Well here goes... Aluminium frame router design

[CharlieRam]

Okay so what if I were to make the spindle plate 30mm thick aluminium or 20mm+ thick steel ( I could grind the steel perfectly flat and parallel at work, well within a couple of microns anyway), would that be strong enough to have the bearing blocks the opposite way round or does that method just not work? in fact a lot of the plates could be made of steel which would be easier for me to grind flat as we use mag chucks, it would add weight but also a lot of strength.

Probably dont want to change too much though as the weight will increase substantially!

[CharlieRam]

Anyone? Is My idea 3 Z axis just not up to it, will changing the front plate to steel help or not? Just the Z plate, not the x rails as I have mounted them directly on top of the rails.

[routercnc]

Hi Charlie
I think you are not getting a reply on the z plate as most people have it set up in reverse (carriages on Y plate) so can't give you an opinion.
If you want to go with the opposite (carriages on Z plate) then you'll need a calculation of 20mm aluminium with reverse mounting (at full reach and home-they will be different which is why that route us preferred) vs forward mounting with 20mm steel and 30mm aluminium (at any reach as they will all be the same stiffness).
if I get the chance I'll have a look tonight and see if you can recover the stiffness through gauge.
What is the distance from the lowest Z
axis bearing to the collet at full Z extension?

[routercnc]

Hi Charlie,

OK, I've made a few assumptions for dimensions and compared different Z axis options for deflection. The deflections are artificially small because the rest of the machine is grounded, but you can compare the numbers with each other to see the trends.

1. Shows the conventional layout (carraiges on Y axis plate, rails on Z plate) using 20mm aluminium plate at full extension. This gives 10um.

2. Shows the same conventional layout as 1 but at almost minimum Z extension. This gives 0.2um. This is 20x stiffer just by raising the job so that cutting is done with the Z axis near the home switch.

3. Shows the unconventional layout (carraiges on Z plate, rail on Y plate) using 20mm aluminium plate. This is drawn at an arbitrary extension because with this design all extensions have the same stiffness. This gives 10um. You can see that it is the same as 1. But if you need to do a deep cut, or hold a fine tolerance etc. then you cannot raise the Z to increase the stiffness. So, onto your question about recovery . . .

4. Shows the unconventional layout using 30mm aluminium. This has reduced the deflection to 3um. 3x better than 20mm aluminium, but still nowhere near the best conventional result (2).

5. Shows the unconventional layout but using 20mm steel for the Z plate. This has a deflection of 3.5um, so about the same as the 30mm aluminium plate. This is also nowhere near the best conventional result (2).

So, if all of your cutting is done near full extension, and your machine is stiff enough to cut well at that extension, then it does not matter if you choose conventional or unconventional (and both with 20mm aluminium plate) as they are the same.

But chances are you might not get the finish you want at full extension as you are cutting at quite a distance from the support bearing, and so with the conventional layout you can raise the job in Z and try again. You can stiffen the axis quite a bit by doing this. With the unconventional layout this makes NO difference (in terms of Z plate bending) so you are stuck with it. You can upgrade the Z plate from 20mm to 30mm aluminium and stiffen it a bit more, or go for 20mm steel (works out about the same), and gain an improvement. But you cannot get close to the stiffness improvement of the conventional layout raised close to the home switch.

I can't tell you which option will be stiff enough for you, but I can say that once you have made the machine if you run into problems and have gone the unconventional route it is difficult to make a significant improvement. Hope this helps.

[CharlieRam]

Hi Charlie,

OK, I've made a few assumptions for dimensions and compared different Z axis options for deflection. The deflections are artificially small because the rest of the machine is grounded, but you can compare the numbers with each other to see the trends.

1. Shows the conventional layout (carraiges on Y axis plate, rails on Z plate) using 20mm aluminium plate at full extension. This gives 10um.

2. Shows the same conventional layout as 1 but at almost minimum Z extension. This gives 0.2um. This is 20x stiffer just by raising the job so that cutting is done with the Z axis near the home switch.

3. Shows the unconventional layout (carraiges on Z plate, rail on Y plate) using 20mm aluminium plate. This is drawn at an arbitrary extension because with this design all extensions have the same stiffness. This gives 10um. You can see that it is the same as 1. But if you need to do a deep cut, or hold a fine tolerance etc. then you cannot raise the Z to increase the stiffness. So, onto your question about recovery . . .

4. Shows the unconventional layout using 30mm aluminium. This has reduced the deflection to 3um. 3x better than 20mm aluminium, but still nowhere near the best conventional result (2).

5. Shows the unconventional layout but using 20mm steel for the Z plate. This has a deflection of 3.5um, so about the same as the 30mm aluminium plate. This is also nowhere near the best conventional result (2).

So, if all of your cutting is done near full extension, and your machine is stiff enough to cut well at that extension, then it does not matter if you choose conventional or unconventional (and both with 20mm aluminium plate) as they are the same.

But chances are you might not get the finish you want at full extension as you are cutting at quite a distance from the support bearing, and so with the conventional layout you can raise the job in Z and try again. You can stiffen the axis quite a bit by doing this. With the unconventional layout this makes NO difference (in terms of Z plate bending) so you are stuck with it. You can upgrade the Z plate from 20mm to 30mm aluminium and stiffen it a bit more, or go for 20mm steel (works out about the same), and gain an improvement. But you cannot get close to the stiffness improvement of the conventional layout raised close to the home switch.

I can't tell you which option will be stiff enough for you, but I can say that once you have made the machine if you run into problems and have gone the unconventional route it is difficult to make a significant improvement. Hope this helps.

Cheers for that, I am in Egypt atm and the internet in my hotel is dodgy as hell. I will have a play with that spreadsheet (is that the stiffness v7) when I get back because I think the distance to the collet from bottom bearing block is quite a. Bit smaller than 150mm and consistent stiffness should give more predictable results. Cheers

[routercnc]

Hi Charlie,

OK, enjoy your visit. I wrote this sheet just for your problem.

I've now added it to v7 and created v8. Posted here so you can have a play.

Thanks

[Boyan Silyavski]

Hi,
Maybe my build log will be of interest to you and especially the way I made the z platehttp://www.mycncuk.com/threads/6457-Sturdy-and-Fast-all-Steel-CNC-my-first-build?p=59627#post59627
I intended to do it from aluminum but instead did the z from steel plate. It's extremely strong extended and compact.
My advice- draw it but do it when you have in your hands the ball screw nut, as it happens they can send you a slightly different housing than at the listing. It happened 2 times to me. Same for the spindle bracket, as they cast it sometimes the holes are not where you expect to them to be. Chinatown ��

[CharlieRam]

Well I'm back I had a play with that spreadsheet and decided to stick with my unconventional layout, the reason being is I have used calculations based on 35mm thick steel plate, this is the 20mm thick plate plus the 15mm thick spacers, I was thinking of gluing and screwing them together but I am not sure if this will give the same results as a 35mm thick plate with a recess machined for ballscrew.

Anyway, the figures from that spreadsheet give me roughly .65 micron deflection, the conventional is slightly better at .19 microns but a lot worse at full extent with 10 microns but if I reduce my cutting depth to just 100mm by raising the bed and spindle by 50mm then I get .19 microns across the full cutting height!

I am a bit stuck at the minute though, am trying to figure out how to drive the X ballscrews from a single centrally mounted motor but leaving the ends of the M/C open for any job overhang I may need. It seems like I may need a lot of pulleys to guide the belt but I am not sure it is a very elegant solution, Any Ideas or am I just going to have to go with twin motors?

I do want to add a 4th axis at some point and I don't want to go silly with the spend on adding unnecessary motors and drivers if they aren't needed, speaking of which, I have drawn the base up now and I'm thinking of having a go at welding but last time I tried that it looked like pigeon shit! How much should I expect to pay to have one fabricated?
I have added some more sketchup images, the steel parts are colored blue.
As before I look forward to your criticism/Acclaim
Thanks, Charlie

EDIT. Also checked the gantry sides stiffness using the calculator using height to top of gantry as 235mm but it assumes the 20mm plate is not braced, Mine has plates on the back bolted to the bearing block plate, the gantry plate and also the gantry beam so can I just reduce the length in the calculator so it reads full length minus the brace plate giving me a total height of 103mm to give a realistic figure?

[CharlieRam]

Just been having another play with the stiffness spreadsheet but I couldn't figure out what to select for my gantry type, I have done the special type but the deflection value isn't shown in yellow as stated and also not selectable in the gantry type drop down menu?

[routercnc]

Hi Charlie
The gantry side stiffness calculation is for simple routers without raised X axes. It is to help determine whether to use plate, box etc for the sides and what thickness
You have the vastly superior raised X axis design ( which wasn't that popular when I first wrote the sheet) and so isn't relevant for you. Your gantry sides will be fine as they are.