Printable View
Server cabinets are cheap enough, the larger ones can't be given away sometimes. I have one dismantled in storage, deep enough to put a tool cabinet in the front and the control electronics in the back.Quote:
Originally Posted by AndyUK
I built my control box in an old computer case, with the computer in another :-
Attachment 26740
It got a bit messy recently when I splashed it with a solution of black ABS, nothing like the mess when a spray can exploded on my garage:upset:
Cheers,
Rob-T
I've been wondering about how to retro-fit stall detection to deal with every 2-motors-on-one-axis-machine owners worst nightmare using DIY rotary encoders. So far I have a bag of 15 optical detectors (under $10AUD including delivery from RS), a nice Postcript file that Google found for me that will draw the encoder pattern and a piece of 0.25mm stainless shim. The plan is to use the 'toner transfer' method beloved of DIY circuit board builders to acid etch my own discs to fit onto my ballscrews and use an Arduino to detect when one axis moves and the other doesn't. You could always use this to detect lack of movement on any axis by comparing the encoder output with the step pulse input.Quote:
Originally Posted by diycnc
This is going to be one of those 'something to do while the varnish dries on my main activity' projects so don't hold your breath, but I will be reporting my success (if any) on the forum.
Kit
I noticed that. I think it's something my beloved CamBam software cannot do and might just be the last straw that kicks my a**e into gear and makes me put the effort into learning to use Fusion 360. I'm sure my machine is rigid enough, or it will be when I finally find a suppllier for the right sized Ally plates to convert my Z-axis from it's current plywood prototype assembly.Quote:
Originally Posted by JAZZCNC
Got a new multimeter and I'm getting 33v from each secondary. Glad I didn't get a 35v transformer or I would be very close to the 80v max.
Attachment 26744
To wire this thing in series, am I correct in thinking I connect the yellow and black?
Attachment 26743
Also the drivers dont have positive and negative marked on the input, both are just labelled AC. I'm guessing that because it AC it doesn't matter which way around it connected?
Yep.Quote:
Originally Posted by diycnc
The drivers dont have positive and negative marked on the input, both are just labelled AC. I'm guessing thats because its AC it doesn't matter which way around its connected?
Does this also apply to the location of the fuse?
Yep often they are labeled R S but can put anyway around. I always connect live to R and Nuetral to SQuote:
Originally Posted by diycnc
Finally got around to putting some test electronics together, just need to connect the motor and flash grbl onto the arduino. Would be awesome if someone could take a look at my wiring befor I do anything too stupid. I'm a bit unsure about the earth at the moment. The negative enable, step, and pulse are all connected together and the ground pin on the arduino. And mains earth isn't connected to anything at the moment.
Once I have one tested and working, will test all three together.
Attachment 26770
Just spotting I have the fuse on the zero volt side of the psu!
If it were me I'd run the enable, pulse and dir. signals as twisted pairs - common the -ve ends up at the Arduino. This approach gives much cleaner waveforms and better resistance to interference.
Maybe because you've used brown for 0V and blue for +68V & hence thought the fuse was in the "live" side?? I tend to try to keep blue and brown for mains N and L and use other pairs of colours e.g. red/black or orange/violet for other power rails, helps to know what's what if the wiring gets "busy". As it's AC and floating it really doesn't matter anyway!Quote:
Originally Posted by diycnc
<deleted - thought I was replying to a recent post about wooden enclosures... >
That's exactly what I was planning on doing, but at some point I managed to wire brown to + and blue to 0! Not really sure how.
When it comes to the finally wiring I'm definaltly planning on using other colours to differentiate. Is there a standard when it comes to control panel wiring?
All three drivers tested and working perfectly. Thanks for everyone's help with this!
Iv got a feeling it might be a while before I make any more progress on this. Some genius (me :culpability:) decided to start decorating three rooms at the same time, so that has to take priority now or the missis won't be too happy if it ain't finished before Christmas
Think on the bright side: You may be able to lose the cost of a tin of Hammerite in the housekeeping paint budget:glee:
Found a little bit of time to wire up the motor on X axis (side to side one on the gantry) to see something moving.
I very quickly seamed to find the limit of these motors+drivers+screws at between 4000 and 5000mm/min (800-1000RPM at the motor). I was hoping for a little more than this. Acceleration is down at 20mm/sec2.
Do these sound like reasonable numbers when compared to other similar machines? Im a little more used to deal with lightweight 3D printers which are happy fly around at 150mm/sec+ all day. Guess a heavy CNC is a totally different animal though.
Spec is; 3.1Nm Nema 23 wired in parallel, 3.08A peak driver current, 68Vac, 1600 steps per rev (1/8th micro stepping), 1:1 belt drive, 1605 screw.
I know 1610 would be much more appropriate, but 1605 was sooooo much cheaper!
Motors are rated to 4.2A, so I could push the peak current upto 3.77A. Would heat dissipation become an issue running at 90% the rated current?
A quick google seams to suggest around 1000RPM is the limit before torque starts to drop of significantly, so im starting to think this is the genuine limit of the motors.
Its bloody loud too! Was expecting digital divers to be nice and quiet! I know increasing microstepping would help. What microstep settings do other people use?
My machine is a very similar size, and also with 1605 driven 1-1 from a NEMA23/3Nm. I get around 5000mm/min as well, although I haven't really tried winding it up much more as that is about as much as the long axis will take anyway. That was my design goal in any case as, as you say, that is about the corner speed of those motors when they start to lose torque. Can't remember the acceleration I'm using - I'll try to check next time I fire it up.
Brian's law of motor current tuning is - "keep winding it up until it starts to get about as hot as you can comfortably hold when it's been running for a while." I'm not sure how accurately those driver current limit numbers are given that they are averaging a pulse output, and the motor current ratings are RMS? Peak? Who knows - the only issue is whether or not they get too hot in practice! Stepper motors are designed to run hot and mine have run using that adjustment technique for a few years now with no apparent ill effect. As for noise - that one's very difficult to judge!
I'd be interested to know where you found cheaper 1605 ballscrews. A quick scan of Fred's BST Automation's site has 1000mm long ballscrew sets of 5 and 10mm pitch at $148 AUD and $135 respectively, including delivery to Australia. Interestingly he also has 1616 screws for those wanting an ultra-fast machine.Quote:
Originally Posted by diycnc
Ebay.
850mm 1605 screws are just £30, and are delivered from the UK so no need to worry about getting stung with import tax.
To my in experienced eye they seem ok. Were packaged well, and dont appear to be to bent.
That might be why things are sounding loud. Are they lubricated well?
Not at all! Just the sticky grease they come with at he moment. Pretty sure the noise is just the motor vibration on the metal frame, not the screw. Might try taking the belt off tonight to test without any load.Quote:
Originally Posted by AndyUK
This vibration could be causing resonance which will affect the performance.Quote:
Originally Posted by diycnc
However, at 5Mtr/min your close to reaching the limit and other factors will come into play so resonance could be clipping your wings a little.
Also depending on the controller then you might be better off dropping the micro-stepping. When reaching limits the controller and quality of pulses starts to play its part.
With 68V and digital drives on a machine with decent rails, aligned and set up correctly with average gantry weight of 30-50kg then I'd expect to top out around 7mtr/min with 5mm pitch. I would tune to run at 5mtr/min.
Would something like a CAT 6 cable be a good idea?Quote:
Originally Posted by Voicecoil
CAT6, CAT5 whatever you have lying about. Try one length of cable per driver with all the screens grounded at the Arduino end as well. If you can keep the cable runs for driver input signals and high current motor outputs separate that helps too.Quote:
Originally Posted by diycnc
Had an experiment with micro-stepping today. Increasing micro-stepping reduced noise, and increased the speed before it started skipping steps.
Im using GRBL for testing, which has a limit of 30kHz, so 1/8 is about as high as I can go and still achieve a sensible speed. Any higher and the Arduino crashes.
What micro-stepping settings do others tend to use when running Mach 3 or similar from a parallel port?
The drivers go up to a whopping 1/256!
A lot depends on the stepper drivers - many max out around 100kHz-200kHz.
My recent tinkering are with a lathe conversion and I'm settled at 8x, though I'm re-using modified drivers that will take 250kHz
These drives are rated to 200khz. Assuming I'm running at max for 5m/min, I could get to 32x with a very healthy margin of safety.Quote:
Originally Posted by Doddy
I'm running Nema 23 / 3Nm steppers driving 2005 ballscrews up to 1000 RPM via 1-1 belt drive giving 5m/min rapids. I use x8 microsteps and it all works fine. General opinion is that somewhere around x8 is the sweet spot between smoother, quieter, motion and losing torque (which happens as you wind up the microsteps). I wouldn't exactly say that my machine purrs but it isn't a bag of nails either!
Pointless because your motors cannot resolve that low and it just stresses the pulse engine, Also if I'm reading correctly and you are using an Arduino then I'm not sure it can output pulses at this rating.? . . . But i don't know much about them so could be wrong.Quote:
Originally Posted by diycnc
The main reason to up the MS is to give smoother running motor but there is, as always a cost, much past 16x is waste and just stresses out the controller and wastes torque.
The max frequency for a GRBL on an arduino is 30kHz, which effectively limits me to 1/8 anyway.
I just tried running a motor off the machine, with zero load, and it's 50:50 if it will run at 900rpm, which is only equivalent to 4.5m/min. I can feel the motor start to vibrate as it accelerates up to speed.
Going to try tidying up the wiring to see if that helps. Got twisted pair for the control side. For the motor side, this might be a silly question, should I run all 8 wires all the way to the driver and connect the pairs there, or join the pairs close to the motor and just run 4 wires from there. For the purpose of testing I'm not using shielded cables at the moment, would that impact motor performance at all?
I'm assuming you mean you have 8-wire steppers? In which case it depends on how you're wiring them. If in series then definitely wire the phases together at the motor - you'll reduce your losses due to the resistance of the wiring that way. If in parallel then its less obvious to me (5:30am in the morning for me) - paralleling at the motor and only taking half of the cables back to the driver suggests greater losses in the motor leads. Ideally short the motor pairs at the motor without cutting the cables and take the pairs back to the driver and terminate there as well (i.e. run the cables in parallel back to the driver). Or just wire in parallel at the motor and use wire with a large CSA to export the motor drive back to the driver. It will make some marginal difference.Quote:
Originally Posted by diycnc
Shielding will not affect motor performance, thought it might help to avoid any induced noise into signal cables.
Could someone help shed some light on the correct orientation of AC bearings?
I thought I had a handle on it, but iv just come to install some pulleys and iv spotted iv got some face to face, and some back to back, so now I think i must have misunderstood something.
My current understanding it face to face is the preferred orientation and to preload them to remove any backlash I shim the outer race.
Face to face is with the larger inner race face pointing towards each other?
A second opinion on this would be great!
Scratch that.
If I'm shimming the outer race, I need them back to back.... with the larger inner race face pointing away from each other...?
Attachment 27145
Yea, for this application back to back is what you want.
I'm struggling to decide between to different belt arrangements for my y axis (the one that moves the gantry). I will have to ball screws driven by one motor, with either one belt connecting all three pulleys. Or two separate belts, one connecting the screws, and one connecting the motor to a single screw.
I have space for either setup, and enough shaft on the end of the ball screws for the two pulleys needed for the dual belt arrangement.
I can see pros and cons for each, but struggling to make a decision.
Any suggestion?
Use the one that allows the shortest belt arrangement. They will all work equally well as each other but shorter belts are easier to find and implement.
This is my preferred setup.
Attachment 27732
Thanks Jazzcnc.
Can anyone recommend a method of reinforcing my bolted connections on my steel frame. Currently have 4 m5 bolts at each connection between the steel box section, and after moving the machine around a few times during construction I'm certain the joints have moved.
I'd rather avoid buying a welder, so that leaves either more bolts, or glue.
There isn't much room for more bolts, so I think glue would be better, but I'm not sure what type would be best. Would a slow setting epoxy be a good idea?
What's the orientation of the sections you're joining? Do you have plates/brackets? (a pic or drawing might help) Either way, 4 x M5's sounds a bit light duty to me, do you have room to go up in diameter? Properly done with the correct surface preparation and curing, epoxy can give a very strong bond - seems to work OK in aeroplanes :-)
I'm not in the workshop at the moment, but take a look at the pic in post 36.
Agree 4 x m5 isn't enough, and I dont think there is room to go up in diameter. The existing bolt holes are a little too close to the side walls to go up in size.
Will try some thin slow setting epoxy on a test piece befor I apply to the frame
From the pic. and the few screws that are visible it looks like there might be a bit of room to go larger - but then again it's not a close-up, so do check it. If you're using slow setting epoxy do check the manufacturers curing instructions. Araldite Precision for instance cures with better strength at a slightly elevated temperature - as well a a lot quicker. And remember that once it sets you can't adjust stuff any more - so make sure everything gets properly lined up!Quote:
Originally Posted by diycnc