Tormach PathPilot.

Advert

New Replies

Report Bug

Help/FAQs

Tormach PathPilot.

This topic has 29 replies, 7 voices, and was last updated 2 October 2015 at 21:32 by Anonymous.

Viewing 25 posts - 1 through 25 (of 30 total)

1 2 →

Author

Posts
17 February 2015 at 23:24 #180146
Another JohnS
Participant
@anotherjohns
This one is new to me – LinuxCNC base, Tormach front end (given back to open source, from what I read)

**LINK**

It replaces Mach3/Mach4. I do know that their CNC lathe runs a modified form of LinuxCNC already.

It's nice to see development in these controls for home machinists – lots of interesting stuff happening with LinuxCNC (stable platform), and especially the "future branch" stuff – machinekit.io

Anyway, if like me you had not heard of this new Tormach control software, you have now!
Advert

17 February 2015 at 23:24 #15065
Another JohnS
Participant
@anotherjohns
17 February 2015 at 23:39 #180148
Michael Gilligan
Participant
@michaelgilligan61133
John,

Did you post the wrong link, or am I being dim ?

YouTube shows a fancy 5-Axis machine.

MichaelG.
17 February 2015 at 23:58 #180150
Muzzer
Participant
@muzzer
Presumably this. In true Tormach fashion, they have published a white paper explaining the system and an installation guide that shows that the FPGA expansion board fitted to the PC is from Mesa Electronics. Not absolutely clear if they are using Linux but it sort of adds up, given the hints to LinuxCNC and the use of the Mesa board.

Very interesting, although doesn't sound as if we will be able to use it for a while. I love Tormach's approach in general. Well spotted!

Murray
18 February 2015 at 00:03 #180151
Another JohnS
Participant
@anotherjohns
Thanks Murray; Michael – yes, probably cut-n-pasted the wrong link. Those Youtube links are not intuitive. In the middle of cooking dinner for the wife, so my focus was not totally on this forum…

John.
18 February 2015 at 00:32 #180153
Michael Gilligan
Participant
@michaelgilligan61133
Posted by John Alexander Stewart on 18/02/2015 00:03:16:

Michael – yes, probably cut-n-pasted the wrong link. Those Youtube links are not intuitive. In the middle of cooking dinner for the wife, so my focus was not totally on this forum…

.

Thanks, John

I guess this was the link you intended

… Enjoy your dinner.

MichaelG.
18 February 2015 at 00:33 #180154
John Stevenson 1
Participant
@johnstevenson1
Bloody hell Deja Vu all over again.

The following was in an email I sent to Art Fenarty, the original writer of Mach3 tonight at 23:14

Ten minutes before the other John S post appeared here but we had been in contact for the previous hour.

**************************************

Quoted email.

Tormach hold the answer to small CNC but they don't realise it.
If they were to do for mill what they have done for lathe with LinuxCNC then there would be a mass exodus. The problem is they don't realise that they have a cash cow with those screens for Linux and the linux guys are that far up their own arses they can't see that the screens are stopping people from using it.

All it takes is for someone to write a M3 screen editor for Linux……………………………………………………………
Hello you there Art ???…………………………………
18 February 2015 at 02:06 #180156
Another JohnS
Participant
@anotherjohns
Posted by John Stevenson onIf they were to do for mill what they have done for lathe with LinuxCNC then there would be a mass exodus. The problem is they don't realise that they have a cash cow with those screens for Linux and the linux guys are that far up their own arses they can't see that the screens are stopping people from using it.

"Human Factors" people are generally the first to be laid off in a downsizing operation – which is too bad, because Engineers can't design for the ordinary person. (I lost a Human Factors person I was using, due to "downsizing" fairly recently)

Saying that, there is a new screen or two – one for LinuxCNC called Gmoccapy, and one for the experimental branch (machinekit.io) called, IIRC, Cetus.

Don't know if they are better or worse than the old LinuxCNC standard screens or not.

Cheerieo – The better looking JohnS. (so says my wife)
18 February 2015 at 08:13 #180167
Michael Gilligan
Participant
@michaelgilligan61133
With the recent release of the upgraded Raspberry Pi …

How close are we to to getting PathPilot into a very small self-contained box ?

MichaelG.
18 February 2015 at 12:52 #180219
Anonymous
Since I have a Tormach mill I am on their mailing list, and got an email about PathPilot last night; looks interesting. I initially ran my mill using a PC with XP, but had a few issues, the primary one being the software would occasionally go into incremental mode during a toolchange. That led to excitement when the tool unexpectedly whizzed off into the middle distance, milling through everything in the way. I now have a Tormach controller, although I will need to check on it's compatability, as it is a few years old.

Mach3 is sort of alright, but the screens are a bit messy and inconsistent. I've also had major issues when running 4th axis code. Basically it doesn't seem to be able to control feedrates on combined linear/rotary moves. I ended up using the bizarre G93 inverse time feedrate feature.

I have signed up to the webinar tomorrow evening, UK time.

Andrew
18 February 2015 at 13:35 #180225
Muzzer
Participant
@muzzer
MichaelG, I've got a couple of RPi's now, including one of the new RPi2's and have been trying to figure out how close / realistic it would be to run a CNC machine on one. I'm unlikely to be the pioneer who does so, so I've been watching what has been happening out there. I've managed to spin a stepper motor but I have to say I'm not a very capable user of Linux or Python so it wasn't a walk in the park for me.

I think the consensus is that the RPi simply doesn't have the IO to be able to drive stepper drives etc directly, so you'd need to augment it with a secondary controller board – something like a Beaglebone Black (BBB). The drivers for the Broadcom SoC are also proprietary and there are currently some building blocks missing.

There are people who have managed to configure the LinuxCNC for a RPi but it's some way from being a functional CNC system. If you want a compact CNC system today, I suspect you would want to get a Small Form Factor PC and run LinuxCNC or Mach3 on it, with (realistically) a set of Mesa Electronics cards and/or one of the CNC controller cards.

However, you can run Linux on a RPi, complete with office applications etc and it has similar power to a PC of only a few years ago ie pretty impressive for something the size of a credit card. The RPi2 is certainly a fair bit faster, so worth getting one of these given that they cost the same. It also makes useful media servers, iPlayer boxes etc. That's probably where the main potential of the RPi is for now, along with the educational stuff. Amazing power but not really ideal for real time control requiring lots of fast GPIO.

Murray

Murray
18 February 2015 at 13:37 #180226
Muzzer
Participant
@muzzer
TBH, I think JS is barking up the wrong tree with Artsoft / Art Fenerty (and hasn’t he retired now?). He’s so established in the Windows / Mach3 arena that I doubt it would be sensible to throw the whole thing up in the air and start again. Mach3 branched off LinuxCNC a long time ago.

Tormach are machine manufacturers above all, not software developers or even electronics resellers. However, Toolpath shows what can be done when a professional installation of LinuxCNC is implemented. It’s something that LinuxCNC cries out for and could be an opportunity for someone to make a go of. However, it’s a different business model to what Tormach are currently using so would probably require a different entity, whether within the Tormach group or someone else completely.

Making your own LinuxCNC installation feels as if someone has taken several sophisticated mechanisms, fully dismantled them and thrown all the pieces into a large box before randomly mixing them up, leaving you to try to piece together a working machine from them. If you aren’t an expert in all matters CNC beforehand, you will be afterwards – or you will have given up. It’s not just the LinuxCNC itself. You also have to become fairly proficient in Linux and fathom out all the IO hardware, stepper drives, limit switches etc and how to install and configure them. There’s a lot of interconnections and configuration to be done beyond simply wiring up the IO board to the various external components. That’s quite a challenge in itself but it’s really not the challenge a lot of CNC converters had in mind to begin with.

If someone were to integrate a semi-universal controller board with the current LinuxCNC / Mesa products and create an integrated Linux / LinuxCNC installation to drive it (boot straight into LinuxCNC), I could imagine there would be a fair bit of interest and possibly enough for somebody with the right drive and vision to make a viable business of it. I’m not sure the existing LinuxCNC gurus would want to do that as they seem to be otherwise employed and may object in principle to commercialising the application, even though this is part of the Open Source plan. However, there’s nothing to stop a modern day equivalent of Mach3 coming about, making a commercial product from the current offering. This would be a step above the ubiquitous “parallel port breakout board” DIY approach but also a lot more accessible to a bigger and more serious market.

Those are my initial thoughts. Looking forward to hearing more about Toolpath!

Murray
18 February 2015 at 16:13 #180237
Michael Gilligan
Participant
@michaelgilligan61133
Thanks for the thoughts, Murray

Given the success they had with the 5MP camera on the GPIO, I suspect it will only be a matter of time before someone figures out how to do the job.

… We live in hope !!

MichaelG.

.

Edit: It's probably worth keeping an eye on this page.

Edited By Michael Gilligan on 18/02/2015 16:32:58
18 February 2015 at 17:23 #180244
Muzzer
Participant
@muzzer
MichaelG – the camera doesn't use the GPIO. Here's an excerpt from Magpi (magazine) on the camera:

"From its first launch the Raspberry Pi has had a connector on it to attach a camera to the GPU (the VideoCore 4 Graphics Processing Unit on the Raspberry Pi). This connection uses the CSI- 2 electrical protocol and is a standard used in most mobile phones. It is an extremely fast connection, which on the Raspberry Pi is capable of sending 1080p sized images (1920×1080 x10bpp) at 30 frames per second, or lower resolution at even higher frame rates. It had always been intended at some point to release a camera module that could use this connection, as the ability to stream high speed video data through the GPU without any interaction with the ARM processor would always make the camera much more efficient than any USB attached webcam. It would also enable the use of the GPU’s ability to encode H264 video, or JPEG images in hardware."

I managed to get a camera installed and working a couple of weeks ago by following the idiot guide provided. Sure enough it works pretty well. Here's the spec.

Looks as if GPIO is being developed but may not be quite there yet.
18 February 2015 at 18:00 #180250
Michael Gilligan
Participant
@michaelgilligan61133
Posted by Muzzer on 18/02/2015 17:23:06:

MichaelG – the camera doesn't use the GPIO. …

.

Sorry, Murray … I wrote in haste, and stand corrected

MichaelG.
18 February 2015 at 23:42 #180291
Another JohnS
Participant
@anotherjohns
Posted by John Stevenson on 18/02/2015 00:33:02:

…If they were to do for mill what they have done for lathe with LinuxCNC then there would be a mass exodus. The problem is they don't realise that they have a cash cow with those screens for Linux and the linux guys are that far up their own arses they can't see that the screens are stopping people from using it.

John – I think that you have made a grave misunderstanding here.

My thoughts after pondering this for the past 24 hours.

LinuxCNC just works. It may not look pretty, but the idea is to drive machines with as close to 100% consistency as is possible.

So, LinuxCNC is NOT designed to be pretty, because, hopefully, you are doing other things than watching the screen!

I did read through some of the Tormach white papers, and I now realize the reasons for the difficulties that others are having wrt Mach3. I've never used Mach3, but did wonder about some of the issues that I have read about others having, and some of the kludges required to ensure that the machine did not end up "going haywire".

To conclude: if the choice is "nice screens" or "absolute reliability", I know which one I'd choose.

I know that you are not against LinuxCNC, so don't treat this post as a criticism, maybe more of me being introspective and trying to put into words what has been bouncing around in my head, and you've given me the catalyst to do so. Thank you.

The other JohnS.
19 February 2015 at 00:54 #180295
John Stevenson 1
Participant
@johnstevenson1
John,

I'll try to take your remarks in the same order.

Firstly I spoke to one of the development guys who is in the UK and quite local to me, known him a long time, in fact I installed both his Tormach machines for him before they were publically announced and had to keep quite about it.

I saw the lathe and software well before it was brought out and will be seeing the mill next week. The lathe came about, or rather the software did because of the fact Mach3 cannot run lathe properly, mainly as regards threading. Linux CNC can but Tormach recognised that the Linux software was written by developers and not engineers so they chose to write their own screens so it works out of the box.

The biggest problem with Linux CNC and I mean to say this as nicely as i can is that it;s written by geeks, for geeks. Now straight away someone will say I use it and I'm not a geek. Sorry mate if you have been inside Linux setting it up with HAL files etc then yes in all due respect you are a geek. Even the UK development guy on the phone today admitted that it took some serious tweaking of the HAL files to get it to all come together.

However what they have now is what is needed, a box you load up with a DVD and it runs this one machine.

Not it's adaptable for any retrofit fit but just runs this machine they make.

The problem is when you come to sell a machine to a punter who has no knowledge of CNC, controllers or programs. Most can find their way around windows and Mach 3 is very customisable by even non programmers to present a controller you can teach easily.

Getting these same people to even look at an operating system they don't understand is impossible and when it looks as weird as LinuxCNC in it base form, no easy buttons for loading a job, start and stop buttons don't look like start and stop buttons and we as kit resellers have lost a sale. Whether it works better is of no consequence as they won't even try it.

People like yourself are not in this group, we have builders [ you ] and users [ the unwashed ]

You say it's not designed to be pretty, I say it's not designed anyway. It's been put together by a group of developers who don't have a clue how industry works. If the Linux CNC way was the proper way we would have industrial machines looking like it but we don't, if anything they look more like Mach3.

Mach 3 got where it was by the literally thousands of users making suggestions and Art listening, one man and not a committee, realising early on he made a screen designer so anyone could have what they wanted. You didn't need to program to use it, all you had to do was understand Microsoft Paint and drag and drop.

It certainly suited Tormach for enough years and even they had their own screens. What has come about now though is the fact they spent time and money getting a lathe to work and now it's only a tiny leap to do the same to the mill to future proof their product. Remember they sell machines not software.

Have you seen the new mill screen layout ? doesn't look like Linux CNC at all except the crude on screen graphics but with tabbed toolbars and even the layout it mimics Mach3.

Now the white paper tells of the problems with mach3 and in my opinion rather unfairly when it was such a good work horse for them for so many years. Yes the parallel port is a problem in this day and age but there are USB breakout boards on the market and have been for quite a few years. Sieg have fitted them in the KX series for 3 and a bit years and seeing as I do world wide support on these machine I can truthfully say that AFAIK there has been not one return on these cards.

I took delivery this week of some 5 axis cards on USB, far, far cheaper than any of the bolt on motion control cards like Mesa and Smooth stepper which also has problems.

Also TBH on support i have never heard of many of the problems that Tormach listed, not saying they don't exist for some, just i have not heard of them.

Tormach do not propose to offer the Path Pilot to retrofitters and the point I was making was if someone was to do these same or similar screens for Linux CNC then you would have the best of both worlds.

Just going back to Mach screens, this is the screen that the Sieg users can buy for a very nominal sum, it's done that way as the users name is hard coded in to try and stop it being stolen but there is nothing to stop anyone doing their copy.

Although it still has the tabbed toolbars as standard this front screen has been written so everything you need to load and run a job is on there. I never go to any of the other screens at all.

Nice square buttons for touch screens, all the tool offsets can be done from the tool information group, MDI for manual moves is at the bottom, feeds and speeds are also in groups and can be over ridden on the fly.

Now if that screen or one very much like it and lets face it LinuxCNC doesn't have ANYTHING like it then you would get far more users. This would answer this statement.

"To conclude: if the choice is "nice screens" or "absolute reliability", I know which one I'd choose"

John S [ the ugly bastard ]
19 February 2015 at 09:53 #180334
Another JohnS
Participant
@anotherjohns
Hello the UK JohnS;

All very valid points in your post, and I do believe that the LinuxCNC/MachineKit.io people do have a couple of ears to the ground. Maybe not enough, but some.

With regards to the default LinuxCNC screen – called "Axis" – it is most certainly getting tired. I keep meaning to try the "Gmoccapy" screen, designed for touch screen input, and the Machinekit.io people have a networked one called Cetus, if I remember correctly. So, it is an acknowledged problem, and some effort is going in. What the end result will be – who knows? (actually, it will be successful – but who defines what successful is?) (smile)

HAL file (Hardware Abstraction Layer – for others info) creation – my latest CNC mill, with a Mesa 5i25/7i76 in it, was configured by a GUI called "pncconf" that comes with LinuxCNC. My little lathe was configured with "stepconf" as it is a software stepper based machine. (I did add in axis backlash compensation for the lathe, and am wiring in my own particular config of an MPG, but, as you say, I'm most certainly a geek).

But, I think your focus is wrong. Not for the last decade, nor for today, but for the future. I see more and more computer controlled machines in the workshop (as do you, I'm sure) because of the benefits. CNC 3 or 4 axis mills are a bulk commodity now; we'll see more lathes, and 3D printers, and bending rolls and tool grinders and 5 axis machining centres and … driven by computers.

Whilst the machinekit.io people are getting their private parts kicked around a bit by the 3D printing community, they will accept the criticism and move forward. This Machinekit.io is being pushed for "future, bleeding edge" stuff, whilst the traditional LinuxCNC is conservative, for those just wanting a stable platform.

Now, the traditionalists may say "no more computers" but lets face it, stepper controlled rotary tables, electronic lead screws, digital readout systems, and so on have quietly been accepted, and now are almost the norm. With dropping prices of computers, thanks to the "mobile revolution" we have cheap small low power computers of incredible power and flexibility. Where will that lead?

The concept of the HAL file, and the inherent flexibility of LinuxCNC/Machinekit.io means that it is poised for future projects, not tied to one commodity item. (Yes, it IS rather difficult doing the HAL coding, at least at first, by hand, but the point is that it's possible – take Tormach's software, it's LinuxCNC pre-packaged – we'll see lots more like that, I'd expect)

Final point – before breakfast over here – for me, it's a LOT cheaper to order in things China->Canada as it is USA->Canada, so I totally agree with your point about the costs of hardware.

It's a changing world, and I thank you for your conversations we've had in private and public over the years.

The bloody-freezing-in-Canada JohnS.
19 February 2015 at 10:12 #180338
John Stevenson 1
Participant
@johnstevenson1
Just one small generic point.

We talk about going forward and everything these days is 'Cloud orientated ' but if you are happy with what you have and use a dedicated machine for your controller you don't need to move on and always be looking.

I use Mach 3 but a very early version as it works for me and the bits added on over the years are not important to me.

I also have a large 3 tonne machine that was retrofitted well before the birth of Mach 3 with an American AHHA system. The screen date of this program is 1993 and runs in DOS.

But it works and works well, it would be easy to convert to Mach 3 but I would gain nothing and even loose a couple of bits.

So OK it's locked in time but it suits me. All too often we look forward for the sake of looking forward, ideal as a learning tool but not so ideal if you need this machine to work.

I have a few pieces of various software where I and other users have reverted to a previous release because the new release has either broken something or took something away. The classic case here is a program that needs to go online to check the license. Not every machine is capable or allowed to go on line.

I have dumped one piece of 3D CAD software for this reason as when it goes on line it updates without permission and if I stop it going online it refuses to run as the licence is missing even though it's paid for until 2020
30 September 2015 at 22:30 #206231
Anonymous
After a wobbly with my Tormach controller on Monday morning I have ordered a more up-to-date controller from Tormach, with PathPilot installed. Should be here on Friday; pretty darn good from the US!

The existing controller died sometime on Sunday night, as I had left the machine on since I didn't want to lose work offsets. Monday morning nothing, no screen activity at all. Reboot simply gave a message implying that the hard disk had pointed its feet in the air. Panic ensued as I have a load of CNC milling to do over the next few weeks. After a couple of hours messing about I did get the mill running again. My suspicion is that it was a failed connection in the controller. It started working after I dismantled the controller to see what type of plug-in board I would need for PathPilot, and took the opportunity to waggle the hard drive cables.

Although Tormach stated that my old controller will work with the addition of a PCI add-in board there is no restore path back to Mach3 and the controller is known to be sluggish in response to user inputs. So I decided to go the whole hog and order the latest controller.

Andrew
30 September 2015 at 22:38 #206233
Muzzer
Participant
@muzzer
Andrew – when you say there is no restore path back to Mach3, you could take out the current HDD and fit a new blank one for the Pathpilot. You could then simply swap the drives back if you really wanted to "restore" Mach3. Can't imagine those circumstances mind.

But there are compelling reasons for upgrading the whole controller. Just don't tell the accountant….
30 September 2015 at 23:54 #206240
Another JohnS
Participant
@anotherjohns
Posted by Andrew Johnston on 30/09/2015 22:30:30:

After a wobbly with my Tormach controller on Monday morning I have ordered a more up-to-date controller from Tormach, with PathPilot installed.

Andrew;

I'd be interested to know what you think about PathPilot. My machines here are all LinuxCNC based, most with a Mesa card like the one that you'll be getting; I think it all works very well, but I'd like to get familiar with the Conversational Programming aspects of PathPilot. (I do conversational programming, but *not* with the PathPilot interface)

John.
1 October 2015 at 08:14 #206246
Muzzer
Participant
@muzzer
Sounds as if Tormach have made a very (typically) thorough job of redesigning the whole LinuxCNC installation so that it is almost unrecognisable – probably how LCNC would really be by now if it were project managed professionally with defined goals.

Trouble is, Tormach seem to have no plans to release it as a stand-alone application. There are no compelling reasons for them to do so and it could be counterproductive or just downright distracting for them as a business. Pity.

The only solution would be for you to get yourself a PCNC. Sounds like a good plan!

Murray
1 October 2015 at 11:00 #206269
Clive Foster
Participant
@clivefoster55965
Drifting off topic a little I often wonder why we at the hobby and small component end of the spectrum need a conventional G-Code based CNC machine control system at all.

Going back to basis a component is ultimately produced by "flying" a cutting edge along a path corresponding to its shape with the velocity and accuracy required to produce a part of suitable within dimensional tolerance. In a similar vein the process begins with a similar flight along the raw stock at a suitable depth of cut along an appropriate path to begin the conversion from raw stock to component. Most components will need several passes of appropriate depth of cut and, probably, different shapes but the principle remains the same.

All the machine needs to fly the cutter along the paths is either a string of stepper motor pulses, if its the common low cost open loop direct drive system, or a comparison system reading axis encoder pulses to send the necessary commands to a more sophisticated servo style drive. Its completely uninterested in how the stepper pulses are generated or encoder comparison work is done. Conventional systems do this live on a computer either dedicated, more elegant but costly. Or use a general purpose machine trying to stop all the operating system cruft drawing things out of sync. Cheaper but high up on the "its not how well the bear dances but the fact that it dances at all" spectrum.

Given the low cost of large amounts of memory these days it would seem perfectly practical to directly load the necessary digital path information onto memory stick and run the machine drive directly via a small buffer. Need one file per axis of course. I imagine a 64 GB stick could hold enough data to give direct stepper motor pulse outputs for a fairly respectable model size component but that is an exceedingly inefficient way of doing things. Count n pulses on one axis then stop, simultaneously count 1 pulse every z pulses on second axis works very well with a compact file. Well it did for me on a scientific instrument control running off a BBC B mumble mumble years back. The B could either run the gear or collect data but not both at the same time and stopping at every data point meant every run took forever.

I imagine Muzzers Raspberry Pi would be well up to generating the path data off line along with doing the essential but tedious homing and raw stock position location set-up stuff on line. Only has to keep up with human operator for set-up things so slow computer is plenty fast enough. Path generator could be via G-Code but probably direct derivation from CAD or via its printer file should be fine. Objectively G-Code hasn't been fit for purpose for decades. Unsurprising as it was originally based on co-ordinate drilling machine control strategies.

Clive.

Edited By Clive Foster on 01/10/2015 11:01:34

Edited By Clive Foster on 01/10/2015 11:02:40
1 October 2015 at 13:20 #206276
Another JohnS
Participant
@anotherjohns
Murray;

Trouble is, Tormach seem to have no plans to release it as a stand-alone application. There are no compelling reasons for them to do so and it could be counterproductive or just downright distracting for them as a business. Pity.

The only solution would be for you to get yourself a PCNC. Sounds like a good plan!

You can get PathPilot for a small shipping charge, and install it on your own. A few good writeups on-line show how it is done. The back end stuff that Tormach did for LinuxCNC is in the official LinuxCNC distros now. Seemingly they have no problem distributing it, but not surprisingly, they don't want to support it on all types of machines. (look in the cnczone website, IIRC)

Looked at getting a Tormach PCNC 770, but do not have the vertical headroom. Did not want to put together another mill, but I did; if the PCNC 440 was released about 18 months ago, I would have picked up one of those.

I did have a good (hands on) look at their CNC lathe, but again, no room!
Author

Posts