Advanced Search

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.

for my 4x A6 Ethercat servoswith 1kw each. and a ethercat catio module, and a mesa card.

I'm aware. I just was wondering, based upon your knowledge, if it will affect me making that part.

Hello rodw,

Thank you very much for your input, i feel more relieved knowing that about ball screws.

Regarding EtherCAT, i feel like it'd be better if i just get a servo drive (it seems to support CIA402) and cheap IO module to get familiar and maybe use it in the future projects.

For the home switches, yes, i initially plan to install limit switches and later on deal with homing through absolute position from drive and displacing the home switches as E-Stops.

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

Answer 0: Yes, thats what i had in mind.
Answer 1: Thank you very much for the offer!
Answer 2: Seems like the drives are CIA402 compatible and they provide the XML file for it so ill maybe get one drive for experimentation and getting familiar with it.
Answer 3: Most of boards seems to be out of stock, and i am gravitating towards 7i95T, having 6 encoder i can use it for feedback and use it for pendant as well, so ill reduce the need IO pins by some.
Answer 4: I wasnt considerint it for real time application, just a method to avoid homing by searching limit switches and reading the position from the drives.
Replly 5: I'll see what i can do...

Hi Guys. I have used RIOcore on a Tang Nano 9k (TangBOB board) a couple of times. I am now wanting to try out Ethercat. I was wondering what my options are for doing this. I want to use a RP5 with the Tang connected via an SPI connection to the GPIO pins on the RP5. My first question is where do I connect my Ethercat Drives? from the RP5 or from the Wiznet module connected to the Tang. I would dearly love it if all my configurations for Ethercat can be done within RIOcore.

Cheers
Andrew

Logthor wrote:

 

File Attachment:

File Name: BBU_part_V...sion.zip
File Size:19 KB

My only concern regarding using my parallel port computer with a breakout board is the delay (hence why I was considering the mesa card). Hypothetically, let's say my jitter isn't a problem. Will my experience making chips be the same as with a mesa card? I want to make some pretty complex parts and am concerned that any amount of jitter could make that impossible. I've attached an STL file of a part that I want to make for a friend. Could this be possible using a parallel port, or should I just spring for the more expensive card?
 

The main difference between a parallel port based system and a system with external Step/Dir generation
hardware is the maximum step rate, This speed will be limited by the real time jitter on the parallel port system.

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

 

So I should probably start with my Dell first and then get a Mesa card once I feel up to it? Just to confirm, you are saying that you don't think it'll be an issue for making that part (assuming the jitter is not a problem)?

leexi wrote:

Please share your comments and experience.
 

1. This card is not clone. This is original chinese design. 
2. Never upload any firmware. The proper firmware is loaded by manufacturer.
3 You have to do "mesaflash --device 7i92 --addr 10.10.10.10  --readhmid". This shows card configuration.
4. There is in your hal file line like "loadrt hm2_eth board_ip="10.10.10.10" config="num_encoders=1 num_pwmgens=3 num_stepgens=5"". Functions ommited in this line are unavailable, but you can use it's pins as regular I/O. Of course you are limited by card design. If manufacturer designed some pin as output you cannot use this as input (and vice versa) because output circuit is one way.
5. As mentioned above dkms driver may help. Also other version of kernel may help. The "error finishing read" means the break of connections caused by driver/kernel bug rather than hardware problems, however check cable.  I experienced this problem even after a few hours of proper work. The good idea is leave opened LinuxCNC for as many time as you can (no matter you do something with, or not ).
6. I tried use PnCConf two times (the first and the last) and not recommend this to anybody. Better option is to find config similar to your needs and change something in text editor.

IIRC the original parallel port could switch up to 1,000,000 times a second. Assuming two switches per pulse, 8000 pulses per rotation, that's 3750 RPM, which is way more resolution than you need (steppers are 200 steps/rot, or 3200 with microstepping), and way faster than big steppers can probably run anyway.
When I was using steppers and a breakout, it was the steppers that were the limitataion.

  My only concern regarding using my parallel port computer with a breakout board is the delay (hence why I was considering the mesa card). Hypothetically, let's say my jitter isn't a problem. Will my experience making chips be the same as with a mesa card? I want to make some pretty complex parts and am concerned that any amount of jitter could make that impossible. I've attached an STL file of a part that I want to make for a friend. Could this be possible using a parallel port, or should I just spring for the more expensive card?