Piston and Valve Rings, Packing, Seals, etc.

I figure this is worthy of its own thread.  Following work done by Porta, Wardale, and others, the general consensus on making steam engines steam-tight is to provide multiple sealing components in series.  For example 10 or 12 narrow piston rings rather than 2 wide ones.  The same logic applies to the packing for the piston and valve rods- use multiple packing elements in series.

Cast iron rings, sometimes in combination with alternating rings of cast bronze, seems to be the standard material for piston and valve rings.  Wardale experimented successfully using high-chromium cast iron to give better wear characteristics. 

An idea that intrigues me that might be worth trying on No. 110 is the use of self-lubricating materials.  Bill Petitjean proposed some years ago that it might be possible to apply such materials to saturated steam engines such that separate forced lubrication systems for the pistons and valves could be omitted.  I've done a little research and there do seem to be some off-the-shelf synthetic materials available that are suitable for the temperatures and pressures involved.  Bill is doing work with small, stationary steam engines suitable for use for home or light industrial power and may have done his own research on this.

Thoughts?

Hugh

I wonder if some of the Spilling oilless-engine approaches might be adaptable.

Graphite impregnated iron or hard bronze, perhaps......why don't we ask Bill Pettijean? Given he's now in the steam oil business, it's possible he didn't find the answers in self lubricating materials.

Dave

Again, if the engine is kept saturated the potential range of materials is greatly increased. 

It may be practical to hard-coat the bore and then choose a ring system to suit.  In theory it should be fairly simple to build a PVD system that can be rotated around an axis that can be translated, and mount it on a rod that runs in appropriate bearings (dummy head at the front, temporary bearing at the rear, good conduction insulation) that can hold the required vacuum and controlled atmosphere.  Block or plate off the openings with appropriate vacuum seal, and fabricate connections for the various power and chemical lines.  A university might assist with specialized equipment and, perhaps, with specialized materials and testing equipment.

In operation you'd pull a fairly high vacuum on the cylinder and outgas.  Heat the cyllinder (extermally) to a temperature that will keep compression on the coating at all operating temperatures (this should help with outgassing).  Coating is applied sequentially by rotating the head while translating it longitudinally to keep the coating relatively even in thickness.

Technically I suppose you could apply the coating to liners or bushings before installing them, which would greatly simplify much of the apparatus.  I do not know offhand whether special care would have to be taken if the liners are to be installed by chilling them -- would not be difficult to determine, perhaps even by reviewing the existing literature.

Note that reduction of the TYPES of oil fed by a lubrication system might be the goal here, not abolition of any oil in the lubricant at all.  This is perhaps massively simplified, again, by running the valves at no more than saturated temperature.  IIRC Bill has worked on this at Green Velvet or whatever the company is now, and can probably advise on materials and processing choices for single-lubricant design optimization without giving away the 'crown jewels' of actual oil chemistry.

One of the things I'm chasing down all the bits and pieces about I can is the bore and line thing.......and the accepted practice seems to be boring out the cylinders, making a liner to be a sweat fit or interference fit to the bored cylinder casting, then either sweating or forcing it in......the liners OD will be critical relative to the bore's ID, but the ID of the liner is generally finish bored to final dimension in place after installation - some form of screwed in plugs are run through cylinder and liner walls to prevent movement in service in most of the processes I've found out about. So, the specialized ID coating won't easily combine with that standard practice.

I'm certainly hoping to learn more...one Australian source seems to use steel liners rather than cast iron, and since I can't find a source of short cutoffs of the heavy CI pipe I used to find, other materials are of interest.

Dave

Here's the URL for Bill's Petitjean's current project/company, called "Practical Steam":

http://practicalsteam.com/Home_Page.php

A very good overview of the different treatments for various piston rings, lands, etc. is the Mahle "cylinder components' book -- ISBN 978-3-8348-0745-4.  This is nominally IC-engine specific, but extrapolation to wet environment and higher thermal cycling is not difficult to make once you understand the need to do so.

You do NOT want hard chrome on the working face of the ring (or -- horrors! -- in the liner bore}.  That's been tried, and the results were instructive (see the steam_tech archives for details).  But you DO want it in the lands and grooves, and on the mating pressure surface of the PLANE of the ring...

Graphite-loaded intermetallics or powder-metallurgy 'sintered' composites might have advantages in this system.

One thing to consider is whether 'one-sided' or double-sided sealing is going to be involved.  Getting a good seal on some piston-valve rings, for example, is going to be different from the situation on a DA power piston...for the latter, you might want flanking conformal rings on either side of a 'bearing' ring...

Or, www.steamenginelube.com for his custom blended lubricants - which are excellent.

Dave

We've touched on some broad ideas about materials without making any specific recommendations.....I think there's so much new stuff out there it's difficult to try to filter things out without a guide and a road map.

So, being no fun and generally miserable, I'm going to narrow the focus a bit: what can be done in a smaller railroad machine shop with an engine lathe, Bridgeport mill, and a couple fairly talented part time  manual machinists?

It's nice knowing what the possibilities are if there were no limitations, but even better to know what the available possibilities might be.

Dave

On to packings........

I've used various kinds of rope packings, 17 has them, but I'd expect we'd have some form of metallic packings on 110. I am familiar with the King and Paxton Mitchell varieties which were common on US railroads, and retrofit a Shay with a very nice packing from France Compressor that used stacks of interlocking segments enclosed by a girdle spring that D&S recommended, which was fine and had the ability to let water pass but not steam. Strasburg uses Teflon packings which they must lubricate at the gland to prevent the packing extruding out along the rod.

I think we've all seen the photo in Red Devil of the Wardale packing......certainly worthy for a high speed efficient mainline locomotive. Not sure it's applicable for 110 due to space and geometry limitations within the gland itself.

Ideas and other examples please.

Dave

Here is a site with illustrations of the different types of France packing.  (It also has some piston ring specific things). 

http://www.dghco-berlin.com/comprossor_02.htm

There are some notes on what geometries and materials work best in given circumstances.  Most of this stuff is NOT going to be fabricated with backshop tools or traditional techniques... which is not to say they can't be fabricated with relatively inexpensive, portable equipment, or great expense.  Here perhaps more than anywhere is an opportunity to 'amortize' quite a bit of the cost, and run up the learning curve both of thought and of experience, in what amounts to an open-shop situation.

Machining grooves and lands correctly is not adifficult exercise.  Hard-plating the lands, however, is going to be more involved.  The rings are made of one material, faced with at least one other material, their pressure face is different from their internally-bearing face, and some of them (the equivalent of 'one-piece' oil rings, for example) will have a complex cross-section.

I notice there is little reference to rider rings in the steam literature.  It seems clear to me that rather than having a great multiplicity of rings in an articulated valve body, it might make better sense to have much of the region between the sealing/oil rings made up as a single elastomer rider, perhaps even oversized in the bore.  (One old design of mine involved a helical-faced rider, essentially acting like a very long spiral piston ring made of non-wearing material backed with a solid hollow cylinder...)

Reading between the lines, I get the impression that the idea of "multiple diesel-style rings" inherently assumes that the rings involved will be primitively machined and made of the usual 'cost-effective' materials steam mechanics are used to seeing.  That is fine as far as it goes, but I think we need to look at where we can leverage current experience.  There is a very mature body of knowledge that is now developed around piston gas compressors that -- to put it gently -- offers much more.

It appears (as I suspected) that polyether ether ketone (PEEK) is the material of choice for self-lubricating rings.  I tried to find a reference for toe 'loaded' type of piston ring (where the physical structure of the ring is sintered metallic or intermetallic material and the PEEK or other polymer is incorporated between the grains at sintering time, like a solid version of an Oilite bearing) but have had no direct luck.  I know how I would make these (HIP sintering in vacuum conditions for the powders being part of it) and I suspect there are people out there who know also.

Note the differences between compressor and steam-locomotive practice that need to be considered.  

For the valves, the effective steam edge is going to be very hearly the ring edge.  This to me makes an L-shaped first sealing ring essential -- well, perhaps not as essential for 110 if she is only going to operate 15 mph, but as a design criterion for future practice.  I am now wondering whether a very thin elastomer 'rider' (really no more than a sealing ring, a few thou thick) could be incorporated between the two faces of a sectional ring, which would immediately eliminate any particular leakage between segments as well as providing a good 'lubricated' yet conformal plane to allow the rings to adjust mutually.  THAT you could make with simple tools.

I'm thinking that universities will have many of the tools (and the staff, and fabrication equipment) to make portable versions of CVD/PVD/HVOF coating equipment.  I suspect that one setup for, say, face-coating rings would serve all the steam-community demand ... install it at Strasburg or Wasatch, for example, and watch the fun start.

I believe Texas Platers Supply (one of my favorite advertisers in the Shotgun News) has all the necessary equipment and chemicals/supplies to do hard chroming.  Very little of the actual valve body needs to be hard-chromed, so the biggest expense is going to be a proper rectifier, and assuring the power to run the thing.  Again, a self-contained portable rig including testing facilities or specialized pre/post treatment tools would be a good resource for the steam community...

Just as a couple of reference notes:  see the 5AT Project sections on steam tightness:

http://5at.co.uk/index.php/modern-steam-2/principles-of-modern-steam/steam-tightness.html

and piston-valve considerations:

http://5at.co.uk/index.php/modern-steam-2/principles-of-modern-steam/valve-design.html

In light of what Dave was saying about packing that "let water pass but not steam" I wonder if Porta's comment about much of the signficant loss being in the form of condensate will apply to packing design and selection.

Dave, can you explain to me exactly why "lubricating" Teflon packing would be necessary to prevent it 'extruding out'?  And what kind of lubricant is involved here?  Filling little asperities and scratches in the rod that tend to move the soft PTFE along?  (Anyway, PTFE is probably little more the right answer for valve-rod packing than it is for rings...)

Thirdhand from Strasburg...I've never used it. They (recollection) have a drip cup oiler that drops onto the valve rod just outside of the gland.

This is years ago information gleaned at a TRAIN conference. They may have changed some things since.

Dave

Isn't Kelly Anderson involved with this project?

We can ask him and get a definitive and detailed answer... on RyPN, or directly, if not here.

I remember some discussions of Teflon on steam_tech years ago -- it might be time to float a new topic there. There was a great deal of (carefully-concealed) consternation involving Teflon in a couple of European locomotives, I think SLM/DLM, IIRC more or less around the time of the Dreadful Acoustic Howling problems.

I don't think Kelly knows we're aliive and doing this over here. He's great and his opinions would always be welcome. One of his machinists was here for a while....but I think he imagined something a bit more fast paced and with a funded experiental program well under way, and he's sort of lost interest. he had some good and large ideas that would have been worth exploring if he'd posted more about them.

I just noticed the 5AT has an (apparently) new page on improved steam tightness for steam locomotives. There might be some info of interest here:

http://5at.co.uk/index.php/modern-steam-2/principles-of-modern-steam/steam-tightness.html

Hugh

Didn't I quote this page earlier?

The issue of 'steam' losses as condensate mass (e.g. from film) was an interesting thing I hadn't thought of. Are there actual statistics that would give us the magnitude of such loss for built (and operated) examples?

You may have indeed quoted it previously, Robert. I was at their site looking for info on combustion air pre-heating and ran across it. I didn't recall seeing it before.

I'd be surprised if there are any statistics on the phenomona other than what Porta himself compiled.

Hugh

In reading this the "new" thing for me was the limiting of the location of leaks to the location of the ring gap, piston and cylinder wall interface, followed by the means for controlling it. It would all be new to anybody who hadn't read Porta's paper in the Camden book. Apart from increasing the efficiency of insulation on the valves and cylinders, I don't think there's a lot we can practically do to reduce wall condensation losses.

I recently posted specs for not only diesel quality rings, but actual diesel rings in sizes up to 15.5", which is almost the 15.75" we will have for bores after relining as currently envisioned and includes the 8" size for the valves. So, if we can figure out how to include as many as possible of these replacing the standard two, and then somehow index them such that the ring gaps don't line up, I think we'll have made some significant improvement.

Dave

I mentioned it previously that both Rolls Royce in the UK and Hudson in the US used "pinned" piston rings and claimed them to be superior to normal "floating" piston rings.  Here's a photo of a Hudson piston; you can see a roll-pin inserted through the top edge of the piston.  This is what holds the rings in position:



Here's a quick/crude drawing showing what the accompanying piston rings looked like:



The rings are installed with the "cutout" aligned with the ring pin in the piston. The rings are installed in alternating fashion so that the ring gaps don't align (admittedly the gaps on adjacent rings would be fairly close, but they're not aligned).  

IIRC, on the Red Devil's piston valve rings, Wardale pinned each ring individually, using a pin inserted perpendicular to the ring sealing face.  The only thing I don't like about this arrangement is it requires that you have a fairly large minimum ring gap, which may negate any advantage to avoiding aligned ring gaps.  

I do seem to recall that Porta said the rings should be installed with the gaps at the bottom of the cylinder (for pistons and/or valves without tail rods), as the piston outer diameter will be pressed against the inside of the bore at this point by gravity, such that ring gaps and alignment become largely unimportant.

One thing I want to check, carefully, is whether the references to diesel-STYLE rings in Wardale et al. mean actual OTS diesel rings - metallurgy, surface finish, corrosion resistance, tribology, etc.

I am concerned that the environment in an IC engine is so different.

I'm going to advocate again that we consider using circulated overcritical boiler water to heat the cylinder block (using some form of small pump or jet pump). This can probably be done with a tap from a Cunningham circulator (which eliminates any difficulties with small size of jet pump if only this circulation were provided. The principle is similar to tracer lines to prevent pipe freezing, but adapted to elevate the temperature of the cylinder block (with greater heat transfer to regions with higher relative mass, if necessary, so heatup stresses or potential distortion, e.g. of bores, will be minimized). These lines would be applied to the cylinder metal, probably with some kind of (graded) thermal-transfer materialto optimize contact heat transfer, and would then be covered by efficient lagging, then insulation, then jacketing. (I believe that commercial multiple-shield nanoinsulation is now commercially available, and while it is probably not as good as mine, it will be fine for the outer layer adjacent to the jacketing...)

Remember that Kurt Greske noted that only the .007" of metal adjacent to the bore cycles with the steam. It follows -- at least for me -- that if the mass of metal adjacent to that .007" is at initial steam temperature, there can be dramatically less gradient across that .007" while the steam is expanding against it. It is my opinion that this will cut down formal wall losses (and port losses) almost completely, and should improve tribology of the rings and seals as well.

It is at least theoretically possible to do something like this for a hollow piston rod and thence to the piston (via heat pipes). This can be done with a non-freezing heat transfer agent, circulated through coils or exchsngers adjacent to the boiler shell or arranged inside a9with appropriate safeguards for maintenance). Two flexlines, or a coaxial line, would link the exchanger to the crosshead. You could do the same for the valve rod, and for the valve bodies if not articulated or the heads if not Trofimov. I don't think these is as significant as the bores, but it helps to know there is an answer if it turns out to be significant.

Something like this, imho, is almost essential on a saturated engine that will be worked hard at times, as this one will be. It should also stave off deleterious effects of nucleate condensation much as superheat would, to some extent (and it would be interesting to see, empirically, what that extent would be).

Pinned rings might be useful, but much the same effect could be had by machining a 'wave' profile into part of the inner face of the ring (making it look a bit like a very shallow sort of Torx recess) and then machining a similar profile into the base of the groove. You would then index the ring at potentially many positions, it would self-align when installed, you would not have stress raisers at the pin, or fiddly weaknesses involved with locating and inserting multiple pins to do the indexing.

I can't imagine putting a ring gap at the bottom of a piston, where it will quickly fill with lubricant and any carryover or corrosion detritus. Imho the ring gap belongs at the top, where 'gravity' will compress the ring smoothly around the groove, and the gap is up where contaminants will be carried away.

Thanks Robert....I'm recalling my not understanding why running the steam through a cylinder jacket heater before admission and dealing with temperature and condensation losses there was better thanjust accepting the losses happening within the cylinder on steamTech many months ago. I finally figured out that it's the same situation as Watt moving the condenser out of the cylinder in Newcomen's atmospheric engine........where the loss takes place matters, not just that a loss takes place. Moving the loss away from the cylinder in which useful work is done is a good thing to do, if it's practically possible. I don't think we have the ability to install large chambers for steam path around the valve and cylinder castings, but a coil for a smaller stream of superheated water....there's possibilities there.

What don't know how to do is quantify it.

I don't see getting into the piston and rod at this point as being practical or unobtrusive. We'll just have to accept some things as given.

Dave

If you have to provide new sleeves anyway, you could "thread" the outside to provide a passage for the heating water.

OTOH, I recall reading something by Porta that said heating the cylinder barrel wasn't that important- he thought heating the cylinder heads was the place to concentrate effort. I can see where that'd make sense, since ALL the steam that enters the cylinder sweeps past the head and stands to lose much more heat there.

Heating the barrel IS less important than heating the ends ... but if we remember Stumpf, we'll remember that having uniformity of heating in all parts of the cylinder is valuable in making the geometry true and hence the tendency either to leak or to wear less pronounced. The fringe benefit is less wall condensation at the center, where the mean heat drop due to expansion is the greatest.

Remember we are not heating to keep the steam 'evaporated'; we are heating so that heat transfer has an initially high 'source' temperature, both to retard heat going out and maximize heat going in...

I had not thought of the idea of incorporating the passages in the bushings themselves. It would certainly work, especially if we were sleeving the effective bores down. The one difficulty might be if the locomotive were left in freezing condition with the lines undrained; in my approach only a few feet of tracer line would be damaged; in the 'threaded' approach, either the cylinder block or the liner could be cracked, or deformed enough to cause interference with moving parts or noncircularity of bore ... any of which are likely to be intensively expen$ive to fix even to an operable state.

That had occurred to me, too.......although i was captivated with the Red Devil pictures of something similar. The situation is, the cylinders are worn down almost to the counterbores on the bottom, and the original bores were just 1/2" D smaller than the counterbores. So, I'm not entirely happy with a sleeve just .250" thick, and Hugh and others pointed out that for adhesion purposes we could well make the bores a touch smaller....so if we go to a sleeve .500" thick we get down to 15.750 for the final bore, and everything works fine, plenty of meat for 2 or 3 bore jobs in the future. This goes away if we cut grooves into the OD of the liner.....

If we decide this is worth doing, outer piping between the casting and insulation is probably the practical option. We could try one side with it and one without, and take indicator diagrams of both to compare how much difference it makes on 17 before committing to 110. So Robert. can you send us your insulation?

Dave