Advanced Search

I feel a spam coming on...

TCP is where the magic (and head-scratching) happens in 5-axis machining. Whether you're using G43.4 style dynamic TCP or delegating coordinate transformations to your CAM post-processor, having LinuxCNC handle real-time spatial transformations keeps feed rates consistent across multi-axis arcs. Super helpful thread for anyone moving beyond standard 3-axis setups!"

Logthor wrote:

Cool, I will probably go with the parallel port in the meantime, then. I should have enough I/O with the parallel port to control my x, y, and z axes and my spindle, right? I'm just looking for a yes or no regarding the timing issue and whether it would be a problem or not in making that part (assuming I have 15000-20000 nanoseconds jitter).

Yes, I have started a video series on a parallel port control board from scratch. This is the first of 3

This is new for me as I've only used Mesa and Ethercat before.
Spindle is working fine with 0-10v control, There is enough inputs for estop, probe, home switches and outputs for coolant, mist etc so it will be complete.

The plan is to redo the spindle to use RS485 Modus and eventually migrate it to Mesa and even Ethercat. This is more of a training video than for a specific machine although I'll eventually build one for it.

 

I use ./bin/rio-flow riocore/configs/TangNano9K/config.json
The automatically generated ini and HAL files cannot output a 1kHz PWM spindle signal. Do I need to edit the HAL and ini files myself to output the spindle PWM signal?

Cool, I will probably go with the parallel port in the meantime, then. I should have enough I/O with the parallel port to control my x, y, and z axes and my spindle, right? I'm just looking for a yes or no regarding the timing issue and whether it would be a problem or not in making that part (assuming I have 15000-20000 nanoseconds jitter).

They just posted this, but 15000$ seems a bit far out, although the only thing close to it is BaxEDM and i do not know what that goes for, so might still be OK price.
I wonder if they implemented the "back off on short" functionality, i know someone on youtube did some 10 years ago.

And they use the open source nature to to their benefit with "no cloud, private, etc" , nice.

I've found having a spindle encoder really nice. Specifically, the spindle-stall detection. It's definitely saved breaking some tools. Depending on how the spindle is set up, you may be able to get that without an encoder (and it may be a non-issue anyway if you use your machine within its limits and don't have a dubiously-tuned servo as the spindle motor).

TheTinkeringMechanic1 wrote:

After posting that image of the g code I believe I see it now. The % symbol is on line 5. So that would explain everything being off set considering the line is starting at 1...
I knew it had to have been something silly.

Update: That was it.... 

Thank you for reporting back.

tommylight wrote:

Upload a picture of the whole screen.

Aciera wrote:

Also, please post the gcode

You did neither, seems you are hiding something...
Yet you ask us for help.

dst wrote:

Hello Tommy, haven't seen you for quite some time so definitely ill visit you!

You are welcomed, always.
Coffee's on me.

Is there a cheap way of acquiring tools for the mill (e.g., cylindrical edge finders, collets, drill bits, end mills, etc)?

Oops, it was in this thread...

I highly recommend NOT buying the cheapest of the cheap cheap no-name stuff. Cheap collets are prone to having high runout. The Sherline is a small machine, so you're probably using small tools. If you have 0.002" runout in your collet and are using a 1/8 2-flute end mill that's expecting a 0.001" chip load... you may well have one flute taking NO cut, and the other taking 0.003"- 3 times as much as you expect. Really easy to break tools "for no reason" like that. Even if you don't, tool life, surface finish, and dimensional accuracy will all suffer.

I don't know what collet types you need. For ER collets, I've found Techniks to work well for me. IIRC they're... $15ish each. I got mine on eBay. You probably only need a couple? 1/8, maybe 3/16 and 1/4"? You're not working with much power, so the larger tools aren't going to buy you too much (non-zero, though- larger tools for facing and/or more rigidity for deeper pockets is useful even if you can't put more power into them)For end mills, I don't use HSS, so can't speak to that. For carbide, any of the following places will get you good tools at reasonable prices:
onlinecarbide.com/
www.maritool.com/
www.carbidetoolsource.com/
routerburr.com/

Double-ended end mills can save you a bit on cost. For drills, I really like the PCB drills. At that last link, you can get them for cheap. The nice thing about them (in addition to being cheap) is they have a common 1/8" shank size, so you can use the same collet you use for end mills to hold them instead of a drill chuck (one less thing to buy, no drill chuck eating a mile of Z height).


For workholding, I've found the screwless "grinding vises" to be excellent. I've heard that even the cheapest of the cheap of these are actually pretty OK, but can't vouch for that. I got mine here: littlemachineshop.com/collections/screwless-vises Just make sure you get clamps that can hold it down.

I haven't tried these, but supposedly the Xin Dian self-centering vises are shockingly good for their price: xindianprecise.com/collections/manual-self-centering-vise If you're flipping parts over, you can avoid having to re-center in one axis at least... two if you also use a work stop.


Other assorted tools: IMO buy as you go and buy nice used stuff. eBay can be good for this, but also if there are industrial surplus auctions near you a lot of times you can get the common stuff dirt cheap, since a lot of that customer base already has the basics you're looking for. Searching "<name of big company around me> surplus" may turn up auctions that a lot of large companies regularly have to get rid of stuff.

Sorry you did not mention for ethercat, mine are for Mesa and parallel port.

Things a machine can physically do are absolutely the same for parallel port and Mesa card, so there are no limitations either way.
Things that are the same:
-you can make whatever you, your CAM, and your machine can do
-any machine, 3, 4, 5 or more axis, robot, lathe all will work
-you can use more than one of any of those, and you can combine them as you need.
-
Things that are not the same:
-pulse speed and subsequently the machine speed
-Parallel port can do approx. 20000-50000 pulses or encoder reads per second
-Mesa can do over 1000000 (1 million, some even 10 million) pulses per second
-Parallel port can have 12 outputs and 5 inputs in OUT mode, 4 outputs and 13 inputs in IN mode, 8 outputs and 9 inputs in mode X. Not all ports can do this, but all built in ones do.
-Mesa can have any of the 34 pins (some cards have more IO) set as input or output in hal, granted it is not used as Stepgen or PWM or encoder or sserial.
-
Or in short, parallel port with LinuxCNC will be enough for any hobby machine and most 24/7 use machines, even for some industrial machines that are not to complex
-
Yes, i do have a lot of Mesa boards/machines (probably over 70 or 80 or more), and yes i still use parallel ports and still have machines in daily use with both for over 15 years (i am fuzzy on the exact numbers and timelines)

@NWE: Re box size... that's what I was poorly trying to say- "get a bigger box than you think you need"

@Logthor: Re cheap tooling... there are few things you don't want to be too cheap with. IMO power supplies, tooling, and motor drivers are 3 of them. ALL of them can drive you crazy and/or totally ruin things.


Power supplies: The PSU is one part of your machine that can easily take out all the rest of your electronics if it fails wrong. Plus super cheap supplies can be fire hazards, and power issues can be a nightmare to debug. Meanwell is a good suggestion.

Motor drives: Listen to @tommylight, he has way more experience than I do. My experience (15-20 years ago) was that the cheap drives I bought wouldn't work well no matter what I did- the steppers whined, missed steps, jerky motion, and eventually a crater literally exploded out of one of the ICs. Spent way too long trying to chase jitter, signal issues, etc., and it was just crappy parts.

Tools/tooling/collets: I'll post a reply in your other thread, since you asked about it there.



Other things:

• I linked to Geckodrive because of their explanation of motor torque/speed/voltage/current. It's informative regardless of what drives you use.
  
• You mention in your other thread wanting to do complex parts. To me that implies multiple tools per part. You can definitely get by without a toolsetter, but I would at least not cut off the option to add one without too much pain. It doesn't take many broken tools and/or scrapped parts to pay for one.
  
• MPG == manual pulse generator. It's a hand-held thing that lets you jog your axes about (and other functions, depending on how fancy) without needing to be at the keyboard.
  
• I forgot to mention ferrules! Way better than putting bare wires into screw terminals, if that's a thing you're doing. I got a cheap ferrule + crimper set off eBay for my machine, and it was a very nice QoL upgrade. Not essential, but nice.

As Peter said,

PCW wrote:

The output hardware does not limit the complexity of the created shapes just the maximum speed.

I suspect however you will find limitations due to the part geometry on any machine.