What are the requirements for the computer to run the reprap?

What are the requirements for the computer to run the reprap? The reason I am asking is that I have a few single core AMD athlon 64 processors and a few motherboards they will fit in and 4gb of the old 184 pin ddr ram. I was thinking that I could bring the parts back home with me when I leave from here and build a dedicated reprap computer. Do you think a 2ghz processor, 4gb of ram and an old 7600gt video card will run it? I would do all the 3d design on my much much faster photo editing rig.

What about opperating systems.

I still have 2 liscences of XP SP2
I have a copy of Solaris 10
and I have an upto date copy of Ubuntu

I barely know my way around Solaris and am not at all familiar with Ubuntu. XP I am a wiz at but use vista ultimate 64 bit on all my personal computers.

Well,
From what I know:
the host runs fine on my single core amd64-3000 with 2gb ram under winxp.
there are no pronlems with my nvidia card either (you'll run into problems with ubuntu and this videocard; at least with the live-dvd)

portable host for winxp:
[forums.reprap.org]

I don't have a working machine yet to test it under building conditions;
but since everything else runs fine I see no reason why that shouldn't work.

'sid

Just for reference, my little EeePC 701 doesn't work. It's got 1 GB of RAM, but something in there causes significant pauses when I try to print from it. It's running Xandros, but still had problems when I booted the 0.6 RepRap Ubuntu DVD.

I don't know about that particular video card, but I've got two laptops with Nvidia cards built in, an old Toshiba and a new Mac, and they both run the Ubuntu DVD fine.

I have been running out of RAM on my 1.8 GHz 2 GB Toshiba when loading large files into ReplicatorG, but that's more of a problem with the software which I believe was fixed in the more recent versions.

Dual CPU's does seem to help a bit; anytime you've got any sort of pause in the computer (like a disk save, garbage collection, whatever), it results in a lump of extra plastic while the extruder is waiting for the next Gcode. Dual processors seems to keep things moving right along most of the time.

Once we get things optimized, you should be able to run a RepRap off just about anything, maybe even just an SD card, but at the moment you need something with a little bit of power to keep things working smoothly.

If it will run Alibre, you're probably good to go.

Wade

I guess I will just build another AM2 rig. I should be able to put one together with onboard graphics for less than $100. 2gb of ram is dirt cheap now. I have 2 old eIDE 80gb hard drives and a 2.2ghz athlon x2 processor I could use. Then to just throw in a cheap sata dvd drive and a copy of xp and would have a running reprap only computer for less than $100. Ill wait until I am almost finished with the build before I build it tho.

Good to know that about the eee pc. I was planning on buying one and upgrading it's ram to 2Gigs. when you were using the 701, did you run it at basic clock speed, or upgraded clock (the processor was underclocked I remember reading somewhere).

I'm looking into getting a really cheap laptop that I can carry around when I'm doing presentations with my reprap. As far as I can see from the progression of the control programs, the requirements of the system in terms of CPU and RAM are going down.

Well, if it's that much of a resource hog to run a build, then it looks like I'll /have/ to arrange something else when I get that far; all of my computers are substantially less beefy than that. And all of them run brl-cad perfectly well, too... Something seems more than a little off if it's taking more than an eeepc to run a build.

I think what they are saying is that without ample processing power and a normal amount of ram that the machine will pause for a second when the cpu switches processes to something else. It happens with anything connected to your serial or USB ports. If I am exporting photos to my external hard drive and I load something like WMP or photoshop the cpu will briefly pause the data going to the external hard drive and use those processes to quickly boot up photoshop.

The eepc was not designed for heavy use. It was designed for the One Laptop Per Child program that President Bush started. Its processing power is limited as in under-clocked for heat concerns. It was deigned to handle basic things like word processing, audio, web surfing, and power point presentations. It was never meant to run resource hog applications like 3d rendering, or photo editing. For what you can buy an eepc for you can get a low end full size laptop that is more than capable of running the reprap.

Also for that an EEpc cost you can build a fairly stout desktop tower. I recently built a new photo editing / gaming / cad rig for under $400 sans video cards. With a mid range video card you could build it for under $420. I have 2 $150 video cards in my rig but those are mainly for gaming.

Maybe we should for the time being create a minimal PC requirement list. It would not take long to make up.

I like to tell again...
The "smallest" machine I tested the host with is an via epia m 900 all in one board (eden cpu 900mhz) added just 512mb ram; inbuild videocard lan usb and such.
Runs fine with winxp and the portable host (in test mode, no machine attached)
ubuntu live dvd is only capable of 800x600 pixel

'sid

Just to be clear, the EeePC ran the RepRap host software just fine (0.6 DVD), but when it came time to print, I had problems with pausing between Gcodes causing big blobs in the print. I was using ReplicatorG as well as the Python Send program, in Xandros as well as a couple flavors of Ubuntu.

I don't think I ever print tested it with the RepRap Java host software, because as at that time, last summer, the RepRap Java Host had pausing issues on all of my computers. Personally, I've found there's a big difference between being able to run the software, and being able to print good objects. I have not tried the more recent versions of the RepRap Java host on it though. All of the software is a bit rough around the edges at the moment, and with my lack of proper coding skills, it's simpler for me to use a faster machine, than try and re-write the software.

Keep in mind, Forrest has been running his Tommelise printer directly off a Pic, no computer attached, and tofletcher has his RepRap running of a PIC32 reading Gcodes directly from a SD card, no computer required either, and I think they both get better performance than I've been getting. The issue is not CPU horsepower, but lots of CPU horsepower can mask the other issues we're having with the software at the moment. I suspect Chris's new Gcode interpreter will work fine with just about any computer than can keep up with 19200 baud on a USB port, but I haven't got it running yet.

Designing parts vs printing them is another question, but I do know the EeePC will run Blender and AoI just fine.

About the EeePC - it's not part of the OLPC program, it's a competitor to it. I probably would have bought an OLPC instead, but they weren't available back when I bought the EeePC. It's actually a pretty kick ass little machine, but my hands are too big to use that keyboard very long. I have been running Blender on it quite a bit, it's got plenty of power for basic editing of RepRap models; I did many of my edits on it while my power hungry laptop was busy printing out parts. I also used my EeePC to help troubleshoot a solar charge controller in a third world school last year; that's not something you can do with a cheap desktop.

Finally, SCphotog; I was under the impression that Nicholas Negroponte started the One Laptop Per Child program (www.laptop.org), and that it's been running off corporate and private support, with zero support from the US government. Maybe you could prove me wrong with a link or two showing Bush's involvement? I couldn't find any references to him on the OLPC page. It's an awesome project, and I'd like to make sure proper credit is given.

Wade

Edited 1 time(s). Last edit at 12/21/2008 11:27AM by Wade Bortz.

Wade,
I was just going off of a report I saw on cnn a few weeks back about the OLPC program and Bush was at an event in 2006 speaking about the No Child Left Behind act and spoke for a few minutes about the OLPC program and IIRC he mentioned that it was part of his NCLB program. I could have heard him wrong tho. I rarely pay any attention when he speaks anymore.

A quick Google search confirms that you are right. Nicholas Negroponte started the One Laptop Per Child program. I was under the impression that the Asus EEpc was part of the OLPC program because of an article in PC mag last year that did a comparison of all of the mini laptops in the program and the latest EEpc at that time was part of the test. I remember this because I was thinking about buying one to serve as a compact flash card dump machine for when I covered long events or irreplaceable photo events like weddings and sporting matches. I wound up buying a purpose built cf card dump hand held that has a 3" lcd and 36gb of ssd storage for about $150.

Hi all, I'm new to RepRap.

I've been looking in on progress now and then since I first found out about it a year or so ago.

I note that some of the controller designs have on board memory, but it's looking like this may only be for buffering, after reading this thread.

How big can the output files get ?

What's the difference in compactness between gcode and it's predecessor ?

What would be a big enough memory to assign for the controller custom hardware, that would comfortably hold the most complex model you could imagine it being practical to build ?

And on the PC requirements side, assuming you can buffer the WHOLE model in reprap memory before beginning a print, then it seems to me you could run a very old PC for both CAD and the conversion program, that runs the algorithms to convert to gcode.

Graham.

Hm, I don't think we have an actual predecessor to Gcode; the old way as I understand it was to generate the commands on the fly from the STL and send em straight to the steppers. Some people have been using a binary command format (Forrest and tofletcher off the top of my head), which should roughly halve the file size, but it's not a huge problem yet. I think the idea was to reduce the processor load, not the file size.

I generally send output files of several MB to the RepRap to build over the course of several hours. For example, the 5 extruder parts together take 2.6 MB. 4 corner brackets generated a 12 MB gcode file. This is at 0.5 mm filament; as you reduce the filament size, the file size is going to increase very quickly.

My Gcode directory has roughly 200 MB of Gcodes piled up in after building a complete Darwin, but that includes lots of tests, mistakes and do-overs. I think you should be able to fit an entire Darwin in less than 100 MB at 0.5 mm easily.

The actual data rate is quite slow though (19200 baud), so with some proper buffering at the Arduino/Sanguino end of things, anything that can keep up with 19200 bits per second should suffice. At least until Nophead gets his warp-speed Darwin variant slinging plastic at 50 mm/s. For what it's worth, he uses much higher data rates on his comms. In the absence of buffering, I've been running at 57600 baud to try and cut down on the delay between codes.

Personally, I'm leaning towards one of Zach's motherboards with a 2GB+ Sd card slapped in there; that should be room for almost anything, and future proof to boot.

Ideally, the electronics package will come with a complete data set and a clearly marked "Replicate" button on the front of the control panel, but that might be a year or two off yet. Maybe not though, things are picking up...

Wade

Thanks Wade,
in the interim, I've posted my thoughts on the electronics platform here:
[forums.reprap.org]
, and I appreciate your information on file sizes

I will hunt out the info on Zach's motherboard, but I'm thinking I need more power on the embedded side, for resolution, and to avoid glitches.

I don't have one yet...

Lots of work ahead for me I think

I believe a dual core EeePC is supposed to be launching some time in the future.

Also the OLPC is designed to be as open source as possible so you can guess that the EeePC which is primarily AFAIK a closed Asus developement is not part of it. Besides that the OLPC team have had problems with Intel recently, they feel that they're (Intel) trying to undermine the program by convincing people/countries to go with Intel based comps (which are alleged to be better but more expensive) instead. I believe that Intel had officially pulled out of the OLPC team a while ago, don't know if they've rejoined since then.

Possibly Wade's problems with the eepc:

I believe it has flash disk instead of a standard hard drive. 1G may be enough memory for a PC file swapping to hard drive, but the issue may be file swapping to Flash disk being slower. They use large sectors, but there are still significant delays with the write times, and I suspect, no facility to overlay writes to very many pages at a time.

This would explain your symptoms, if the print process is using up large chunks of memory... You could talk to those on the print development side, and see if they can do a tweak to process in smaller chunks at a time.... and to reduce repeat accesses of data out of the addressed chunks

I don't know how the memory system optimises data access..