Reading through the forums there is a great number of engines that have finially died through the years due to water seal failure.
Robbie\'s post on another thread reminded me to ask this question :
Quote from: RobbieRX-7Mine died at 49,000 miles. Water seal failure. After chatting to carl he made a good point about imports being started from cold at the docks, moved a few meters then switched off. Fuel sitting in the combustion chamber soaks into the water seals causing premature failure.
How come another material is never used for the waterseals? Could Copper not be used instead as this could stop seal failure for the above quote? (as well as possably other areas?)
I often wondered about this. We use copper for sealing cylinder covers on the steam loco I work on (full size, i\'m not talking models here!). The covers and cylinder deck both have grooves machined into them (not unlike the rotary) to accept the seals.
To make a seal, all we do is get some small (3mm) round copper rod, bend it to the right radius, cut to the right length, then lap the joint so that rather than being a straight butt joint, the ends are at 45 degress to the long axis of the rod, and when the cover is bolted down it squeezes them together.
Anyone know what the cylinder pressures get to in a rotary? We\'re working at 250 psi on the loco.
Or even better, has anyone tried this? It\'s not an easy job, and takes a bit of patience, but anyone with a little skill should be able to do it, and the seal you get is extremely hard wearing.
The only problem I can see is that the ali rotor housings may not be hard enough to squish the copper rather than distorting a bit themselves. However, provided the right grade of copper is used, and its annealed properly, it ought to be ok. Anyone got some scrap housings and plates we could use to try it out? ;)
If I had some scrap parts to play around with I\'d fit a copper seal, torque the thing up, then section it with a big band saw from the outside to the centre, to make it into cake-like slices, with each of the bolts running through the centre of each slice. The seal could be inspected all the way round, and even hopefully through the joint in the copper, to see how well it squishes into its groove and makes contact with the plates and housings. Slices could then be dismantled to see what the effect on the housings and plates might be.
Purely hypothetical, but I think it would be an interesting experement to carry out. I seem to recall that ali Rover V8s modified by TVR had copper rings that sat in machined grooves in the cylinder head to make a sturdy seal, but I could be completely wrong. I\'ll have to rake out the book I think I remember reading this in (tuning Rover V8s or something like that) to confirm, but if that is the case it would be a good pointer that it may work in another application.
Opinions?
might find they were wills rings in the rover V8, same as most highly tuned hillman imps, head is machined to take wills rings which are gas filled and expand with engine heat to seal, no head gasket is used only an outer gasket around the water jacket
There are quite a few alternative water-seals on offer in the US of A. When Kerim rebuilt my engine, I opted for Teflon coated seals as these are in theory much less prone to heat and corrosion.
if you need any other bits mate i have loads everything! :)
Interesting, so at least in theory it is possable :mmmm
I think age is a big factor too. So many heat cycles. My water seals were brittle as hell when i took em out. So was the stationary gear O ring, it broke into 3 pieces!
Indeed - this is hopefully something that using copper can also have an advantage I would hope?
Doh! Looks like I forgot to post my reply before I shut the PC down yesterday. Muppetry award for me.
Anyway, cheers for the offer of parts for experimentation! I\'ll get back to the relevan people when I\'m ready to set the whole thing up.
Robbie, Ricco - Sounds pretty bad. Copper shouldn\'t do that, though. I\'ve certainly never seen steam seals give up, even after the terrific abuse they get thermally and mechanically. Can\'t wait to see how it turns out in the rotary!
Red lines on the drawing represent saw cuts from outside to inside.
The idea behind this is to gain an idea of how the seal sits in the groove after the bolts are torqued up. This can be done with only one plate and a housing.
Another valuable experiment would be to use a full set of housings and plates, leaving the rotors, shaft etc out so you we can get inside and have a good look.
Ideally, assemble it with no seals, and torque up the bolts, get inside and measure the distance between the plates at various pre marked points. I\'m pretty sure I have and inside micrometer for this. If not, I\'ll buy one.
Then take the thing apart, make up seals, install them, reassemeble, torque up, and take the measurements again from the same points. If they\'re different, then there may be a problem with the seals not squishing correctly, but we\'ll cross that bridge when we come to it.
If all is well and good, the coolant jacket can be pressure tested by bolting on a water pump with filler cap, blocking off any other passages, and pumping the thing up to pressure with a tester. Could even be filled with hot water before as another test, but i\'ve no idea how I\'d heat that much water. :rolleyes: Maybe I could jury rig the cooling system from my land rover to run through the test rig! Lol
I\'ve got a feeling my motor is coming up for rebuilding, so if this looks even promising with scrap parts I\'ll do it for real. It\'ll be an interesting experiment!
m4drx, amoungst others - you still got some bits available? I\'ve got some time over the next couple of weeks, and I\'m just dying to try this out!
MMMMMMmmmmmmmmm....interesting reading boys.You have to consider that when the housing and plates meet,they are in acyual fact mated together.The copper ring would have to be off exact measurements to compress into the grooves(square),then we have the expanding and contracting feature,plus the blow by possibility...........:Typing Keep working on the theory guys........:bandit
Something like......
This....
I\'m sure you can all figure it out but i should maybe have explained those drawings a little better. They show a section through the plate and housing, dark grey for the iron plate, lighter grey for the ali housing, and orange for the copper. Number two shows what should, in theory, happen when the plates are forced together by tightening the bolts if the copper is properly sized.
Another factor is the chemical reaction between the copper and ally (oxidation etc )....?
What is the expansion rate of copper to ally and steel ? Once copper is formed that is it ?...............interested in the concept,but \'devils advocate\' as well :guinness
Quote from: oldoneWhat is the expansion rate of copper to ally and steel ? Once copper is formed that is it ?...............interested in the concept,but \'devils advocate\' as well :guinness
Expansion coefficients:
ali 22.5
cast iron (grey) 10.8
copper 16.5
rubber (hard) 77
Steel 13
Bolts are made of steel, plates are made of iron, housings made of ali, seals made of rubber. But what kind of rubber?
Now, a bit of dead reckoning...
Assumptions:
The ali housings make up 60-70% of the length between the bolt thread where it enters the front plate, and the bolt head clamping down on the back plate. We are left with two housings, and the middle and rear plates sandwiched between the front plate and bolt heads.
We can largely discount thermal expansion of the front plate, as it does so mostly outside the bolts ability to control its expansion. IE it should expand forward of the front housing.
We have three expansion coeffiecients londitudinally - bolts, housings and plates.
So, some more dead reckoning. Ali and iron in a 70/30 mix down the length of the motor should yield an expansion coeffiecient of around 17
Steel is 13, so in theory, the bits the bolts clamp together expand a little bit more than they do, thus creating more stretch in the bolts and greater clamping force. Great so far.
We\'re also smack on target for coppers expansion rate of 16.5. BUT, as the housings and plates are held tightly by the bolts and their own thermal expansion as we have already established, they can\'t be forced apart by the expansion of the copper, and it ought to make a better seal.
However, if the rubber seals normally used do have an expansion coefficient of 77, then we\'re far short of that with copper.
The biting question is; is this a problem? Or will the small amount of thermal expansion in copper be enough to make the seal, coupled with its relatively high strength and ability to be pressed into and stay in shape
Regarding reation between copper and ali, This crossed my mind too.
There are copper rings that seal things like fuel lines, sensors etc. all over most engines, so i\'m tempted to say that it shouldn\'t be an issue.
Sorry for the long post. Really interested to hear a few more opionions.
And thanks for making me think, Oldone!
One more thing guys to throw into the mix is the fact using dissimilar metals in a water based environment you may well get electrolysis. It could be quite embarrassing if the copper became a sacrificial anode. Its too many years since chemistry lessons and too early in the morning for working this one out so over to you guys.
Kevin
I thought that water seal failure was because of corrosion of the plates between the waterways and seal grooves, Rather than the seal failing?
Also when rebuilding my engine I used the the Rotary aviation seals which have a lot less give than Mazda seals(but more than copper). I was warned by a reputable engine builder that he\'d seen these seals break the land between seal groove and waterway.
Quote from: FD3UKI thought that water seal failure was because of corrosion of the plates between the waterways and seal grooves, Rather than the seal failing?
No, it\'s usually because the rubber becomes brittle and breaks down and flass into bits. I had a 3cm long piece of water seals missing from mine when we tore the engine down!
Quote from: kev-yorksOne more thing guys to throw into the mix is the fact using dissimilar metals in a water based environment you may well get electrolysis. It could be quite embarrassing if the copper became a sacrificial anode. Its too many years since chemistry lessons and too early in the morning for working this one out so over to you guys.
Kevin
Good point.
You can get copper head gaskets for a lot of engines which have an iron block and ali head, so I can\'t see this being a problem though.
how would you join each side of the copper togeter when they meet? so there would be no leeks? after putting it into the grove?
Quote from: FD3UKI thought that water seal failure was because of corrosion of the plates between the waterways and seal grooves, Rather than the seal failing?
This is true, but I would still expect the copper to last langer even if the plates have corroded?
Quote from: yamasterhow would you join each side of the copper togeter when they meet? so there would be no leeks? after putting it into the grove?
The tried and tested way of doing this is to use a kind of lap joint in the copper.
The (poor) illustration shows a section looking through the groove, with the plate on the bottom, and the housing on the top - pre torquing down. The copper is proud of the goove, ready to be squashed. The white line represents the lap joint which will get pressed together to form a tight seal once the bolts are tightened. This can be initially cut with a hacksaw, then finished with a file before the seal is inserted into the groove.
It should easily be possible to supply them in pre-cut, pre-jointed lengths with a former for bending each size around to make the correct shape, should we ever get that far!
Quote from: MartinyThis is true, but I would still expect the copper to last langer even if the plates have corroded?
This I\'m not too sure about, and it would be a difficult one to test. Copper should withstand an awful lot of punishment, and problems like nikp\'s where section are totally destroyed would be almost impossible. But, it\'s not as elastic as rubber, and doesn\'t expand so much, so it might not be able to fill a large hole left by a piece of plate that corrodes away.
However, how many engines fail through corroded plates rather than seals breaking? Is it a common occurance, and how many fail in this way with the proper mix of coolant in the system?