I had a read up and I think it's a similar thing however it didn't seam to make any difference at all that I could figure I tried up to 4000.Quote:
Originally Posted by EddyCurrent
Printable View
I had a read up and I think it's a similar thing however it didn't seam to make any difference at all that I could figure I tried up to 4000.Quote:
Originally Posted by EddyCurrent
Mach De-bounce isn't used with Csmio. More advanced Noise Filtering is done in the controller thru the Plug-in.Quote:
Originally Posted by EddyCurrent
Not to be trusted so run a Earth wire from both the frame and the Gantry back to the Star point in the Controller.Quote:
Originally Posted by charlieuk
When you say the ali frame is bolted to the floor that's OK but it would still need to be bonded the star point. Also what voltage are you running the sensors at. .. CliveQuote:
Originally Posted by charlieuk
Edit the mistro beet me to it:star:
Had a feeling that was the case jazz
thanks guys will do that and report back voltage at the switch is 23.9/24v
Thinking about it then I really need to earth each of the axis? If so can I earth the z to the y then y to the main frame then back to the control box ?
Charlie ring me I'll explain been in workshop since 6am drilling and tapping holes so can't be arsed to type all that's required as my fingers are killing me.!! Did try ringing you earlier today and just now but no answer.Quote:
Originally Posted by charlieuk
I have now earthed every thing together and it seams to have improved it in that the filter can now be set a lot lower however I still have the same problem. I did all so find one ferrite ring I put onto one of the signal wires at the control unit end but that didn't make any notable difference. What the best next step to take try some better switches?
Set up the soft limits and slow zones, section 10.8 of csmio manual I'm looking at.
Mach3 General Config, try 'Home SW.Safety' in both states to see what happens.
Follow csmio manual section 10.11 and go through Mach 3 settings as shown.
Will take a look at that, just to let you know I have separate home and limit switchesQuote:
Originally Posted by EddyCurrent
I took a look at all of the above and no luck so I have just taken a new switch and wired it straight in to the csmio and it would appear by just testing it by hand it cure the problem which would suggest that perhapsthe switch or the csmio doesn't like the long cables !
long cables generally add capacitance and that will affect the signal, you can get low capacitance screened cable if you do a search.
As an experiment use some single core unscreened cable and temporarily wire the switch back to the input using that.
It won't be the Csmio Charlie it will be a dodgy switch for sure. The Csmio are by far the best controller I've ever seen at handling noise, even my high frequency Tig welder doesn't bother them, no other Card I've used could handle this amount of noise.!Quote:
Originally Posted by charlieuk
To test I've just connected one of those cheap prox switches to a 100mtr drum of cable which I put next to my VFD running at full speed and worked fine so i'd be changing the switch.
What kind of cable ?Quote:
Originally Posted by JAZZCNC
Yea my guess is the Chinese switches too! Can you recommend any good ones? I looked on rs components but couldn't tell the difference they were between £20-50 each though! Would love to get some to try before the weekend though as tony is coming back down from Leicester to help tinker with it on sat. In the mean time I will swap the x limit swich and see if it's any better but I have a feeling they all need to be swapped to better ones
I have just switched the x limit switch over to a new one and no difference so thinks it's going to need them all the way around. At least I will have some for a little machine I would like to do some time.
3 core CY same stuff I fit on machines and same stuff Charlie got on his machine I presume seen as I told where to buy it.?Quote:
Originally Posted by EddyCurrent
Yes it was from the link you sent. The only thing was that Tony used the pice of 12core I got to run accessories on to do all the switches so he used common + and - and then separate signals for each of the 6 switches. Should they all have there own separate screen cable which I do still have a roll of.?
Dean - Just a thought, I don't know much about the CSMIO, but do you remember the issue we had with proximity switches and the PMDX? The BOB accepted both NPN and PNP switch inputs but was not auto detecting the correct type. We fixed it by adding resistors between the input terminals on the BOB. I doubt it's a similar problem, but just thought i'd throw that suggestion out there incase Charlie is using similar switches.
You having a laugh Si how could I forget that . . . . . . Cost me a nights sleep away from home and the remaining few strands of hair I had left. . LolQuote:
Originally Posted by firetrappe
Csmio inputs don't work like the PMDX and doesn't need to detect switch type as your free to use any logic at any voltage upto 30vdc. Your problem was with the PMDX rather than the switch and how it handles PNP logic.
Oh well, it was worth a shot. I guess it's the switches then, unless the length of the wire runs are causing some issues. Might it be worth trying a spare switch separately on shorter wiring into the csmio to test?
It was worth it though to see you dancing around my garden when we got it working though. (yes folks, Dean occasionally does a little dance when he's happy lol).Quote:
Originally Posted by JAZZCNC
Thanks for the suggestions anyway
so what do we think the next step is? Run a 3 core wire and try that instead of having them all together?
Try some different switch and if so which ones? Are Any from rs sutable as it would be handy so I could pick one up to try tomorrow.
That's what I suggested earlier but using single core wire. Capacitance acts like a time delay and that's what you are getting.Quote:
Originally Posted by charlieuk
Better still use a piece of CAT 5 twisted pair, it's imunity to noise is good.
Sorry I must have missed that I will give it a try.Quote:
Originally Posted by EddyCurrent
We'll I tried 3 core shielded the full length and no luck I actually had to set the filter higher but still had the same effect. Running out of ideas, perhaps try a short run of cable just long enough to go straight from the box to the switch? Or try a different type of switch
I'm not giving in! Just wired a short length of cable (2m) straight from the csmio to the same switch and it works perfect so I ether need to some how change the wiring or the switches.
For limits I prefer this type, mechanical with a roller arm. I assume your Home switches are close to the control unit ?
http://www.ebay.co.uk/itm/ME-8108-Mo...algo=undefined
Also looking at some earlier photo's I don't like the look of your 'target' for the sensor. It looks like a stainless cap head screw, try a piece of flat mild steel instead or a large steel bolt head for example.
The home switches are all on the gantry the only or that could be made closer to the control box is the x home. The triggers are at the moment a stainless bolt in a alli washer I will try all steel now
just tried just a steel bolt and it still had the fault
Did it have a hole in the middle like the cap head screw or was it solid metal right across and at least the same diameter as the proximity switch ?Quote:
Originally Posted by charlieuk
This is from Baumer inductive sensors catalogue.
Baumer
"Maximum cable length A large cable length is a capacitive load for the output circuit and increases the influence of interference. Length should therefore be kept as short as possible."
They don't quote a figure for max length though.
This is why I suggested using single core bits of wire just as a test over the distance, the capacitance is going to be low if they are kept well apart.
Save yourself the grief and fit mechanical switches for limits, you won't have these problems then.
It was a m10 flat head bolt. The stainless bolt works fine on the short cable
But thats masking any issues not resolving them.! Charlies wire run isn't long, it's just longer than your average CNC machine but for industry 10-12mtr isn't long.Quote:
Originally Posted by EddyCurrent
Mechaincal switches are OK for Limits I agree but what does he do then about Home switches.? . . Shouldn't have to put up with lesser accurecy by using mechanical switches.
I'm sure the problem lies with the fact these Cheap chinese switches obviosly don't like the longer wire run and any capacitance thats coming with it.! . . . . . That or he's got some other issue he isn't aware of yet.?
No, it's about finding the right product for the job, I know that mechanical switches will be unaffected over the length of cable he's using, plus I'm trying to find a solution that's cheap and easy to implement. If Charlie want's to keep going to resolve the issues that's fine, I don't have a problem with that.Quote:
Originally Posted by JAZZCNC
which is why we used these connected via stanadard armoured cable, okay up to 200m according to Schneider.Quote:
Originally Posted by JAZZCNC
http://uk.rs-online.com/web/c/automa...004,4294876081
No, that's why I mentioned in an earlier post about his Home switches being near to the control box, it would make sense to have the Home position at that end.Quote:
Originally Posted by JAZZCNC
That is the best conclusion now, but with fault finding it's best to start with the easy things first and in the case of DIY, the cheapest too, and I don't like guessing, it's far better to work through logical steps then we all know the 'correct' answer if it comes up again.Quote:
Originally Posted by JAZZCNC
Don't talk daft Man course it's masking the problem if you haven't found out why.!! Also you can't say proximity switches are the wrong product for the job of sensing it's what they are made to do for gods sake.! Only thing that could be questioned is the quality for intended application.Quote:
Originally Posted by EddyCurrent
That's ok for one axis what about the other Two. . ? Still got to accept lesser accurecy just to make it work M-Switch.! . . . Shouldn't and doesn't have to be that way.Quote:
Originally Posted by EddyCurrent
We started with Logical steps and eliminating obvious potential issues which don't cost money and now it's got to the point where it's time to swap components to see if fault or weak parts are the cause. Swapping out for inferior accurecy components shouldn't be even considered IMO. Esp with home switches.!!Quote:
Originally Posted by EddyCurrent
You keep on about people picking you up wrong but now you're doing it yourself.Quote:
Originally Posted by JAZZCNC
My conclusion is that it's capacitance of the cable so using mechanical switches is not masking it but using a product that will work in that situation.
I never said specifically that proximity switches were the wrong product for the job, but maybe 'those' ones are.
I know full well what sensors of all kinds are made to do having been in industry electricals for over 40 years.
I think the home switches will be okay because of the slow approach speed.
All other aspects I think are what is to be expected with DIY.
No not really got it wrong at all.! Have you considered Your Conclusion is wrong.? Swapping for MS is masking the issue has the capacitance is still there and needs dealing with rather than avoiding (masking). Charlie as tried different cable and it's still there, was worse actually, yes could still be capacitance but I don't think so at only 12mtr length and 24Vdc. But my conclusion could be wrong also.?Quote:
Originally Posted by EddyCurrent
The home switches are not Ok even at homing speeds this has been tried.! Also how can they be trusted if we haven't conclusively found the cause.?
Your method while yes maybe getting the machine working isn't the correct one IMO. It's still a mask for the real cause and who's to say this cause isn't going affect other aspects of the machine later down the line.? . . . .What if it is a wiring or grounding issue that isn't found now but surfaces at a later date when newness wears off or other attachments get added like probes etc which could suffer from same issue.?
Best resolved now and after all obvious causes have been eliminated then the next logical choice is to replace suspect components for same type or like in this case with same type but being of better quality. Changing to a compleltely diffeant switch type will in this case potentially mask some other cause or fault.!
Thanks for all the input guys I went to rs but they didn't have anything suitable in stock and it would be 3-4 weeks the chap was looking at some by omron ? So unfortunately walked away with nothing. Tony has just found two to try from a place near him in Leicester and he is going to bring some micro switches down tomorrow as we'll just for a test per poses, hopefuly a second brain will help find out what's going on so fingers crossed. Many thanks for all the help it's much appreciated having people to bounce ideas around with.
The solution to the capacitance issue, if that's what it is, is to buffer the switches locally. It's not clear to me where the switches are located but a simple transistor or IC buffer would fix it at very low cost and could handle multiple switches if they are close together.
That's far too logical a suggestion for this discussion ;)Quote:
Originally Posted by irving2008
If charlie's keen he could put differential signalling ICs (e.g. MAX488) on both ends and may alleviate problems later on. They're low cost and easy to implement since the circuit's in the datasheet...probably excessive though so personally I'd try a transistor first as it'd take 2 mins to test.
The switches are all on the gantry on the moving side of each axis, just let me know what I need to do to try your idea and I will give it a shot if I can get the stuff as I don't really understand what you are saying at the moment.