Chaz It is in the manual about page 37 and it clearly shows how they are connected. so the data is going via the cat5 cable so you are not limited by the 5 inputs on a PP
Printable View
Chaz It is in the manual about page 37 and it clearly shows how they are connected. so the data is going via the cat5 cable so you are not limited by the 5 inputs on a PP
Chaz I think you are getting bogged down with all this. When you home the machine the software knows that you are homing and therefore treats the switches as (homes) when the machine has homed the software then treats the switches as limits.Quote:
Originally Posted by Chaz
In your case you will have plenty on inputs so why not keep them separate. (it is also OK to use just one switch for home and one end limit)
I agree with Clive in his post above (Hi Clive !)
However my personal preference on big powerful machines is to have the home switches inside the boundries set by the limit switches, and have the limit switches trigger powering off of the drives. Also to have the limit switches within the boundries of mechanical stops that limit travel before ball screws over travel.
With a big servo driven machine like my CNC Beaver Partsmaster Mill, or the Traub lathe I rebuilt last year (now sold), the servo drives are a few killowatts, so if they run away damage WILL be done without limits.
However that said, my CNC Plasma Table is set up sharing home and limit switches driven by Mach3 - it's how the chap who originally built it made it, and it wasn't easily altered when I re-built it. But the drives are steppers and only Nema 46's so in the worst case if it runs away not much breaks !
Thanks. The penny dropped, just never knew this before.Quote:
Originally Posted by Clive S
Clive has just answered the home/limit question - Mach3 knows what you are doing so treats the switch inputs accordingly. There is a minor gotcha with this. If you use a switch with effectively no offset between on and off positions, you can get problems (to be more accurate, I had problems) with this. Z homed, then X and Y started moving. The Z switch had swapped from homing to limit function, and the slight vibration caused the Z switch to trip and stop the machine ("limit switch event"). Fortunately, the IP/M has the capability to move the axis very slightly away from the home position immediately after homing which fixed the problem.
The limit on number of inputs only applies if you are using a parallel port. That statement on the Mach3 config page is a bit misleading. There are only five input pins on a parallel port as all the others are output only, so the limit is a physical one and nothing to do with Mach3. You will be fine with the IP/A and are free to use all the ports. I doubled up my switches (combined upper and lower) mainly to reduce the number of wires needed and I still have some unused inputs on the IP/M.
Don't worry about voltages. Desertboy was talking about Arduinos and they are definitely 5V only. The IP/A will be very happy with 24V signalling, and in fact this gives much better noise rejection. The CSMIO kit is built to use the industry standard 24V signalling. No pull-up/pull-down resistors needed, but watch the wiring instuctions carefully so that you get it right. Less flexible motion controllers only have one input pin but there are two per channel on the IP/A. There are plenty of examples of how to wire NPN n/c proximity switches in series if you do a Google search - I have had four in series working reliably in testing although in practice I only wire them in pairs on my machine.
Awesome, thanks.Quote:
Originally Posted by Neale
Suppose now I need to decide if its proxy and mech or only proxy. I could do mech as primary home / limit and proxy as limits via E Stop circuit.
My drive will likely support limits too although Id have to undo some of the wiring already done to accommodate this.
On my IP/A setup i used mechanical limits and mechanical homes, two limits and one home per axis.
My limits are a high power circuit (24v contactor circuit as opposed to a logic circuit) as they are wired to a PILZ relay for the safety circuit which cuts all power to the drives if tripped - I really did not want to risk runaway blowing the end brackets off my table :)
The homes are inboard of the limits and i use home-offset to set the locations.
It all works well here.
When using mechanical switches, get quality ones with Gold-flashed contacts, these will not degrade under logic levels, normal switches rely on an arc to keep the contacts clean - these will fail pretty fast at logic voltages.
How do I know if they have gold flashed contacts? Quick look at RS doesnt seem to show that level of detail and their details are normally fairly generous.Quote:
Originally Posted by Davek0974
In my opinion standard switches will be fine. (remember logic levels are 5V) the op is using 24 V.Quote:
When using mechanical switches, get quality ones with Gold-flashed contacts, these will not degrade under logic levels, normal switches rely on an arc to keep the contacts clean - these will fail pretty fast at logic voltages.
On the whole it's the current being switched that matters rather than the voltage itself. A friend who worked on older systems that used extensive mechanical switching reckoned that increasing switch current to something around 10mA improved contact reliability. Sticking in pull-up/down (depending on how you arrange the switching) resistors of around 2K2 for a 24V system would be about right.
However, consider the failure modes. For limit switches, we are forcing contacts apart. It is very unlikely that contacts will get welded together to the point that they will not separate although it's possible that contact corrosion will stop them "making". This is at least a safe failure mode. I have no idea what the likely failure modes of proximity switches are, but I suspect that they give more repeatable homing positions. This is all pointing at mechanical limit switches and proximity homing switches - which has already been suggested.
On my own machine I have gone proximity switch all round, but I'm only using Nema 23 3Nm steppers which will probably stall before doing major damage.
Yes it needs a little current on a normal switch, i think gold flashed ones will be listed as suitable for logic levels or PLC inputs, sometimes have a very low max current as well, unlike the usual power switches.
With servos, we have index homing so accuracy in the switch is not needed - it works really well and insanely repeatable. On steppers as you said, damage is unlikely but the torque from a servo is frightening. :)
Argh, broke my long 10mm endmill and did some damage to the Z Axis, not too bad, just cosmetic but cant finish a slight hole enlargement I need to do. Pics to follow.
Microswictches can be very very accurate
http://www.baumer.com/us-en/products...ical-switches/
Accuracy is 70 times thinner than a human hair ;) For me though that level of precision is not needed lol but maybe one day.
Neale have you ever had a problem with stray chips triggering the sensors? I assume your using inductive not capacitive sensors.
I'm wondering if the honeywell microswitches might be what I want (They're gold contacts, £3 each), or rob an old lady for servos lol.Quote:
Originally Posted by Davek0974
I'm pretty certain I've sent working servo stepper motors to the scrap yard because I didn't understand what I had at the time lol.
Quick answer is no, but full answer is that virtually all my work to date has been with wood and only a tiny bit of aluminium However, the switches are mounted in fairly well-protected locations. I don't have any compressed air to blow chips around either. Yes, my switches are inductive. My only reservation about them is that the cheap Chinese switches I am using seem to have virtually no on/off hysteresis which makes them slightly less suited to home switch use. It's not a parameter that is ever quoted, though. Still, the IP/M allows me to work around this very easily. I do get very good home position repeatability.Quote:
Originally Posted by Desertboy
Well designed switch elements in micro-switches (or thumping big panel switches for that matter) are designed to 'wipe' so that as the contacts open they also slide a tiny bit in what is effectively a cleaning action.
A higher whetting voltage on contacts will overcome small amounts of oxidation but 'normally open' contacts that have been in previous heavy use, but are then left un-operated, will build up a troublesome oxide layer.
My Traub TNC350G CNC 5 axis lathe had sat unused for several years before it fell into my arms (not literally - it weighs 4.5 tonnes!) and I had no end of trouble with oxidised contacts - got there in the end and it's now back in industry making widgets, but those contact problems took up quite a lot of time eliminating.
Many thanks to AndrewMawson for the chiller. Yet to to and make it work.
In terms of man cave, I've seen nothing like it on a 'hobby' level, very inspirational. Many thanks for the tour.
Can anyone help with some Fusion 360 and post processing.
I use the standard Mach 3 post processor and override G28 in the options.
However in the code I keep getting this. This command tells my machine to home the Z axis but always hits the top limit. So I manually delete it each time (which is a pain to do).
Why do I get this and what does it mean / what is the purpose?
I refer to the G43 line at the bottom. It seems to use the machine coord not the Z that I zero manually.
7007)
(T2 D=10. CR=0. - ZMIN=-29.99 - FLAT END MILL)
G90 G94 G91.1 G40 G49 G17
G21
(ADAPTIVE3)
M5
M9
T2 M6
S4000 M3
G54
M8
G0 X41.253 Y-56.367
G43 Z15. H2
But the G43 is a tool length compensation, so perhaps something is set wrong in your Fusion Tool Table ?
More likely your Mach3 tool table.
G43 does not use machine coordinates. But it does change Z zero to correspond to the length of tool #2 (H2) in your tool table.
It uses the one height set in the operation, but seems to ignore that it's not machine values and basically tries to go past the machine 0 and of course hits the limit.
I'll snoop around later and do some testing.
Thor's Z Axis is alive. Short vid of it running on my very wonky table at 5000mm/min - 25mm cycles via the PID tuner. No tuning has taken place yet, Ill do this once mounted and limit switches have been installed.
https://www.youtube.com/watch?v=lZ9888Zv6Ck
Vertical Video Syndrome !!!!!! :
https://www.youtube.com/watch?v=Bt9zSfinwFA
Lol, sorry.Quote:
Originally Posted by AndrewMawson
No offence meant Chaz. I hadn't realised myself why some videos were like that until I happened to trip over that explanation
And non taken.Quote:
Originally Posted by AndrewMawson
Thanks for that Andrew, I hadn't realised this until I just read your post!Quote:
Originally Posted by AndrewMawson
I thought videos displayed in three vertical stripes like that were simply some new technological fad intended to irritate some of us old farts! Lol
Maybe people shooting vertical videos are often unaware of this, I certainly would have been.
Let's hope the manufacturers improve their default settings to mostly phase out this mode of recording.
I need to order cable for limit switches (3 core for Proxy) and then 4 core for the Spindle / Earth.
Do I need either of these cables to be shielded? 3 Core cable is common, shielding adds a fair bit to the price. I dont mind paying if its absolutely needed however no point if its just a 'nice to have'.
Any suggestions for cable for the 4KW type Chinese spindles?
Thanks
Personalty I would use 4 core cy cable 1.5mm for the spindleQuote:
Originally Posted by Chaz
Thanks, what is CY cable?Quote:
Originally Posted by Clive S
http://uk.rs-online.com/web/p/produc...FUQz0wod-YEKnAQuote:
Originally Posted by Chaz
any electrical wholesale should have it
https://www.cse-distributors.co.uk/c...creened-cable/Quote:
Originally Posted by Chaz
You can buy it by the Mtr.
Thanks both.
And on the limit switches, screened or not?
Suppose I can just buy 50 meters of this stuff and use for both the limits and spindle, done.
Its not a twisted pair but at least its screened.
As with Clive I always use shielded.
Pity you didn't mention it when you were over here at the weekend - I bought full rolls of SY cable for the limits and steppers on my CNC Plasma table and have a fair bit left
No problem, thanks. You were already giving me free stuff, felt bad enough taking that, nevermind loads of other stuff too.Quote:
Originally Posted by AndrewMawson
How well does SY cable bend in energychain? I'm only used to using the 32amp 3 phase 5 core variant but I know that would be a bitch to fit unless you you used the same energy chain I see on skelly trailers (8" wide)Quote:
Originally Posted by AndrewMawson
I regularly throw energy chain away this size if it would be use to you Chaz, far too big for me lol
SY cable and CY cable are different (I might be wrong here) SY is armoured cabling and I would think not suitable for moving cables in a cnc machine although perfect for the main feed. We used to use SY cabling for our 4 poster ramps (7.5T each) hardy shit but I still managed to break it lol.
In case your wondering drive over SY cable with all 3 axles of a trailer locked on will do it ;)
SY in the 4 core / 6 core / 0.75 mm CSA and 1.0. mm CSA sizes works fine for me in energy chain on my CNC Plasma
Here it is being installed:
Attachment 21875
Attachment 21876
Attachment 21877
Chas As Andrew has stated. Here is a link to see the difference http://www.control-cables.co.uk/. But if you are ordering you can get 3 core 0.5mm will be fine as there is hardly any current to worry about in limit and signal circuits. Everybody will have their own opinion of course but these have been tried and tested by many.