[cincron]

Hello everyone.

First post after lurking and reading for quite an extended time.

How does the X6-2200L go with cutting small details in metal? (Brass/aluminium etc) and when I say small, say details under a mm (1/32?) in size? 《I'm from a land of metric so my imperial translation may be rusty)

How far can you go in working small decoratively speaking?

I am sure there is no way on accuracy to be close to good enough for watchmaking level - but with the right cutters can you get good/reasonable results at the tiny end of working with these machines?

I cut 360 brass and T6 aluminum. How fine do you want to go? I've used .02" ball mills on brass.

X and Y I can hold a .001". Z wanders a bit. .005".




Sent from my iPhone using Tapatalk

[Navarre]

Many thanks Cincron :-)

That was helpful seeing the pic of small work.

[extent]

Has anyone done a lot of repeatability testing doing xy probing? I find that as I repeatedly probe the same position (without zeroing the dro) that the DRO will slowly drift off zero and it seems to effect the actual work coordinates. I can setup a test indicator and see that it's actually stopping in the same place within a few tenths, but after probing about a dozen times in a row I can get the DRO to read like 2 thou off. If I actually g1x0 then the zero that it has saved will be 2 thou off from where the test indicator is set and the DRO was originally zeroed. I can g1 to the same positions over and over again and hit the same spot each time, so it's not like I'm losing steps or something mechanical, it only creeps in when I use G31.

Doesn't seem like a mach problem or I'd think more people would have noticed it, maybe something with the USB motion controller or the mach interface plugin?

[RCaffin]

Yes, I have spent a lot of time doing g31 testing with a Z-probe I was developing. Of course there was noise in the results, and sometimes there was drift.

But the noise was of the order of a few steps of the Z-axis, and the steps are about 0.8 microns. (DC servo, 512 line encoder, 3:1 reduction, 5 mm pitch). Given that I was not going dead slow for these tests, I thought that was reasonable.

The drift was another matter. Total drift might have been of the order of 10 microns: that is the Z position would decay by up to 10 microns down to an equilibrium value.The results did look like a classic decay curve, and I suspect that was dirt, contamination or oxide layer on the contacts. In some cases the contacts were Ti rod, and that does have an oxide layer. The drift was significantly reduced when I replaced the weak spring with a stronger spring.

You mentioned 2 though: that's 50 microns. That's a bit high (depending on the machine): I would want to know a bit more about your probe, and whether there could be anything on the contacts.

Cheers
Roger

[extent]

it's not drifting in physical position though, it reliably (or reliably enough) stops in the same place each probe which the physical test indicator confirms. It's the DRO that is drifting. When it doesn't drift I get a repeat reading within about 2 steps (6 micron on a stock setup, at least for x/y)

My probe is more than questionable, but probing variation shows up on both the DRO and the test indicator, so I feel like that's pretty well eliminated for the size of offset that I'm seeing.

What controller are you using on your machine? What it feels like to me is that the motion controller is handling the probing routine internally, like it does for homing, but it is sometimes slow to update mach3 when it has actually stopped. Thus allowing mach3 to overrun it's position internally and slowly lose it's home position. The DRO drift is always along the direction of travel, where regular trigger noise is +- a few steps.

[dharmic]

Don't forget that G31 just tells the thing "begin the stop process when you sense", and the program stepping to the next instruction will be some mildly variable length of time after that based on deceleration curves etc. Add to that the fact that your probe input is being polled (AFAIK there's no interrupt on the probe sense) via software on a jittery architecture and I'm not surprised it's moving around a bit.

What happens to this phenomenon as you change the feedrate of the G31 command? I noticed big lag and jitter at 500mm/min and very, very little at 20mm/min. Which is why a lot of probing routines go to a sense at speed, back off a bit and try again at a creep.

[RCaffin]

What controller are you using on your machine?
Dell Vostro, WXP SP3, NO network connection to other machines, mach3 .062, ESS (latest driver), Gecko 203V drives to DC servo motors, 512 line encoder feedback.
Probing speed was about 50 mm/min, which is a bit fast, but it was reproducible.
Note that AFAIK it is the ESS in my system which senses the probe, in hardware (no polling), and collects the data, which it then feeds back to Mach3. This feedback is NOT time-critical. I have found it reliable, but I have also heard strange stories about other Chinese controllers.

I would NOT try probing at 500 mm/min - at least not without a LOT of over-travel allowance on the probe!

I have never seen any Mach3 DRO drift.

Cheers
Roger

[madeinoz67]

Just received my X6-2200EPL (with standalone controller) and happy with the quality, accuracy is good in my initial tests on some PCB isolation routing with copper cam, 0603 smd with 15deg v-bit. have yet to do some real cuts on aluminium. I've yet to tram the machine and PCB was milled without any bed levelling.

Is there a procedure for tramming these machines?

[dharmic]

1. Use really really tiny bits so you don't have to worry about it

More seriously: between the collet size and the minimum realistic spindle speed of about 6000RPM on these machines means 12mm end mills are about as big as you're going to go. At that size, bed and beam flex are going to cause more surface finish issues than the spindle being out of whack by half a degree. Furthermore, anything you do to tram in the spindle is going to have to be re-done every time you loosen the clamp and move the spindle which, given the fairly limited Z travel on these, I've found to be at least once or twice for each new job.

I will probably maybe look at getting mine sorted one day - pulling up the bed, bracing it, re-seating it and getting everything squared up. In the meantime it's been good down to about 0.1mm with light cuts so I haven't bothered and I'm the first to admit that if I need better than that kind of accuracy I'm probably looking at spending my time and money towards a proper mill instead of throwing it at what may be better than most cheap routers, but nevertheless remains exactly that - a cheapo chinese router.

[Mmpie]

You can't tram it as it is. The spindle mount is larger than the spindle and contacts only in a few spots, that leaves play for the spindle to move around inside the mount.

I've tried shimming the inside of the mount by wrapping tin foil around the spindle although I'm yet to do any sort of test to see if it helps.


Sent from my iPhone using Tapatalk

[madeinoz67]

Thanks for that, 12mm is more than I was looking at seeing I've come from a very light model cnc that could only do 3.175mm max so even 6mm is a bonus. For me

How about bed levelling? while I haven't checked mine as yet, can they be levelled reasonably manually before surfacing spoilboard?


Sent from my iPad using Tapatalk

[dharmic]

The beds on these aren't great. The extruded t-slot panels are pretty much unsupported down the length and they aren't held together very well on their long edges. At present I have a 0.7mm step between one and its neighbour in the middle of the table, most likely caused by a lack of support plus my getting a little carried away clamping something down on one side at some point. Ignoring that, I get pretty major flex on the dial gauge just leaning on the middle of the table.

Somewhere way way back on what's become quite an unwieldy beast of a thread now, someone has removed the bed and added bracing before returning the bed to the table. On the list of one day things to do for my machine, at which point I'll also do some shimming to get the table level to X and Y and the t-slots accurately parallel with Y (I haven't yet measured to see if they are or not). That'll let me make up a couple of lock-in fences which I know are parallel to X or Y for some of the clamping I want to be able to do etc.

[madeinoz67]

Yes I gathered that most of the aluminium extruded beds on these type of machines would have some flex, however I don't know why they resorted to panels and not a single bed, well I guess I do is all about cost, but even some of the thicker single piece extrusion would her even better than what was supplied. I purchased one of these very cheaply for a smaller CNC I have and is very solid, much thicker than one installed.

[Mmpie]

Sorry, probably me who turned this thread so unwieldy :drowning:

Anyway, my bed is unparallel also. At the far end I have already drilled out the poorly positioned tapped holes and use low profile head bolts going straight through the extrusion into nyloc nuts below (this also allows me to clamp my spoilboard into a tnut right at the very end of the bed), I don't know what I'll do about the other end as you can't access the underside. I'm planning on doing the steel bracing sooner rather than later and also filling the channels between each piece of the bed with epoxy or cement to make it a more solid single piece.

The thin strip of the extrusion that the bed is bolted to and rests on at the far end strikes me as a big weak point which will give potential for some real flexing. I'm thinking maybe an extra strip of steel running underneath that will help sure it up.

[dharmic]

Don't apologise, there's a wealth of information here now. Just a sad fact of life of forums that it gets hard to find as a thread grows.

Please try (if you remember in the heat of the moment) to grab some photos as you go and share them with us!

[RCaffin]

I know that making machines from extruded aluminium sections is very popular - perhaps because it is so simple to cut the bits up. But the stuff is not all that strong or stable. It is great for building partitions and display screens where accuracy does not matter.

Don't believe me? Find a bit of the stuff and put a straight edge across it. Not along the length, ACROSS it. You will quickly see that the surface is not flat: it is slightly concave. This is deliberate as it guarrantees that sections will mate easily. But at the same time, it means that when you tighten up the connecting bolts, one of two things will happen. If you do the bolts up to a 'reasonable' torque level, the joint will be weak and will flex. If you do the bolts up tightly, to eliminate this flexibility, you will deform the extrusion cross-section.

Needless to say, quite a few mfrs will disagree with me. That is to be expected.

Your choice what materials you use, but You Get What You Pay For.

Cheers
Roger

[A_Camera]

I know that making machines from extruded aluminium sections is very popular - perhaps because it is so simple to cut the bits up. But the stuff is not all that strong or stable. It is great for building partitions and display screens where accuracy does not matter.

Don't believe me? Find a bit of the stuff and put a straight edge across it. Not along the length, ACROSS it. You will quickly see that the surface is not flat: it is slightly concave. This is deliberate as it guarrantees that sections will mate easily. But at the same time, it means that when you tighten up the connecting bolts, one of two things will happen. If you do the bolts up to a 'reasonable' torque level, the joint will be weak and will flex. If you do the bolts up tightly, to eliminate this flexibility, you will deform the extrusion cross-section.

Needless to say, quite a few mfrs will disagree with me. That is to be expected.

Your choice what materials you use, but You Get What You Pay For.

Cheers
Roger

This is true, but it depends on how and what you make out of the extruded material. For frame it is OK because if you are using right angle joints then your angles will be right, i.e. 90 degrees, if you square your machine. If you just drill through and join using T-nuts then you might get problems. Also, the material is not the most optimal as table material, unless it is surfaced, but it may not be as important for everybody, so maybe not everybody is bothered. I think that how the machine is depends not only only on the material, but also on how careful it was put together and which other materials are used. But of course, the better the material is the easier it is to make a better machine. That's obvious. Also, remember that there are many different aluminium extrusion qualities as well, so not all are the same. Never the less, of course, one major reason why aluminium is popular is the fact that it is easy to work with. But... it is not the cheapest material, so probably it is not just about money, but more like money vs. time. vs. tools needed to work with vs. what is the target product vs. a lot of other things. So it is not that simple to answer why aluminium extrusions are used by whom and why they are popular, but sure, my guess is that the simplicity of using it is a major factor.

[Mmpie]

To throw my 2p in to the ring. I think the lower frame could and would be much better made of steel, either welded or bolted together like the current frame. As mentioned in a previous post, at one end the bed is mounted on a thin lip protruding out the side of the main extrusion, this must offer serious potential for vertical movement in the whole bed. The long side arms which the rails sit on are just long t-slot extrusions and the linear rails are bolted in to t-nuts. Surely a steel frame with properly positioned holes for bolting the rails into would offer much better support and stability to the rest of the machine and also be cheaper to manufacture?

These machines don't even have the build quality, upon taking mine to pieces I found several misaligned holes with bolts forced in, some bolts very loose (which could have happened in shipping) but some were over tight.

[VisualConduct]

Did anyone machined copper and/or sterling-silver on this machine, i was thinking of 1mm plate.
What minimum size of bit would you suggest and what quality?
Would love to hear any of you experiences. :-)

Many thanks, Robert

[cincron]

Did anyone machined copper and/or sterling-silver on this machine, i was thinking of 1mm plate.
What minimum size of bit would you suggest and what quality?
Would love to hear any of you experiences. :-)

Many thanks, Robert

145 copper machines nice. So does 6061 and 360 brass.

I've used 1.250" inserted facemill and I've used .02" diameter ball mill (diameter, not radius)

On X and Y I can hold .0015". My Z axis wanders, aprox .005"





Sent from my iPhone using Tapatalk