Emco PC 55 mill conversion sugestions

More
07 Jul 2018 05:06 #113633 by Aaroncnc
Hello,

So i recently got a Emco pc mill 55 with original PC and controller board all all the software disks to go with it.
However it would fail to boot about 50% of the time and recently the HDD has given out. (CD drive also dead)
The system ran windows 98 with winNC.

I am faced with 2 choices
1. Rebuild the PC with vintage HDD (something around 1-10 gig IDE HDD) and reinstall the OS and all the software.
2. Replace the brains and use one of my more modern PCs (intel core 2 duo E8500 or E8600 rigs with built in parallel ports) and keep the original VFD and stepper drivers as they are 5 phase drivers and i dont want to replace the motors as i am on a budget.

The pros and cons im not quite sure of them all yet

PROs of a new rig are more stable and longer lasting hardware. More modern control software. Ability to run linuxcnc.
CONs Seems like the brain replacement costs can vary wildly between around 50$ and 5,000$. Unsure on what i will be gaining or losing depending on what part of the spectrum i choose.

PROs of fixing the original system is possibly cheaper as long as nothing else dies. I dont have to reconfigure any the software and will be setup correctly out of the box
CONs System may lose another part and if something like the mobo goes with its ISA slot finding a replacement will be costly.


My thoughts on the new machine was to use the E8600 dual core with 4 gig of ram and a 120GB SSD with no networking or audio enabled. Using the mobo parallel port to drive a breakout board. The break out board i was looking at is Breakout Board and maybe add a pci parallel port SYBA SD-PEX10005

So time for my question,

Do you guys think its a good idea to keep buying vintage used HDDS and PSU's to repair a machine this old?
Any one also running hardware this old care to comment or how long i could expect it to last?
Better to just build and go with the newer?
Is my hardware i picked out going to limit me in any way?

I am somewhat cost limited here so as much as i would love to go with some of the higher end cards for the new system i just dont have it in the budget.
More
07 Jul 2018 22:37 #113731 by andypugh

Aaroncnc wrote: Do you guys think its a good idea to keep buying vintage used HDDS and PSU's to repair a machine this old?


No, because you retain a closed system that will never advance, and where bugs will never be fixed.

I guess that the original Emco board consumes G-code and outputs steps? Do you have any idea what the maximum step-rate is?
The following user(s) said Thank You: Mike_Eitel
More
08 Jul 2018 01:22 #113750 by Aaroncnc
the max was around 30ipm.

Just am still on the fence with this machine.
I know your right in that the new software will only get better and what i have now is out dated.

I think the main thing holding me back is the idea of going to linux. Not that it is bad just the idea of starting over and having to learn everything again.
More
08 Jul 2018 01:33 #113752 by andypugh

Aaroncnc wrote: the max was around 30ipm.


Do you know the pitch of the screws and the steps/rev of the motor. From that you can work out the required step rate.
The reason this matters is that if it is less than about 50kHz you can use the Parallel port and the hardware is approximately free. If it is much more than that then you are going to have to pay $200 or more for a Pico or Mesa (Or General Mecahatronics etc) card.


I think the main thing holding me back is the idea of going to linux. Not that it is bad just the idea of starting over and having to learn everything again.

I did it. I was a Mac user until I decided to install LinuxCNC to run my CNC machine. (It was that or Windows / Mach and that was even less tempting)
More
08 Jul 2018 02:02 - 08 Jul 2018 04:59 #113753 by Aaroncnc
I got this info from browsing the web and seems to be accurate.
From the documents I have, the original max feed rate was around 30 inches per minute on each axis. This was with a 2.5mm pitch ballscrew and a 2:1 belt drive reduction. Given this, the original motors were driven to a max speed of 609.6 RPM ( ((30 ipm * 25.4 mm/inch) / 2.5 mm/rev) * 2).
A full step is 1.8 but can be driven lower.
Last edit: 08 Jul 2018 04:59 by Aaroncnc.
More
08 Jul 2018 12:13 #113800 by andypugh

Aaroncnc wrote: I got this info from browsing the web and seems to be accurate.
From the documents I have, the original max feed rate was around 30 inches per minute on each axis. This was with a 2.5mm pitch ballscrew and a 2:1 belt drive reduction. Given this, the original motors were driven to a max speed of 609.6 RPM ( ((30 ipm * 25.4 mm/inch) / 2.5 mm/rev) * 2).
A full step is 1.8 but can be driven lower.


So, at least 2kHz wit x1 microstepping, perhaps 8kHz with x4 microstepping.

This is within the range that the parallel port can be used, so maybe start that way and with a cheap BoB from eBay or Amazon, this sort of thing:

www.amazon.co.uk/Powlance-Stepping-Inter...eakout/dp/B07DPJQVDQ

If you don't want to use the parallel port then you might want to consider store.mesanet.com/index.php?route=produc..._id=290&search=7i76e (connects via ethernet) but that is a much bigger cash outlay.
More
08 Jul 2018 19:35 - 08 Jul 2018 19:35 #113831 by Aaroncnc
Could you explain the 50khz limit to me?
Is that the limit at the parallel port to send instructions?
So going cheap for now only limits my max ipm worst case?
Or I just turn down my microsteps and go closer to full steps and get more ipm is that correct and lose a bit of resolution?
Last edit: 08 Jul 2018 19:35 by Aaroncnc.
More
08 Jul 2018 22:35 #113848 by andypugh

Aaroncnc wrote: Could you explain the 50khz limit to me?
Is that the limit at the parallel port to send instructions?


LinuxCNC relies on a real-time kernel to make the step pulses at the correct time.
Most pCs can't manage much better than 20µS jitter on the thread timing, so 50kHz ends up as about the limit.
(Some machines can manage <5µS so they can step a bit faster)
If you go to an external step generator card then you can step at MHz rates. This isn't necessary but does mean that the step-rate _resolution_ is a lot finer. (Imagine the 20µS thread, at the top speed you can do 16kHz, 25kHz or 50kHz and nothing in between)


So going cheap for now only limits my max ipm worst case?
Or I just turn down my microsteps and go closer to full steps and get more ipm is that correct and lose a bit of resolution?


Indeed, and you can always upgrade the hardware later.
More
10 Aug 2018 13:35 - 10 Aug 2018 17:08 #115862 by Aaroncnc
Well, i decided to due both options.
I rebuilt the PC and got it running stock.

I came to the conclusion that it would be much better and faster to do a retrofit if i had a working system to probe out and learn from.

So i have ran into a bit of a snag that i think i have a solution to but not sure.

The system uses 2 signals 1 normal and 1 inverted signals for step and dir on all axis. (0v,5v)/(5v,0v)
So it uses 12 pins in total.
So i decided to use a Hex inverter ic to take and generate the other signal i need.
My only worry is that i feed the one signal directly into the driver board and then it goes to the hex inverter then goes back to the driver board to the differential signal input.

Do you think there will be any timing issues with the differential line being a tiny bit behind the other due to only using 1 ic?
I understand that i could add a non inverting hex buffer to equalize it but i dont have that part right now.
Last edit: 10 Aug 2018 17:08 by Aaroncnc.
More
11 Aug 2018 21:20 #115918 by andypugh
Do you have enough IO to generate the inverted signal from inside HAL?
Moderators: cncbasher
Time to create page: 0.080 seconds
Powered by Kunena Forum