The switch.Quote:
Originally Posted by Radu_Andrei
Printable View
The switch.Quote:
Originally Posted by Radu_Andrei
NPN NO inductive proximity senzor, 4mm detection distance, 6-36VDC,300mA,3 wire (brown,blue,black)
Attachment 25179
I have all the limit switches and estop wired in series (NC) to the main contactor that powers everything except the breakout board. One signal wire goes through a spare contact of the contactor to the breakout board to tell mach3 to stop the g-code.Quote:
Originally Posted by Radu_Andrei
The homing sensors are connected to the same BoB input in parallel as I don't use simultaneous axis homing but need the spare inputs (actually I use one more input on X as I have dual motors and need to square the gantry). The homing sensors are configured as home and limit in mach3.
In 5 years I've never hit the limit switches one single time as the sensors are working as limits as well and never failed.
Thanks for the reply.
I don.t get why is the contactor involved in all of this. When you interrupt the signal wire (open the contactor), you also kill power to everything else, except the BoB. Why? Why not use just Bob input terminals?
Could you please post a picture with your limits and homes setup? If not to much trouble even with the whole cnc machine. It never hurts to see a new approach.
Another thing that gives me a hard time, is that I have some ballscrew whip. It.s 20mm dia, and 1500mm long, one fixed and on simple end support. On the y axis, on the last 300-400 mm moving towards the stepper, I get some whip. At 3000 mm per min (118inch per min) it.s quite strong. Reducing the speed to 2200 (86) it is very little. I assume lowering to 2000 (78) will kind of fix it. I would prefer 100 ipm.
I found two ways of fixing that online. One is a spring support at the middle of the screw, that reduces somewhat the whip, the second is a change in design to make the ballscrew stationary, and rotate the nut (A LOT OF WORK TO CONVERT TO THAT).
Has anyone encountered this kind of problems, and what solution did you come up with?
P.S. Regarding the pictures with the cnc machine from post 204, please feel free everyone to post some photos with your cnc routers, especially those which used them for some time, as I.m sure you.ve updated your machine design to solve all or most of the problems. This includes you too, Rob. :) You.ve posted a photo with your beautiful limit switches, but I did not have the honour to see the entire cnc, and as you always seem satisfied with its performance, it is of paramount importance to my existence to see your cnc rounter. :)
The BoB input alone is not reliable enough to be used for an emergency stop.Quote:
Originally Posted by Radu_Andrei
Estop, the emergency stop, that big red mushroom button should be there for emergencies. Water is leaking through the motor electrical connections and on the floor so you could be electrocuted or something in the machine catches fire ... You will be hitting the Estop while running for the door or the fire extinguisher. In those cases you want to kill all power to the machine with that button.
The reason I've put the limit switches on the same loop with the Estop is that I have the sensors for the limit. If the sensor is not stopping the axis, then something should be really bad with mach3 or the motor drive and I want the end switch to stop everything before hitting the end of axis travel.
For minor "emergencies" like loosening the workpiece or hitting a clamp you have the mach3 Stop or Feed Hold buttons.
Hi Paulus.v
I also have an emergency estop which cuts power to everything, except the pc and the BoB, but I have never used it. (this being the hardware stop). I did, however used the software estop (input terminals on BoB) a few times, when I messed up gcode programing, and it worked quite well.
Your representation of the work conditions in which the cnc will function are a bit dramatic for a work shop, as I.m quite sure it is not positioned under a waterfall, but as I cannot help notice the flag situated at the left side of your user name it.s quite similar with mine, I think such a Estop described by you, should be very helpful if assigned to our prime minister, leading party and most of the political machine that is running our country. :) Best whishes, and happy national day for 01.12.2018! :)
I have read now what problems you had with the grounding and EMF. Are you not afraid to install the homing sensors in your system? :devilish:
Please do not wire the limit switches as I did! Better wire them in series to a relay coil and from the relay to BoB estop and the stepper drives enable pin if they have it.
If I were in your place I would try to convert all the control circuit to 24V if possible. 24V is the industry standard as it is more reliable and resistant to noise and EMF than lower voltages.
How far is your workshop from the high voltage power lines, measured horizontally?
ps. Our government deserves more than an estop, I wish them many years in jail, and I hope those times will come.
A lot of the noise problem was due to improper grounding, which was my fault, but thanks to the patience of the people on this forum, I managed to solve that and many other issues.
I was considering to run the limit swithes on 24v or at least 12v, but they seem to work fine, from the low voltage perspective. They however are very cheap lever switches and tend to have some parts wobble or come loose. I will upgrade to some profesional ones soon in the future.
My workshop is right next to the high voltage power lines. Like 3 meters away. The building existed before the high voltage tower was put in place. I asked a few guys at the City Hall about this issue, but they do not seem to have a problem or an interest in that.
I.m a bit concerned about some aspects of the cnc build, but not afraid. Crashing it, does not seem a big deal in theory, with single drive for each axis.
And it would not be the first time I mess up something in this project. Doing the math, I think I spent around 2500 euro for this project until this point. I bought the components bit by bit, during a few years, with the intent of building it someday. But from this amount, 5-600 euros was the tax for stupidity, representing materials or components, which from lack or knowledge or impatience etc, I ruined. But for the experince aquired, it worth every penny. :)
What I am somewhat afraid, is that, in order to reach a certain profitability with this kind of machine, you need a decent cutting speed, and I might not be able to get that properly. Perhaps I should have bought higher quality components, and make the cnc smaller and lighter. Who knows. I have a few components left from this build so maybe in the future I.ll build a smaller version.
You are 3 meters away from the double power lines in the picture above? No way!!! There are regulations for a protection corridor of around 30-50 meters along the high voltage power lines. I hope you are not living near... For health reasons it is recommended a distance of 150-300 meters from the HV power lines.Quote:
Originally Posted by Radu_Andrei
Maybe those power lines are used only occasionally and coincides with the time your CNC starts going crazy. Are you hearing any hum under those lines?
Don't think for one second, that I am satisfied with my machine. It is very 'entry level' A4 machining area, Unsupported rails, trapezoidal screws and delrin nuts but its' most damning feature is that it was made by MD. I have done what I can to stiffen up where possible without major rebuild. If you go carefully with it, you can get OK results.Quote:
Originally Posted by Radu_Andrei
As far as the inductive proximity sensors is concerned you need at least 12v, preferably more, for them to work reliably.
Paulus v.
As I said it.s 3-4 meters away from the workshop. My newly built house it.s like 10 meters away. The construction authorization was released by the owner of the towers, in this case Enel, and the City Hall. If I remember right, the minimum distance required was 7 meters. I did not bribe anyone. It is just the way it is. I do not hear any hum, but I.m quite sure they work all the time. I am not sure, if the are high or medium voltage power lines. My cnc works fine now after fixing the grounding. There is still room for improvement though.
As for the health issues, this electrical tower is here since my grandfather was young, and no one in the family had any serious health problems, and counting aunts, uncles and cousins, it is a big family.
There are people here alive which worked in the uranium mine, and they are alive today, and hardworking strong aparently healthy men that died of cancer. You have 60 years old smokers that do just fine, and 3 years old with leukemia. So, excuse my scepticism toward the recomendation of the experts, but they miss something. I am NOT saying that care and resposability are of no use. In fact they are great and make your life a lot better, but I would love to see for once, attached to a claim (regarding anything, health, nutrition, cancer, electromagnetic radiation etc) to see the study behind it, and I mean place, date, names of the people involved, telephones, emails, experiments, conclusions, reports etc. This is not a reply pointed at you, it is just about a principle. The scientific method principle. First let.s prove it, and then share it. :)
Sorry for the long reply. A couple of years ago, I did a personal study about nutrition and health, mainly about what foods should you eat, why, in what quantities, and listening and reading all the sides involved in the debate, these including vegans, vegetarians, frutarians, raw eaters, only meat, no meat, no grains, no eggs, no legumes etc, It lasted a few months and in the end, I guess the only safe way is to go out and chew grass, or eat nothing. Very little studies existed for all this claims. Anyway, if my health goes bad anytime in the future, I will keep you posted. :)
Rob.
An MDF, A4 machining area, Unsupported rails, trapezoidal screws and delrin nuts machine that works, it.s better that a fancy one that does not. And better than all is the attitude. I honestly though you have a high end hiwin rails, high pitch ballscrew, or tensioned double pinion and rack, 2.2 kw water cooled spindle, aluminium extrusion frame, vacuum table machine. All the bells and whistles. Perhaps more than a better machine, I need a better attitude towards it.
I just finished this evening installing the homes, and they seems to work just fine with 12v supply. I will however post a video these days, with the aparently last problem :)) (I know, don.t laugh), of the y axis wobble, or whip, I am not sure yet, maybe you see something I don.t. It.s not that bad and it is manifesting a bit strong only on 10-20 cm of travel.
My little mistake, is that seduced by the almost great functioning of the machine, I did a few works and sold some, and now the interest is quite big, especially for religious carvings, and I get a bit of pressure from the people in the area, and I cannot deliver. Not yet. Until I figure out if this whip is something that can ruin a ballscrew very quickly or it is not a big deal. 20mm 5mm pitch, one fixed one free end, 1500mm (4.92ft) long. We shall see.:indecisiveness:
My X axis ballscrews are also 2005, about 1.7m long, and fixed/floating bearings, so it's a very similar setup. I do not get any significant whip below 4.5m/min which is my max stepper speed (around 900RPM). You do need to make sure that everything is very carefully aligned. That is both bearings and ballnuts - these must be bolted to a surface that is exactly at 90deg to the ballscrew. If anything is slightly out of alignment it can encourage whip, even if it seems to turn OK at lower speeds. I tried to include as many adjustments as I could so that everything could be aligned on assembly. This took some time, but it seems to have worked OK. Don't know if this helps or encourages you, but it shows that your setup can work very well with a bit of setting up and adjusting!
Let me correct you, it is worse than MDF, it is an aluminium frame by Marchant Dice :lemo: It has a Kress router.Quote:
Originally Posted by Radu_Andrei
Rob. I still don.t get why you seem so happy about the Kress router. I assume yoga might be your secret. :))
Hi Neale.
Truth be told, for the y axis I kind of hoped for the best in the setup. I tried some techniques I found online to align the rails, and the ballscrew, but it seemed to not work so well as advertised, so in the end, I just did a feel run, like aligned one end of a rail, than bolted that, that the other one, and used some shims to adjust any issue, and when it felt like it was running smooth, I mounted the ballscrew, in the same manner. One end first in one plane, them attached a pice of rod to the coupler, and turn it with the tips of my fingers, and bit by bit, fixed all nuts until I could turn the ballscrew with the tips of my fingers for the whole 1420mm run.
With other word, beyond the language barrier: I have no clue what I did from a technical perspective. It just felt good enough. So could you please guide me towards a reliable way or alligning the y axis rails and ballscrew. :)
The secret was in the MD bit. It is a good bit of sarcasm from way back. :friendly_wink:Quote:
Rob. I still don.t get why you seem so happy about the Kress router. I assume yoga might be your secret. :))
Agree with you 100% Andrei, I am totally sick of hearing "The evidence shows . . . . . " going unchallenged!!!Quote:
Originally Posted by Radu_Andrei
Downunder "they" have a thing about MSG, which to me is the best thing since sliced bread so I use it whenever necessary.
Where's the friggin evidence?
Similarly, I've been deprived of eggs and butter for so many years because of "their" hysteria!
Hi AndyGuid.
I get a bit irritated about this, as I have tested some claims on my own, and I would say, good luck working as a carpenter on a vegan or frutarian diet. Or going out on a date, with your newly updated friend, called digestion. :) This approach to information creates a lot of confusion and frustration, and even if most debates kept by this new series of experts with the diplomas released by facebook or history channel, are harmless, it becomes a bit concerning when it becomes political. There are many people here nostalgic for the fall of communism and they would like it back, and if not, they listen a bit of anti immigration discourses and they embrace the national socialist ideology, and between Hitler and Stalin, I choose Britney Spears, cause I. sure she.ll figure out that you can close/control the borders without the need for a dictator. Same regarding health, between chemotherapy and special treated magic water, could we please see, that people are not doing well with either. Evidence shows, experts say, scientist say....first their name, location, time and date.
Clive S. I.m getting there with the big mysteries of the world: origins of men, why women exist :P , british sense of humour...
Back to business, now:
Please find below 3 video with the whip/shaking problem described. It is quite subtle, and can be best seen in the cable carrier, in my opinion.
https://www.youtube.com/watch?v=E-ti2wfkMkY
https://www.youtube.com/watch?v=4Km5oIXvfdY
https://www.youtube.com/watch?v=5BuU5OufHlk
Another thing which is a bit unclear, is why in this picture Attachment 25223, although the wood went through the thicknesser, and measuring it with the calipers , it.s equal on both sides of the board, if you look closely at the frame, the upper left side got enough detail and all the rest, especially the lower right side lost a lot of it.
I did not run a gcode facing program as I assume the thicknesser is quite reliable in delivering flat boards of solid wood. This is the reason I stopped production for the moment, as the result are bad.
P.S.
Neale , I have no clue how you can reach 4500 mm per min, as mine stalls at 3400-3500. How much does the gantry weigh? Could you please post a photo.
If I told you my secret, it wouldn't be a secret any more. There is nothing wrong with a Kress router, but it is noisier than a spindle and VFD. They are good quality, but my next build will have a water cooled spindle.Quote:
Originally Posted by Radu_Andrei
MSG per gram has less sodium than table salt - it is the sodium that causes problems- so MSG is better than NACl !Quote:
Originally Posted by AndyGuid
Sorry for my uninformed reply. On a quick google search it looked like a Dremel.
Your carving lack of detail at the bottom edge, looks as though the wood was not set level to the XY plane. So I would check the bed is parallel to the X and Y axises.
I just found out about the tramming thing and will do so in the next days. The bed was milled with a 10mm end mill, so I guess it should be somewhat parallel.
Andrei, take a look at this thread. I have described my machine there, so start with that and ask questions if there is anything else that you would like to know. Quick answer to your question is about 40Kg, I think, but the mass of the gantry is not that important to speed but it does affect maximum acceleration (and deceleration).Quote:
Originally Posted by Radu_Andrei
I spent a lot of time aligning the ballscrews and the linear rails.
Just finished reading the build log: That.s a great library of information. Thanks for that.
Unfortunately, I did not find what I was looking for. My gantry as seen on the first link from post 218, sits on two identical 24mm aluminium plates, so no need for epoxy levelling my my case, although in the absence of a flat reference surface, this sounds the way to go.
I find straightforward the alignment of the rail as paralel and square to the other axis (4 holes test), but I have no clue how to make them coplanar. I just cut the steel tubing tiny segments that hold the one upon which the rails are mounted, as equal as I could, then when only the rails were fited I adjusted the whole thing until I did not feel any resistance, but they might stil not be complanar, as is the linear bearing have quite a tolerance. Except with a construction level, I have no clue how to check that. Then mounted the ballscrew , added the coupler and a 14mm rod on the other side, and turned it with the tips of my finger the whole distance, while adjusting the up/down amd right/left screws.
It seemed ok at the moment.
It is failry easy the release the bolts and redo the calibration, but I don.t know what I would do differently.
Self correction, updated for today. Although the rails seemed a while back paralel enough, they are not. I have a 1-1.5 mm error on a 1.5 meters run. Anyway, I.ll take the y axis apart in the next days, and assuming you are in my shoes, and you plan to allign the y axis from scratch, in what manner would you do it? Where to start first? Paralel, perpendicular, coplanar? In what order? And how would you do it? The method, please. :)
The requirement is for the two rails to be absolutely parallel, both horizontally and vertically. To deal with the second of these first - I ensured that the two rails both sat on "perfectly" level surfaces. This meant that I could adjust each side separately because "level" is a measurable quantity. I actually used a precision level bought from eBay for the purpose. Probably a bit more sensitive (and therefore more difficult to use) than necessary, but it needs to be better than you can achieve with a builder's level. I had used epoxy on my rails, but did not achieve a perfectly level surface all the way along so i ended up using shims to compensate. With the aid of the precision level, this was not too difficult. Note that I have talked about leveling each side individually. I deliberately built in the capability of dealing with two rail surfaces that were not coplanar. Other people talk about using a bridge between rails to achieve coplanar surfaces which is a good idea if you can do it; I preferred to allow for a small error in height (about 1mm, measured with a builder's level).
The other requirement is that the rails are parallel in the horizontal plane. Actually, getting them parallel is the easy part. The difficult part is bolting down the first rail absolutely straight. I used a straightedge as best I could, although it was not the full length of the rail. However, for a woodworking machine I decided that this was good enough. If I figure out how I can do better, I can always adjust it again. Once the first rail is in place, you can use the gantry to set the position of the second rail. Lightly bolt it in place so that it can move sideways, set up the gantry, and slide it along. This will bring the second rail into alignment and you can tighten the bolts as you go. This is actually one of the methods described in the Hiwin catalogue, if you cannot machine accurate reference edges into the machine structure. When you have finished, you should be able to move the gantry the whole length of the rails with very little effort. Note that the bearings actually move more smoothly if they have some load on them. These profile rails and bearings are amazing. When I was setting up my machine, at one point I had the bearings fitted to one end of the gantry while I was adjusting the second rail. I tried to support the gantry weight while the second rail was not in place but I realised that I had forgotten to do this. The end of the gantry (probably roughly 20kg?) sagged by no more than a couple of millimetres, while it was held at one end only by two Hiwin bearing carriages.
I also spent some time aligning the ballnuts and ballscrews. I started by ensuring that the ballscrews when held in the ballnut were parallel to the rails both vertically and horizontally. Gantry in the middle of its travel and ballscrew screwed into the ballnut so that it was more-or-less balanced, and no bearings fitted. I could then measure the level of the ballscrew (spirit level) and distance from the frame (as the rail was pretty much parallel to the side of the frame). I then converted these measurements into the size of shims I needed behind the ballnut mounting plate to bring the ballscrew into alignment. Once this was done I could fit and adjust the bearings at each end by turning the ballscrew (I actually used a cordless drill to wind it backwards and forwards) to bring the ballnut as close as possible to the bearing, then tightening the bearing mounting plate fixings with the bearing positioned by the ballscrew. Repeat at each end, then test. It took a while to achieve all this, but it did work.
Hi Neale. Thank you for the comprehensive reply. It is so valuable, as I did think about this issue all day and could not come up with a method as yet. I was thinking of taking the time to build a large precise construction square, so I can figure out if the gantry is perpendicular with the y axis rails, but then, in order to get the most precise reading, I needed to have a tool that touches both x and y rails at the same time, and it is a quite a pain for me to build such a tool with enough accuracy, as there is a significant difference in height between this two sets of rails.
Regarding your post, ˇˇThe difficult part is bolting down the first rail absolutely straight.ˇˇ, before doing that, is it not necessary to set the y axis linear bearings/rails absolutely perpendicular to the x axis rails?
All holes for the bearings, rails, steel tubing, aluminium plates, are a bit larger than they were necesary, so sideways adjustement is not a problem, but I.m not sure if I should set the bearings in relation to the x axis first or the y axis rails and do the bearing afterwards.
A precision level cost here 100+ euro, and even if I.m tempted to discuss this option with Santa this year, I.m a bit concerned that, except for this job, I.ll never use it again. Thanks for the tip, though, I did not even know this tool exists until today.
I don't know if this would work or even what level of accuracy you would get, but what about a water level ? You could make one out of clear plastic tubing https://en.wikipedia.org/wiki/Water_level_(device) it is easy enough to make one that will do 4 places, just by adding in T's and more tubing. That way you could put your X rails coplanar, for a start.
Hi Rob. I had the same thought initially. A water level. Although old fashioned, for long distances, I do not know a better method today.
But still, the question remains, what do to first. Rails or bearings?
An easy way to check the coplanarity of the two rails is by stretching two diagonal strings and check if they touch in the middle. With thin fishing line and eventually compensating the thickness of the line for the one above you could get very good accuracy. When you have the ends coplanar then you can straighten the rails with a straight reference (the hardest part).
For the perpendicularity of the gantry to the x axis, draw with the cnc the corners of a rectangle, and measure the diagonals which should be equal. The larger the rectangle, higher the measured accuracy.
One more great ideea. Thank you paulus.v
Work from bottom to top i.e Frame, X rails, Gantry, Y rails, Spindle, then you are not so likely to be re-iterative with your adjustments.Quote:
Originally Posted by Radu_Andrei
Ok. Thank you.
Actually, getting the gantry square to the other rails is one of the easiest adjustments, and it was almost the last thing I did. I am assuming that you are able to adjust the gantry - in my case, I can loosen the bolts between the gantry and the plates which carry the bearings so that if I turn one ballscrew, I adjust the "squareness" of the gantry. So, get it as close to square as you can - I used a carpenter's square - and then cut a test piece. As someone has suggested, drill four holes on the corners of a square and measure the diagonals. I used the shanks of drills as pegs in a piece of MDF, having used the same size drill to make the holes. With a little bit of schoolboy trigonometry, you can then work out how much you need to turn just one of the ballscrews driving the gantry to bring it into square. Tighten bolts and check again. I wrote a short piece of gcode to drill the holes so that I could ensure that I always approached the holes from the same direction to remove backlash from the process.Quote:
Originally Posted by Radu_Andrei
My very final adjustment was to use the machine itself to skim the strips of wood I use as bed supports so the the bed was reasonably flat once I screwed down a sheet of plywood. Not good enough for precision metalwork, but this is a woodworking machine. I machine a smaller spoil board if I want more depth accuracy for a particular job.
Good luck!
Hi. Happy new year! I know it's february, but after holidays and furniture orders, I just managed to get back to the cnc project last week. Hope you are all well.
So I took the whole thing apart, removed the wooden frame it had underneath, bought some steel tubing and made a new support for the cnc.
I managed to get the rails coplanar, with the two strings method, and a cheap usb microscope.
I have used a straigh edge to do one rail on each axis, then by moving the gantry, got paralel the second. It took a while but it worked.
Today I got the y and x square with the 4 holes method. All methods, mentioned in this thread, so thank you everyone.
I only have left the ballscrews allignement.
Reassembling the cnc, I noticed there were some screws loose. It might be that I did not tighten them in the first place, or they came loose while the cnc functioned. Anyway, I used only bolts and plain washers, and it might be not good enough, so I intend to upgrade this with something else, before I get the cnc working again.
Doing a bit of reasearch, I found a few option, the best one being crazy expensive. Some nord lock washers. But still there are a few which might work. Tab washers, safety wire, and locktite (I do not like this idea, as I am not sure if I can unscrew the thing). It might not be a must to upgrade all the bolts, but some for sure, like the ballnut bracket ones.
What do you use in order to prevent screws from coming loose due to vibration? And is it a must to use any special type of bolts, or the cheap silver one will do?
I use tensile bolts with either spring or serrated washers. Where I need nut & bolt I use nyloc nuts.Quote:
Originally Posted by Radu_Andrei
No bolts have came loose.
Examples:
https://www.ebay.co.uk/itm/SHAKEPROO...qgg:rk:10:pf:0
https://www.ebay.co.uk/itm/M2-M2-5-M...OicKyTw5sf01Co
https://www.ebay.co.uk/itm/M4-M5-M6-...ID805KSc9mzpls
https://www.ebay.co.uk/itm/A4-STAINL...nSZw:rk:3:pf:1
Hi dazp1976. Thanks for the reply. In the end I did use plain and spring washers, as they were easily available. Not the cheap silver springs, as they go flat. I used tensible bolts for bearing cases and nut cases, but not for rails, as they are quite expensive. Will upgrade in the future if necessary. So problem solved for now. The cnc is working fine and did quite a few medium sized carving with it.
I ran the machine 5 hours straight at most until now, but there is a job that is on my mind, a 4x2.5 feet carving, which will be a table top with some glass over it. Anyway, the job will take apx 11 hours to carve, and
I have no clue for how long can this cnc machines function without a break. For a 5 hour cut, the x motor which does most of the work, gets quite hot, like you can hold your hand on it for 5 seconds max, but I did read in some previous posts that it is quite normal. It is nema 34 1200 oz.
What do you think? Should I break the job in two parts or do it in one go? Did anyone used the cnc for such a time or longer?
Hi.
In march I finally managed to get the thing going, and was super exited and did quite a few jobs with it, then for the last month and a half, I was stuck with some other things and did not touch the cnc, and last week had in mind to get back to it, only to notice that the pc (gigabyte ga-8vm800pmd-775-rh, ddr2, 533mhz) which worked perfect for some time, including 10 hour straight in some cases, now turns off without notice. I did not change anything to it.
The workshop is a bit dusty, but its enclosure is mostly sealed,and the little dust that gets in through the fan holes, I blow it away with the air compressor (from a distance and gently) every week.
I did not manage to install the windows on an ssd for this pc, so I just cloned the one it had (win xp, sp2? - very old, does not receive an antivirus) on its hdd, and changed the ram memory, and it worked fine. Until now. It just turns off, randomly. No error apears, nothing, sometimes a few times a days, or a few time per hour. Any ideas what could be?