Valve Gear Simulation

A special section just for steam engines and boilers, as without these you may as well fit a sail.
User avatar
fredrosse
Full Steam Ahead
Full Steam Ahead
Posts: 1906
Joined: Fri Nov 20, 2009 5:34 am
Boat Name: Margaret S.
Location: Phila PA USA
Contact:

Valve Gear Simulation

Post by fredrosse » Mon Jan 31, 2022 6:06 pm

For those of us who might like to see many steam engine valve gears, with motion simulation, get these free programs for Windows at ;

http://www.billp.org/Dockstader

Thanks to Mr. Dockstader for producing these programs, they let the user see animations of all sorts of previously invented valve mechanisms. One can adjust all relevant parameters/dimensions to match your engine, or to design a valve mechanism with great detail. It is quite simple to just view an animation, with motion speed fully adjustable. Beyond that, one can get into the details, with valve diagrams, etc. etc, and could occupy many hours of study.

I have downloaded these files today, McAfee malware inspection passed, and the programs work on my Windows XP computer as well as my Windows 7 computer.
User avatar
fredrosse
Full Steam Ahead
Full Steam Ahead
Posts: 1906
Joined: Fri Nov 20, 2009 5:34 am
Boat Name: Margaret S.
Location: Phila PA USA
Contact:

Re: Valve Gear Simulation

Post by fredrosse » Mon Jan 31, 2022 6:17 pm

For many years I have evaluated steam engine valve gears, and have often wondered why the Stephenson Valve Gear Mechanism is so prevalent. Does anyone know why?

It seems to me that the Stephenson mechanism would be obsolete before the end of the 19th Century, considering that it is perhaps an order of magnitude more difficult (and thus expensive) to build than other valve gears of the later 19th century, many of them providing better valve timing events.

The Stephenson mechanism is directly suited to engines having their valve rods in line with the crankshaft, which might have been considered an advantage, but all other useful valve gear mechanisms can accommodate this valve stem location with far less complication and expense, some requiring the addition of a simple pinned lever to transfer valve motion to the centerline of the crankshaft.

The main disadvantages that are present with Stephenson Reversing Linkage, as applied to typical double acting engines with typical piston valves or slide valves:
D1. Stephenson requires two eccentrics, taking extra length in the crankshaft, and expense to fabricate.
D2. Stephenson requires a curved sliding element, which is difficult to machine, prone to wear, and difficult to renew when worn.
D3. Stephenson machinery parts are much more numerous and difficult to produce than many other mechanisms. It requires two eccentrics, one curved sliding element, four or more pinned journals.
D4. Stephenson produces valve events (timing) that is less than ideal.

For example, consider that the Marshall (or named Bremme, or Klug) has these advantages;
A1 The Marshall mechanism can perform almost perfectly with respect to valve timing events, both forward and reversed,
A2. The Marshall mechanism needs only one eccentric, no sliding elements, and four pinned journals. Fabrication of the parts is clearly simpler than for the Stephenson mechanism.

The Marshall valve gear on my steamer, the Margaret S., provides very nearly equal valve events all the way from 30% cutoff to 85% cutoff. This range of cutoffs is very adequate for any performance objective, and is virtually equal for both the head end and crank end of the cylinder, forward or astern, as illustrated in the attached diagram.
Attachments
MARSHALL BREMME CUTOFFS.JPG
MARSHALL BREMME CUTOFFS.JPG (108.42 KiB) Viewed 8955 times
User avatar
barts
Full Steam Ahead
Full Steam Ahead
Posts: 1069
Joined: Wed Nov 18, 2009 6:08 am
Boat Name: Otter, Rainbow
Location: Lopez Island, WA and sometimes Menlo Park, CA
Contact:

Re: Valve Gear Simulation

Post by barts » Mon Jan 31, 2022 6:43 pm

For many years I have evaluated steam engine valve gears, and have often wondered why the Stephenson Valve Gear Mechanism is so prevalent. Does anyone know why?
I've often wondered the same. The Marshal gear has the practical advantages you mentioned: that there are no sliding links or blocks; all motion is rotary and so can be easily made renewable, even in the field, if desired. The only reason I can suggest is that the Stephenson mechanism is perhaps easier to understand. There's a forward eccentric, and a reverse eccentric, and when the link is in completely in forward or reverse, the other eccentric doesn't substantially affect the motion of the valve.

I wish the source for those computer programs was available. It would be nice to see them rewritten to be more portable.

- Bart
-------
Bart Smaalders http://smaalders.net/barts Lopez Island, WA
User avatar
cyberbadger
Full Steam Ahead
Full Steam Ahead
Posts: 1123
Joined: Thu Nov 07, 2013 9:16 pm
Boat Name: SL Nyitra
Location: Northeast Ohio, USA

Re: Valve Gear Simulation

Post by cyberbadger » Mon Jan 31, 2022 11:09 pm

fredrosse wrote:
Mon Jan 31, 2022 6:17 pm
For many years I have evaluated steam engine valve gears, and have often wondered why the Stephenson Valve Gear Mechanism is so prevalent. Does anyone know why?
Abner Doble (of steam car fame) considered Stephenson valve gear:
(...) the most universally suitable valve gear of all, for it can be worked out for a long engine structure or a short one. It can be a very simple valve gear and still be very accurate, but its great advantage is that its accuracy is self-contained, for the exact relationship between its points of support (eccentrics on shaft, valve crosshead, and link hanger arm) have but little effect on the motion of the valve. Its use on engines in which all the cylinders lie in one plane, represents, in the belief of the writer, the best choice.
https://en.wikipedia.org/wiki/Stephenson_valve_gear
I wonder if everything in this wikipedia article is actually correct and a fair assessment...

I know Stevenson's valve gear is not the best valve gear technically for several specific applications. (Walscherts for locomotives, Maybe Baker Valve gear for Traction Engines, Corliss for stationary power)

I think Stevenson's valve gear success is partially for being a next logical step from gab gear - so Stevenson came fair early in steam engine history - and for being a reasonable choice for multiple applications which lead to popularity.

I think of each eccentric as being a program for the valve. It's neat that you are basically shoving a different program physically into place inline with the valve steam.

I do enjoy the operation of the Stephenson valve gear on my 6HP Two Cylinder Toledo steam carriage engine on my launch Nyitra. Wes pointed out in, it works well with piston valves, which is what I have in my Toledo engine.

At the 1901? US Automotive Show there was a public demonstration of a 6HP Toledo engine without a load at 2000RPM. At that speed, they pulled the reversing lever. The engine was said to have not been damaged by the stunt.

I've never tried to run mine at anything near that speed... Usually I try to reduce throttle slightly before I pull the lever. But when I do move the reversing lever - it is as smooth as butter, and the engine responds immediately. It does takes a little time to reverse the inertia of the shafting and propeller. If I reduce the throttle to the bare minimum, I can use the reversing lever to control the engine angle almost like a servo. I can move the engine forward half a turn and then have it turn half a turn back. -- This is probably why twin steam engines where chosen to be used as remote servo's to control the steering in large steam ships like SS William G. Mather.

-CB

P.S. If I had $20K and a machinist to make any steam launch engine I wanted right now, I very well might choose Marshal valve gear. I would prefer better control of power through the valve gear.
User avatar
fredrosse
Full Steam Ahead
Full Steam Ahead
Posts: 1906
Joined: Fri Nov 20, 2009 5:34 am
Boat Name: Margaret S.
Location: Phila PA USA
Contact:

Re: Valve Gear Simulation

Post by fredrosse » Tue Feb 01, 2022 3:49 pm

Thanks cyberbager for this info from another post on this forum:

Doble considered Stephenson valve gear: "(...) the most universally suitable valve gear of all, for it can be worked out for a long engine structure or a short one. It can be a very simple valve gear and still be very accurate, but its great advantage is that its accuracy is self-contained, for the exact relationship between its points of support (eccentrics on shaft, valve crosshead, and link hanger arm) have but little effect on the motion of the valve. Its use on engines in which all the cylinders lie in one plane, represents, in the belief of the writer, the best choice."

I am glad the Doble quote indicates this is his opinion, rather than hard fact, as this distinction is so relevant in the engineering world. Opinions are just that, whereas hard facts are indeed hard facts, IFF the engineers making statements do their analysis correctly. I would consider that every advantage Doble delineates above could also be attributed to the Marshall-Bremme valve gear as well.
User avatar
PeteThePen1
Full Steam Ahead
Full Steam Ahead
Posts: 553
Joined: Thu Nov 19, 2009 11:53 pm
Location: Aberystwyth, Wales, Europe
Contact:

Re: Valve Gear Simulation

Post by PeteThePen1 » Wed Feb 02, 2022 12:46 pm

Hi Folks

I am always interested to read these discussion between you folks who know about such technical matters. I also wonder if my very lack of knowledge might perhaps be an exmple of one of the reasons for the apparent popularity of Stevenson's valve gear. If most people know about it then maintaining it, whilst perhaps a bit of a pain, becomes feasible where ever the engine happens to be since everybody has come across Stevenson's.

Following on from that, if Marshall is the best design for launch engines, how about the idea of drawing up and publishing in Funnel (or similar) an article on using Marshall on one of the popular Stuart Turner engines? I have not analysed the Steamboat Register to see which engine is the most popular. We have had technical articles on various valve gear but nobody has done a 'how to install it' article that I can remember.

Regards

Pete
TriangleTom
Full Steam Ahead
Full Steam Ahead
Posts: 134
Joined: Mon Oct 31, 2016 5:57 pm
Boat Name: No Boat Yet
Location: Yukon, OK

Re: Valve Gear Simulation

Post by TriangleTom » Mon Feb 21, 2022 5:48 am

barts wrote:
Mon Jan 31, 2022 6:43 pm
For many years I have evaluated steam engine valve gears, and have often wondered why the Stephenson Valve Gear Mechanism is so prevalent. Does anyone know why?
I've often wondered the same. The Marshal gear has the practical advantages you mentioned: that there are no sliding links or blocks; all motion is rotary and so can be easily made renewable, even in the field, if desired. The only reason I can suggest is that the Stephenson mechanism is perhaps easier to understand. There's a forward eccentric, and a reverse eccentric, and when the link is in completely in forward or reverse, the other eccentric doesn't substantially affect the motion of the valve.

I wish the source for those computer programs was available. It would be nice to see them rewritten to be more portable.

- Bart
I actually sent a letter a few years ago to the address listed on Mr. Dockstader's website. Unfortunately, the news I got back was that it seems that he unfortunately passed some time ago, so whatever source code was once available is now regrettably most likely lost.

Fred, I've probably watched your video on YouTube on the Marshall-Bremme valve gear you constructed 100 times. Do you happen to know of any resources aside from Mr. Dockstader's program that deal with the design of Marshall valve gears?
User avatar
fredrosse
Full Steam Ahead
Full Steam Ahead
Posts: 1906
Joined: Fri Nov 20, 2009 5:34 am
Boat Name: Margaret S.
Location: Phila PA USA
Contact:

Re: Valve Gear Simulation

Post by fredrosse » Thu Feb 24, 2022 4:50 am

I wrote an EXCEL program that evaluates the Marshall-Bremme motion, every input value that is in a boxed cell can be varied by the user. The valve ellipse, showing admission and exhaust points for both the head end and crank end of the cylinder is presented vs % stroke. The motion graph is based on 360 calculations of one degree each. After messing with mechanism dimensions (god only knows how many times), I was able to arrive at a mechanism that gave virtually even cutoffs throughout the valve forward and reverse notches.

PM me if you would like the EXCEL file, and you can design your own mechanism. In decades past I spent countless hours drawing valve motions on tracing paper, before EXCEL was available, and now one can perform valve dimensional evaluations in one second, which took several hours work in the past. This gives real appreciation to the efforts of 19th century designers who had no computers to solve the quadratic equations necessary to evaluate the valve motions.
Attachments
MarshallScreenshot.jpg
User avatar
PatJ
Lighting the Boiler
Lighting the Boiler
Posts: 37
Joined: Mon Mar 05, 2012 5:23 pm
Boat Name: To be determined
Location: Central USA
Contact:

Re: Valve Gear Simulation

Post by PatJ » Sat Mar 12, 2022 11:26 pm

I got pretty involved in studying valve gear design a few years ago.

I ran across Docksteder's programs, and was impressed with what he had done.
I emailed him, and convinced him that he should really share the source code for those programs while he was still able to do so.
He did send me some source code, but boy is it some terribly convoluted stuff.
He told me where he got the base equations that he used (some book or books).

His source code is not written in any modern programming language.
I took one look at that and threw in the towel.

There is another fellow out there who does valve gear programs, and perhaps it is the same guy who is a member here; german guy as I recall. (Edit: Ranier Radow maybe?)

I have read a number of old books on valve gear design, and for the most part, they generally say that the "patent" gear designs are junk, mainly because although they are great in theory, they can't be maintained in the field, I guess to do wear, etc.

I have noticed many steam locomotives with I guess Walsherts (spelling correct?) gear, and there does seem to be some popular and successful valvegear developed after Stephenson.
Some of the more modern gear seems to be very similar to Stephenson.

To answer the question about why the Stephenson link is so popular, it seems like you are more looking for reasons to discredit the Stephensons, with your mind already made up, rather than to look at why it works so well.
That being said, as I understand the Stephenson link (not completely by any stretch), it has to do with how it automatically advances timing as you move away from full gear, in a fashion similar to the advance on a gas auto engine ignition. This is one critical aspect of the Stephenson link.

Another aspect is simplicity, ie: at some point the Stephenson link was very widely used on locomotives, and everyone understood to some extent how it worked, how to make it, and how to repair it.
The patent gears can be very complex, and I doubt if many understand the geometry of them.

And there was a British guy who wrote a program, and he showed how to get nearly equal valve movement in both forward and reverse, and across the midpoint-full gear transition.
One of the keys to getting equal valve movement is the position of the pin, which is not exactly at the center of the link curve.
This pin position is very critical, and the guy points out that few ever get it right.

The support arm for Stephenson links is often on one side or the other of the link, but the British guy points out that without a center support of the link, you will never get equal valve travel.

And some Stephenson links have eyes outside the link, and some have the two eyes ofset under or over the curve.
This can be critical.

Remembering from many years ago, there are "launch links", "crossed and uncrossed rods", and some other things.
A cross rod arrangement acts much differently than an uncrossed one.

I will try and dig out some of these articles.

As for Docksteader's program, I finally decided to check valve gear design using the motion analysis of Solidworks.
I can mock up a valve gear in Solidworks, run a motion study, zoom in, and note the valve travel over the port openings as I manipulate the link from full gear to its centerpoint.
And thus I do an end run on Docsteader's programs, which run many many pages for a single valve gear type, and just use the geometry programming that is inherit in Solidworks.

Smarter not harder as they say.
I am not trying to prove my mathmatical prowess to anyone, I just want to verify that the valve travel is almost exactly symmetrical for full gear, half gear, and centerpoint gear (I guess the valve does not move much at center gear).
You can get near symmetrical valve travel for a wide range of gear positions with the Stephenson's gear, if you know how to do it.

I want to build engines that operate well, not write lengthy computer programs.

I think the Stephenson gear worked very well, it worked reasonably well over time (ie: even when the joints wear), it was well understood by the locomotive folks, and it was a consistent performer, unlike the patent gears that were tried. One of those things, if it is not broken, don't fix it.
Eventually the Stephenson was replaced by more modern gear types, but it is still the gold standard as far as consisten in-the-field performace.

.
Last edited by PatJ on Sat Mar 12, 2022 11:33 pm, edited 1 time in total.
User avatar
PatJ
Lighting the Boiler
Lighting the Boiler
Posts: 37
Joined: Mon Mar 05, 2012 5:23 pm
Boat Name: To be determined
Location: Central USA
Contact:

Re: Valve Gear Simulation

Post by PatJ » Sat Mar 12, 2022 11:32 pm

Don Aston has done a lot of work with the Stephenson link, and has a spreadsheet.

I am not sure how to attach stuff here, or what format is accepted, but I have these items.

.
Post Reply