EE Advice Needed NA

[Henning]

I have an Atlas shore power supply box. What this box does is takes whatever voltage & phase power in, rectifies it to 600VDC then inverts it to 208 60Hz 3 phase out to supply the ships power panel.

We blew out 2 capacitors on the output board. They were 250v 9000microF capacitors. Atlas shipped us capacitors for the input board which are 450v 3300microF capacitors (this was before the Aussie technician came to diagnose the box). With the blown caps, we couldn't fire up the box to get the diagnostic computer on it. Atlas said "Go ahead and put them on, they're at their limits, but should last "a while" (trying to pin down the definition of "a while" proved unfruitful) and you can get the diagnostic. I said, "Until I know what "a while" is, I don't want to use this box (we're on generators right now) because I don't want a capacitor to blow up and ruin some of these $5000 circuit boards they're next to". So my engineer called an electrical guy here and he said that safety wise for the capacitor we're better off with the 450v 3300uF capacitor, but we'll have dirtier power (no option of going to a 450v 9000uF capacitor due to space restrictions) which won't cause you problems unless you have a bunch of fiddly electronics. So I basically have 2 diametrically opposed versions of whether the capacitor will last or not, one coming from the manufacturer and one from the local electrical guru. Then I also have to figure, "I have a sh-t load of fiddly electronics.". How dirty is this power going to be? How much worse than generators could it be? I have the proper capacitors on their way, but shipping from the states takes a week.

In the mean time I'm left wondering what should I do about this. Should I leave the 450v ones in there (one person says they're stronger, the other says weaker) and if I do leave them and use the power, how bad will it be on over $100ks worth of electronics hardware?

Help.

 

[RJM62]

Has anyone thought of attaching a scope to the output to see how dirty it is?

-Rich

 

[Henning]

Has anyone thought of attaching a scope to the output to see how dirty it is?

-Rich

Don't have one, wouldn't know what to do with the info if I did, but I'll look into doing it. Question is though, how dirty can it be?

 

[jhausch]

Here is my take on this, I can't speak to the actual "strengths" or cleanliness of one over the other, nor the likelihood of a catastrpohic failure, but I work with VFDs so do have some experience until you get an EE to give you some better info.

First - Going against the mfr may cause other issues, so having the agreement of the people that send you the spare parts is important.

If you can get them to agree (in an email) to the caps recommended by the local, that would be great.

Regarding dirtier power, does the output of the inverter go through a sine filter or load reactor? If so, that would clean things up a bit.

 

[jhausch]

Regarding your electronics - what voltages are the stuff you are are worried about? If you are worried about the nav stuff, isn't that all running on DC? If so, no worries about dirty AC. If it is some high end AV stuff; well, then it depends on the quality of the power supplies inside those boxes. You may find they are quite tolerant, too. Frankly, I'd be most worried about lighting and climate systems because those tend to have the cheapest power supplies for their electronics.

 

[RJM62]

Don't have one, wouldn't know what to do with the info if I did, but I'll look into doing it. Question is though, how dirty can it be?

I was trained as an electronics tech, not an EE; so I really don't want to speculate on that. An EE would do the math. I would break out a scope.

Basically, DC is flat line at the correct voltage, and AC is a perfect sine wave with the peaks and troughs at the correct +/- voltages, with no noise, in the best of worlds. In the real world, not so much.

A real crude approach would be to see which power source produces what seems closer to a perfect tracing, and use that one. A classier approach would require an EE's math or the on-site presence of a tech to look at the thing and test the output.

A possible stopgap to protect your equipment could be to use a power-conditioning UPS between the power source and the equipment, provided you can find one of the right voltage and sufficient capacity for your ship's system at your present port of call.

-Rich

EDIT: Looking back at your original post, I personally haven't seen a power-conditioning UPS for that voltage. I suppose someone makes them, but I also suppose that if so, they're pretty pricey.

 

[nyoung]

Caps are all about storing energy and smoothing a voltage rail. In a typical circuit, they resist voltage swings by storing or sourcing current.

As an aside, when thinking about analog electronics, I like using plumbing analogies. IMHO - caps are a lot like a pressure tank sitting behind a well pump. The pressure tank gets charged up by the well, and then when the user turns on the faucet, the pressure tank provides the initial few gallons at a relatively constant water pressure. As the pressure depletes, a separate circuit will turn on the well pump. The well pump will inject water into the plumbing at a high pressure, which would cause a huge rise in the flow at the faucet. Fortunately, the now (somewhat) empty pressure tank helps to absorb some of the initial pressure wave of the well pump by absorbing extra water flow. In analog circuits, output smoothing caps are used in almost the exact same manner to maintain a consistent voltage (pressure) by absorbing/sourcing current (water flow).

Anyway, I agree with the electrical guy you consulted, it probably is ok to use the lower value caps in the circuit, but that will result in a lot more ripple in whatever power circuit these caps were filtering. It is really tough to say how much, and without having a schematic of the circuit, I cannot say how much risk is involved.

If they sent you enough of the 3300uF caps you could wire them into the circuit in parallel (capacitance adds in parallel). 9900uF is within 10% of 9000uF and would filter as good as the original requirement.

If the original circuit was designed with 250V caps, then using 450V caps should be a non-issue. Going the other way (250V caps in a 450V circuit) would matter.

I have an Atlas shore power supply box. What this box does is takes whatever voltage & phase power in, rectifies it to 600VDC then inverts it to 208 60Hz 3 phase out to supply the ships power panel.

We blew out 2 capacitors on the output board. They were 250v 9000microF capacitors. Atlas shipped us capacitors for the input board which are 450v 3300microF capacitors (this was before the Aussie technician came to diagnose the box). With the blown caps, we couldn't fire up the box to get the diagnostic computer on it. Atlas said "Go ahead and put them on, they're at their limits, but should last "a while" (trying to pin down the definition of "a while" proved unfruitful) and you can get the diagnostic. I said, "Until I know what "a while" is, I don't want to use this box (we're on generators right now) because I don't want a capacitor to blow up and ruin some of these $5000 circuit boards they're next to". So my engineer called an electrical guy here and he said that safety wise for the capacitor we're better off with the 450v 3300uF capacitor, but we'll have dirtier power (no option of going to a 450v 9000uF capacitor due to space restrictions) which won't cause you problems unless you have a bunch of fiddly electronics. So I basically have 2 diametrically opposed versions of whether the capacitor will last or not, one coming from the manufacturer and one from the local electrical guru. Then I also have to figure, "I have a sh-t load of fiddly electronics.". How dirty is this power going to be? How much worse than generators could it be? I have the proper capacitors on their way, but shipping from the states takes a week.

In the mean time I'm left wondering what should I do about this. Should I leave the 450v ones in there (one person says they're stronger, the other says weaker) and if I do leave them and use the power, how bad will it be on over $100ks worth of electronics hardware?

Help.

 

[gismo]

I think Nathan covered it pretty well but IMO there's little or no risk using the smaller value caps. For the caps themselves there are two issues, voltage and current. The smaller replacements have a higher voltage rating so that's definitely not an issue here. And the lower capacitance will actually reduce the current through the caps so as long as they were designed to handle similar currents (you may need low ESR caps for this kind of application) the smaller caps should be fine WRT current as well.

As to the effect of increased ripple on the circuitry downstream I wouldn't be too concerned. First of all worst case you'd only see about a 50% increase in ripple and chances are very good that that's within the original design margins. In addition, properly designed power electronics will not self destruct as a result of excess ripple and I would expect that the engineers who created this $100k product were fairly competent or there would be all sorts of problems.

One thing that's not clear from your post is whether the caps in question are filtering the rectified 50/60Hz line voltage although this seems likely from your description of the equipment's purpose. If that is the case I'd definitely give the caps a try. I would attempt to check their temperature after a few minutes of operation and if they were heating up I might abort. What's confusing me is the bit about converting shore power to 600 VDC. A 450v rated cap (let alone a 250 v one) wouldn't work there. What is the actual RMS AC voltage of the shore power you're connected to and what is the maximum allowable input voltage of the inverter?

 

[Palmpilot]

What's confusing me is the bit about converting shore power to 600 VDC. A 450v rated cap (let alone a 250 v one) wouldn't work there.

Good point. Without seeing a schematic it's hard to give an intelligent opinion.

 

[river_rat]

Good point. Without seeing a schematic it's hard to give an intelligent opinion.

He did say they were on the "output board" so that would be only be 208VAC. I only do 3 volt or less ones and zeros so I don't know if those caps could be used to filter the output.

For reading in your free time.
http://www.freepatentsonline.com/5804953.pdf
I skimmed it but I don't think there is anything in there that helps with this question.

I would be more interested to know why both caps blew and if there is some other issue that is going to make the new ones blow out soon after they are installed.

 

[EppyGA]

He did say they were on the "output board" so that would be only be 208VAC. I only do 3 volt or less ones and zeros so I don't know if those caps could be used to filter the output.

For reading in your free time.
http://www.freepatentsonline.com/5804953.pdf
I skimmed it but I don't think there is anything in there that helps with this question.

I would be more interested to know why both caps blew and if there is some other issue that is going to make the new ones blow out soon after they are installed.

http://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=1215451&userType=inst

 

[Henning]

One thing that's not clear from your post is whether the caps in question are filtering the rectified 50/60Hz line voltage although this seems likely from your description of the equipment's purpose. If that is the case I'd definitely give the caps a try. I would attempt to check their temperature after a few minutes of operation and if they were heating up I might abort. What's confusing me is the bit about converting shore power to 600 VDC. A 450v rated cap (let alone a 250 v one) wouldn't work there. What is the actual RMS AC voltage of the shore power you're connected to and what is the maximum allowable input voltage of the inverter?

I am not positive if the the capacitors are pre or post IGTs, but they will be seeing 208v either DC or inverted AC at 60 hz. I haven't really looked for a max input voltage, but typically what we get on docks is limited to 480 volt 50 Hz 3 phase or less. Why do they rectify it and step it up to 600 volt? I haven't the foggiest. I know the purpose of the box, to convert any input power into an output of 208 60 hz 3 phase, that's all I know. Why they did things how they did them, I don't know, I haven't talked to the guy who designed the box. As for why they blew, the tech said they are 3-5 year items depending on use, and they've seen a lot of use since install in 2007.

 

[Palmpilot]

He did say they were on the "output board" so that would be only be 208VAC. I only do 3 volt or less ones and zeros so I don't know if those caps could be used to filter the output.

For reading in your free time.
http://www.freepatentsonline.com/5804953.pdf
I skimmed it but I don't think there is anything in there that helps with this question.

I would be more interested to know why both caps blew and if there is some other issue that is going to make the new ones blow out soon after they are installed.

The patent shows a harmonic filter (Item 58 in Figures 2 and 2A) on the output, and the capacitors in question could be part of that. A harmonic filter is needed because the way the output AC is generated could cause the output to have strong harmonic components that would seriously distort the waveform from the desired sinusoidal shape.

My guess is that the higher voltage rating of the substitute capacitors is fine, but the smaller capacitance would probably raise the cutoff frequency of the filter enough to seriously distort the output waveform. Whether this would confuse the synchronizer or the voltage regulator (Items 60 and 66a in Figure 2A) enough to cause damage to the power supply itself is something I couldn't even guess at. I would be reluctant to do it without a written commitment from whoever is suggesting this that they would repair any damage that might result.

And I would be VERY reluctant to use the supply to power over $100,000 of electronics with the substitute capacitors, since I don't know how forgiving those electronics are of harmonics in the AC power.

If the capacitors in question are being used somewhere else than the harmonic filter, then I couldn't even guess as to the effects of the substitution.

 

[David]

I am not positive if the the capacitors are pre or post IGTs, but they will be seeing 208v either DC or inverted AC at 60 hz. I haven't really looked for a max input voltage, but typically what we get on docks is limited to 480 volt 50 Hz 3 phase or less. Why do they rectify it and step it up to 600 volt? I haven't the foggiest. I know the purpose of the box, to convert any input power into an output of 208 60 hz 3 phase, that's all I know. Why they did things how they did them, I don't know, I haven't talked to the guy who designed the box. As for why they blew, the tech said they are 3-5 year items depending on use, and they've seen a lot of use since install in 2007.

Henning,
The voltage changes, up or down, are done with a transformer on the AC side - not DC. It's easy and cheap to convert AC voltages with transformers. It looks like they just want a standard voltage for the VFD so they step the voltage up to what they want to input to it - transformer T1. For the output, in your case 208V-3phase, they step down the output of the VFD - transformer T2. The other two transformers T3 and T4 look like CPT's - control power transformers.

 

[jhausch]

Henning,
The voltage changes, up or down, are done with a transformer on the AC side - not DC. It's easy and cheap to convert AC voltages with transformers. It looks like they just want a standard voltage for the VFD so they step the voltage up to what they want to input to it - transformer T1. For the output, in your case 208V-3phase, they step down the output of the VFD - transformer T2. The other two transformers T3 and T4 look like CPT's - control power transformers.

Don't you suspect the VFD is putting out a PWM to simulate the voltage and freq desired and T2 is probably just a load reactor to smooth out the sine wave?

 

[Tom-D]

I have the proper capacitors on their way, but shipping from the states takes a week.
Help.

What ever happened to any where in the world over night?

 

[Skip Miller]

What ever happened to any where in the world over night?

You lose a day crossing the date line, and so it is "over two nights" to Sidney. Now... off to Cairns? Not the end of the world, but you can see it from there.....

-Skip

 

[Henning]

What ever happened to any where in the world over night?

No worries about it getting in country over night, but Fed Ex doesn't control customs. It usually take a week to get things from the time they land to the time they make it to the boat.

 

[gismo]

No worries about it getting in country over night, but Fed Ex doesn't control customs. It usually take a week to get things from the time they land to the time they make it to the boat.

Screw customs. Have the shipper attach a small chute and do an air drop to you on the high seas.

 

[Henning]

Screw customs. Have the shipper attach a small chute and do an air drop to you on the high seas.

Sure, in plain view of NSA satellites so they can track me and inform my next port who will treat me like a terrorist or drug smuggler....

 

[RJM62]

Sure, in plain view of NSA satellites so they can track me and inform my next port who will treat me like a terrorist or drug smuggler....

Unfortunately, that's not even far-fetched.

 

[Henning]

Unfortunately, that's not even far-fetched.

Dude, you have no idea how realistic it actually is....