Foam XL hot wire cutter

[Boyan Silyavski]

Hi, after adding some hot wire attachment (http://www.mycncuk.com/threads/11038-DIY-Plunge-Foam-heat-wire-cutting) to my main machine wich showed excellent results cutting 1200x600x50 foam sheets, i decided to build something bigger. And better.

At the moment i use DEVFoam LE to generate the paths and i believe this is the best program at the moment. Yep, tried them all before i purchased this one. Its worth every cent.

Basically i would like to be able to cut 1200x2300 foam blocks 10-1000mm thick. After considering this and that (http://www.devcad.com/eng/devcncfoam_f1.asp), i decided to purchase another application DevCNC Foam and drive the contraption using Arduino Mega and Ramps 1.4 which connects via USB to PC. Then from DEV CNC Foam on the PC i could visualize what's happening and most of all hit a button and all axis home and square. Aint that great. The big deal is that Dev Foam CNC has and upload feature to the Arduino, so hopefully i will not mess with firmware.

As you can imagine after calculating this and that price, for sake of sanity and simplicity i will have Hiwin 20 on all axis. Not Hiwin, but the copy they sell from Germany on ebay. Hopefully it will slide ok.


The challenge is what to use for linear movement. I wanted ball screws but price wise decided on lead screws as speed will not be big. So choice still is between 10mm and 8mm lead screw which will spin a whopping 2500mm length. I need 200mm per minute cutting speed which i find is the best quality wise, when cutting with 0.16mm constantan wire. So with 2mm 2 start lead screw / 4mm per turn/ i will need to spin the screw mere 50 times per minute. Anyway i decided to upgrade to 24V the Ramps board and use DRV8825 drives that will be coupled with the most powerfull Nema17 Step Motor 42BYGHW811 4800g.cm 2.5A 2Ph. may be even at 1:3 for the sake of resolution. thought at 50 steps per mm resolution will be 0.02 and if i use some micro stepping even better
I believe the motors have to be good enough to move the "gantry" at that low speed and low acceleration. I am somewhat more worried of the natural sag of the long lead screw. i will design and print the nut, so backslash is not a worry.

here is the initial design. It could be wide as much as i like in future by just changing a couple of aluminum pieces . i am still contemplating to fully separate the sides and connect only the " gantry bar" which is a must for everyday use as i will be cutting mainly letters not wings. Profile for the sides and gantry legs will be 30x60 Bosch. It must not vibrate and be strong even if i am radiation cutting, as if it vibrates in any way, then i have to sand. And that's a tedious process ... better to be perfect quality result from start


I have contemplated another linear solutions, offline controllers, bigger motors but at the end i believe this is the proper way to do it. FYI end result will be better than 6k foam cutter which dont even use supported rails :thumbdown:

Hopefully i will even avoid separate PSu for the wire and heat it from Ramps

22460 22461

[Robin Hewitt]

I have a need for cast iron cannon axle boxes, a sort truncated cone with the inside removed. I have large lumps of elm wood in my garage waiting for them. I did get some bod up North to make me a pair but he decided to keep them in his garage rather than sell them to me. More than a bit odd. However this machine you are proposing could cut me a new pattern for moulding such things, easy peasy.

[Boyan Silyavski]

I have a need for cast iron cannon axle boxes, a sort truncated cone with the inside removed. I have large lumps of elm wood in my garage waiting for them. I did get some bod up North to make me a pair but he decided to keep them in his garage rather than sell them to me. More than a bit odd. However this machine you are proposing could cut me a new pattern for moulding such things, easy peasy.

Once is ready i could cut you whatever you like. but now you have got me- You have a canon at home???

[Robin Hewitt]

Doesn't everyone?


http://www.robinhewitt.net/Falkirk.wmv

[Boyan Silyavski]

Any idea for possible cheap solutions for the long rails ??? I decided on 1200 and 2200mm travel .

I am currently exploring options for the way to do it and its seems it is hit and miss

-The oozenest C beam seems perfect but they stock it up to 1500mm. And i need 2500mm

-Mjunit is ready made system but such lengths from China will be 1500euro in total to receive it here after customs

-Similar problem with Motedis, too long travel. I will wrtite them now but i see the 6mm steel bearing rod that fits

-meanwhile talking with Fred from BST

-The german CNc Discount does not answer me or i would have bought the rails from him


Any idea how to achieve that travel on the cheap? And best with stuff from Europe? And if not on the cheap, just on the reasonable side? I will not buy Chinese rails with Hiwin Price tag from EU. Now if they are cheaper ...

I am starting to think people don't want to take my money :cower:

[charlieuk]

I have been trying to come up with basically exactly the same thing for myself to cut eps block to put on my cnc to cut my surfboard however I need to be able to cut down 16' x 4' x 2' blocks.

I have been thinking about it for a year or so trying to figure a way to get some cheap linear motion. I was wondering about just using a wheel that runs on the ali extrusion but then thought maybe that would not be enough. then there are some of the methods available on cnc router parts but in the end I always come back to just trying to find some cheap low quality square rails as they will probably be better than all the above and a lot easier to set up.

my plan for the drive was a belt, it should be more than accurate enough

looks like your going for just two axis rather than 4? I would love 4 axis but think I will probably go 2 for simplicity.

[charlieuk]

the one in this vid uses booth belt and wheels straight on the extrusion giving something cheap and simple.

http://www.youtube.com/watch?v=-kWGnj5FeI8

[Boyan Silyavski]

the one in this vid uses booth belt and wheels straight on the extrusion giving something cheap and simple.

http://www.youtube.com/watch?v=-kWGnj5FeI8

Thanks for link, but i don't like bulky things


Deadly accurate, dependable, simple. Price starts from 16.900$ :joker:

Thats exactly the machine type and design i hate. Big servos, where nema 17 will do the job, 10 times bigger than needed aluminum profile. In other words, overbuild thats showing absolute lack of understanding. Not me.



I have been trying to come up with basically exactly the same thing for myself to cut eps block to put on my cnc to cut my surfboard however I need to be able to cut down 16' x 4' x 2' blocks.

I have been thinking about it for a year or so trying to figure a way to get some cheap linear motion. I was wondering about just using a wheel that runs on the ali extrusion but then thought maybe that would not be enough. then there are some of the methods available on cnc router parts but in the end I always come back to just trying to find some cheap low quality square rails as they will probably be better than all the above and a lot easier to set up.

my plan for the drive was a belt, it should be more than accurate enough

looks like your going for just two axis rather than 4? I would love 4 axis but think I will probably go 2 for simplicity.


Mine will be 5 axis in fact. But it will be used as 2 axis most of the time as what i will be cutting is letters mostly. Plus i dont have the place to fix it properly for now. So it will start as 40cm wide 2 axis, that could be transformed to 2m wide 4 axis for under an hour, when need arises and when i find bigger place.

To sum it up for the linear movement and structure:

1. Motedis failed me this time. Blah Blah Blah, we dont cut and ship anything larger than 1980cm. Though it says on their web if one needs such thing to contact them to arrange. Now the funny thing is that i found DOLD Mechatronik in germany, that is their partner and does that, plus he has 3m screws and is willing to ship them to me. WTF. I am in the process of buying the profile from him.

2. I prefer the leadscrew TR10x2 for now, simply put - it will be more reliable and cheaper. Screw is 20 euro per 3m . Double row angular bearing pack is 2 euro from china. All else will be printed or machined from plate of aluminum. It could be made also with the Gt2 steel belt , it will be same money but will take me more time to design it, as i don't like much long double belts.


3. I have changed all structure from 3060 to 2040 slot6 , as it will be good enough for the job. I know, not typical for me to not overbuild, but it's not needed here. Final length will be 1300x2600mm travel will be around 1200x2400, as i remembered that locally i have 600x2400 foam sheets

-The original Hiwin for 6x1300 and 6 carriages / yes i will be but connecting them/ is around 800$ and after all will be around 800 euros. Quite expensive

-The Chinese MGN12 6x1300 and 6 carriages will be 300$ and say 60 euros import tax. Probably they will do the job, if something does not move right, will open it and change the balls, as i have seen people on Youtube do. So thats the winner

-The Mjunit is around 1500euro after taxes . is very nice looking but at the end of the day is just aluminum profile with steel rods and U bearings. But again i will say it looks very professional.

-The C beam from Oozenest Looked nice but is only 1500mm so again has to be attached together. Will cost me as much as do all from Chinese MGN12, so better not deal with rollers/ if they are not angular contact?/

-I lost patiance trying to explain Chinese that i need it for a slow foam cutter, otherwise there are some that could make me the whole linear movement cheap from 1204 ball screw , alu profile and square rail. Shipping would be then 300$ for that lenghts .


4. Now some interesting stuff. Long ago where there was no 3d printing and i had crappy belt driven machine, i have tried to make something with delrin ball that could slide on top of Bosch profile. Nowadays its even easier. Couple that with 2040 profile and we have a winner. 500 balls are 20$, each carriage has around 18 balls i believe. So i could use them for even something else. I will buy the balls even if i go with Hiwin

No plans here, but i have figured it out with 8mm balls:


https://www.youtube.com/watch?v=7ksZgqKZnZo


Another one, looks nice and have downloaded the files. Waiting for him to see if he will pass me Step file, no just STL:


https://www.youtube.com/watch?v=0IUlMtbzhR0

Check how they compare:

22682 22683


With the preload feature i dont see a reason why it will not work perfectly for a foam cutter especially where they will not even make a noise., Not to speak of that i could print them in POM. I will wait a day though as i have another idea inspired from i don't remember which bearing block.as i simply like them less bulkier.

There was a bearing that looks like hiwin block and inside it has 4 big spheres that enter the rail channel in a very similar way as abobe. But inside the block there is arrangment of very small bearing spheres that the big sphere lay on. So basically no recirculation as the above but a kind of. The only benefit will be to lessen the bulk and make them look more professional.

[charlieuk]

i wasn't talking about the motors or the price just the simple drive system, and if it does the job then ....

I cant see the point in the printed carriages they will never last under any sort of load and ether the plastic will fail or the surface of the aluminium will deteriorate. I would rather just use something like a cheap set of skateboard bearings and make a arrangement that runs down the groves of the aluminum that way at least the have a larger contact area and the aluminum should not deteriorate

8 x 4 eps blocks are not light but if your frame is not supporting the foam and on a solid floor you can probably get away with a smaller frame I wouldn't even consider anything lighter than 40x40. you need a relatively high tension in the wire to reduce sag and drag if you want straight lines and want any sort of reasonable speed without to much curf. and if you are ever going to do 4 or 5 axis you will need stiffness in the frame to get that, unless you build a complex frame that will likely cost as much.

[Boyan Silyavski]

any idea of the force needed to tighten the wire for radiation cutting? i think its less than 1 kg in mass, if weight is used. i cut now using 0.16mm wire and a small spring, for sure less than 1kg pull.


I see the benefit of the big aluminum profile and 20 size rail, as that was my initial design, using 3060 profile. But that's more for industrial use where somebody will work for you and will not take care of the machine. Doing it for myself, i think 2040 is enough. Still have not bought either yet. And have a day or 2 to think it over. Price wise will be just a bit more expensive.


By the way cages could be printed from Delrin, so they will be ok for the job. I have read reports on a printer after 1500h was perfectly fine

[charlieuk]

any idea of the force needed to tighten the wire for radiation cutting? i think its less than 1 kg in mass, if weight is used. i cut now using 0.16mm wire and a small spring, for sure less than 1kg pull.


I see the benefit of the big aluminum profile and 20 size rail, as that was my initial design, using 3060 profile. But that's more for industrial use where somebody will work for you and will not take care of the machine. Doing it for myself, i think 2040 is enough. Still have not bought either yet. And have a day or 2 to think it over. Price wise will be just a bit more expensive.


By the way cages could be printed from Delrin, so they will be ok for the job. I have read reports on a printer after 1500h was perfectly fine

I'll see if I can figure it out, it's more than you think as with a manual one you are normally pressing down on to templates which tenshions it further. The larger the block your cutting the more tension you need or the slower you have to go. It's a combanation of having the right temperature, the right tenshion and the right cutting speed so as to get a clean cut and not too much burn back

At a guess I would say mine is probably 2-3kg tension but I'm cutting relatively slow and and not too hot.

The cost saving is minimal on the Ali, I know what I would do..

I see less of a problem with the printed cages than I do with the balls them self. If using steel ball on Ali you don't need to be a materials scientist to know which will be the winner!

[Boyan Silyavski]

The balls are Delrin, so would be the printed cages. There is no steel on aluminum.

Agree about the aluminum profile, there is false economy there.

I have cut till now at most 100mm thick. I cut at 200 mm/min and i don't push the hot wire in the block, rather the heat makes the cut, so no stretching or effort at that speed. I am not in a production environment, but even when i would be, what matters is the cut quality, not the speed. And it seems at that speed all details cut well.

[charlieuk]

that's incredibly slow for such thin material and no use if you want to cut big blocks you will be there all day. With a cnc and no templates you obviously have to be very carful not to push the wire as if you do you will not get a straight cut when it comes to curves this is also where tension is very important so the higher the tension the less likely you are to get problems. I think with time even hard plastic would start to leave a mark on aluminium but by all means try it, personally I don't see it being accurate or stiff or durable enough.

If there is a slightly cheaper linniar rail option for the long axis I would be happy to pay for it the vertical axis could probably get away with maybe something like v bearings certainly if its only two axis but like you say you need smoothness or else it shows up very clearly in the surface finnish

[Boyan Silyavski]

" Incredibly slow" ??? I have no previous experience with foam and was trying to find info on internet what is the normal speed to no avail. So i did some experiments heating more the wire and going faster, slower and so on. 200mm/min gave me the best quality/speed relation .

What you say is the normal speed for this?

Why i have to push the wire into the foam, this means the cut will lose precision? I will not be cutting big blocks all day, i am cutting intricate shapes. if i would be cutting big blocks all day i will make a cutter with 20 wires heated at once, not one wire.

[Clive S]

. if i would be cutting big blocks all day i will make a cutter with 20 wires heated at once, not one wire.

Good luck with the power supply then you would need about say 36V at 100A

[Robin Hewitt]

Good luck with the power supply then you would need about say 36V at 100A

You think a minor thing a 3.6kW power supply would put Boyan off?

In my one foam cutting experience, I found I was trying to balance the wire bowing in to an arc against the width of the cut becoming uncontrollably large. It was also different depending on the material, maybe I had the wrong stuff.

[charlieuk]

im cutting 30" blocks at about 500mm/min and I would consider that still painfully slow at a guess 100mm blocks would be well over 1000mm/min

like robin has said getting the wire not to bow is the hard thing and the longer the wire the harder that is. You will also need to take into account that the wire stretches when it heats up so you will need to take this into account.

[charlieuk]

22684If I can't find some linniar rail at the right price I will probably go for a aragement like this but with a deralin caridge that houses the bearing but allso as a guide even though the bearings are almost a perfect fit width wise. I will find it hard to find the cost saving of linniar rail viable over the time it will take to design build and set up my own method

[Boyan Silyavski]

No, no, i have prepared the money for the Hiwins or the chinese MGNs. I was just exploring options as its a low demand machine.


As i said quality is the Top priority, not speed. So i will not use a thicker wire than i use now 0.16mm. I could go even 0.10mm. I am using constantan wire. As when you cut only and mainly letters, later you need to repair the cut on an A for example. So thats time, and time is money. When i cut so much letters that 200mm will take me all day , that will mean i make money enough to make a second machine

[charlieuk]

I will be interested to hear which rails you go for then, it seams crazy to use genuine hiwin for the amount of use it will get compared to my cnc but it's hard to know what you are geting from Chinese suppliers

[Boyan Silyavski]

Well, 260euro for 6x 1300mm + 6 MGN12H, just payed it so is done.


That will not stop me buying the Delrin balls though. Just for the experiment. By the way as the guy originally designed the bearings not in parametric, i took my time yesterday to redesign them from scratch. I did not like the not squarish design. I have also some other idea about the preload and how will look at all. This is the result for the moment. More purification when i have the balls and have tested the length so the balls fit tightly and exactly

22690

[charlieuk]

that's a very good price what supplier did you use?

[Boyan Silyavski]

that's a very good price what supplier did you use?

https://www.aliexpress.com/item/6PCS-12mm-Linear-Guide-MGN12-L-1300mm-linear-motion-rail-6pcs-MGN12H-Long-linear-carriage/32830411520.html

[charlieuk]

I started drawing my thoughts up is there any difference running the linear rails on the side rather than top of the extrusion?

[Clive S]

I started drawing my thoughts up is there any difference running the linear rails on the side rather than top of the extrusion?

I don't see any problem with that

[Boyan Silyavski]

I started drawing my thoughts up is there any difference running the linear rails on the side rather than top of the extrusion?

I think there is not absolutely any need the foam block to be supported by the structure . It could easily be simply on the floor or lifted by rests of foam. or a bed formed from foam leveled by the wire.

Same could be said for my design, but the cross bracing supports the smaller alu profile i am using against twist.

[charlieuk]

The problem comes when you have to move the blocks of polystyrene around, the blocks are not all that light and I'm dealing with 16' x 4' x 2' blocks so having a smooth and flat table to easily side them on to make it posable just about by myself were your method I probably could not. I allso will be putting mine in a 20' shed that has a wood floor so I need to build some extra strength into the design so it is self supported.

[charlieuk]

https://www.aliexpress.com/item/6PCS-12mm-Linear-Guide-MGN12-L-1300mm-linear-motion-rail-6pcs-MGN12H-Long-linear-carriage/32830411520.html

just wondered if you had received this yet? any feed back on quality if so?

[Boyan Silyavski]

Still waiting and yesterday night started pushing them as the number they provided is not correct. Will update as soon as i know.


There is a very nice detail that 2040 and MGN12 have holes that coincide with the channel so i designed a nut for the TR10x2 screws. and a bearing plate from 2040. https://www.thingiverse.com/thing:2527055

[charlieuk]

at the moment I'm still going to go for the 80x40 10mm slot kjn aluminium, partly as all the other extrusion I have is 10mm slot so have the jigs and extras to go with it all.

with the 10mm slot I'm leaning towards the 15 or 20mm rails, 12mm doesn't give much to sit on the aluminium.

I have drawn my gantry at 40x80 also but may go for 40x40 on that.

the vertical axis I was thinking would be belt driven however I want to link the two sides together which may be easier if they were screws

I think maybe a lego mock up is needed2278422785

[Boyan Silyavski]

yes, there is absolutely no point of going with 12 size rails given the size of your extrusion. 20 size will be best. 15 size may save you some $$

i went with screws as on that length price was same, but i had to go with bigger belt, at least Hdt3 15mm wide, which would mean bigger motors, bigger drives. And at the moment i am hoping to drive all with 30 euro ramps 1.4 and drives, 4 motors 10 euros each and the DEVFoam CNc which seems best for the purpose

[charlieuk]

I'm probably not actually even going to make mine cnc to start with, until I decide to go 4 axis there is not a lot of point as any wings I would need are tapered and then I would probably just build a small machine for them. most of what I do is just cutting square blocks and I can do that without so will just have some sort of motor and gearbox and some basic controls. I also decided to make the machine the full 6meters ( full length of alli) so that means ether belt or rack.no I do have some spare rack here that was left over that I could use but its going to be hard to get it to work without having it on booth sides so I'm going to probably have to use a belt to drive that.

[Boyan Silyavski]

What would you cut at 6meter length? i mean, arent all foam blocks 2mters long?

if i was going to make such a serious machine i would have probably used belt on belt drive which will again simplify things. In that scenario there is no need of very big size belt as the stretch is "none"

[charlieuk]

the same thing stuff that I cut on my 6m long cnc. The blocks I normally use are 16' x 4' x 2'. belt on belt is a good idea I will just have to figure where is best to place it, it will probably mean that I will need to have one ether side and two separate motors but that shouldn't be too much of a problem. unfortunately I cant see myself being able to join the gantry underneath due to the length and trying to support the frame.22786

[charlieuk]

so I remembered I had this idea a while back but forgot and that was to make it so that in the future I can add cnc and convert to 4 axis. It wont really add anything to the cost of the frame but would require two extra lengths of rail and two more carriages so I will at least do the frame like this from 40x40

2280122802

[charlieuk]

thought I would get some quotes for rails, china want £638 plus shipping for 15m of linear rail and 6 carriages odiously its not going to be real hiwin were as cnc4u is £989 for the same but real hiwin. This is just to do a single rail each side for two axis which would obviously be nearly double to do 4 axis and have the two rails each side.

ether way I think I may be back to looking for possible cheaper solutions, I was wondering about round rails if I did the two rails each side but not sure there would be that much saving in cost.

[Boyan Silyavski]

On such a big machine round rails-No. What about exploring the 3d printed +Delrin balls scenario? It will be the best pricewise.

[charlieuk]

I'm thinking maybe these and 3 per side for the main axis and maybe stick with the linear rail for the vertical. It sucks I really hade to cut corners but for the amount it will be used I cant justify 1k on linear rails.

other wise I guess It would be using some sort of bearing carriages to run on some steel guides. like cnc router parts do.


22803

[Boyan Silyavski]

I'm thinking maybe these and 3 per side for the main axis and maybe stick with the linear rail for the vertical. It sucks I really hade to cut corners but for the amount it will be used I cant justify 1k on linear rails.

other wise I guess It would be using some sort of bearing carriages to run on some steel guides. like cnc router parts do.


22803

Most probably that rollers will not do. Inside there must be angular contact bearing or a bearing with no clearance. I am 99% sure thats not it. Plus they need to be turned, not molded plastic

[Clive S]

Charlie I think you need to be realistic regarding the tolerances here, I don't know but surly you won,t need more than a 1mm for cutting foam you are going to get sag in the wires etc. I think something like what you are planning should work ok.

[charlieuk]

yea 1mm is fine its just trying to make it nice and ridged and reliable really without getting to complicated or expensive to manufacture. I think if I just go with two axis the wheels would be fine, I could maybe turn some on the lathe if I wanted I guess not sure if I its worth it. I'm going to look into some bright steel strip to see what lengths I could get that in maybe to see if that would be a option for running some skates down.

[charlieuk]

spent a while looking into other rail options. I spoke to the guys at hepco however even there basic utili track is going to be more than genuine hiwin so that's out!

I think its going to be between plastic rollers direct on the extrusion or making some sort of v wheel of bearing setup that runs on steel.

I have seen a few options for this there is one were they are using angle steel bolted top and bottom to which the bearings run on which would probably give the most stiffness if I were to want to run it as 4 axis (sorry pic is a little small)

other wise I think I will maybe try and make my own plastic wheels ether by turning them on the late from something or perhaps 3d printing them maybe ?

[Boyan Silyavski]

Just received the Chinese MGN12 rails and carriages. Well, in one word:Not so bad. It will do the job. Great for the money .


In detail:

All was very well packed.


6 rails 1300mm and 6 long carriages for 330euros including the import taxes i paid here /60 euros/. Should have payed 30 more to the transport company to do the import documentation, but as i spend quite some time to learn how to do it by myself, save that money for an hour work. The Hiwins would have been around 800 euros.

All carriages came preloaded. 2 of the kits are same same as Hiwin, if i close my eyes and move the carriages. 3 of them are a bit more stiff. lastly on one of the rails and carriage combo at a certain point the rail is like stopping the carriage. But nothing very bad. All will do the job. maybe it helped that i bought them from bigger company that makes them, as in China everybody makes them, so difference from quality is very possible.


Rails are definitely steel, not iron. And well hardened at that, tough to cut with angle grinder. No hack and chop saws! Only abrasive cutting will work.


Rails bottoms came dirty with marks from scotch tape, so had to clean them with acetone.



Charlie, i feel that MGN15 for sure, or even MGN12 will do the job you are looking to be done. Maybe some extra carriages/10 euros each/ and nothing to worry about. Anyway i bought rail and carriage combo so i have 2 left. You could do the same. Double the carriages on the long axis and thats it. There is no better and cheaper option i think. Or why not even double the rails? I did not buy the MGN20 for the sole reason that weight more and shipping was 200 for it. So at the end of the day some calculation should be made.
The problem you are facing is that the MGN 12 will enter the slot, as your slots are bigger. hence check that also. Or drill and tap.


One more thing. There seems but joining such small rails like the 12 size will Not be a problem. I was also wondering that when i was deciding. All that saves on shipping costs.

[charlieuk]

Thanks for the update glad they are ok and will do the job. As much as I want to use slides at probably over £700 by the time they arrive just to do two axis I think I will give the roller idea a go first and do a few quick mock ups and see if I can get the stiffness to do what I want at least for starters. The aluminum on its own is the best part of £500 so to do such a simple job that im after a machine that would end up at around £1500 is going to be over my budget.

[charlieuk]

been looking around again and have been wondering if 16mm bright mild steel rod could be used with some sbr bearings if I was to make some sort of supports however I did then find 2 x 2500mm sbr16 rails and 4 x sbr16 bearings for £112 which for the time it would save may be a good compromise assuming that they should but join together ok. The only thing I guess is it would mean ether making a interface plate to mount the rail or drilling and taping the extrusion to mount the rail.

http://www.ebay.co.uk/itm/2-sets-SBR16-2500mm-Liner-rail-Guide-4-pcs-SBR16UU-block-bearing-/262453480496?hash=item3d1b724c30:g:cTUAAOSwTdJXRW-h