Newbie - Which CNC router to chose ? - build or buy ?

Hi everyone.
I am totally new here and this is my first post.
I am just about to embark into the world of CNC routers.
I have done a lot of reading, googling, pouring through the posts on OpenBuilds, etc...
I feel somewhat overwhelmed and not able to easily chose which route I should go down with.
I do not mind spending money as need (within reason as I am retired) but I want to make sure that I buy the right kit so as to avoid finding out a fem months down the road that I should have bought something else.
Some of the things that spring to mind that will affect my decision include.
==> Should be robust enough to route aluminum (I understand that smaller passes will be needed to achieve this compared to a proper mill and that it will take much longer to mill a piece of aluminum)
==> I also plan to do wood, acrylic, etc...
==> I am guessing that somewhere around 1 meter by 1 meter , plus or minus, should give me enough work area for what I want to do, and space in my workshop.
==> I understand that lead screw systems are better than belts (I am happy to be told that this is not the case)
As I live in the UK, I am sort of drawn towards the WorkBee 1 meter by 1 meter lead screw driven CNC router from Ooznest, but that is based on googling and researching and very limited knowledge of CNC routers (but learning quickly hopefully).
I also note that you can order the same kit from Aliexpress for a lot less money
I am also confused as to wheter I should go for GRBL or Mach3/4
I am also a bit confused about 2.5D and 3D and whether that is all down to the software or if the hardware chosen affects that ability (It would be nice to have 3D as I plan to do carving.
I do already have a metal Lathe and CNC mill ( the Warco ones) so I am able to turn and create straight forward milled items if needed for the CNC router.
I would really appreciate it if those much more experienced people could help to guide me a bit so as to enable me to buy the right CNC router so that I can hopefully avoid all the newbie pitfalls of buying / building the wrong CNC router and then having to revisit the whole thing a short while later
Thanks a million guys. Your help is really appreciated.
ps. I love this forum. There is so much good and helpful information here.

 

Hi GrayUK
Many thanks for your prompt and helpful reply.
Interesting that you mentioned fixed gantry / moving bed machines, as that is my view as well since the gantry can be so much more robust from what I can see and I did look around a bit on that , but could not find any such machines/kits around.
It also means that you could have a very long platten quite easily. One of the things in the back of my mind on the wish list of things I would like to make is wooden kayak oars which would need a working space length of about 2.5 meters so that it can be made in one piece.
Yes, it does require more space but I could live with that if it gave me that length of working space
Is that what you have ? If so what is it and where did you get it ?
I had a look at the USBCNC you referred and it says that you do not need Mach3. Your comment suggests that you are using Mach3/4 with the USBCNC. Could you expand on that please ? I will have a look at the user manual a bit later as well
Thank you for the description of 2D/2.5D/3D which does make it clearer.
What I was also asking is whether the hardware has any bearing on whether you would use 2D, 2.5D or 3D from a technical point of view, or does it not matter and it is just down to the software ?

 

Thank you for that Craig.
In your opinion, would you say that this is the "sturdiest/most rigid" CNC machine that one can buy/build currently (and with easy access to parts. aka openbuilds store, etc...in the UK
In other words, the best suited CNC machine for doing aluminum and similar harder materials work ? (quite important for me as a fair amount of my work might be with aluminium, though a lot of the work will also be to cut out routing guides/template out of acrylic for manual free routing with a standard router) Though I understand that it will be much slower than using a commercial mill with sky high prices
I note that it is labeled as CNC mill as opposed as CNC router, which I assume is intentional to reflect the rigidity of the machine and hence suitability for harder materials.
I note that that machine says max cutting size 750mm x 330mm. Is that 750mm on the Y axis ? ( that is my impression looking at the picture)
I might be able to work with that, but perhaps a tad too small for quite a few of the projects I have in mind
If so, it sounds like increasing the x-axis length should not create any reduced rigidity apart from the whipping issue you mentioned beyond 1000mm x axis travel ?
You mentioned that you are designing one with a fixed, double gantry for strength, but only about 1000 x 1000mm. Does your decision for the double gantry for strength indicate that the one in the kit is still not quite adequate ?
Are you keeping your design to 1000mm x 1000mm due to workshop space constraints or because of the lead screw whip issue ? If the latter, would it take much to increase the lead screw to a larger diameter to address that ? My thinking is that if I could push it close to 1500mm z-axis work area then I could do an oar half and half at a time, keeping the center of the oar as a reference point to keep things level when doing the second half, then finish the center bit by hand as that is less critical.
I did not mention spindle or router as I had not yet done sufficient research on this, but whilst we are on the topic, I would be grateful for your recommendations on this. Whilst researching a bit on this, I noted that some members had to build some electronics to control their router/spindle down to about 3000 rpm for acrylics otherwise they were just melting the material on higher speeds. I note that the obvious Makita/Dewalt routers suggested do not go much below 12000 rpm.
Is that a fundamental difference between spindles and routers ?
For the varied work I would like to use the machine for, what would you think would be best ?,Spindle or router ? ,and which model would you think would be most up to the job I am after and "matched up" the rigidity of the CNC mill you referred to above, or perhaps more to the point, to the more robust one you are currently designing as I might like to piggy back on that new more robust design if that would be OK for you ?
Also on the same vein, what power nema motors would be best to match the new design framework you are working on ?
Would it be best to use the more powerful ones so that there is plenty of power in reserve in case one wants to mod the setup later and require more power without having to upgrade the motors later ?
I will lookup USBCNC / Mach3 / Estlcam and Trochoidal milling on youtube, etc.. to understand those a bit more.
Sorry about all the questions I am sending your way, but I am now inspired by the higher rigidity of the CNC mill you referred to or more to the point, to the more robust one you are currently designing.
I was originally thinking of just going for a straight kit to make life easier, but to gain the additional rigidity,I am considering going down your new design route if that is acceptable with you.
Fortunately, I do have a reasonable selection of tools, power tools, a lathe and mill. I am also familiar with arduino and general electronics, so that should all help in building a CNC mill, though a bit of a daunting process as it is totally new to me !
Many thanks

 

Hi GrayUk

Thanks again for all the time you are taking to answer my queries. It is really appreciated.

As you say these are your personal ideas and experience, but I think we are both on the same page on everything we have discussed so far, so that is very encouraging for me as it reinforces my thoughts on all this and it is great to be able to discuss these sort of things with like minded people.

I should have said "best of the bunch" of what is out there. Ie What is the best result I can achieve with what is out there including any mods that makes it even better.

I agree that the gantry is the weakest point where most of the leverage occurs due to the distance from the cutting point to the gantry attachment point causing torsion in the gantry which will be worst when the router is is midway on the Gantry. At least with the fixed gantry, some of the issues of leverage and torsion have already been eliminated.

I follow what you are saying about the X and Y axis sizes and issues.

I agree with your definition of what the X, Y and Z axis are. I was bit thrown off as in the Ooznest description of the XL CNC mill, it says ..."cutting area - (X x Y mm) 330 x 750" which I suspect is the wrong way around as the X-axis seems to be the longest one based on the photos and bed dimensions. You have just confirmed that my suspicion was right and that their description is incorrect.

Have you already had experience with the XL kit ? or is your current CNC something else ?

Are you basing your new 1000mm x 1000mm CNC on the XL kit ?

In the 1000mm x 1000mm machine you are designing, Am I right in assuming that this is the bed size you are talking about as opposed to the XL kit which is 1000mm x 500mm and you are leaving the x-axis size as is but doubling the Y axis.

Am I right in estimating that this would give you a cutting area of (X x Y mm) 750mm x 830mm with the moving Y-table protruding out 170 mm at full travel, front and back.(same as the XL model).

I can see where you are going with the dual gantry C-beam and cross bracing with a linear rail as that would reduce torsion by a very large factor on the main gantry beam and would be very effective.

For added strength and reducing stress on the "junctions" of the gantry supports and gantries, would it be an idea to also have some additional linear bracings between gantries at 45 degrees to make that double gantry rock solid (a bit like flat roof steel trusses are made) ?

With the longer y-axis, would anything need to be beefed up in the moving bed axis rails or table movement mechanism so as to prevent deflection when cutting in the x-axis as a bigger leverage will occur when the table is at mid-point of it's travel than with the standard XL kit. I guess we have two deflections to address, Vertical downward deflection and x-axis deflection. I can't see how we could beef this up easily as the table gantries wrap fully around the table rails, other than using bigger rail cross sections. Maybe I am worrying about something that is not an issue for a 1000mm x 1000mm bed ?

How strong are these wheels and v-groove mechanisms ? At what point do they become an issue once we have beefed up the Z axis, and anything else that can be beefed up ?

I am still pondering over how to handle the larger projects like carving 2500 mm oars. Yes, I can see how using the step process would work, but on a fixed gantry system, a long piece of approx 2500 mm would need to be fixed to the moving deck and canter levered out into space, causing uneven loading on the table. I guess this could be offset with canter levering a similar weight on the opposite side, which is a possibility, but somehow does not feel quite right.

I may try this first to see how well it works in practice, but I suspect that longer term, I might land up with the "beefy" plate cutter we have been discussing and then a second larger custom CNC to handle larger wood projects and as you say possibly with belts to handle the length of travel, etc...

I can see where you are going with the water cooled spindle. I generally get the impression that people eventually move to a water cooled spindle sooner or later for more serious work, so your comments also confirm that. I had a look at the one you provided a link for and the price is not bad. Do these really give you 0 to 24000 rpm or is there a "sweet spot" range that they realistically work best in. Can those also be controlled from the software (ie on/off and speed). I am probably leaning towards going down that route from the start as it seems a far more superior way to go.

Another "mod" that I am pondering over is as follows... I sometimes have a need to mill the end of a piece of brass, say 100mm x 25mm x 25mm, for use as the "burning bit" for a branding iron. ie milling a logo in the 25mm x 25mm end of that piece of brass. I have done this crudely in the past on my manual mill, but it is very cumbersome to do this like that manually, but the result is a bit crude. This is what sparked me off initially to look at CNC mills. It is possibly feasible to convert a normal mill into a CNC mill but there are a lot of drawbacks to that as well so I decided not to go down that route.

The issue will be how to do that as there is not enough z clearance to stand a 100mm piece upright.

I guess the gantry could be raised to provide the required clearance, but you then introduce more issues with deflection. I was pondering in making a hole in the moving bed just big enough for that sort of size material in a part of the bed where it would not interfere with the travel of the bed and that should enable me to "sink" that piece of brass into the bed, but still clear the underlying table. A small vice could be attached to the moving table over the hole to grip the piece of brass during milling.

I note that some people have used an aluminum plate for the main bed to provide a more solid bed, then add the MDF board(S) on top of that, which gives you a bit more clearance for working Z travel


If you had to build that new 1000mm x 1000mm bed CNC mill that you are working on from scratch, I would be interested to hear your thoughts on your make/model choices for the motors, electronics, software, router/spindle, etc...ie everything you would chose to build that new machine. As you clearly are way ahead of the curve compared to where I am on this, your experience and thoughts on this (very valuable to me) plus my google research should hopefully enable me to get the right kit from day one rather than revisiting this later and buying more kit later (better to spend a little more money now than lots more over time because of poor choices now)


Out of interest, where are you based ? I am near Petersfield in East Hampshire (UK)


Sorry about the load more questions, but this is now getting very interesting as it looks like we are both heading for a very similar build and have very similar views on this !

It should be real fun to work on this project in collaboration and sharing our experiences !

 

Nice sketches ! What did you use to create that ? It is great to be able to rotate that in 3D so that you can inspect it from all angles.
You did a great job of that !
Hmm! I like it lots !!! No wheels in sight. Yipee !!
Again, it seems we are thinking along the same lines in that the wheel system is a weakness from my point of view as per the concerns in my last post to you.
I like the use of linear bearings in that they can be secured all the way along as needed to a very rigid structure, so you could virtually achieve a zero deflection in all axes at right angle to the direction of travel.
If I may add a few comments ( that is me being OCD about the rigidity). I would be inclined to make the following mods and would be interested in hearing your thoughts on this.
==) underneath the moving table --- a few more braces to support the linear bearings so that they cannot flex in the vertical or horizontal direction (ie at right angles to the direction of travel of the linear bearings. Some of these braces to be parallel to the one you have shown in the middle and some X cross braces at 45 degrees to prevent the Y-rails from deflecting, especially as we near the 1500mm length (I like that x-axis length, very useful !! and with your overall design, I think it can be achieved !) Again, just like those flat roof steel trusses
==) On the side of the two vertical gantries --- brace them on the front and back at about 45 degrees to the horizontal Y- axis rails to stiffen up the vertical gantries and prevent any high stress on that gantry to Y-axis junction causing possible deflection ( I guess the higher the better)
==) On the back of the double Gantry. brace in a similar way to the underneath of the table with parallel and 45 degree braces.
==) The one thing I would like to also brace, but I can't see any obvious way of doing this , is to brace the top of the gantries so that x-axis cutting forces do not try and push out on those gantries and cause some deflection. Perhaps this would need some outriggers on the base that such brace could be attached to at 45 degrees to the gantries. Hmmm needs more thought.
Is there a reason why you have chosen to only have one motor for the X and for the Y axis rather than two per axis ? I ask as I have read a lot about people having issues with jamming or snagging with only one per axis. Perhaps two motors is unnecessary with your design as you have two sliding bearings per sliding rail a good fair distance apart. I do wonder if that would cause some issues when you are cutting quite heavily in say aluminum very close to one Y-axis rail whilst the table is moving. Would that perhaps cause uneven rotational forces on the table, causing premature wear on the linear bearing / slides ? (Again, I am probably being OCD about this with my engineering head on and no practical experience on this yet, though that flag went off in my head based on comments I have seen whilst researching).

Whatever we do, I think this will be the most robust and rigid hobbyist CNC mill around !!
I don't mind forking out a bit more money to get that extra rigidity.

Would you be cutting and tapping all your C-beams and plates yourself ? or would you get someone to manufacture those for you so as to ensure that they 100% accurate

You have come up with a great design !
I love where this is going, and I have virtually all but decided to go that route rather than the OpenBuilds wheel based XL kit.

 

Yes, It would be great for people to jump in and comment as the more views and comments we have, the more likely this will succeed !

Tolpuddle is only about 2.5 hours drive from where I live.

My workshop is quite small too (an outbuilding) and I already have a mill, lathe, drill press, and quite a few other things in there, so I am scratching my head as to where I would fit this 1500mm monster as I guess I would need about a 3000mm x 1200 area with some space to get around it. Some serious reorganising of my outbuilding is needed. I do have quite a large greenhouse that I use as a "work area" now, so it might have to go there !

What do you plan to use as a table / bench to sit that on top ?

I like the self retracting window blind idea to protect the screws.

I had not heard of Acetal before. Sounds quite useful as a material.

I take on board what you said about the stepper motors and power. The Nema 23s sold By Openbuild come in three power ratings, 175, 265 and 345 Oz. Are you suggesting the 175 Oz ones or one of the more powerful ones for a build such as ours.

I will have to think about that hole in the moving bed. I would think that a hole in a aluminum bed ( about 50mm by 50mm ) probably should not degrade the strength of that bed at all.

May I ask why you think that this would negate an aluminium bed ?

I will go out and re measure my workshop to see how I can reshuffle things !!

 

Hi GrayUk
I had a good measure up in my workshop and I think with a reshuffle I should be able to fit it in there.
I would need to reduce my "walkway" width a bit, but it would still give me about 700mm to walk between the lathe mill (these are back to back) and where the CNC mill would go.
I was trying to estimate the overall space needed.
Assuming the design in terms of overlap of the moving bed is similar to the overlap of the Openbuild XL kit at 170mm, and that one is using a 500mm lead screw, ours at 1500 ball screw with a moving bed overlap of 170mm at each end would give us a maximum working y-axis length of 1090mm (750mm + 170mm + 170mm ).
I guess the working x-axis length would be as per the XL kit at around 750mm, so overall max working area of x-y 1090mm x 750mm.
Without overlap of the bed, it would be x-y 750mm by 750mm.
I think I have got all that correct ? unless you had different thoughts regarding overlaps, etc...
So workshop space for the unit including overlap would be minimum 1840 mm (1500mm + 170mm + 170mm) by 1000mm
Say allowing for some space between the unit lengthwise along the wall of say 100mm and 600mm access space at both end, I reckon I would need a floor area of 2700mm x 1100.
I guess I would need to have some form of lockable wheels on the table where the CNC mill would go so I can pull it away from the wall to access the side facing the wall for maintenance !
Now to look for a suitable rigid table for this, minimum 1600mm by 1100mm to allow 50mm all the way round the base.
Ok, space sorted out but a few things will need to find a new home out of my workshop. Small price to pay for acquiring a very useful CNC mill. Lol !!

 

Hi Gray

1) I have a spare laptop with Windows 10 on it that I can use for that, The way I see it, I can take it indoors to work on designs, etc... then just take it to the workshop and plug it to do the cutting.

I need to do a bit of research as to whether to use Linux or Windows. I suspect Linux might be better as it is more reliable for control environments as Windows has a habit of doing it's own thing from time to time and possibly halting the process, which might not be an issue for this type of process as the process would just stop cutting and then resume. I will go with Linux if Windows gives me any issues.

A raspberry Pi would also be a good "computer" to use for a permanent processor with the CNC. It is small, quite powerful and runs a port of Linux Debian. I use them for my automation controllers running as a server. Only about £35.00 or so

2) Loads of learning to do now and preparation of the workshop area

3) As building this new rigid monster will take quite a while, measured in many months I am sure, taking into account all other commitments, learning, sourcing all the parts and then making/remaking any brackets and fittings accurately and I do not currently have access to a CNC, I am still considering purchasing one of their C-beam kits to use to manufacture all the components needed for the new big beast as doing this by hand will take forever and not be so accurate so that will save me a lot of time better spent of the new build. It will also allow me to start straight away with the learning curve of machining with a CNC and getting things made.

This way, I do not have to rush the new build and can start manufacturing bits and pieces straight away and ensure that the new build is a really robust and rigid CNC for the future.

 

Hi Gray
I have been a bit quiet, busy googling and getting to try and get to understand all that CNC stuff !!
Yes, I agree that a water cooled spindle seems to be the way to go ...
Would that spindle you suggest (0.8 kw) be powerful enough for aluminum work ? or would it be better to go for the 1.5 kw one.
I like that it has a huge range of RPM, especially low RPM for plastics, etc...
What would be the most practical method for circulating the water cooling ?
I have seen some Heath Robinson ideas of a bucket of water and a fish tank water pump, but I guess it does seem simple and effective ?
Thanks Gray

 

Thanks Gray.
1.5kw is also the "gut" feeling I have. I will research further on this.
I think I will definitely go down this water cooled spindle route as I have read quite a lot of issues with standard routers ion that their run outs are bigger and they are not designed for continuous use as in a CNC environment.
Thanks for pointing those water cooled spindles out to me. Also, they go down right low in RPM with a very good range of RPM

 

I was looking at the weight of the Dewalt d26200 that seems to be the norm shipped with kits (or Makita). It weighs1.9 kg.
An .8 Kw water cooled spindle seems to weigh in at 2.6 Kg (65mm x 210mm) and virtually the same weight for a 1.5kw.
Once you jump to 2.2kw, the weight almost doubles to 5kg.
So from a standard d26200 to a 1.5kw water cooled spindle, there is approximately 0.7 kg more weight.
What is the overall weight of the whole moving Gantry including spindle, Nema motors, hardware, etc .. I am guessing at 25 kgs (stab in the dark as I can't seem to find those figures at the moment).
so the percentage more weight would be about 2.8% more weight ? Would one even notice that ? I doubt it.
I suspect the weigh issue is probably not the actual weight itself of the whole gantry (unless we place an elephant on that), but rather the inertial weight (ie acceleration / deceleration). I would think that making sure we use powerful enough stepper motors, we can "swallow" all those differences. I think there are tables for the nema motors that provide all that kind of information, but pragmatically, when you see the low power nema steppers used on most machines, if you just fator in to get a slightly more powerful one, I would think we should not need to worry about any such minimal additional weight.
For me, I am happy to go for the 1.5kw spindle as there is virtuall no weight difference to teh 0.8kw spindle
These are my opinions based on my thought process, so please do research this further to make sure you are happy with the thoughts.
What we must not forget is what power outlet you will be using for your CNC setup.
4 x Nema motors(about 40w), spindle (1500w), driver electronics + power supply (guess 250w), laptop guess (150W), vacuum cleaner(average 1000W) , water pump for cooler (guess 100w), lights for your CNC surface. All these mount up quickly.
As a quick count,I make this at around the 3000w mark. A 15 amp socket at 240v will give you 3600w, so it should just about do it , (unless my guesses above have been undercooked) but you will not run much else on that socket (bear in mind that a 32amp ring main in a house is the total for all sockets on that ring main and there are often close to a dozen of these around, so you need to also watch what else you have plugged in and running on those socket !

 

ScottyOar
There seem to be different adverts with different figures.(Maybe different makes and models) This one on ebay shows the 0.8 and the 1.5 as being 65mm diameters and the 1.5 as 8mm longer than the 0.8
I guess they are using the same armature with a few more coils, so if that is true I can believe they are similar in weight. With the same armature, the cooling path would probably be the same as well
Having said that, I will want to confirm those details before ordering just in case they have messed up the specs !

1.5KW WATER COOLED SPINDLE MOTOR 1.5KW VFD INVERTER HIGH SPEED MILLING WELL MADE 283961322354 | eBay

 

Out of interest, doing more research, I am still wondering about getting a 2.2kw water cooled spindle rather than the 1.5kw one even though it is probably twice the weight of the smaller ones
A thing to consider is that at the higher RPMs, yes, you seem to get getting most of the power out of the spindle so a 1.5 kw is probably more than enough, but when machining aluminum, slower RPMs are a lot better but that is where the spindle will have least power/torque. I have read in a few places that at those low speeds, the smaller spindles can easily stall, which is a bit pointless if what you want to do is to machine aluminum.
Yes,some people seem to be machining aluminum at 8000 RPM or higher, but in their comments, they say that you need to go for a deeper cut to create bigger chips so that you do not melt the aluminum, which is fine, but I am fairly sure that you will then be generating more deflection and hence introduce more inaccuracy in your cuts as a result.
The post below makes interesting reading ... Also I note that he is using the Hitachi X200 inverter , which I have read in a few places as being much better than the Chinese ones.
I still think that this 2.2kw water cooled spindle should be fine on a standard gantry of a kit like the WorkBee ( same C-beam profiles as the Sphinx or C-Beam machine) which seems quite beefy compared to other and older hobby CNCs, so the extra 2 or 3 kgs of static weight is probably minimal in impact overall compared to the sideways forces of milling which could be hundreds of pounds (I am guessing at these forces but I have seen comments in that ball park)
So, my current point of view is that I might very well go for a 2.2 kw water cooled spindle and see how I get on.
I probably will be buying the WorkBee 1000mm x 1000mm as a generic CNC for average tasks as I will need to make a new set of kitchen cupboard doors and this will be big enough to do that job ( and thus pay for itself) and uses screws throughout rather than belts. Also being a moving gantry, I can use "pinning" to make longer pieces such as kayak oars, etc... and milling ends of branding irons. This WorkBee also seems quite capable of aluminum milling as per the following YouTube video so that will enable me to make the plates, etc... for a second CNC that will be purpose built for milling aluminum and perhaps a little steel from time to time.
That will take that heavier load off the WorkBee so as to reduce wear and tear on that. The new custom CNC will be made just big enough to make what I need in aluminum but no bigger so as to keep rigidity to a maximum (probably a similar size to the small C-Beam plate maker kit, (but wayyyyy tougher and ridid)
If I find that the 2.2kw spindle is "too big" for the WorkBee, I will use it on the new custom CNC build as that will definitely be built to handle that or even bigger spindles!
I hope this helps you guys a little.
It is good to share our research so that we can all benefit from that !!
==================================================================================
"Chinese" spindle minimum speed
Hi Hirudin,
Sorry for the late response.
I've just received my new 2.2KW water cooled spindle.
Before purchasing I had exactly the same dilemma, as many times when machining steel, or even aluminum with large diameter HSS tools you need rather low RPM.
Now I have this spindle, and a Hitachi X200 inverter.
The very first tests that I did was seeing at what lowest RPM I can run it.
To my amazement, it was no problem at all running even it as low as 200RPM.
At first it had very little torque at this speed. After playing with some parameters of the inverter, I managed to throw the full 10Amps while turning 200RPM.
Of coarse it will start to heat up rather quickly, but this is the advantage of water cooling. If you have sufficient pump and radiator, there is no problem.
I have about 1 liter water reservoir, and a small radiator with a fan.
I can run it at 200RPM, 10Amps for more than half an hour. It becomes pretty warm to touch, and the reservoir temperature climbs to about 45 centigrade, but that's the hottest it gets. I can continue to run it, and it won't get any hotter. I think that putting a bigger radiator would reduce the temperature even further.

I tried to mill steel (low carbon) with 3/8 carbide endmill, 0.1mm DOC at 800RPM. It ate it without any difficulty.
Aluminum is even better.

 

I called Ooznet about their WorkBee CNC and they recommended not going above 3Kg for the router spindle, which seems odd to me considering all the other huge milling force, but .. that is what they recommend, so better to stick with that for now.
I asked them about milling aluminum and they said they use the Dewalt 26200 at it's low speed (I think this is 16000 RPM) using an 1/8" or 1/4" carbide bit with a 0.4mm cutting depth and a cut width of the size of the bit and a cutting speed of 600 mm per minute.
He said they have also done mild steel at 0.1mm depth at 180mm per minute quite successfully.
Catch 22 for a water cooled spindle right now in that I would need to fabricate a collar to fit the spindle onto the WorkBee before I can use it on that. (I guess I could do a crude wooden one to get going, but I may just start off with the D26200 and work like that whilst fabricating the new dedicated aluminum milling CNC machine and then chose as powerful a spindle as possible for that then and keep the D26200 permanently on the WorkBee, so as not to have to keep switching them around and having to do all the realignment everytime.
At the moment I am looking into spare parts for the D26200 and those chinese water cooled spindles, specifically for the wearing parts especially bearings so that I can service these myself as the bearings will wear out, and quicker when milling aluminium.
I am guessing that getting spare for the Chinese spindles may be very difficult, and if so, it means a new spindle when the bearings wear out.
I have found a site that sells spares for the Dewalt router and the two bearing are about £4.67 and £7.83, which is not too bad, so we know that router can be maintained by ourselves. Their link is
Dewalt D26200K Type 1 Router Attach Spare Parts | Miles Tool & Machinery Centre
I also plan to investigate a separate robust spindle shaft and motor with a selection of reduction gears so that I can keep the motor spinning at optimum speeds and control milling range with the gearing or belt pulleys, a bit like it is done on real mills.
I will keep you posted as I find out more

 

Gray
Thanks for that .
I will have a look at that book