Which path to follow.
|Duff Machinist||13/07/2020 04:02:54|
|12 forum posts|
A bit of a long one this, so apologies up front.
I'm new to CNC and am in the early stages of learning. After too long, I'm now in a position to delve into CNC to undertake a number of projects that can only be done effectively on a CNC mill. This thread could equally be in the CNC Build section.
My dilemna is which path to go down - do I convert (and possibly recondition) a manual mill (new Chinese, or second-hand Swiss - I can dream) to CNC or do I upgrade/recondition a second-hand dedicated CNC mill (e.g., Emco)? On paper, I realise there are advantages and disadvantages to both but I'm unclear how they would apply in the real world. (I don't currently have a manual mill by the way).
I should also add that I'm restricted by cost, lack of know-how (at this stage), and a lack of and access to machinery which may have a bearing on which path to follow (my Myford is currently in bits and my diddy shaper is limited in what it can do). Cost, and ease of upgrade or conversion are the main considerations.
The work required of the mill, if manual, will necessitate it to be in the 200 to 500kg (ish) range (e.g. Seig SX4, Aciera F3), a working area of 300mm (x) 150mm (y), and 50mm (z) minimum, and materials machined will be mainly aluminium but occasionally tougher stuff such as 41xx steel. Most work will require precision, +/- 10 microns at the very worst. Speed is of much lesser importance. Again, if a manual mill, then an infinitely variable spindle would be preferable over gears but I'm open-minded on this.
Is anyone able to impart their knowledge and experience comparing the two paths?
My apologies if it all this sounds a bit newbyish (cos' it is ). You can call me out on any naivety on my part, and I stand to be enlightened on basic aspects of CNC.
18323 forum posts
Lack of machinery will be a problem doing a conversion as you are bound to need to modify parts to fit ball screws and nuts plus there will be motor mounts to fabricate.
As you say you are mostly working aluminium then you may find many manual mills lacking in useful speed, I know you say it does not matter but do you really want a machine and compressor/coolant pump running for 4 or 5 times as long as one that can run and feed at a decent rate? One option if converting may be to ditch the mills spindle/motor/gears and replace with a nice water cooled spindle maybe with ATC using ISO 20 taper tooling if you are going to want to be changing tools a lot during a job.
I was able to go down the dedicated CNC route and plug and play was easy for me.
I also think your 10 microns or less may be beyond a basic budget.
Edited By JasonB on 13/07/2020 07:58:13
|Barrie Lever||13/07/2020 08:02:52|
|653 forum posts|
I would say that with the exception of Tormach that many CNC machine's at the hobby/semi pro end of the market are a conversion of a manual mill.
The small EMCO's don't have the X,Y working range that you require.
A couple of points to clarify from your post, do you require the best tolerance to be +/- 10 microns? and restricted by cost is rather open ended, even a Formula 1 team is restricted by cost, just it is just a different number to most other peoples !!
Spindle RPM or lack of is a big factor in manual mill conversions when doing smaller work in aluminium.
Is your primary objective to make your projects or are you more interested in tinkering getting the machine working?
Like Jason I went down a plug and play route, that is not to say that I am not capable of tinkering but that is not my main interest.
|410 forum posts|
I would imagine that you would need a dedicated machine and for the capacity and accuracy you require you would need something like a Tormach or a Haas, as Barrie says the Emco machines just don’t have the capacity.
|1545 forum posts|
The Emco UM630 will meet all of your requirements with 5 axis simultaneous capability.
The Denford Triac (cnc mill) almost makes the 300mm X axis only misses out by 10mm and they do come up occasionally with ATC using BT30 or BT35 tooling but may not meet your 10 micron spec, but there are not many machines available that will hold such close tolerances.
|Michael Gilligan||13/07/2020 11:08:40|
15886 forum posts
___ Subject, of course, to the availability of adequate funds
|Andrew Johnston||13/07/2020 11:41:37|
5558 forum posts
If worst case 10 micron accuracy is really needed then you're into mid to high range professional machines. Converting a manual mill on the cheap simply isn't going to cut the mustard.
For a start you'll need high end ground ballscrews and ideally low friction ways. Conventional ways have static friction that will exceed your accuracy requirements - ideally you would need a closed loop system with an independent measure of position.
JasonB makes a good point on spindle speed, for aluminium you could easily make use of 10k+ rpm.
I have a Tormach CNC mill. I think they claim something like 1.3 thou per foot basic accuracy based on the ballscrews. I reckon I normally achieve 1 to 4 thou depending upon cutter, toolpath and material. I think the smallest step on the Tormach is about a tenth of a thou, but there is at least half a thou of lost motion due to stiction and probably another thou or so due to backlash.
Some years ago I had a heatsink machined by a professional machine shop as I didn't have the time to make extra ones. Unfortunately they made rather a mess of it. So I ended up making detailed measurements of one made on my Tormach and one made on their professional CNC mills. I had errors around 0.03mm to 0.05mm whereas they were a bit better at 0.02mm to 0.03mm, but still a long was from 10 microns.
You will also need a fairly sophisticated controller. Quite large errors can be caused by changes of direction in machining if the controller and/or the CAM software don't look ahead and change feedrates or make changes to the toolpath to compensate. A simple case to consider is machining a rectangle. The table can't change direction instantaneously at each corner but will tend to overshoot slightly.
One other point, the specified Z movement is far too small. It won't accomodate varying work pieces or simple tool changes.
|David Colwill||13/07/2020 11:56:35|
|640 forum posts|
I may have a Denford Triac for sale
1376 forum posts
Plus 1 for a Denford Triac.
They're a good machine of about the capacity you're looking for and can be tweaked quite easily to get more out of them.
I haven't systematically done any tests for accuracy, but I think I'm getting around about +/- 0.01mm on the parts I make on mine.
I've added a counterweight to the head on mine for less stress on the stepper motor drive and I've also added a spacer between the column and base to give me more working height.
Mine is running Mach 3, but I want to upgrade to PlanetCNC as soon as possible as that's the controller I use on my lathe and I'm very pleased with it.
|Barrie Lever||13/07/2020 12:38:12|
|653 forum posts|
That +/-10 micron does need clarifying, there are not many machines that you can jump on straight away and get those kind of tolerances at the first hit.
There are many things that effect a dimension requirement like that, if we were measuring across a 100mm flats for example it means the cut is looking at being within 5 microns as 2 x 5 micron errors and you are bumping on the tolerance.
+/-10 micron is possible on good hobby machines but you have to be hovering over the job like a hawk, the machine has to be loved and the job setup well and it will not be done with a collet and cutter set off of Banggood.
As Andrew said ideally the machine would have linear bearings as these have much less stiction than dovetails and decent ball screws (German or Swiss are usually good) , there are a number of options to get high speed spindles mounted up on the type of mill you describe. I have a 50,000 rpm spindle option (jager) sat under the bench and that mounts in place of the conventional mill head but only has a 1/8" capacity.
On reflection I think you meant your most demanding requirement would be the occasional foray into the 10 micron territory.
|Michael Gilligan||13/07/2020 15:09:46|
15886 forum posts
It would certainly be easier if you are right, Barrie ... but the statement does seem pretty explicit:
“Most work will require precision, +/- 10 microns at the very worst.”
|Barrie Lever||13/07/2020 16:10:01|
|653 forum posts|
It all hinges on "the very worst" and which way that was coming from, will make a huge difference to the final approach.
|John Haine||13/07/2020 16:33:59|
|3181 forum posts|
It would be good to have clarification on whether the OP meant resolution, precision or accuracy...as well as what "worst" means. The lack of current machinery indicates an upgrade of an older CNCmachine would be better.
There quite a number of Denfords around, Novamill and Triac, which might meet the requirement. If you are lucky they would be without the old electronics which is very outmoded and cheap to replace with modern hardware, software and drivers. Denford machines were often bought for education and may well have had an easy time given the lack of trained teachers!
In a similar position to the OP but with a working lathe and big mill I bagged a Novamill without electronics on eBay for 600 squids and built new electronics for less than 200.
|396 forum posts|
If your starting out don't touch retro fitted machines. A manual machine was designed to be just that. There are plenty of training machines about, but I would recommend the Seig KX1 or KX3. I've had a KX1 for over ten years and wouldn't be without it. Once commissioned (you need some IT skills) the conversational programming is easy to master and you can stitch any number of different operations together to build very complex programs.
|Ian Johnson 1||13/07/2020 17:27:52|
|274 forum posts|
I agree with mick above, conversions always seem to me to be a compromise between a purpose built CNC and converted manual machine.
I also have a Sieg KX1 and think it's a great little purpose made CNC mill. I've not inspected every single piece of work it has done, although everything has come out pretty darn good, it is advertised at 0.01mm repeatability. So just under half a thou isn't too bad, not far off the accuracy you are after.
A purpose made machine looks like the sensible option.
|Barrie Lever||13/07/2020 17:37:17|
|653 forum posts|
Where do you get a Seig KX1? Ketan stopped selling them and Axminster don't list it and also look to be low on stock of any Sieg CNC.
|Duff Machinist||13/07/2020 17:43:56|
|12 forum posts|
Thanks everyone for the big response - very, very much appreciated.
With your advice and observations, I'm now strongly persuaded to go down the dedicated CNC path. Tinkering is fun and educational (machinery allowing, which I clearly lack) but getting the jobs/projects done is more important for me. To clarify on some points brought up (which may now be moot)...
Yes, I will need precision as stated but the precision will vary. In many cases the tolerance will be +/- 0.03mm, with the 0.01mm precision less frequent but very much needed - it would be self-defeating to buy a machine that only does 95% of a job. However, I could mitigate against this by using two machines...
...the minimum size of table was given to incorporate the larger job sizes. Therefore, I'm basically asking a machine to do one and all jobs. In reflection, I could 'delegate' the jobs to two separate machines - a dedicated CNC to do the smaller, precision jobs, mostly aluminium, (about 90% of jobs) while the other can take on the larger jobs (about 10%). If I do this, I can reduce the working space of the CNC mill to about 150x140x40mm to incorporate all my needs. With this in mind, I'm also wondering if a CNC router (e.g., a Wegstr, etc.) might be good enough for those smaller aluminium jobs.
Having a look at my projects, the larger jobs can be done on a manual mill. It'll just take a lot longer, with a fair bit of fiddling and thinking outside the box but it can be done so perhaps I can consider this side of the issue to be solved. Also, the smaller projects are the more important so this can wait a while longer anyway.
Costing. Well, as little as possible of course. At present, £3500 at the very outside. This may be broadened if I can sell a few things.
I shall look at the Denford Triac but with my revised requirements of the working space for a CNC as stated above, would a CNC router be a viable option? I'm thinking of one where the table moves and the gantry is fixed as these appear to be more stable but there's very little feedback on the interweb on the build quality and ability of these machines. Any thoughts? (I'm also conscious that I'm going off-topic on my own thread by asking about this!).
As an aside and certainly off-topic, I've heard that, given the right set-up, a CNC router (within the price range given above) can be used for cutting 41xx steel. I'm unconvinced of this but would be interested to know if I'm wrong.
Thanks again everyone.
Edited By Duff Machinist on 13/07/2020 17:47:14
|Michael Gilligan||13/07/2020 17:44:35|
15886 forum posts
Mmm ... I read it as +/- 10 microns being the least stringent tolerance [i.e. the worst product]
but I suppose it could be read as being the most stringent ... if "the very worst" means "the most difficult to achieve"
Ho Hum : The joys of the English language !!
|Duff Machinist||13/07/2020 17:56:28|
|12 forum posts||
Sorry, it wasn't overly clear...
I meant that the precision of the machine is no worse than 10 microns, so basically the ability of the machine to consistently perform within or at 0.01mm of the precison/measurement required as the parts need to be within that tolerance. I also realise it's a tall order for a machine that costs less than a small house.
I'm not sure I made it any clearer. Words is difficult sometimes
|Michael Gilligan||13/07/2020 18:00:29|
15886 forum posts
Thanks for the clarification
Please login to post a reply.
Want the latest issue of Model Engineer or Model Engineers' Workshop? Use our magazine locator links to find your nearest stockist!
You can contact us by phone, mail or email about the magazines including becoming a contributor, submitting reader's letters or making queries about articles. You can also get in touch about this website, advertising or other general issues.
Click THIS LINK for full contact details.
For subscription issues please see THIS LINK.