[GITRDUN]

One thing that I would do is putting a fan directly on the supply board and turned it on before power up, just to see it would help. Other thing is checking for the daisy chain of power lines to the machines, and rewiring them to star wiring.

htrantx

What is the purpose of putting a fan on the board before powering it up? Is it to cool the NTC resistors and create a slower power up time or something? I talked my maint man this morning about checking the NTC thermistors on the machines power supplys to make sure they are in working condition and he doesnt recall any of them having NTC thermistors on the power supplys.

[htrantx]

I talked my maint man this morning about checking the NTC thermistors on the machines power supplys to make sure they are in working condition and he doesnt recall any of them having NTC thermistors on the power supplys.

You have my sympathy, if your maintenance guy is correct. You are driving an old car, which has no seat belt. Your problem is staring at you in the face all the time… Inrush current kills.

Your guy probably has replaced the old cap with a low ESR cap, or the new board has the wrong cap. This is one time that too good a thing is bad.

htrantx

PS: You can buy main line NTC devices and put them at the main power lines as shown in the attached.

[RCaffin]

Most data loggers has a RMS input stage, and since the tail of the RMS filter is much longer than a cycle, you cannot see the true sagging magnitude of a glitch in milliseconds.

That may be true of antique data loggers, but it is not true of modern micro-processor driven mains data loggers at all. They have had special glitch detection circuits for 20 - 30 - or more years now. As GITRDUN mentioned in his posting.

Cheers
Roger

[GITRDUN]

Question. Can i put a 440V 20A power line filter on 208V? If yes, would it also be rated for say apx 40A on 208V since the voltage is half?

[Al_The_Man]

No, usually the components are rated for that particular current, the other issue is depending on what the nature of the filter, the voltage peak suppression components will only become effective at and after the 440v rating level.
Al.

[RCaffin]

Can i put a 440V 20A power line filter on 208V?

Yes.
But you can't put a 208 V rated unit on 440 V!

If yes, would it also be rated for say apx 40A on 208V since the voltage is half?

Not a chance! The current rating is separate from the voltage rating.

Peak suppression components: these are not normally found in that sort of Power Line Filter. Just several chokes and capacitiors.

Cheers

[Al_The_Man]

The Mitsubishi ones I used also uses VDR's or MOV's so if the voltage is halved, they are not so effective.
Also precludes using other than the correct maximum voltage.
Al.

[RCaffin]

Mitsubishi ones I used also uses VDR's or MOV's so if the voltage is halved, they are not so effective.

Ah - those are surge arrestors. I don't think the standard Power Line Filters include them, just a pi network of Ls and Cs.
Yes, I have used MOVs as well, but they can come later, if needed.

Cheers

[wizard]

Ah - those are surge arrestors. I don't think the standard Power Line Filters include them, just a pi network of Ls and Cs.
Yes, I have used MOVs as well, but they can come later, if needed.

Cheers

The big problem I see here is that one really needs to handle the problem at the source if at all possible. It isn't even clear to me that this problem is coming from outside the plant. If it was you would likely have the whole local community complaining and I've yet to see indications that the problem exists outside these plants.

[RCaffin]

The big problem I see here is that one really needs to handle the problem at the source if at all possible.

I agree with the concept, but my own experience has been that 'the source' is really 'all the rest of the community'. Cleaning up every single source of spikes and noise out there is simply impossible in practical terms.

It isn't even clear to me that this problem is coming from outside the plant. If it was you would likely have the whole local community complaining and I've yet to see indications that the problem exists outside these plants.

Well, not so fast. What does the average house have in the way of susceptible gear? Light globes? Ah, they just wear out. Motors (fridges, washing machines, etc)? They are largely immune. Computers? If anything would be susceptible it might be these, but ALL PCs have power supplies which have power line filters built in. Guess why? (It's in the Standard.) Little plug packs or wall warts? Those things are so inefficient or low power they wouldn't notice the spikes. Or if they do, one simply sighs and replaces them.

Yes, of course it is possible that some of the problems come from within the plant. Even probable. The best approach for handling this is to put a filter on every machine, plus a small filter at the input to the susceptible power supplies. Surge suppressors on distribution boards for power tools. Confine and protect.

In the ideal world the grid is a source of perfect sine wave power. Dream on. In the real world it is a nasty, noisy, spiky, variable source of unclean power at varying impedances, often highly reactive. Live with it, 'cause the alternative is worse.

Cheers
Roger

[wizard]

I agree with the concept, but my own experience has been that 'the source' is really 'all the rest of the community'. Cleaning up every single source of spikes and noise out there is simply impossible in practical terms.

I would agree that cleaning up every source of noise isn't possible. However the problems here are so frequent that one has to suspect that something internal to the plant is an issue. This is why I've asked about what else is happening in the plant at startup time.

Well, not so fast. What does the average house have in the way of susceptible gear? Light globes? Ah, they just wear out. Motors (fridges, washing machines, etc)? They are largely immune. Computers? If anything would be susceptible it might be these, but ALL PCs have power supplies which have power line filters built in. Guess why? (It's in the Standard.) Little plug packs or wall warts? Those things are so inefficient or low power they wouldn't notice the spikes. Or if they do, one simply sighs and replaces them.

Well that might be the case if it wasn't for the reported cases of cell phone and FM radio failures. The radio is one issue but the failure of a cell phone charger on the same day is more interesting. Most of these chargers have universal inputs permitting a wide range of AC voltages and are fairly rugged. If whatever was happening was coming through the public lines it would seem to be significant enough to impact the rest of the neighborhood. Honestly I don't know if that "neighborhood" is residential or industrial but the frequency of failures here would likely be noticeable by the rest of the neighborhood if it was a supply problem.

Yes, of course it is possible that some of the problems come from within the plant. Even probable. The best approach for handling this is to put a filter on every machine, plus a small filter at the input to the susceptible power supplies. Surge suppressors on distribution boards for power tools. Confine and protect.

We will have to disagree somewhat. I will support proper electrical installations which means in this case proper use of surge suppression in the machines controls. I can't even imagine control panels built by credible CNC suppliers that don't build this into their controls. However if you have a source of interference it really is best to try to locate it and neutralize the interference there.

In the ideal world the grid is a source of perfect sine wave power. Dream on. In the real world it is a nasty, noisy, spiky, variable source of unclean power at varying impedances, often highly reactive. Live with it, 'cause the alternative is worse.

Cheers
Roger

That may all be true but the one fact we do have here is that power line monitoring isn't finding anything extremely odd. Now the power line monitor could be crap, but it could also be a sign that the problem isn't AC supply related.

This problem is strange, I work in a rather largish plant and frankly can count on my hands the number of power supplies for the CNC controllers I've replaced over the years. Often the failures are rather well defined, such as worn out electrolytics. This includes power supplies with sensitive crowbar over voltage protection that takes very little to trigger.

This has to be one frustrating problem and frankly it will cost a fortune to throw a bunch of new hardware into the plant to try to address the problem.

[RCaffin]

Hi wizard

I don't think we are disagreeing all that much.

I can't even imagine control panels built by credible CNC suppliers that don't build this into their controls.

I agree, I was a bit surprised myself. But GITRDUN has a largish number of CNC machines and they do not seem to have the necessary mains input protection. Now, I can't guarantee that sticking large power line filters at the input to every machine will totally solve every main-related problem in the plant, but absent magic diagnostics I can say they will be an excellent start. If it turns out there are still a few problems, well, then one looks at them.

You mentioned dying electro's. Yeah, been there, and there could be a few of those still in his machines - stressed by previous events.

A 50 A 480 V PLF costs about $350. If fitting one of those to a machine only halves the number of failures per year it has to be a very cheap solution compared to the many costs inherent in failures, repairs and downtimes. A good 1st step.

Cheers

[neilw20]

I am highly suspicious that the supplies are failing at switch off. No amount of power logging will find that.
With many regulators, it is important that the output voltage does not exceed the input voltage.
This can occur, say in a 5v regulator, the 5V rail has some filter caps.
When the main supply goes off, and the 5v rail is above the now fallen input supply, the regulator absolute max specs are exceeded.

Solution: (and this is always included in well designed circuits)
On a positive regulator, a diode (anode) from output side to input side (cathode) will ensure that the regulator specs are not exceeded, and the output side is discharged.

On a negative regulator, a diode (cathode) from output side to input side (anode) will ensure that the regulator specs are not exceeded, and the output side is discharged.

Simply installing a diode, as above, won't be detrimental to normal operation, as the diode is reverse biased. No harm can come.


Are there any circuits of the failed supplies? Post here. Or just some photos of both sides of the failed boards.


Once the regulator has failed, it will (by design) become a short circuit to protect the electronics it supplies.
The output of a regulator, is usually a ZENER, a bit higher than the output spec and is designed to melt on overvoltage or reverse voltage, or possibly other failure mechanisms.

A CROW-BAR type circuit on the input side of the regulator will usually destroy the regulator if the diode is not included!!
Does a crowbar fire on shutdown? Quite possibly.

It is about time this ongoing saga is resolved!!!
Buy me a ticket, feed me, and I will come and fix it, for once and for ever.

Neil

refer to link..
http://www.cnczone.com/forums/genera...ml#post1407862

[RCaffin]

Hi Neil

This can occur, say in a 5v regulator, the 5V rail has some filter caps.
When the main supply goes off, and the 5v rail is above the now fallen input supply, the regulator absolute max specs are exceeded.

Yes, I know about that problem - have known for 30+ years. But I would be cautious about assuming it is what is causing all the problems here, for several reasons.

The first reason is that there are a lot of machines and they have been running for many years, with only occasional failures. If the circuit design was such that it was easy to reverse-bias the regulator, then I would have expected far more frequent failures, like once every few days, and that these would have been happening from day 1.

The second reason is that in most PS circuits the big filter cap before the regulator is usually much bigger than the capacitance after the regulator. You might have 2,200 uF before the regulator and a series of 1 uF and 0.1 uF caps after it. Abnd the circuit is always pulling the output of the regulator down. So this reverse bias problem is seldom seen in practice.

Third, I find it a little hard to believe that the manufacturer would not have been aware of this basic electronics problem. The machines come from Japan, not China. Commercial regulators usually include the reverse diode anyhow (these days). It's possible, but it's a bit of a stretch.

The next problem is the size of some of the power supplies. Some of the spindle supplies put out over 20 A at a fair voltage. These aren't toys, and the power supplies are usually rather rugged beasts. They have to be! And that is before we get to the ability of the supply to handle the reverse current when stopping a motor - oftem worse than powering up. That implies a reasonable amount of electronics engineering skill.

On the other, it may be that some of these machines were built without power line filters at the inputs, which does not look so good for the manufacturer's skills (or my reasoning). In this context it is worth noting that more recent machines which use COTS PC power supplies are not amongst those which have blown up. PC power supplies all do have power line filters at the power input: they have to in order to meet the PC industry Standards.

I still could be wrong. We would need to see circuit diagrams with details of what components failed.

Cheers
Roger

[GITRDUN]

Ok ive done alot of experimenting the last few months and i figure its time for an update.

Here is whats changed as far as power up break downs. October had at least 6 power up break downs, Nov had 8, Dec had 5, Jan had 2, Feb had 2, March had 4, April thus far had 2 identical machines blow the exact same components on the same cuircut board on the same day at power up after a severe lightning storm the night before.
This is what we have done to the machines. In January we added small EMI filters inside the machines on the power lines that feed the spindle drives control boards and also the lines that feed the servo drives control boards. The filters also have capacitors in them, may be insignificant but might not be. I also added small inline fuses to the power lines feeding our most troublesome board in our vertical mills. Am still waiting on another one of them to fail to see if the fuse pops or not but so far none of them have failed again yet.

One other note. We have 5 Haas vertical mills and have had lots of trouble with them randomly tripping the main Tx breaker when they are powered up in the morning. It seems to have only started maybe 6 months ago or so but it happens quite frequent. Last week i instructed all employees to open the control cabinet and check to see if the breaker is tripped before they switch on the machine in the morning. This morning 2 of the mills had the main Tx breaker tripped before the machine was switched on. So we have an AHA moment. Caught them in the act. Neil hit it on the head. At least for the mills they are tripping the main breaker when they are turned off at the end of the day. About 6 months ago is when we started pushing E-stop before turning the machines off so im betting that has something to do with it but who knows. At least we know when its occuring.

Ive bitten my tongue for quite awhile as i figured its just all coincidence and didnt want to jinx myself. But 6 months of logging every break down and every experiment i think is starting to pay off as the trend is getting better. I pray it stays that way. I have more experimenting to do to the ground wiring.

[neilw20]

I'm still looking for a trip.

[GITRDUN]

Well i think i can finally put this riddle to rest. Problems are all but gone.

Around mid Nov 2014 our problem just stopped. We went an entire 6 weeks with not one single problem. That had not happened in years. Since mid Nov we have only had 1 cuircut board blow. I heard through the grape vine that the local power company found a problem in our area and fixed it about the same time our problem went away. I have not personally talked to the POCO but a city administrator has spoken to them about our problem and they are denying anything and everything about finding a problem of course. I have heard from a local electrician that they found a bad neutral somewhere on the main line.

No way to know anything positively but all i know is things have returned to normal around here. Finally.

[RCaffin]

Hi GITRDUN

I have heard from a local electrician that they found a bad neutral somewhere on the main line.

Yeah, that would do it. You could get all sorts of random voltage glitches that way. Mind you, getting an official acknowledgment of the problem might be way beyond the bounds of probability.

I seem to remember saying, years ago, that the problem was coming down the line from the grid. Could even be true.

But leave those filters in place anyhow!

Cheers
Roger

[GITRDUN]

Just one last update to this nightmare that is now long over. Our problem seemed to disapear for a short period and then came back again in spades. As an absolute last resort we put isolation transformers on many of the machines that could not do without. Instantly the problem was gone again, so added more isolation tx's. Its now been about a full year and not a single glitch. Nothing. Problem is totally gone. Ive almost forgot what a daily headache it was to deal with all this mess and without spending a fortune on repairs business is back to doing well. I still do not know what the actual source of the problem was. There are lots of possibilitys.

[RCaffin]

Hi Gitrdun

I seem to remember saying, years ago, that the problem was coming down the line from the grid. Could even be true.
An isolation transformer is a wonderful thing. It will transmit the 60 Hz power just fine, but it will block most of the crap coming in from the grid. Combine it with a filter (as you have) and you should be smiling all the way to the bank. Of course the PoCo will deny everything - they don't want to be liable for the cost of all your repairs!

There is only one step beyond what you have: a ferroresonant isolation transformer. That adds voltage stabilisation to all the filtering. They are very heavy and not cheap at all. I use one for our network, initially because of all the mains-born problems we were having but also because I had a suitable unit for free.

Glad to hear of your success. Go make chips.

Cheers
Roger