Page 4 of 8 FirstFirst ... 23456 ... LastLast
  1. #31
    To avoid chicken egg problem, why not start from the Z. Usually the weakest spot on any machine. THEN this will command the gantry design. Or you have it clear there?

  2. #32
    George, I forgot you were cutting stone so maybe 4 bearings per side is not overkill, my mind is currently filled with routers for wood Will wait for drawing.
    Last edited by EddyCurrent; 18-12-2013 at 09:24 PM.

  3. #33

    Quote Originally Posted by silyavski View Post
    To avoid chicken egg problem, why not start from the Z. Usually the weakest spot on any machine. THEN this will command the gantry design. Or you have it clear there?

    i started with the plan that the z should be able to handle different spindles+additions, without differing what forces it can handle in both +-xy. most of the moving gantry designs, like the one you linked to, favour one direction down X, and rely on weight of the garntry to offset the stability looses you get by having to mount the spindle forword of the Y, on the drawing you link to, every mm forword the z mech is away from Y, the force you can apply at the bit tip drops substantialy, and would be different X+ than X-.

    so, the problem/limit for z, when considering deflections , is all in the Y, when you factor in a 4th axis, again the differing forces between X+ and X-, start loosing accuracy , the longer the jobs running, the worse it gets, the bigger the job is down the x axis, the faster this will happen. .

    but, if you have 2 of them Y axis you linked to, facing each other, and the Z in the center, Zs force is equal on X+- and Y+-. the forces Y can take, could be ajusted by spacing th 2 bearings per rail further apart

    then, Z drop, is the next place you start looking at, will try and get it in sckechup tomorrow, failing that, i will take a photo of my drawings and post that

  4. #34

  5. #35
    Hi George,
    i don't know from where you dig out that boards, as i am also at this point but see them for first time.

    The first one has its outputs 3v, so you have to check out if compatible with the drivers you intend to use
    The second one seems I/0 board only, unable to generate train pulse for motors if i am not mistaken
    The third, seems some problematic chinese board

    It seems money is an issue, then take a look at :
    - usb boards with proprietary software instead Mach3, if you still don't own it
    -if you want to go Linux, look at the Mesa FPGA boards, search the Linux forum
    -CNCdrive - motion controls

    Of course depends how many axis you will like to drive,

    Then comes the Ethernet smooth stepper + BOB / PMDX126/ or a CSMIO 4 axis board/ or the expensive 6 axis, which in fact has all that one should look for in a board altogether/

    Another possible solution i am looking at now is the KFLOP, which for the moment seems to be my choice, money wise for what it offers

    Pokeys is also a good money wise .

  6. #36

    Moneys not a problem, but am bulding 2, test one first, then production one, so its ok if i make the odd mistake. If i find a board that does everything i want, and someone else has used it, i will buy it.

    am looking at the solutions you pointed to, and this looks good

    More axis the better, if possible, i would like to stick to Mach3/4 and Linuxcnc, just becuase of the learning curve and macros. I will read up on that one this weekend.

    and as you guessed, i didnt read option 2 before i posted,

  7. #37
    Another option i am considering which looks very promising is the Centipede KSI Labs, LLC .

  8. #38
    Quote Originally Posted by silyavski View Post
    Another option i am considering which looks very promising is the Centipede KSI Labs, LLC .
    It's a very nice bit of kit but be careful because I think the developer has thrown the teddy out the pram with Artsoft because he doesn't like the direction they are taking with Mach4.?

    Regards the Kflop and Galia etc then unless your intending on using Servo's with high count encoders then your wasting your money has other than quality they offer no more advantage than the ESS or other Sub 200 cards for steppers.
    If your using steppers and don't need slaved Axis then the Csmio-IP-M is by far the best I've ever used and use's industry standard 24V differential signals. there's No Bob's needed and very high quality units in a neat package.

    If I needed higher pulse rates for servo's then I wouldn't hesitate to buy the higher priced Csmio-Ip-S which will give many of the other much more expensive high end motion controllers a good run for there money.

    The Kflop is a nice bit of kit but it's still using USB and Ethernet is so much better and noise immune compared to USB. I've never been a fan of USB for Bob's etc and while I didn't have too much trouble with using the USB version of the SS I did get some troubles. With Ethernet I've not had any issues with noise at all other than a Dodgy card and can run machine from in the house with a laptop sat on my knee if I like using the network.!! . . . . I don't but I could.!!

  9. #39
    i need 5 or 6 axis minimum , am fairly sure i will need to slave a 2nd x at least, I need to stick to bobs that can run mach3/4 and or emc, just because of the support you can get if you get stuck. I need at least spindle control and 2* encoder reading capability's from the electronics. As well as spindle on the y ganrty, i want to build in a turret into the machine frame, and have a sepotate 4th axis, and seporate b/c head. i could do it all with 5 axis, just mean swapping the leads around when i want to set the machine up different,

    i had guessed through reading the other threads, that a smooth stepper is a good idea, and you have just confirmed that, am looking at the other cards mentioned on here, am going with steppers not servoes

    Am up for any recomendations, ?

    and can run machine from in the house with a laptop sat on my knee if I like using the network.!! . . . . I don't but I could.!!
    Ha ha, not without a really really long estop cable.

    Tell you what though, you have just got me thinking, i wander how secure the cards are on a network and how they would react to being probed,

  10. #40
    Sorry, don't want to hijack the thread, but here is what i think.

    I agree with all that Dean said. But the reason i am still digging for BOBs is axis count and money i would pay for what i would get. Similar to you i will start with 4 motors. Then i want a minimum 1 axis for rotary table, possibly 2. So -6 axis until now. Later i am thinking of separate axis maybe for plasma and may be one more for a device, that pushes MDF sheet while routing and lifts / i saw this somewhwre on internet.
    I agree that all can be done with 6 axis and some thinking it out. But also i am living in rented house and in foreign country, so may be one day if i have to change country will be best to sell the machine, so axis count could be better

    OK. After continuous digging and analyzing if multi axis tralalala is desired , here are my personal conclusions:

    1. No usb boards. They will be obsolete in an year or two. PCI-maybe. Ethernet-best.

    2. only 3 boards/ combinations/ for me are worth the money, which support 6 or more axis:

    -ESS +PMDX126+PMDX-107 - quite cheaper than the CSMIO, 500Mhz step frequency, mach3, 180$+40$ shipping + 174$ +$57.00 +~40$ shipping = ~410 euro +eventual tax
    -CSMIO- 5-24v differential I/O, 6 axis, mach3, 549eur+23% vat= 675 euro +??shipping= ~700euro , so, very nice but quite expensive in fact

    -Mesa 7i80+7i76 or 5i25+7i76 . More and more seems Linux is the way. First combo is 119$+150$ +~ 40$ shipping=~250euro , second combo is 200$+~40$shipping =~180 euro.
    Basically the first combo is similar to the CSMIO with ethernet, 5v-32V, differential I/0, etc + the added benefit of aditional 5 axis for only 100$ more+ this baby has 4 slots, so 20 axis or aditional things can be put there
    The second combo is PCI and again axis can be added later.
    And not to speak of the drives and other boards that Mesa have and can be seamlessly integrated for cheap.

    The other boards from the previous posts i ignored , due to USB, due to limited support, price per what they offer, lack of differential 24v I/O

    I wonder why nobody wrote a serious plugin for the Mesas and Mach3. This is f+==g crazy. Furthermore they have a new board still not supported in Linux -7i76E which is Ethernet 5 axis for 199$ that can be further expanded.

    What i will do is start to play with Linux to see if i am up to the task and save some money.

    PS. If i knew just before a week what i know now, i would have bought the Galil 4080 for 540eur at ebay just on 31. December. But somebody knew this so he bought all 3 of them. Bloody ignorance of mine.
    Last edited by Boyan Silyavski; 06-01-2014 at 12:03 AM.

  11. The Following User Says Thank You to Boyan Silyavski For This Useful Post:

Page 4 of 8 FirstFirst ... 23456 ... LastLast

Similar Threads

  1. Design Help Pt2 Required for CNC design/Build
    By MikeyC38 in forum Gantry/Router Machines & Building
    Replies: 38
    Last Post: 21-07-2014, 01:05 PM
  2. Design help etc required with DIY CNC Router Design / Build
    By MikeyC38 in forum Gantry/Router Machines & Building
    Replies: 12
    Last Post: 21-10-2011, 03:50 PM
  3. Opinions
    By p1lts in forum General Discussion
    Replies: 8
    Last Post: 03-02-2011, 07:47 PM
  4. RFQ: Manual Lathe Repair & Syil SX3 CNC Re-build Required.
    By Chas in forum Projects, Jobs & Requests
    Replies: 0
    Last Post: 07-11-2010, 08:28 AM


Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts