The Steamboating Forum

Mike,

My only means of running a simulation is to CAD model the engine which is an enormous task and it requires hundreds of measurements from the engine or plans. I CAD modeled my entire 6A engine down to the last nut and bolt before I cut the first chip, this way I didn't waste a single minute making parts that were not going to be perfect and fit with everything else. The CAD model can be presented on screen with a section view cut through the model at any point and at any angle. The CAD model can then be rotated just as the real engine is rotated with all parts moving together in harmony. With the CAD model of my 6A I can change any dimension and see the resultant effect. For example I can change the eccentric timing, steam lap and exhaust lap on the valves etc.

In answer to your question, I am afraid that I don't have time to CAD model your entire engine. But you can easily make a study of your timing events by charting the timing at different shut off positions on the Stephenson's link. This can be a little challenging if you have a piston valve but it can be done and I am sure that you would find it to be an enlightening process.

Many people talk about the 6A blowing the valve off it's seat due to compression. I can say for a fact that this is caused by the HP eccentrics being too far advanced and the exhaust lap being too great in order to compensate. The shut off on the exhaust happens earlier as you adjust the Stephenson's link to shut the steam off earlier. So what you find is that before you get the economic advantage of shutting off the inlet steam earlier, the engine starts to bang and rattle due to over compression. Most people will say that they adjust the shut off until the engine runs it's smoothest. What they are really doing is finding the location where the poor valve events are causing the least negative effect. It would be true to expect the 6A to run more quietly with a piston valve, what is really happening is that the over compression on the HPis being masked by the piston valve (as it can't jump of it's seat) and not corrected, as it is a timing issue, not a slide valve versus piston valve issue. When I modeled the 6A with all dimensions exactly as per Stuart's drawing and adjusted the shut off to 50%, one side of the HP cylinder was closing the exhaust at 50% of stroke, and that is an awful lot of compression, especially with a compound where there may be 60PSI or more in the exhaust side of the HP cylinder when the exhaust valve closes. I did a lot of reading about valve timing and looked at quite a number of examples of timing on commercial marine engines, the HP side of the 6A was not at all similar to other examples. As you have mentioned in another post, there is a longevity advantage in balancing the HP slide valve in the 6A, this unfortunately adds to the problem of the valve being blown off it's seat. I changed the HP timing on my 6A from 40Deg advance to 30Deg, more in line with the commercial examples that I studied, the simulation shows massive improvement. With the inclusion of the center suspended link and careful calculation and testing of the position of the suspension pin, I can now change the shut off any where from 78% of stroke to 30% of stroke, and ALL events remain in balance between both ends of the cylinder within 2% across the range, in both forward and reverse. Not that reverse matters much in a boat. Good railway engines managed to get this down to 1%. A standard 6A with the valve gear made according the the Stuart drawings has balance errors of as high as 35% .i.e. the opening and closing positions of the ports can vary between one side of the cylinder and the other by up to 30% of the full stroke length when the engine is linked up. So you can set the 6A to be reasonably correct at one link position, but change the link setting and it all goes terribly out of balance. It sounds impossible, until you see the simulations.

As a matter of interest, my CAD software allows me to simulate and measure the dynamic balance of the engine, as the mass properties can be loaded for each material used in the engine, thus the software calculates the weight and balance of each individual part and combines and calculates them in the assembly.. This means that I can plot the correct counter balances necessary for the HP and LP, and yes they need to be different. I can change the material properties for say the pistons from CI to aluminum and instantly see the change in balance on the crank. Fun stuff, if you are naturally patient

From Mike: "If i send you the information on my valve gear will you run it on your simulator? It's single expansion single cylinder and runs well but I'm consumed with curiosity."

To simulate your valve gear there is excellent freeware, form Charles Dockstadder. His software can simulate almost all forms of valve gear, with great precision and with full user inputs allowed for the geometry of your particular engine.

I have also written EXCEL software for simple engine valve gear, in my case Marshall-Bremme, (or Klug in some countries), allowing the user to vary any dimension for every variable. I experimented with this geometry about a billion times (OK, I'm exaggerating) to get my valve gear that is almost perfect with all realistic cutoffs in both forward and reverse,

I have 9 notches forward and nine notches astern, with virtually the same cutoffs ahead or astern, head end of the cylinder as well as the crank end, from 30% cutoff to 85% cutoff. It makes me wonder why more engines do not use this type of valve gear?.

Sitting here thinking about this problem. Probably thinking too much but here goes.

It seems to me that if the the lifting link (that's what it's called in locomotive designs) is long enough, it shouldn't make much difference whether the lifting link is connected to either end of the Stevenson link or the middle. I can see where if this link (maybe called a hanger sometimes?) is very short and connected to the near end of the Stevenson link, it might cause problems. On my engine, the lifting or perhaps we should call it the connecting link, is split so it goes on both sides of the Stevenson link. In other words, it is connected to the far end. It is perhaps twice as long as the split or Stevenson link.

Or am I wildly misunderstanding the problem?

Mike,

Given the number of engines out there with side/end suspended expansion links, it seem obvious that like you, many designers of small marine engines, and also a couple of locos designers thought that making the lifting links longer and side/end suspending the expansion link would be better. They were horribly wrong. Like we often see with Chinese copies of successful products, they copy something with what may seem like slight modification,but it doesn't work well because they didn't understand the science of what is going on within the original design.

There is some very precise science going on with the Stephenson's link that is very difficult to grasp unless you simulate it in some measurable way. Not only is it important to suspend the link by the center with respect to the arc of the slot, but the distance or offset between the center of the suspension pin and the center line of the slot is critical. For example, with my 6A. If I change the offset distance of the pin position by just 0.5mm, the balance of events will go out by 3% of the stroke length. The 6A has different eccentric timing on HP and LP, so I had to make the suspension link position slightly different for HP and LP in order to achieve correct timing for both cylinders.

As the expansion link rocks though it movement, the offset of the suspension link causes the expansion link to move slightly sideways. The sideways movement shifts the die position along the slot ( which is at an angle) and provides a necessary adjustment to the valves position relative to the ports.

The side/end suspended link completely stuffs the science up and creates big errors rather than providing small and necessary improvement.

A lot of people think that the mission is to set the valve gear up such that the die slides as little as possible along the slot as the engine is running, this is in fact completely wrong. When the Stephenson's valve gear is correctly set up the die moves back and forth along the slot more than one would expect, but it needs to do so for the science to work. If you look at all of the big admiralty engines, I think you will find that they are all center suspended.

Lionel

Ok Lionel, I now get what you are on about with the centre suspension, and it's subtle but evidently important. I also see why many designers have missed the point. I have a more fundamental question, what about the radius of the quadrant, from where is that derived? I see many that I think (what do I know?) are too flat or too curved.

Mike,

The radius of the center line of curved slot must be the same as the distance from crankshaft center line to the center of the slot. So if the eccentrics had no offset, i.e. they were concentric, the die block would not move the valve at all when shifting from forward to reverse with the crank shaft stationary. I hope that makes sense.

The total angular swing of the expansion link should not exceed 50Deg from full forward to full reverse in order to avoid problems moving from full forward to full reverse.

The distance from the center line of the slot to the eccentric rod pin attachment points on the expansion link should be kept as small as possible to reduce errors.

The lifting links should be as long as practicable.

The pivot point of the lifting links should be at 90 deg to pivot hole in the die block with the valve centered over the ports and the expansion link in mid gear.

Lionel

Thanks Lionel, yes that makes sense, and is exactly what I thought.

There are an awful lot of engines out there with dodgey valve gear then!

This is contradicting the designs of some very brilliant designers and engineers, like George Whitney who designed this engine around 1896. As you can see in the drawing pic, how much the radius is effected by shortening the suspension arm and how it relates to sideways motion of the die block which should be minimized for wear and why these should be kept as long as possible. It works in your CAD model, but I think it is something you'll regret having an excessive amount of.

No offense, but the 6A is being presented as some sort of metric to make broad generalizations about all steam engines. The 6A is known to be physically small for it's stroke. Which causes excessive loading on the crosshead guides, short coupled valve linkages etc.

I don't know, but I would think some geometry compromises were made on automotive steam engines, where compact arrangement was most important.

One can look into the Dockstadder freeware (screen shots shown below), which shows an animation of the engine and its valve motion, plus several graphs of valve events to see where shortcomings may become manifest. The user can change any dimension, set the engine speed, the position of the reversing lever, etc.

Far better time spent here than on conventional TV programming, in my view.

Do these simulators take into account the amount of lag at the port at higher RPM. What I have observed setting valve eccentrics to the crank, running at higher RPM, the eccentric needs to be advanced, just like with an I/C engine. The difference is stark.

What's a "TV"? Oh, I used to have that about 12 years ago, it came through a cable in the living room and there were about 150 channels and nothing to watch.. I did see one interesting show one night and then I realized it was a re-run I had seen before then as I sat through 17 commercials in a row (I counted), the decision was made to cancel it Network TV is going the way of the Dodo, these younger folks don't watch TV anymore. Yes, talking and learning steam stuff is way more interesting.

-Ron