Machsupport Forum
Mach Discussion => General Mach Discussion => Topic started by: simpson36 on November 19, 2008, 06:36:23 PM
-
I need some advice or direction on the correct proceedure for grounding my CNC conversion.
There seems to be a lot of potential for problems if the mill is grounded to the computer to the breakout board to the 5v source to the stepper power supply, etc , etc.
For example, what goes to earth ground?
-
I am certainly no expert on the subject, hopefully Jeff Birt will look in as he seems to know what he is talking about with grounding procedures. Heres my take on it and so far following this I have never had any issues.
All of your shielding wires go to ground(earth) and only one end of them seems to be the preferred method. All of your ground wires should go to ground at one point if possible. You should not confuse a 0V with a ground and you should never use a Ground as a 0v.
Hood
-
Another thing I should add is if at all possible try and use 24V on all of your limits, home switches, external buttons etc, it will mean you need to convert them to 5V for interfacing with the PC but it will help eliminate noise in these wires as they are usually long runs and often pass high voltage wires or potentially noisy equipment such as VFD's. All of the I/O on my Lathe and Mill are 24V and the reason 24V is better is the difference between Hi and Low is much greater where 5v it is much closer so a slight amount of noise can be seen as a signal.
Hood
-
The most important thing is to HAVE a ground, and only one. The body of the machine should be securely bonded, through a dedicated wire, to the power ground of your building. In the US, this is the GROUND terminal present on any self-respecting outlet. You want to have a single ground for the electronics, and connect everything to this point. This can be a buss-bar, heavy terminal strip, a metal chassis, or something on that order - something with essentially zero resistance between the many connections. NEVER ever daisy-chain grounds. For shielded cables, ground the shield at only one end, usually the power source end. So, for example, encoder cables should have the shield grounded at the end opposite the encoder, where the power supply is. NEVER use a shield as a power or signal return. Always provide a dedicated power or signal return wire, or you render the shield largely useless. You never want to setup a situation where you can have current flowing through a shield. You never want to have a situation where a signal wire can serve as a power return.
Regards,
Ray L.
-
I copied most of the following from another post of mine on the forum:
OK, 'grounding'...'(earth)ground IS NOT equal to DC common. They do not mean the same thing, unfortunately we all throw around the term 'ground' very loosely. 'Ground' is this context refers to 'Earth Ground' which is a safety device. The incoming mains voltage is referenced to ground (through the 'ground' rod(s)) and so is your equipment. This provides a path of least resistance to shunt the voltage away from important things like people in case something goes wrong. 'Ground' SHOULD NEVER carry any current, it IS NOT a common return path for all circuits. That is the job of DC common, and your system may have more than one DC common which is not a big deal. Generally speaking you should not tie your DC common(s) to (earth) ground anywhere. Some power supplies may do this internally with a small resistor to pull the DC common up off the ground plane noise.
The problem is that folks think 'ground' is this universal reference for EVERYTHING in a system and try to measure voltage to it from any given point, which is wrong. Think of it this way, if I were to nail three pieces of wood together at odd angles and ask you to measure their length how would you do it? Would you pick the bottom of the closet piece of wood and measure from there to every other piece? Or, from the floor (ground) to each piece? Nope, because that would not tell you a thing. You would run your measuring tape from end to end on each piece of wood (so your reference is the beginning of each piece of wood and you are measuring from that reference to the end of the wood to find its length.) Measuring voltages is the same idea, you are measuring from a reference point (common) to some other point in the same circuit (same piece of wood).
It is very important that you tie your machines (Earth)ground to your mains earth gound, even if you have a local ground rod. This prevents nasty shocks from things being (earth)grounded at diffrent physical locations. A differnece of 100' could create a 50V potential difference.
-
This is how I do it for UK use. We have Live(L), Neutral(N) and Earth(E). This Earth is either a local earth rod or Supplier provided earth. All domestic systems have an earth leakage detection system fitted, an RCD - residual current device (detector) typically 30mA with a 40ms tripping time so that any mains fault to earth breaks the mains circuit. This is also called an earth leakage circuit breaker (ELCB)
If I need say a 12 volts DC, L and N are fed to the primary of an isolating transformer, the case or laminations of which are connected to E. The secondary voltage is fed to a rectifier and smoothing capacitor and the 0 volts line has a star point to which all 0 volt connections in this 12 Volt circuit are returned. The star point is in turn connected to E. This is to stop the secondary voltage floating above E, for example, the secondary could be at 200/212 volts above earth and the circuit would still work but would be an electrical hazard. All machines, computers, cases etc. have a direct connection to E.
This system has served me well in many control circuit installations and is very robust, with no earth induced problems.
I should point out that I am not a qualified electrical engineer, only a Fitter/Turner apprenticed in the days when it was a requirement that safety in the use of electricity should be understood. So, you cannot take any of the above as a recommendation for use in your circuits. Any qualified quys out there prepared to comment?
Regards,
Ian
-
Thanks for the info, guys.
I was looking for a little more specific 'how-to' for CNC rather than general rules. You know "Grounding for Dummies" type of help.
I am adding a fourth Gecko 203V to my control box right now and I want to know that I have the grounding set up correctly. Gecko says to ground their driver frames and I assume that means to earth ground, even though the are not at mains and have a 5v side. I have them all mounted on a .090" aluminum plate as a sink and I plan to attach that plate to earth ground via the unused ground wire in a 16/3 power cable coming from the 36vDC power supply's earth ground terminal which goes directly to the earth ground center prong on the wall plug. I know the ground is good in the outlet. Gecko says to feed each drive separately all the way back to the power supply so I've added the necessary cables to do that.
Also the system has a minaric variable speed DC drive which is installed in an separate aluminum box along with it's associated relays, pot and a CNC interface board that (supposedly) produced the control voltage for the drive to follow. The box is fed 115vac from the wall outlet and also 12v for the relays from a two wire 12v outside source and 5v which is supplied via the breakout board in the control box directly to the CNC speed controller. hot and common from the 115VAC goes to the board terminals and the green ground wire is bolted to the chassis and these run in a 16/3 cable to a grounded outlet.
So far, with the exception of the Gecko's, which are isolated, I have kept the 12v and 5v stuff off of the earth ground as I'm imagining that a short to ground on the 115 side would fry any 12v or 5v (including the laptop) that was on the same ground. I don't think this is correct and I'm wondering what the correct procedure is to ground the laptop and the breakout board, or even if they should be grounded. The geckos are all opto isolated, and the breakout is opto isolated on everything except the specific gecko connections (it is a gecko specific board).
So I take it that all of the 'big' stuff just goes to earth ground individually. My remaining question is how exactly do I ground the 'electronics' side. There is no terminal or screw that says 'this one goes to earth ground' and that the level I need help on, unfortunately.
-
As has been said below as long as you do not use earth as a return supply line and you connect all 0 volt lines to a star point which is earthed, you should be OK
I would suggest that you leave the motor speed control in manual mode until you actually need Mach to control the speed. There are a lot of posts on this site discussing speed control - do a search. Most DC speed controller speed pots are not isolated from the mains supply and you need a good breakout board to make sure that the control voltage to the pot is safely isolated, ask your BOB supplier, most like CNC4PC have diagrams on their website to show how connections should be made.
You mention a laptop; most people use a standard cased PC for Mach because they have found that laptops have issues involving power saving problems and on board video card interrupts that can cause Mach to seem unreliable.
Regards,
Ian
-
Thanks Ian,
I read all of the posts carefully. I have all the power grounds done and tested. I put a ohm meter between the machine frame and the ground in the actual wall outlet and got 100%. Did same for control box and spindle control box (both aluminum). So that's done.
The 0 volt thing is what I am still unclear on. There are no grounding terminals that I can find and no instructions on grounding for the electronics that I can find.
It occured to me that the BOB is probably already grounded as it is mounted via aluminum standoffs to an aluminum plate which is grounded to earth. But I don't know if the mounting pads on the board are connected to ground. There is no separte specific ground wire between any of the electronics boards that I know of.
As to the laptop, I already had the issue with power saver feature. On this particular laptop, it is defeatable in the bios. I've done that and also I run it off wall current thru the charger, so it is not succeptable to the battery going low. It is an el cheapo laptop that I bought only to use with tuning and diagnostic software for an '82 Vette and I don't particularly care if it gets ruined by chips and oil.
I'm just in the process of building my first CNC. I'll put a permanent PC in place one everything is final and I have appropriate protection (including proper grounding) for it.
The spindle controller is already connected and working. It just does not provide a very consistent speed and I have not looked into that yet.
As to 'isolation'. There was only passing mention of it in both the BOB and the DC drive docs. I had no idea what 'isolation' was and consequently fried two DC drives and the BOB. After I fried the first DC drive, I explained in detail to a Minaric rep what had happened and he sold me a replacement unisolated drive . . which of course also fried and again took the BOB with it. Minaric made good by selling me new isolated drives for only the difference in retail price, so I can't complain too much.
But now you see why I am a little paranoid now about grounding.
-
and you connect all 0 volt lines to a star point which is earthed,
This is absolutely wrong (not to step on anybody's toes).
DC common (i.e. OV) is not the same as earth ground. You should NOT connect your DC commons to earth ground. As I have mentioned before many (most) switch type DC power supplies will bond their DC common to ground internally through a small resistor. This pulls the DC common up off the low level noise a 100mV or so. If you run around and bond the DC common to ground you have defeated this and will cause yourself problems.
Anything marked with an earth ground symbol should be bonded to a single point (star ground), you should run a large cable from your star ground to your main ground and an external ground rod (if you have one), any shield braids should be grounded at one end only (usually the controller end).
-
I appreciate the replies, but I still don;t think I have an answer.
I get the chassis grounding and protecting people from high voltage, etc.
I get that the DC common is 0v and should not be connected to earth ground. All of the 0v (common) wires from the Geckos go to com terminals on the BOB.
I have a 5v (to the BOB) and 12v (fans and relays) supply coming from an old PC power supply. Each is only two wires. I am not taking any power from the PC and it is connected by a shielded, all wires connected parallel cable.
Here is the question as succinctly as I can put it:
Are there any earth (or otherwise) grounding requirements that must be done on the electronics side i.e BOB, indexer board, charge pump, spindle controller, PC, etc that I need to specifically address?
Sorry to be a pest, I just don't want to fry anything else due to my lack of electrical knowledge.
-
I appreciate the replies, but I still don;t think I have an answer.
I get the chassis grounding and protecting people from high voltage, etc.
I get that the DC common is 0v and should not be connected to earth ground. All of the 0v (common) wires from the Geckos go to com terminals on the BOB.
I have a 5v (to the BOB) and 12v (fans and relays) supply coming from an old PC power supply. Each is only two wires. I am not taking any power from the PC and it is connected by a shielded, all wires connected parallel cable.
Here is the question as succinctly as I can put it:
Are there any earth (or otherwise) grounding requirements that must be done on the electronics side i.e BOB, indexer board, charge pump, spindle controller, PC, etc that I need to specifically address?
Sorry to be a pest, I just don't want to fry anything else due to my lack of electrical knowledge.
Just as for the power ground, all your power supply commons should be connected together at a single point. You are probably accomplishing this, sort of, through the BOB, but that's really not ideal. You should bring the commons from all supplies to a single point, and bring a dedicated wire from each device (BOB, drives, etc.) to that common point as well. You can determine whether the mounting lugs on the BOB are tied to common by using an ohmeter. Your supply common may, or may not, be bonded to the power ground. I would suggest not. Since it sounds like everything is mounted to a metal plate, that metal plate should be bonded to the supply common.
Regards,
Ray L.
-
Thanks Ray,
Based on the advice I got here, the metal plate that the electronics are mounted on now has a separate wire running directly to earth ground via the ground pin on the wall outlet.
The only shielded wire I have is to the indexer sensor and it's shield is connected to common on the controller board. It came wired that way.
So any boards that are grounded thru their mounting holes will be grounded. Everything is working fine, so I guess I'm good to go.
-
Simpson36,
That's good news! Now the debugging starts. I usually put a screen over the stepper driver wires and earth them only in the control box at the star earth point. This is because the signals in the wires are relativly high power and high frequency and can cause odd induced voltages in other control wires.
Most DC SCR speed controls have their speed pots connected to the neutral line, so unless you have opto isolation in the control or the BOB you are going to have the problems that you had.(Put it down to the learning curve.)
By the way I like your simple 4th axis solution in the other thread.
I hope to clarify earthing the 0 volts star point with Jeff see my later post.
Regards,
Iaqn
-
Jeff,
I wrote :-
"and you connect all 0 volt lines to a star point which is earthed," and I gave my reasons for doing this and regard this as an essential safety requirement.
You wrote :-
This is absolutely wrong (not to step on anybody's toes).
I am always interested in improving my knowledge, can you give your reasons for making this statement please? ( Even at my age I am still light on toes ;D, I might be able to keep them out of the way!)
Regards,
Ian
-
I am always interested in improving my knowledge, can you give your reasons for making this statement please? ( Even at my age I am still light on toes ;D, I might be able to keep them out of the way!)
Ian, I think Jeff explained the reasons in his last post. He gave one example with the
switching power supply common.
Scott
-
Ian, my primary purpose for the warning is to get folks to stop and consider what they are doing before wiring everything that we typically call 'ground' up to earth ground. Switch type power supplies are a prime example, for instance see: http://www.astrodyne.com/smartcat/pdf/AD155drl.pdf, notice the PS has both ground and DC common terminals. Internally they are most likely bonded with a small resistance to 'pull' the DC common up off the ground plane noise a bit. externally bonding defeats this (and most probably the way the switching regulator was designed to operate.) The 'proper' way of bonding DC common to earth ground is largely dependant upon the design of your system (and the components therein). Most folks will get a bit concerned as with a floating DC common you cannot measure from chassis ground to DC components and they interpret this as a problem; of course the only problem is that they are trying to measure a voltage by using the improper reference point.
Here some good references that briefly cover some of the issues involved in proper grounding:
http://ecmweb.com/mag/electric_resolving_grounding_issues/
http://www.ese.upenn.edu/rca/instruments/misctutorials/Ground/grd.html
While it is common (he, he...electrical joke) to earth ground DC common some thought needs to be put into doing so. Consider how your DC power supplies are wired ( is the DC common bonded internally to earth ground?), and how it might effect isolating your controls from noisy equipment, such as VFDs.
-
Jeff,
(With appologies to Simpson36 for hijacking his thread)
I think we have some common ground in this discussion. :) and I thank you for your considered reply.
I agree with you that the case should never be considered as a reference point to measure circuit volts - a hangover from the days of early radio and television where the chassis was used as a common signal return path. Much confusion exists with regard to using the terms ground and earth interchangeably. One even sees the earth symbol used in automotive applications!
Your references were very interesting but I do not see the statement that under no circumstances should the 0 volt line be connected to earth. My problem is your adament statement that connecting the low voltage DC 0 volt line to earth is "absolutely wrong." Your third reference Figs.7 and 9 actually show this connection. My guru in this aspect of circuit design was a M.I.E.E and as far as I was concerned his word was law.
Practical experience, designing and building switch mode power supplies, a touch sensitve electronic piano and power audio amplifiers, taught me the advantages of star earth points and earthing the 0 volt star point from several internal power sources, but the overriding point is safety, the transformed, rectified voltage should never be allowed to float up and this is avoided by connecting it to earth. Floating secondaries are allowable where the equipment is double insulated or, as one of your references stated, the circuit is designed so that finger contact is impossible, not really practical with the circuits we construct to drive our CNC machines.
Bernard Babani publish many books in the UK for the amateur circuit builder and one their most prolific authors is R.A.Penfold. (Never sure if he actually writes all of the books attributed to him!) in his "Power Supply Projects", without exception, all transformer DC 0 volt lines are connected to earth.
You are quite correct in emphasizing the importance of correct ciruit design and the isolation of control signals from noisy wires like VFD's and even microstepping power lines to stepper motors. Metal cases and braided tubular copper shielding for control wires should be considered essential to minimise these problems.
You raised some valid points and interesting references but we really could do with an Authority or Electrical Regulation that would give people reading this thread the confidence to build circuits that are safe. (Not me, I am only a impecunious machinist!)
Ian
-
I whole heartedly agree with Ian on the subject of grounding all systems to a common earth point, after implementing industrial PC based systems for over 25yrs using this procedure with positive results, I am not ready to change now.
Not so much the safety aspect, with low voltage systems, but the avoidance of spurious noise problems.
Siemens also has an excellent publication on their site on Equi-potential bonding in control system installations.
Max.
-
Hood, you mention that you convert 24v to 5v for interfacing with a pc. How exactly are you doing this?
Thanks
Rimmel
-
Previously I have used various methods from relays to optos. Nowadays I use CS-Labs controllers on my machines and they have 24v I/O.
Hood
-
Got a simple diagram handy by any chance?
Thanks
-
No diagram but easy enough, for example a relay with a 24v coil would be used for Inputs to Mach, the coil would be powered via a limit switch and when broken the contacts would open (NO contacts on relay) and the contacts side would be connected to the BOB just like you would any other input.
For outputs it would be the opposite, the coil would be 5v and the contacts would have the 24v.
Hood
-
Ah yes I see - I prefer the solid state/opto route. Which opto's are suitable, I can find many but many seem to have a minimum 24v switching capacity.
-
Never used Solid state relays, cant recall the optos I used I am afraid.
Hood
-
Opto LTV847
-
Brilliant thanks - any idea how to wire it for 24v control?
thanks
-
I see SSR's with triggering voltages of 3-32v often.
-
I see SSR's with triggering voltages of 3-32v often.
Do you also see dead people :-)
-
Brilliant thanks - any idea how to wire it for 24v control?
thanks
Actually, yes I do.
What I do not have is patience for questions asked in layers with zero effort in between.
Good luck with your project.
-
Ah so a knob-head then. Why suggest anything at all in the first place??? - forums are better off without people like you.
-
If you want a bank of I/O optos, look at the Opto22 racks, they have anywhere from 4-48 opto boards with various termination types and also 5v or 24vdc control as well as all degree's of AC/DC voltage inputs.
Plentiful and cheap on ebay.
N.
-
If you want a bank of I/O optos, look at the Opto22 racks, they have anywhere from 4-48 opto boards with various termination types and also 5v or 24vdc control as well as all degree's of AC/DC voltage inputs.
Plentiful and cheap on ebay.
N.
Thank you very much.
-
I do like the look of the LTV847 though as it has 4 opto relays on it. The anode/cathode/collector/emitter parts is easy enough to understand, just struggling with how to get the 24v down to 5v for the input. Looking at the datasheet the max is 6v input. Found a couple of diagrams but they seem to be 5v circuits using a 470ohm resister.
-
http://lmgtfy.com/?q=how+to+use+optocoupler+with+24v
Hood
-
OK I have been researching and come up with a diagram, but before I connect it could someone have a look and tell me if I'm anywhere near? (don't want to blow my breakout board up).
(http://www.freerabbitcontrol.co.uk/images/opto_01.jpg)
The forward current should be limited to 40mA@24v. Now this is where I have no idea - will the LTV-847 be ok at that? I don't understand the info in the datasheet
http://www.alldatasheet.com/datasheet-pdf/pdf/86746/LITEON/LTV847.html (http://www.alldatasheet.com/datasheet-pdf/pdf/86746/LITEON/LTV847.html)
Then again I've read to use a 3.3k resister and don't need a diode???? ???
thanks
-
Apparently a replacing the resistors with 580ohm will give me 20mA@24v
-
Built it, tested it, working great. Used 590ohm resistors in the end which gives me around 20mA@24v.
Thanks to all (well nearly all) for the help. Mucho mucho appreciated.
Rim
-
There is a lot of difference between inexperience and laziness. 'People like me' have a great deal of patience for inexperience. We also are often the inexperienced ourselves as no one person is expert at everything. Laziness is a different animal.
Since you have put in some effort, I will fill in some blanks for you. You have already figured out that you do not actually 'convert' 24V to 5V, so no need to cover that.
You do not need a diode unless you are driving a load that will create a surge that needs to be dumped. Mechanical relays have coils which generate surge when they are shut off and the field collapses.
You will encounter diagrams that show a device identified only as a 'relay' with a diode installed around it. This is typically not necessary with solid state devices.
The IMPOTANT current spec to design to is the MAX load on the DRIVE I/O and to be aware that some drives list max PER pin and some list a TOTAL for all pins. Obviously, do not exceed either. At 20ma, undoubtedly you are below the max on your drive, but be aware of the total if that is how your drive is rated. The LTV can handle enough current to fry most drives, so just limiting enough to protect the chip is not enough. Your choice is very conservative so you are good to go.
I use a 400 ohm power resistor on my Mitsu interfaces. You may want to consider the type of resistor you are using as it will probably get very hot. Power resistors can handle the current indefinitely, provided you space them up a bit off the PC board.
Take the 5V + and the 5V ground off the BOB. Do not use the PC power supply or any floating source for the 5V signal source.
Check the BOB inputs for the presence of pull-up or pull-down resistors and choose your active high or active low appropriately. This info will be in the documentation that came with your BOB. Often you can choose the state with jumpers. You can wire the 5V side of the LTV for either.
If that last statement is not clear, just call 1-800-KnobHead and I'll see if anyone can help you . . . :)
-
I copied most of the following from another post of mine on the forum:
OK, 'grounding'...'(earth)ground IS NOT equal to DC common. They do not mean the same thing, unfortunately we all throw around the term 'ground' very loosely. 'Ground' is this context refers to 'Earth Ground' which is a safety device. The incoming mains voltage is referenced to ground (through the 'ground' rod(s)) and so is your equipment. This provides a path of least resistance to shunt the voltage away from important things like people in case something goes wrong. 'Ground' SHOULD NEVER carry any current, it IS NOT a common return path for all circuits. That is the job of DC common, and your system may have more than one DC common which is not a big deal. Generally speaking you should not tie your DC common(s) to (earth) ground anywhere. Some power supplies may do this internally with a small resistor to pull the DC common up off the ground plane noise.
The problem is that folks think 'ground' is this universal reference for EVERYTHING in a system and try to measure voltage to it from any given point, which is wrong. Think of it this way, if I were to nail three pieces of wood together at odd angles and ask you to measure their length how would you do it? Would you pick the bottom of the closet piece of wood and measure from there to every other piece? Or, from the floor (ground) to each piece? Nope, because that would not tell you a thing. You would run your measuring tape from end to end on each piece of wood (so your reference is the beginning of each piece of wood and you are measuring from that reference to the end of the wood to find its length.) Measuring voltages is the same idea, you are measuring from a reference point (common) to some other point in the same circuit (same piece of wood).
It is very important that you tie your machines (Earth)ground to your mains earth gound, even if you have a local ground rod. This prevents nasty shocks from things being (earth)grounded at diffrent physical locations. A differnece of 100' could create a 50V potential difference.
Hi Jeff. I realise that this is a very old thread but it's highly relevant to the issue I have. I have just bought a brand new CNC router from Pacific Tooling (via Roger Webb). The router has arrived (from China) and mostly looks ok. I haven't fired it up yet for a few reasons but one reason is that I have noticed that there is no earth from the mains power to the chassis of the machine. I believe that I can fix this very easily by running a suitable earth wire from the rear of the mains socket on the machine to any part of the machine chassis firmly fastened of course.
Having read the above thread, I would just like some reassurance that this cannot affect any of the DC components or cause any interference. I would have assumed that the factory would have done the earth as a routine and necessary part of assembly. The fact that they haven't makes me wonder if there is an appropriate reason for the absence of a mains earth.
Your thoughts would be much appreciated
-
Just a thought, but regardless of whether it affects the machine or not, ANY fixed or portable electrical equipment with exposed metal surfaces should be securely grounded using a conductor capable of carrying at least the full current of the protective device fitted to the supply.
Now, if that affects the machine, there is more work to do because it shouldn't.
China has very lax attitudes to grounding - both my 3kw water cooled spindle motors were not grounded to the 4th connector pin, my chinese laser cutter was not grounded either.
-
Just a thought, but regardless of whether it affects the machine or not, ANY fixed or portable electrical equipment with exposed metal surfaces should be securely grounded using a conductor capable of carrying at least the full current of the protective device fitted to the supply.
Now, if that affects the machine, there is more work to do because it shouldn't.
China has very lax attitudes to grounding - both my 3kw water cooled spindle motors were not grounded to the 4th connector pin, my chinese laser cutter was not grounded either.
Thanks Dave, I agree.
-
Ian wrote (long ago):
""and you connect all 0 volt lines to a star point which is earthed,"
I suspect this has been misunderstood by a few people. What it means is that you should connect all the 0v lines from power supplies (and other things if needed) to a single point (eg a brass bolt on a large aluminium plate supporting the drivers and interfaces), AND that plate should be earthed with a separate wire back to the mains input.
That's my understanding of the SAA Wiring Regs in Australia anyhow.
Cheers
Roger
-
Thanks Roger :)