Router Build number 2

[charlieuk]

how is the CSMIO/IP-M now with dual motors? I know there was talk about it a few years ago

[Neale]

how is the CSMIO/IP-M now with dual motors? I know there was talk about it a few years ago

I was one of the people asking about it, because I was trying to decide whether to go one big motor/long belt or two motors, one per ballscrew. In the end, I went the two-motor route with an IP/M. There are probably two issues - initial homing and keeping sync when running. Second point first - I have had no problems with the two motors staying in sync when all is running well. I set up the IP/M as per the manual and it does what it says on the tin. I use EM806 digital drivers with stall detect and they trip if one motors stalls (which has generally been when I've been pushing the machine a bit). Fault signal goes to safety relay which signals IP/M and everything stops with no damage.

Now - homing. I spent some time initially squaring the gantry, using the "four pins in a square and measure diagonals" method. I have inductive proximity switches as home switches on both ends of gantry; one is wired to the IP/M and the other is not. Well, it is really but not configured in Mach3. When machine is first switched on, I home all axes as usual (configured so X and Y home simultaneously to save time). I then hit estop which cuts power to motors. I tweak the slave axis pulley by hand, simulating the Mach3 homing process and watching the led indicator on the switch. That is, I wind the screw until the light goes out, then slowly wind it the other way until it relights. Reset estop, re-home just to make sure (only takes a second or two) and I'm set to go for the rest of that session. Unless estop is triggered for some reason I never go through the manual process again for the rest of that session; on the odd occasion when I've checked it's been bang on. Generally it's close enough even after shutting down and restarting later. Because of the motor cogging effect, I'm probably no better than the nearest full step doing this. So, that's to within 0.025mm at the end of a 1000mm gantry. Probably good enough for a machine used primarily for wood.

I've thought about adding a little bit of complexity which would do a proper master/slave homing job. With a lesser motion controller with the usual 5V single-ended signalling this would be easy but the IP/M lets me use differential connections to the drivers and 24V signalling which complicates putting an extra bit of electronics between the two. On balance, I would rather keep differential signalling and the manual homing. And yes, I know it's not that difficult to put the electronics together, but it's more effort than just sticking an Arduino board in there. Or I could have spent nearly three times as much for an IP/S...

[charlieuk]

cool thanks for that, so if you were to do it again what would you go for?

[Neale]

Good question. Just for context, mine is a floor-standing, all-welded-steel, construction. Mostly 50x50x3 but X rails are carried on 100x50x3. Cutting area approx 1750x750 (quarter-sheet with a bit to spare). 20mm Hiwin rails all round, 2005 ballscrews for X, 1605 for Y and Z.

IP/M, EM806, Pilz safety relay (new-but-obsolete model from eBay for £25), HuanYang VFD - all work fine. IP/M, once I had fixed the correct firmware version and "upgraded" Mach3 to a slightly earlier version, all work fine and support my touchplates (one movable, one set in bed) for zeroing and a couple of macros adapted from elsewhere for initial tool-setting and tool-change. Nothing there I really want to change, although I could wish that IP/M included that extra little bit of firmware from the IP/S to do proper master/slave homing. I even looked at putting a custom Mach3 config together to trick the IP/M into homing X and A separately-but-together (a bit like simultaneous X and Y homing, which works fine). Mach3 was easily confused into doing this, but it seems that you can only have one IP/M configuration on the machine and it didn't want to play.

Things I would do differently (and might still do so, one day):

Use better pitch ballscrews. I bought the hardware a couple of years back, at least, so I had exact dimensions to build to. I didn't want to go to China for that (although I would do so now) and my lathe would not take that size ballscrew to do my own machining (although my latest lathe will). So I bought what was available off-the-shelf at the best price from CNC4YOU (as I had good experience with them) and that meant 2005/1605. No 2010/1610 available. 2005 with my ballscrew length gives a theoretical critical speed of around 1000RPM (give or take a bit) which gives 5m/min. 1000RPM is also about the corner speed of my NEMA 23/3Nm steppers, so with 1-1 belt drive, this was a good match. In practice, I've had to drop to 4.5m/min on X because I get the odd stall on rapid otherwise. The gantry is pretty heavy (all-steel) although theoretically it should be OK. If I were doing it again, I would go for 2010, probably, with a change of pulley ratio and aim at nearer, maybe, 9m/min. The original design speed seemed like magic after my old MDF-built machine with its max 900mm/min, but it's less than optimum by a long way. Works fine, but could work better.

I have a love-hate relationship with my inductive proximity switches. I am using cheap-from-China items, around £12 for 4 kind of level, and they have one big annoying feature. There is virtually no hysteresis between switch-on and switch-off points. So on homing, gantry moves towards switch, switch trips, gantry reverses until switch untrips, and that's home. Mach3 now sets switch as limit not home switch. Any subsequent machine vibration can move the trigger just enough to operate the switch - "Panic! Limit switch trip!" says Mach3 and does an e-stop. This happens a lot on Z and gave me no end of trouble until Dean pointed out that you can configure the IP/M to move the axis a tiny amount away from the switch once it has homed and set zero there. This fixes the problem in practice, but it still annoys me that I have to do this!

One thing I did get more-or-less right was to include all the bits and pieces that I thought that I would want before I started using the machine. So, home and limit switches fitted (each axis to IP/M separately for simultaneous homing) although home and limit switches wired in series per axis. Works fine. Touchplate connections provided. VFD start/stop and analogue speed control from day 1. Two-speed water pump - nominal 12V pump that gets 12V for initial priming but then gets switched to 5V for long-period running - the spindle doesn't need a lot of cooling and I often don't use it at all for quick jobs. It's too easy just to "test it to make sure it works" without, say, full limit switches, and then you hit the wrong arrow in your haste and the gantry tries to run off the end of its rails. I've done this, and was really grateful that limit switches were there to save me. I'm also using industrial-style switchgear from ElectricCenter (branches everywhere and they were keen to give me "trade discount") which is much more solid and suited to the environment than, say, the stuff that comes from Maplin with the same nominal rating but rather less built-like-a-brick-outhouse feel. Worth the extra few quid at the outset, along with a ferrule-crimping tool that means that every wire is terminated with a ferrule. You wouldn't believe how much easier this makes inserting wires into terminal blocks in awkward places, and removes almost all possibility of stray strands causing impossible-to-find shorts.

Today, I might consider using the UC300ETH (I would always go ethernet rather than USB, and forget parallel port altogether) in place of the IP/M. Similar capabilities on paper, but I have to say that I like the feel of the bullet-proof, 24V, IP/M.

[charlieuk]

that's interesting sounds like your machine is similar to what I want to build size wise and you had similar problems to my first machine using cheap sensors which I also use a ip/m on. sound like I will stick with the ip/m then and go slaved unless there is any reason not to I want to stick with similar component to my first machine so I'm not having to learn lots of new stuff slave on the ip/m would be great.

two things I realy need to sort out for my old machine and the new one will be a touch plate and some sort of pendant but when I saw the cost of the one on cs labs it scared me a bit!

[Neale]

I'm using a Logitech wireless keyboard with built-in touchpad at the moment. Haven't had any problems with that (a lot of people say that you shouldn't use wireless keyboards as they aren't too reliable in a noisy environment) but a wireless MPG would be good. I keep looking at the Chinese wireless units for around £100 which would make zero-setting while leaning over the machine much easier. Still, finishing the swivelling monitor bracket so I can actually see the monitor properly while zero-setting would be a good step in the right direction...

[charlieuk]

I'm using a Logitech wireless keyboard with built-in touchpad at the moment. Haven't had any problems with that (a lot of people say that you shouldn't use wireless keyboards as they aren't too reliable in a noisy environment) but a wireless MPG would be good. I keep looking at the Chinese wireless units for around £100 which would make zero-setting while leaning over the machine much easier. Still, finishing the swivelling monitor bracket so I can actually see the monitor properly while zero-setting would be a good step in the right direction...

I'm not to fussed about the wireless thing, I can almost reach the keyboard from the machine but would make life a lot nicer if I could just have something to jog it around easily. do the one's that state mach3 just plug straight in to the pc? or do they go into the csmio ?

[charlieuk]

cool ill have to look in to those a bit more.

so moving on for my goal of cutting alli is 2005 ball screws for the two on the x and 1605 for y and z the right ones to go for?

Any idea on the ali gantry, I know jazz uses them a lot on his but not sure if they are all designed for the same purpose? it would certainly make my life easier if I didn't need to use steel.

[Neale]

5mm pitch is a bit low for cutting wood but plenty for cutting aluminium, or at least this is true on my machine. However, I've only just started trying to cut aluminium so haven't much experience of this yet. On balance, as I said earlier, I probably should have gone for 10mm pitch but as I do a lot of small work and 3D machining, it's acceleration more than absolute speed that improves job time so in practice it's not that much of a problem - for me. There's a bunch of engineering trade-offs in there and you have to decide for yourself where the priorities are.

I have no experience of aluminium gantries. I've gone from MDF to welded steel, so can't make any intelligent comments about aluminium but there are plenty of people out there who do use aluminium (my feeling is that this is the most popular material by far for gantries) and there are plenty of build logs and other comments about it. In practice I don't think that the weight difference is particularly important (which is not intuitive, but then neither is the fact that the rotational inertia of the ballscrews is a major factor in limiting gantry acceleration) so it's down to material availability/price and facilities available for machining and assembly that will probably drive the decision. I have a MIG welder and milling machine, which was a major factor in my choice. YMMV...