctly in Device Manager with no warnings?
Win10 can be very temperamental with any drivers that are not signed.

Sorry to say this but unfortunately you are the perfect example to others why you should never try build CNC machine on the cheap or without asking advise first.
I've seen this approach countless times and nearly always ends up costing twice the price of good components. Worse still, as is nearly case with you can lead to folks throwing in the towel giving up on CNC.

It really doesn't need to be that difficult. People just have to understand and be prepared that building a Good reliable CNC machine can't be done cheaply. Trying to do so just leads to frustration and wasted time & money.

So again GOLDEN RULE. . . . NEVER BUY ANYTHING WITHOUT ASKING ADVISE/OPINIONS.

I was assured it was a genuine UC-100 but without a known good one I didn't have a comparison. It should have worked on my 32 bit XP Toughbook. It went back at their expense.
I tried the BOB on a very old desktop (not mine) and it seemed to work from the LPT. i.e inputs and outputs functioned I didn't connect it to the stepper drivers.

The Chinese motion controller was a gamble but why do they bother to say it works with Mach3 and then change their mind when pressed surely they will just get them all sent back! It was well made and appeared well designed for, EMC and isolation. It went back at their expense.

The SODIAL claimed to have a compatible Mach3 plugin but the .dll simply wasn't seen by Mach3 so clearly it wasn't. Not as well built and lacking good EMC design. The supplied USB cable even had fake ferrite cores on it (just plastic blobs with no ferrite inside!). It's going back at their expense.

The NVUM is properly made and looks to have sensible EMC design. (inputs are a bit highZ which could give interference problems but that's easy to prevent.) It appears in device manager under Human Interface Devices as HID-compliant vendor-defined device and as USB input device. No errors with either. The plugin option appears when you start Mach3 as: NVUM_F-Novusun-Plugin-Ver 2.0a. It is enabled within Config Plugins but Config then takes me to a dialog that shows:
DDREAM
ddream device buffer 200ms with ?? in the button
jog buffer 200ms again with ?? in the button
They say that doesn't matter !

The instruction say you should install plugin NVUM_SK.dll but NVUM_F.dll is all that is supplied. They say that is correct in which case the instructions are wrong.
I get a strong feeling they don't know anything about the product they are selling.

Anyone got a working .dll plugin for an NVUM ?

I don't want to go down the route of an old desktop with an LPT. Grinding dust, coolant mist and swarf don't mix with a fan cooled PC. I could look for an old silent (fanless) PC but then I need to find a sealed monitor and somewhere to put it, a keyboard and mouse too.

I'm beginning to think Mach3 is too based around obsolete tech; LPTs and 32bit OS
Should I give up on it? Mach3 and Mill Wizard looked a promising setup but they have been totally unhelpful.

Is there a sensible priced, established, reliable, simple to learn, alternative I should look at?

JazzCNC,

I completely agree, after consulting n people I had at least n+1 options! I did consult a lot of people and did my research. I admit being caught out by the USB powered BOB was down to me but every other controller has been based on recommendations / reviews and a detailed set of questions sent to the supplier. Sadly the sales departments of most companies simply say what ever the customer wants to hear and claim you must have misunderstood when challenged after the sale. That's why I record all phone conversations. It won't cost me a penny for products that don't work.

Quote:

---

Originally Posted by ManicMetal

JazzCNC,

I completely agree, after consulting n people I had at least n+1 options! I did consult a lot of people and did my research.

---

Please don't get me wrong here I'm not having a go at you just pointing out to others these things need research and acceptance that it can't be done on shoestring budget.

Now you say you had advice. Then I suggest you change who you ask because can tell you without a shadow of a doubt no one here would recommend any of those choices you made.
Those that know would never advise USB and nearly always point you to Ethernet-based controllers unless cost was the issue, even then would issue caution regards USB because it's flaky. Or would point to other very capable alternative systems like Linux CNC and Mesa cards etc.

While Mach3 is aging and starting to lack behind others in some areas of software it isn't the problem you have been experiencing. Mach3 is very capable software with 10,000's of users working on whole variety of different types of machines and hardware.

Your problem comes from the fact those cheap or fake controllers have no support for the plug-ins and Mach3 over the years has had many updates which break these plug-ins which are no longer updated.
Any Device is only good as the support from the supplier regards the Plug-in. Without this, it's just becoming an expensive paperweight.

Mach3 does work on 64Bit systems and on Win10 perfectly fine. It's the parallel port driver that doesn't so is restricted to 32bit but PP is dinosaur and only fitting it should go the same way.!

Now in order to help and guide you, we need to know more about the machine and what your needs are.

The ?? would tell me either the translation is missing, or it's trying to use characters that don't exist on your computer. It shouldn't stop things from working.

There are 3 different plugins listed here - http://www.nvcnc.net/nvum.html
Probably worth trying them all, to see if you any connect to the controller.

And just to cover the basics, you have connected a suitable 12-32VDC supply to the controller?

Quote:

---

Originally Posted by ManicMetal

The SODIAL claimed to have a compatible Mach3 plugin but the .dll simply wasn't seen by Mach3 so clearly it wasn't.

---

Plugins are usually specific to a small range of Mach3 versions, often just one ;-)

yes I did have a supply connected but in fact it isn't needed for the controller to communicate (that is USB powered). The external supply is for the isolated I/O.


It's an Axminster ZX30M Mill
I need 4 axis:
X and Y at up to 43kHz to give me 2000mm/min rapids on 16 microsteps
Z at up to 11kHz
A doesn't really matter.

Step, Direction and Enable opto isolation is provided at the stepper drivers. Polarity doesn't matter.

Inputs should be opto isolated but if needed I can do this external to the controller.
I already have signal conditioning for all inputs.
I'd like plenty of inputs:
4 for combined limit and home
1 for a Z tool offset
1 for an X, Y edge finder probe
1 for Estop

Outputs don't need to be isolated, as they will be external to controller.
10v PWM for motor VFD
1 for mist coolant
2 for driver enables to allow staggered power up and to allow X,Y release without losing Z.
(I can easily achieve this external to the controller if needed).

In short nothing that just about every motion controller doesn't claim to offer. As said above "It doesn't need to be that difficult".

The problem with good support is it tends to vanish once they have your money. That's why I mostly bought from UK suppliers but in fairness the Chinese made more of an effort to solve the problem than anyone else.

As for 'Those that know', for every post from someone saying don't touch USB, there are others saying that it works fine. Same is true for every option! They can't all be right.

Quote:

---

Originally Posted by ManicMetal

As for 'Those that know', for every post from someone saying don't touch USB, there are others saying that it works fine. Same is true for every option! They can't all be right.

---

Well it's all about experience really and there lies the problem.? Most only have experience with one product or technology on one machine and if by some chance they don't have an issue they claim it's best thing since sliced bread and everybody should use it.!

When I tell someone something is good it's because I've tried tested and fitted or used on multiple machines over period of time. Like wise if I call something Bad it's because I've had multiple bad experiences and trust me I've pulled my hair out with USB many times.

Quote:

---

Originally Posted by ManicMetal

It's an Axminster ZX30M Mill
I need 4 axis:
X and Y at up to 43kHz to give me 2000mm/min rapids on 16 microsteps
Z at up to 11kHz

---

Ok well first your wasting your time really with 16ms because most steppers can't resolve to that degree if your chasing resolution, which isn't really resolution in true sense. Only other advantage then becomes slightly smoother running motor but will come at cost of torque.
Then you have the fact some motors don't like running at those frequencys so could have issues in that department with them not playing nice with drives regards resonance etc. Again depends on quality of drives/motors.

You'll have more stable machine with x8 ms or x10ms.
If your wanting smoother running motors then I'd suggest looking at 3 phase steppers as they run much smoother

I'm assuming that you must 2.5mm pitch screws if you have calculated 43khz.

Now I'm going to suggest a card that I know will fit your needs perfectly for Mill and it probably has the Best Software in the business which as been developed over 30yrs and runs $$$$$ machines so is very well supported and rock solid.
However this slightly breaks my rule on recommending because while I do have one of these cards here on test I've not actually fitted any to machines. But I do know plenty about them and have several contacts who like me are very experienced and have fitted to many machines.

The card is Acorn from Centroid and you'll find it here. Again not cheap but it's the best in it's class when figure in the software/backup.
http://www.centroidcnc.com/centroid_...ontroller.html

You're right 2.5mm pitch screws.
I've already had the X and Y drives running (from a signal gen) up to about 60kHz (2800mm/min on 16 ustep) Where the power supplies run out of voltage to drive the 2.8mH inductance. I loaded the bed with my heaviest vice, and angle plate.

It isn't a matter of resolution 8 ustep would give me 1.5um resolution (way more than the tolerances of the machine) but 16 ustep does give a noticeably smoother motion than 8. There is hardly any reduction in torque (this is more a factor of adequate drive voltage). I've specified 2000mm/min at 16 ustep to stay within reasonable rpm for the motors while keeping to a frequency the motion controller should be capable of, I can always reduce this if needed (commissioning trials will prove the limits and determine the maximum ramp rates).

I'll take a look at the Acorn system, seems a shame to abandon Mach3 and Mill Wizard, I like their simplicity.

The accuracy gains between 8 and 16 microsteps will be marginal if any. The only time a stepper will produce full torque is when on a full step. In between those full steps, you're relying on two magnetic fields balancing the rotor over the magnetic detent. Stiction, friction, and cutting forces will all have an effect on the exact position the rotor ends up balancing.
Motor torque is directly related to current, not voltage. The reason voltage comes into the equation, is because as motor speed increases, so does back EMF. The higher the available voltage, the faster the motor can spin before the back EMF becomes too high for the supply voltage to push enough current through the motor windings. Once you reach that point, the current flowing through the windings falls off resulting in reduced torque.

Quote:

---

Originally Posted by ManicMetal

Step, Direction and Enable opto isolation is provided at the stepper drivers. Polarity doesn't matter.

---

Maybe not electrically, but it does for the step/direction timing.


I use USB on all my machines, but I'm well aware certain installations can be plagued with problems.
The big problem is most USB motion controllers will cause Mach 3 to lockup if there is any kind of glitch/communication failure with the USB, at which point the only solution is a reboot, as there is no graceful timeout or failure.

That's spot on m_c. I'm not concerned with the accuracy, it will be plenty good enough and other issues such as backlash, drive belt tooth engagement, quill bearing precision etc will all be significantly greater.

There is very little, if any, torque reduction from 8 ustep to 16ustep it is just a bit smoother at low speeds. I could go down to 4 ustep, and it's still acceptable, if the operating frequency proves to be a problem. At high speeds the momentum of the rotor and load make the microstepping irrelevant and the driver reverts to full step to get a bit better torque.

The stepper drivers are current regulated so they deliver what ever voltage is needed to achieve the desired current up to the maximum the power supply can give. The higher the rpm needed from the motor the higher the frequency from the drive and the higher the reactance from the motor stators. Higher reactance then requires a higher voltage to reach the desired current.

The point about the step/direction/enable signals being opto isolated means I can wire them common anode or common cathode to suit the controller outputs, they will all be the same so no effect on timing.

USB manages to communicate totally reliably for many other devices, other than poor EMC design, why should a motion controller fail to talk properly on USB?

What motion controllers have you had least problems from?

Quote:

---

Originally Posted by ManicMetal

I've already had the X and Y drives running (from a signal gen) up to about 60kHz (2800mm/min on 16 ustep) Where the power supplies run out of voltage to drive the 2.8mH inductance. I loaded the bed with my heaviest vice, and angle plate.

It isn't a matter of resolution 8 ustep would give me 1.5um resolution (way more than the tolerances of the machine) but 16 ustep does give a noticeably smoother motion than 8.

---

Time will tell on this because things change when chips start flying and resonance's start taking affect. Also are you 100% sure the motion is smoother or is it just the change in motor sound that makes you think this.?

Quote:

---

Originally Posted by ManicMetal

I'll take a look at the Acorn system, seems a shame to abandon Mach3 and Mill Wizard, I like their simplicity.

---

Doesn't get any easier than Acorns software and the built in Conversational Cam more than matches Mill wizzard. Another thing I can tell you with 100% certainty is that the motion is much smoother than Mach3. The motion planner is much more advanced and with the Smoothing feature enabled motion is noticably better.

Only USB device I've had any success with is UC300 and even then didn't fully trust it due to odd unexplained happening on several machines. Gave it benefits of doubt thou because the machines that did occur on used the same Dell Pc's which I've had issues with regards USB.
To be honest the PC is the biggest issue for USB and even when optimised in windows still get issues. Mostly I think quality of PC plays big part.

With Ethernet never get any issues of any kind and I've got dozens of customer machines that use them. Some of them in some pretty noisy environments, ie: next to plasma machines. Try that with USB and your doomed.!

My point being here I'm trying to help pass on my experience and not to slag off USB. If you or others choose to ignore this advice then that's fine by me, it's your money you waste it any way you like.

The Frequency/Ms setting etc is irrelevant really because so many variables come into play and each motor/drive/controller/machine combo will be affected slightly differently. But my experience is that high Ms just stress's the controller for virtually no gains. What you gain in one area you loose another and to some degree depends on how your using the machine.! . . .Low down you'll see the smoother action but up high the motors/drives feel the pain and become open to resonances.

However It's just flick of few switches to change so no Big deal provided you've got the frequency to start with and Most External controllers will have more than enough any way so again no Big deal.

MM: I don't have the NVUM, but do have the NVEM (the ethernet variant), and can say that they are largely okay, but come with issues, the first one being that they have a fixed pulse width of 2.25us output to the stepper drivers, from memory (and that ain't great) negatively switched but capable of sourcing or sinking current. The default installation in the manual suggested switching the positive input to the stepper driver and grounding the negative input - with my cheap stepper drivers that lead to the LED in the optocoupler being "On" significantly longer than "Off", coupled with the low performance optocouplers (4n25) and saturation on the phototransistor meant that the stepper drivers were missing steps (scoping input vs output showed a grim transfer function). So, word of warning on the NVUMs (assuming the motion controller is largely the same) - check that you're not in a position described above, changing the arrangement above can help (as can better stepper drivers or replacing optocouplers.... but you're a way from that yet by the sounds of things).

I've never tried USB on a CNC machine (only PPBoB and Ethernet). I did, however, once have a USB interface to a RepRap 3d printer. That was temperamental (occasional missed instructions which lead to offsets in the print space) and a shared view on the inter web pointed at the reliability of USB. Fortune smiled upon me and I managed to shear off the micro USB socket from the board by tripping over the cable - the easiest fix for me was tacking a bluetooth serial interface directly onto the Melzi board. Strangely, BT was infinitely more reliable than USB in that environment. I think USB and steppers and potentially disimilar grounds don't co-exist well.

The NVUM ??? plugin buttons - I have those with the NVEM. Not figured out which button does what but again, from memory, there's only a couple of very basic options to play with on the config screen, so long as those "stick" then the two buttons are likely OK and CANCEL - press one and check the config sticks. The ddream driver - sounds about right - and I think indicates this is likely a rip-off controller, but works to a fashion for me on a NVEM.

Open up the NVUM - you should find the inputs are Optocoupled already so external conditioning is not really needed. Just be aware that the ABS injection moulded case - the receptacle for the screws doesn't butt against the mating face on the other half of the case - so it's easy to overtighten and rip the plastic apart. Tighten to 0 white-knuckles loading.

What exactly doesn't work for you with the NVUM? I guess you've tried jogging with no joy?, I know with the NVEM driver on first start-up I can find (though never really bottomed it) that I can't jog until I home the machine. Maybe a driver issue, but I usually have a 2 minute tussle persuading the machine to start moving - either a G1 X1 feed through the MDI or a home resolves and I then get distracted and never investigate further.

Quote:

---

Originally Posted by Doddy

I know with the NVEM driver on first start-up I can find (though never really bottomed it) that I can't jog until I home the machine. Maybe a driver issue, but I usually have a 2 minute tussle persuading the machine to start moving - either a G1 X1 feed through the MDI or a home resolves and I then get distracted and never investigate further.

---

Mach3 has the same "feature". A bit of a pain if you've parked the gantry at the far end of the machine at the end of the previous session as homing speed is slower than usual rapid speed. I found that it was because I have soft limits enabled - Mach3 errs on the side of caution and doesn't allow jogging until homed as it can't check soft limits otherwise. Don't know if it's the same on your controller. Just a thought.

Quote:

---

Originally Posted by Neale

Mach3 has the same "feature". A bit of a pain if you've parked the gantry at the far end of the machine at the end of the previous session as homing speed is slower than usual rapid speed. I found that it was because I have soft limits enabled - Mach3 errs on the side of caution and doesn't allow jogging until homed as it can't check soft limits otherwise. Don't know if it's the same on your controller. Just a thought.

---

Cheers - you might be right. I'm doing a lot of work on a plugin for my machine at the moment and learning an awful lot of the weirdness of M3 (like default feed rates on initialisation).

That one really narks me - first thing after startup these days I overwrite the feed rate to 3000. Initial jogging at 6mm/min is awfully tedious...

Some time I'm going to be doing a CNC conversion on my own vertical mill, so I'm finding these discussions of various control options very interesting. One thing I know is that it's unlikely to be Mach3! Seems to be some good options available or close to available.

I'd try and make one of your USB solutions work.

The Nvum controller seems a good place to start.

Join this forum then read this thread back to back, you are not alone.
https://en.industryarena.com/forum/h...-315210-5.html

This guy has an updated beta driver
http://www.opalsfromaustralia.com.au/cnc/

He has same controller as you, should be able to help you get going and is a forum member on industryarena.

Quote:

---

I had a lot of the same issues. Went to win 7 32bit and a lot of them stopped. Still have a few issues but the router now runs for hours no problem.

Before I went back to 32bit operating system I tested on another 64 bit system and had the same issues. I was about to buy a different controller but decided to format and try again. Turned out to be something in the 64 bit operating system. Took me months of frustrating fault finding after work.

The nvem now happly runs my Alpha 1300x2500 router and has paid for it's self several times over.

---

Quote:

---

Originally Posted by Neale

That one really narks me - first thing after startup these days I overwrite the feed rate to 3000. Initial jogging at 6mm/min is awfully tedious...

---

This is what annoys me when people knock mach3 when really it's the user ignorance. If you go into General Config you'll see initialization String just enter the feed you want into there and will always be set to that when Reset.

I thought that I had tried that (it's obvious that there are initial settings and I did look for and find where they were set) but it didn't work. So, probably user error rather than user ignorance in this case! I'll have another look - thanks. It's only a nark, hardly a show-stopper, and less important than not being able to dual-axis home with my IP/M. But I used one of those with my eyes open and I can work round that easily enough. Mach3 works well enough for me but there are always odd items where you think, "I wouldn't have done it quite like that!"

Quote:

---

Originally Posted by Neale

hardly a show-stopper, and less important than not being able to dual-axis home with my IP/M. But I used one of those with my eyes open and I can work round that easily enough.

---

This can be done with IP-M if you want to do it. Just takes little out of the box thinking but agree it would have been better if Cslabs made it possible in software.

This one's frustrating because they clearly have the code to do it in the IP/S - maybe the IP/M processor is at the limit of its memory? Maybe it's a marketing decision to make people buy bigger?

I did wander outside the box. One solution would be to add in a little bit of electronics to do the homing. However, that would have to sit between the IP/M and the EM806 and receive/transmit differential signals. That's not a big deal but it's just a little bit more complicated than "just" sticking an Arduino in there with 5V single-ended signalling and it wasn't worth the effort. I also looked at setting up two different machine profiles, with a second one that treated the two slaved axes as separate axes. You can configure Mach3 to home these simultaneously, and I think that this would have worked except that it looks as if you can only have one IP/M profile that loads for every Mach3 profile. Because the IP/M profile knows about slaving, you can't do it.

Current workaround is home the machine normally, then hit e-stop which takes power off the steppers. Manually turn the slave axis pulley watching the LED on the home proximity switch, and in effect manually go through the Mach3 homing sequence. Not as accurate as it probably only goes to the nearest whole step (because of stepper cogging effects) but given that that's max 5/200mm error (5mm pitch ballscrew), about a thou in old money, and at one end of a 1000mm gantry, that's not a whole lot out of square.

But if there's a better way - I'm not too proud to nick other people's ideas!

A couple miniature relays, and some simple logic could be used.

Something like this flow should work-
Initiate homing.
Axes move towards home sensors at usual speed.
When first homing sensor is triggered, it powers a relay which disconnects the step signals to that axis.
When second homing sensor is triggered, it does the same, but both home sensor signals are also summed to trigger the IP/M homing input.
Once the IP/M input is triggered, both relays get de-powered.
Homing then reverses the axes, and the process repeats.

Admittedly just stopping the step signal may lead to a bit of an abrupt stop, but you could switch the drives enable signal on the higher speed approach, so each axis coasts to a stop, then switch the step signal on the retract, which should happen at a much slower pace so an abrupt step stop shouldn't be a problem.

For the relays, something like a EC2-5NU (5vdc coil, or EC2-24NU for 24vdc) would work nicely for switching the step signals.
You'd probably need a couple more for the logic, but that with a couple flip-flop type configurations, and independent homing should be implementable externally with the controller thinking it's just homing normally.
If you can't figure out the required wiring, let me know, and I'll add it to my list of things to do.

Quote:

---

Originally Posted by m_c

A couple miniature relays, and some simple logic could be used.

---

Yes your right M_C that's exactly how I do it and was going to post all this but really this is not the place and distracting from the OP question so think it's best left alone or if somebody's bothered enough start another thread.

Yippeeee!

Finally I have electrons going in and out of the right places on the NVUM motion controller.
It was the plugin. The supplier kept assuring me they had provided the correct one. Big surprise, they hadn't. Got hold of the v2.1 plugin and it leapt into life. I've offered to tell them how to get it working if they settle my invoice for about 5 hours of my time!

The bit of signal conditioning I referred to earlier is simply due to the highZ inputs. The opto inputs only sink about 1.8mA and switch on at around 2v, that could be a problem from noise pickup. So a 680R pullup and 100nF capacitor to GND plus a 3v zener to raise their transition voltage should make them much more immune to spurious signals. As observed before there are no isolation or common mode problems with opto inputs.

Now just need to connect it to the rest of the gubbins....... (and find out if USB is as flaky as suggested).

MANY MANY Thanks for all the help and suggestions.