Plumbing question:
I have a mix of copper and steel plumbing lines in my house. When we purchased it the guy doing the building inspection said this could cause problems with corrosion on the steel pipes/fittings. This in fact seems to be true in some areas (greenish-white scaling).
What is the proper way to deal with the connections between copper and steel? (Converting to all copper is probably not practical).
I would sincerely appreciate any thoughts.
Tips for picking the right paintbrush based on paint type, surface, and personal comfort.
"I have learned so much thanks to the searchable articles on the FHB website. I can confidently say that I expect to be a life-long subscriber." - M.K.
Dig into cutting-edge approaches and decades of proven solutions with total access to our experts and tradespeople.
Start Free Trial NowGet instant access to the latest developments in green building, research, and reports from the field.
Start Free Trial Now
Replies
Use a dielectric connector. It's a fitting with a plastic insulator in it that prevents direct contact between dissimilar metals. They're usually used at the top of wter heaters where the copper piping meets the steel body of the heater, or anywhere else two different metals meet. Your local plumbing supply house should have them.
Bob
I believe it's also permissible to use a brass nipple (6"? 8"?) between copper and galv pipe.
When laying new pipe runs, I like using 6" of brass to separate copper and steel pipe. Seems like you'd still still have a battery, looking at the electrical potentials of each metal, but the kinetics are such that the brass doesn't participate. A short brass nipple helps, but there is still a little corrosion between steel and copper an inch apart.
For a retrofit, I'd use the brass if there was a way to work it in (for instance, you are sweating a copper joint anyway). But you will likely need a union to finish the job, in which case, making it a di-electric union solves the whole problem.
Off the top of my head:
3/4" di-electric union maybe $6-8
6" brass nipple about $3-4
Either works fine. The dieletric union means some of your piping is not grounded, so consider if/where the electrician tied into the plumbing. The union allows for easier servicing / tie-ins later.
David Thomas Overlooking Cook Inlet in Kenai, Alaska
Looking at these photos (link below) might make one rethink the whole brass thing. I have seen problems in my travels, usually with brass valves for transitons as opposed to nipples. Not sure why.
Admittedly, they are trying to sell you dielectric fittings, and a lot of this appears to be for steam...
"Brass to steel problems far outweigh that of copper to steel. For the overwhelming majority of examples shown, a high corrosion condition exceeding 5 MPY was also found - greatly accelerating the galvanic effect."
http://www.corrview.com/tech_p_10.htm
Not sure why they would report, "Brass to steel problems far outweigh that of copper to steel" because it is contrary to my experience.
But 1) they commenting on high temperature systems and made the kinetics get active enough that brass is no longer unreactive the way it is in household piping.
2) as you point out, they are trying to sell di-electric unions.
And 3) in a more corrosive environment (high temp, more acidic, open system) no one is going to spec copper pipe anyway. So you'd generally only see brass/steel in bad situations, not copper/steel.
"I have seen problems in my travels, usually with brass valves for transitons as opposed to nipples. Not sure why."
Maybe because of the shorter valve length (1" to 2") compared to a 6" nipple. No one proposes to use 1" or close nipples for isolation. 6" seems to work well. In marine engineering, they definitely factor in distance (along with relative surface areas, reactiveness, and galvinic voltages from the salt water table).David Thomas Overlooking Cook Inlet in Kenai, Alaska
I'm a DIYer who ripped out the galvanized pipe in my old house. I'm also a chemical engineer who spec's materials for corrosion service, so I know a thing or two about galvanic corrosion. In short, though it may be better than a straight galv to copper connection, don't count on a brass nipple to save you. Brass is still principally copper, and the joint between the two has a galvanic potential of over a volt- that's a lot of driving force which goes straight into eating the less noble material (i.e. the galvanized pipe).
Although the main runs of galvanized pipe were completely intact, every brass or copper to galvanized pipe joint in my house was a wreck, inside and out, for at least three or four pipe diameters. Hot worse than cold, but there were plenty of cold water lines which were badly affected too. Internals half blocked with rust, externals covered in scale. The exteriors were even worse where a sweat solder to FNPT coupling was used for the copper to galvanized transition- the acid flux left over after the solder job went to town on the zinc on the galvanized pipe.
The best solution for domestic water is to get rid of the galvanized and replace it with copper or PEX. If that's not practical, the next best solution is galvanic unions at any galvanized to copper or brass transitions. If it's a house old enough to have galvanized pipe, make sure you inspect every piece of pipe for grounding wires- you'll need to disconnect these from the various parts of the piping system and reconnect any that are still in use to the main inlet piping that's buried in the ground (i.e. before the first galvanic union). Otherwise, not only will your wiring not be properly grounded, but you'll run the risk of secondary connections between your piping which reduce the effectiveness of your galvanic unions.
Industrially, lots of people use bronze valves on galvanized lines without galvanic unions- but that's generally for treated cooling water, loaded with corrosion inhibitors. It's bad practice even for those services IMHO.
Molten: We agree about carefully checking for electrial grounding connections. We agree that all copper or all PEX or copper/pex is better than a system with steel pipe in it and much better than mixing steel and copper.
We agree on the need to do something when joining steel to copper but disagree on the best thing. I suspect we also have a philosphical difference on the value of theoretical versus practical knowledge.
Early in my engineering career (I also studied chemical engineering, although I took my PE in Civil), I saw debates between engineers and technicians/mechanics. The engineers would have some clever new idea of how to plumb an air compressor aftercooler or about running compressed air in PVC because it was pressure-rated high enough. And the techs would say, "No, that's a bad idea." Eng: "Why?" Tech: "I don't know, it's just not done that way" or "that not how air compressor like to be treated" or some other seemingly subjective thing. Then the engineers would get all quantitative and fling around numbers and equations and "prove" they were right. The techs would do as they were told while muttering under their breath.
Later the air compressor would melt down or the someone would tap on the plastic pipe and send sharp shards flying around at 250 mph. A few of us engineers climbed down from our ivory towers and re-examined what we really know versus all the things we don't know. Emphasis on "few".
You and I have both seen the same charts on galvanic potential. You've torn out the plumbing of one house. I've plumbed about 80 houses for solar, built a few houses, replumbed a few others and constructed a few dozen toxic waste treatment systems. I think we're both newbies. I give a lot more weight to what Wet Head Warrior says or what a third-generation master plumber like Peter Hemp recommends than on my own imperfect knowledge and theory.
The frustrating thing is that there probably could be an all encompassing body of theoretical and practical knowledge that would show we are all correct. Taking into account the galvanic voltages, solubilities, kinetics, water temperature and hardness, corrosion inhibitors, age/condition/schedule of metal piping, QA/QC of dieletric unions, etc. And that would explain why, in some installations, dielectric unions perform better. And why, in many residential applications, 6" or 8" brass nipples do better.David Thomas Overlooking Cook Inlet in Kenai, Alaska
David:
I work in a design/build operation. I supervise the fabrication and installation of everything I design- and I work side by side with the tradesmen and technicians during installations. 90+% of our work is for repeat customers- and if we do something wrong, we hear about it! And my father is an auto mechanic. Few trades get better first-hand experience of the "failures of imagination" that engineers are so often prone to. No engineer should be able to get a job, IMHO, without the necessary educational experience of working with their own hands, side by side with tradesmen. Kids who go into the consulting business straight from engineering school without a lick of practical experience are the bane of our profession.
I definitely pay attention to what's commonly done and what works. The problem is that "what's done" and "what works" are subjective things. "What works" in one set of circumstances can be a disaster in other sets of circumstances- and "what's done" has to change over time to accomodate new materials and methods. A theoretical understanding of the underlying principles is what permits people to make informed judgments about what to do in new sets of circumstances. Are these judgments infallible? Hardly. But there'd be no technological progress if people were afraid to try new things based on their best understanding of how the world works.
Engineers who ignore the commonsense and hands-on experience of tradesmen deservedly suffer the consequences- and hopefully learn from them. I can think of many examples of tradesmen who extended a good engineering solution to a particular problem inappropriately to new circumstances without appreciating the underlying principles- and they and their customers have suffered the consequences. I also struggle daily with engineers and tradesmen both who are afraid to try new things, period- and who cost their customers huge amounts of money unnecessarily.
If WetHeadWarrior says that brass nipples work in domestic hot water service, that's great, and I'd trust the result. He's got lots of hands-on experience serving customers in his particular neck of the woods, and I'm sure his customers give him a blast whenever a choice he's made was the wrong one. It's far better than just jamming a copper NPT fitting onto a freshly-threaded piece of galvanized pipe, that's for certain. But I'm also sure that it's not protective in all waters.
I too have heard bad things about dielectric unions- mostly related to leaks which develop over time. I'm sure their performance depends on who made them and how they're installed- and that any union you install just for galvanic isolation is a potential leak point you've added to the system. But we've used them successfully in lots of services- and they are far better than a non-isolated joint between copper and galvanized pipe. Unlike the brass nipple, they provide complete electrical isolation between the two sides, eliminating the potential for galvanic corrosion rather than just slowing it down.
I also know that a galvanic couple in excess of a volt in water service is bad news- and that both brass AND copper can produce a nasty galvanic couple with galvanized. I know it from both hands-on experience AND my knowledge of corrosion chemistry. We agree that kinetics and the other factors you've discussed are of key importance too, and probably explain why the (long) brass nipples last. They can probably de-zincify for years before you'll have any trouble with them- and as long as the pipe remains intact and doesn't leak, the customer's happy and well-served by the choice.
We also agree on what amounts to my key rule in selection of materials for corrosion service: when you have a choice, it's best to choose a material which is IMMUNE to corrosion rather than one which is merely resistant. When the two material choices cost the same, or the immune one is cheaper, it's a no-brainer!
the details of this galvanic corrosion still escape me, but i am learning a lot from this particular thread. i still have a couple of questions some of you may straighten out for me.
if, as was warned earlier, by installing a dielectric coupling you break the continuity of a house's water pipe bonding with the electric ground system, then you would have to install a jumper over the connection to ensure the continuity. Since a jumper would also be copper, would that not simply negate what you just tried to solve by installing the dielectric union in the first place? Or does the water have to be in contact with both metals for the corrosion process to occur?
I am not a plumber nor an electrician, but I have yet to get a good answer from the plumbers and electricians that I have asked.
thanks
brad
Installing jumpers would defeat the purpose of the dielectric union. If an electric installation in your home uses the water pipes for ground, the solution would be to wire a different route to ground rather than using the piping.
my question about the jumper was more a general one, it is not for my house. i do remodeling and many older houses did indeed use the water piping as the ground for the electric system. my understanding is that this practice is no longer allowed, the electric system must be grounded with an approved rod and connectors. however, the metal piping system is still required to be bonded to the ground system, just in case. so, in remodeling, where old steel often meets new copper, it is required to have a jumper to ensure the continuity of the bonding. my electrician is always adamant about this. so, my question is which takes precedence here, avoidance of potential corrosion or avoidance of un-grounded piping? anyone have any ideas or am i fundamentally misunderstanding something here? thanks,
brad
Let's be as clear as possible: what you have to do is disconnect the ground wire from the pipe, splice a new piece of wire to it, and connect it with an approved clamp connector to the first piece of (usually copper) pipe coming into the house (i.e. right before the water meter, before the first galvanic union)- or to an approved grounding rod or grid. A jumper from pipe to pipe will defeat the protection of the galvanic union, and simply disconnecting the ground wire or leaving it connected with galvanic unions installed and no jumpers could get somebody fried.
thanks for the response, but i am still a unclear. (my wife often tells me that i am a little slow on the uptake, so it could be just me). what you describes sounds to me like the procedure for using the water service supply as the ground for a house's electric system. am i right? again, i am not an electrician so my knowledge of the code is somewhat sporadic. and, i am not asking this question for my house so i am not planning on performing any work based on info received here. my question in more general and just to let me understand this better.
according to my electrician, new construction, service upgrades, etc, can no longer simply be grounded to the water pipes, it must be done to a rod driven into the ground. whenever he does work for me on a remodel he drives a new rod with appropriate wire back to panel to ground the electrical system. however, as i understand it, the water pipes, being metal, must also be bonded/connected to the ground system in case there is ever a short or emergency somewhere in the system i.e. water heater short, whirlpool short, etc. since a short could happen theoretically anywhere in the water pipe system, having dielectric unions anywhere in the system could leave some sections of pipe isolated and therefore potentially "hot" in a failure. that is why i am wondering if jumpers or some equivalent are not required around a dielectric union even though therefore they would defeat the purpose of the union. often in remodeling when adding on a bathroom, replacing fixtures, etc, new copper is spliced into/attached to galvanized and it is not always possible to replace the whole house with copper.
i realize i have effectively hijacked this thread and turned it into an electrical question, oops. i throw myself at the mercy of the court.
thanks,
brad
You're right, ideally a new rod would be driven wherever a new ground is needed. But grounding to metal piping was widely done in the last 110 years and will continue to be done, to a lesser extent, because it is so easy.
Anyone dealing with steel to copper connection clearly has a house that had work done by different people in different decades. That increases the chances of an electric ground (or much worse - a current carrying electric neutral) being connected to random point(s) in the plumbing system.David Thomas Overlooking Cook Inlet in Kenai, Alaska
Grounding the plumbing should continue to be done so that if a hot wire ever comes in contact with it, the breaker will trip instantly rather than making all the faucets hot. I think it's a good idea to ground any large metal parts of the building that people might touch.
-- J.S.
Good point. There's a reason NOT to use dielectric unions! Adding a jumper defeats the purpose of the union, while not adding a jumper can possibly result in things taking a shorter path to ground through somebody's foot rather than through the pipe.
Not an electrician myself, so dunno if the codes require that every piece of plumbing piping be grounded. And items like electric water heaters etc. already require a separate connection into the wiring ground system, which as you've mentioned is best done via a ground rod or grid rather than the old way of using the building piping, so maybe the grounding of every piece of pipe is overkill. Again- the guys who live and breathe this code will know better.
"We also agree on what amounts to my key rule in selection of materials for corrosion service: when you have a choice, it's best to choose a material which is IMMUNE to corrosion rather than one which is merely resistant."
It is my experience that there is no material that is "immune" to corrosion. The ship that I work on has an inert gas generator with a salt water scrubber that's made of stainless, with all the salt water piping in the system stainless as well. Yet it still corrodes, it still fails.
Heat exhangers on a ship are usually made of bronze. Plate type heat exchangers are often made of titanium. Yet they all corrode over time. We install sacrificial zincs or swedish steel to protect them. It's usually cost prohibitive to fabricate all the piping out of brass, so we make it out of steel and expect it to fail eventually. To get 10-15 years out of a salt water pipe is good.
Marine engineer
Marine Engineer- we agree that all metals and alloys only resist corrosion to varying degrees, depending on what's in the water. Some are essentially immune under a given set of conditions, whereas others fail rapidly and predictably. However, most "plastic" piping materials are essentially totally immune to aqueous corrosion. Tough to build high temperature/pressure shell and tube heat exchangers out of plastics, granted- everything's got its limitations.
There are good "plastic" piping/tubing materials available for just about any aqueous services to temperatures beyond the boiling point- and there are plastic-lined metallic piping systems for situations where you want the mechanical damage-resistance of steel pipe.
So my point stands: when both are applicable and you can choose between the two, why would you choose corrosion resistance over corrosion immunity?
Sounds like we similarly value both the practical and theoretical aspects of our own work experiences and for engineers and techs in general.
It would be lovely if a discussion like this could bring together enough empirical data (e.g. when/where brass nipples work well versus not; whose dielectrics provide long service versus which ones leak. etc) to reconcile the theory with field experience in all cases. I don't have those specifics to offer, but I'm going to play more attention to the specifics of each leak and system I tear apart in the future in the hopes of figuring it out.
"I also struggle daily with engineers and tradesmen both who are afraid to try new things, period"
I thought of touching on that in my post as well. The one that stands out in my mind was when trying to discharge treated water through ABS. An Oakland, CA building department official said (in 1989), "we require cast iron becuase it's worked fine since 1910." Arrgh! He probably also burns books and PEX tubing.
David Thomas Overlooking Cook Inlet in Kenai, Alaska
why would a wire clamped to the dry outsides of two pipes (one steel/galv. and one copper) to maintain the continuity of an electrical ground have any effect whatsoever on their plastic insulated interface on the inside of the pipes that is exposed to an electrolytic solution (water)? in other words, if you jump the dielectric union with a ground wire, the steel and copper pipes still don't touch where it matters. yes, there will eventually be galvanic corrosion under the copper wire clamp on the steel pipe, but in a reasonably dry environment, not enough to ever matter. are you saying that a ground wire jump will actually cause the steel pipe to corrode on the inside? if so, at what point? right on the inside of the pipe where the clamp is attached? at or near the end closest to the copper pipe? the entire distance between the two? seems counter-intuitive to me but i'd like to hear your explanation. thanks.
m
I don't know if it is possible to predict exactly where it would fail, but the inside of the steel pipe will be giving up its electrons as the "battery" is formed. If a steel pipe is threaded into a brass valve, it will almost certainly fail at the threads.
If a brass flange is bolted to a steel flange with a rubber gasket in between, the bolts connect the two pipes electrically in a situation like you described with the dielectric union with jumpers. Here, you cannot predict where the steel pipe will fail; it is usually at a weld, or if the pipe has a bend in it, somewhere in the bend.
A pinhole will appear and I will put a soft patch on until I can renew the pipe. When the pipe is examined, it is always paper thin in the area of the pinhole.
Marine Engineer
i guess the concept is fairly straight forward, i just don't understand why there wouldn't be a natural, reasonably predictable path of electron flow- hence corrosion-in a simple arrangement like the ground wire example (assuming a homogenous, more or less amorphous steel pipe without welds, bends or other molecular stresses or irregularities). if nothing else, i'd expect the inside surface of the steel pipe to be relatively evenly corroded all the way around from the dielectric back to the ground clamp (assuming even contact on the clamp all the way around the pipe). perhaps moving the clamp well back of the joint- using a long ground wire- would sufficiently dissipate this effect over a large enough area to make it acceptable? does the direction of flow of the electrolyte solution affect it? for example, corrosion is worse if the steel is downstream from the copper in the water line? thanks.
m
I think if you had a perfect piece of steel pipe, it would corrode evenly. I imagine that at imperfections it corrodes easier, and once it starts in a place, it continues there more so than anywhere else.
If I could predict it, my job would be easier. Usually, the pipe is painted, and looks perfect until you see a brown rusty spot dripping, or if it is an out of the way pipe, until it starts to squirt. Then you have to deal with the leak in an emergency situation, for instance if it is the generator cooling overboard line, or if the valves don't hold.
Saw a rusty brown spot on the skin side of an overboard valve. Whadaya do? The ship isn't scheduled for drydock for another year and a half, drydocks cost millions of dollars, you can't just go in for the weekend. Yet if this lets go, the ship could sink. We built a wooden box around it, filled it with cement, and hoped for the best.
Marine Engineer
The resistance within the copper and steel pipes and the jumper is extremely low, so the exact location of the jumper attachments is pretty much irrelevant. The significant resistance is in the water, so you find most of the damage within a few inches of the jumpered union. It takes electrons flowing all the way around the loop to cause the problem.
-- J.S.
The Dilbert principle applies here: if the upside of choosing a new material or method is saving "the company", or worse still, "somebody else's company" or "the customer" millions of dollars, and the downside of making the wrong choice is losing your job because you screwed up, why would you EVER choose to do something new? That's why people stumble down the same path over and over, rather than thinking about it and trying something new. Our sales guy has a cartoon in his office: there's a king in a battlefield tent, and his knights in armour are trying to get his attention so that he'll come over to see what a salesman has to offer. The king says, "I've got a war to fight- I don't have any time to see any damned salesman!". The salesman stands proudly beside his product- a heavy-calibre machine gun...
As far as balancing the empirical data against the theory for brass nipples versus galvanic unions, that would be wonderful. But I doubt we'll sort this one out. There'd just be too many anecdotal reports of both of them failing, without sufficient analysis of why the failure happened.
On the Eng-Tips.com forum, there's a debate raging over whether or not copper is corroded to failure in deionized water. Some people list examples of DI and other water systems they've built which corroded to failure- but others point out that there are millions of miles of copper steam trace tubing in service, and what's condensate other than boiling "deionized" water? Though it's definitely true that if you deionized the water because the thing that's using it can't tolerate any copper ions, you shouldn't use copper- it's far from a certainty that copper is incompatible as a material for use in deionized water and can be expected to corrode through in a hurry...Every anecdote takes careful study and examination of potential causes to sort out the true cause of the failure. So we agree that this materials selection business is a tough one to figure out sometimes! One thing's certain, though- nobody's had a plastic piping deionized water system which was polluted with metal ions or corroded through...
"The Dilbert principle applies here: . . . . , why would you EVER choose to do something new?"
Long before Dilbert, it was observed that, "No one ever got fired for buying an IBM system." Maybe not the newest technology or fastest processor. But you won't be left hanging either.
"examples of DI and other water systems they've built which corroded to failure"
I've heard that about Sierra snow melt water in the Hetch Hetchy pipelines going to San Francisco. Such pure water that the corrosion rates are exceedingly high.
I'd already heard of that example when (in P-Chem?) the driving force for an ion to hop into solution was discussed and calculated. It is impressive how corrosive pure water is.
David Thomas Overlooking Cook Inlet in Kenai, Alaska
Edited 11/13/2003 8:48:28 PM ET by David Thomas
Nothing new about it, David- I'm just wondering why the copper ions don't hop vigorously into the millions of miles of copper steam tracing out there and eat it through in a few weeks- in practice, it lasts many, many years, and usually fails from the outside inward if corrosion is implicated at all- usually it fails from mechanical damage long before you have to worry about corrosion. Granted, condensate is generally neutral (before it returns to the boiler) and doesn't contain much dissolved oxygen, but it's otherwise pure, hot water.
I too have seen lots of anecdotes about copper systems, potable and purified, which have failed, but there's not enough analysis of each anecdote to permit a sensible, general conclusion about corrosion resistance of copper in "pure" water from them.
Corrosion resistance of metals has to do primarily with the integrity of the passive film, rather than the specific metal content of the solution passing through the pipe. If the passive film remains intact, the metal will stay put. Whether the water has zero or 1 ppm or copper in it is pretty much irrelevant to the stability of the passive film. Dissolved oxygen content, temperature, velocity, pH, or the concentration of ions which can disrupt the passive film are more important and can determine the difference between success and failure.
I see you beat me to the punch... I gotta stop being so long-winded. :-)>Kevin Halliburton
"I believe that architecture is a pragmatic art. To become art it must be built on a foundation of necessity." - I.M. Pei -
People have told me that about you....lol.
Bob
i've never heard of using a brass nipple as a proper substitute for a dielectric union. brass is mostly copper, too, but the zinc (or something) in it does reduce its tendency to corrode. anybody else heard of this?
m
Mitch: Peter Hemp discusses using brass nipples in his excellent book "Plumbing a House". Short story: it works.
Does this forum get some special allowance from Prospero that "nipple" doesn't get ****'d out?
David Thomas Overlooking Cook Inlet in Kenai, Alaska
Edited 11/10/2003 5:08:59 PM ET by David Thomas
You could start a thread: What do your favorite nipples look like?
you know, just to see how long it would last.
; )Any jackass can kick down a barn, but it takes a carpenter to build one.
OK, my wife's. No details beyond that.
On topic... the brass trick seems to work over the long term. Dielectric unions I find to be leak prone. I dislike them.
"You could start a thread: What do your favorite nipples look like?"
I was in a big box and the clerk had to call back for a price check. I told her it's a 3/4 x 2" nipple. She couldn't bring herself to say "2-inch nipple" over the store loudspeaker system. So she's prattling on, "It's a short pipe, threaded on both ends, 2 inches long, . . . "
On dielectric unions, brass nipples, and favorite nipples, I'll just say, ditto to WHW.David Thomas Overlooking Cook Inlet in Kenai, Alaska
funny you never mentioned your wife was in the plumbing supply biz before. is that how you two met?
m
I like my nipples close....that's not a mistake, it's rustic
My experience is that dielectric fitting are a temporary solution at best.
Having replumbed (galv to copper) our house in stages (over three years) , I installed dielectric fittings when switching materials. When I went back to remove the unions, there was significant corrosion. It almost looked like the union itself was trapping hard water deposits within the nipple, more so than the galvanized pipe. We live in a hard water area, so I am not sure how much that contributes to the problem, but I would never count on dielectrics for a long term solution. Certainly not overhead!
I've also had dielectric unions fail and corrode on a water heater. 8" brass nipples seem to work fine there, they corrode, but they last about as long as the W/H itself, so it's not a problem.
-- J.S.
Did your inspector indicate whether or not the copper and steel pipes were separated by dielectric unions? If they are, you really shouldn't have a problem, but if the copper and steel pipes do in fact connect directly to each other you need to break the potential for electro-galvanic reactions between the dissimilar metals with dielectric unions to prevent any further problems.
You should be able to see what a dielectric union looks like by inspecting the connection from your hot water heater to the copper supply lines going into the house. If the water heater was professionally installed I can virtually guarantee you will see a dielectric union at that connection.
Basically when two dissimilar metals (copper/steel) are placed in an acidic solution (most domestic water supplies are slightly acidic or slightly basic) you essentially create a battery that erodes one of the metals, leaving it as deposits on the other metal.
A dielectric union interrupts the electrical connection with plastic insulators preventing this battery like electro-galvanic erosion. Check all steel to copper connections for the presence of dielectric unions. If they are missing have them installed as soon as possible and you should have no further problems.
Kevin Halliburton
"I believe that architecture is a pragmatic art. To become art it must be built on a foundation of necessity." - I.M. Pei -
Dialectric or none, don't expect a lot of life from those galvanized pipes, particularly on the hot lines.
It has been many years since galvanized was last in fashion, and it is typically near 'end of life' at this point.
The stuff installed 30 years ago is outlasting the stuff from 10 years ago. Hmmmm....