G Scale Model Train Forum banner

1 - 6 of 6 Posts

1 Posts
Discussion Starter #1
Gentlemen – Hello!

I have enjoyed this live steam forum over the past few years, but have never posted until now.
I would like to begin an engineering discussion with like-minded people to revisit the modifications to the Accucraft GS-4 valve gear as developed by member Charles and assisted by others.


This is NOT:
- An attempt to undermine in any way the excellent work done by Charles and his colleagues in developing and implementing this modification. This was good work based on the admirable premise of improving the efficiency of the Accucraft GS-4.
- An attempt to undermine in any way anyone with commercial interests in these modifications. My interest is to understand and clarify the impact of the modifications from an engineering design perspective.
- An attempt to question the modifications to the exhausting system or steam generation system on this model – I am only interested in the valve gear modifications.

I am stating these things in advance as I have noticed that this forum has some members who seem to have a very “sensitive” nature. If you are offended in any way by an engineering analysis and investigation of these modifications please simply ignore my inquiry. I am trying to better my understanding of steam engines and hope that others may also enjoy the same pursuit.

OK – brief background about me. I am a design engineer and I love machines. I especially love steam engines, especially marine applications. And the most interesting part of the engine for me is the valve gear – the “brain” of the engine if you will. It is for this reason that I followed the Accucraft GS-4 valve gear modifications closely and that I am now compelled to write this.

Based on my analysis I believe that the modification of the Accucraft GS-4 valve gear to include a working combination lever has done little or nothing to improve the efficiency of the engine of this particular model. I believe the modification does not provide variable cut-off as was desired. I make this claim based upon simulations utilizing tools available on the internet, specifically the Valve Gear Simulation routines by Charles Dockstader. Ref. http://www.tcsn.net/charlied

My discussion from here is specific to the valve gear on the Accucraft GS-4 model and does not apply to Walschaert valve gears in general, as Walschearts valve gears designed from the start to incorporate a working combining lever certainly do benefit from its use by providing variable cut-off and expansive use of the steam supply.

The basic premise of the valve gear modification is a good one, as the simplified valve gear in the model as delivered from Accucraft does not provide for the operator to vary the cut-off by notching up the reverser. The simplified valve gear in the model as delivered is a compromise and requires that the outside admission slide valve have little or no lap as I understand it. I am referring to steam lap here and throughout, not exhaust lap. It is this lack of steam lap that I believe renders the modification to a working combination lever – well – non-working.

The combining lever is meant to induce “lead” into the motion of the valve and to overcome the “lap” that usually exists at the beginning of the expansion stroke. But the lap works at the end of the expansion stroke too. Lap is what allows the valve to provide “cut-off” of the steam supply and enables the variable expansion and thus improved efficiencies (under certain running conditions) that are desirable. Without this lap, any combining lever will not be able to provide the cut-off before the piston finishes its stroke. This is the basis of my argument.

Again, I applaud the work Charles and crew have done, but I do not believe it has produced the desired effect. If the valve would have been also altered to provide steam lap, I would have a different opinion.

Can anyone find a blatant error in my premise? Does anyone have some engineering analysis to counter my premise? Please let me know. Again I am only trying to generate some discussion around the design and operation of steam engines. I am glad to provide the specifics of my analysis if this post generates any interest and discussion.

There are some other very interesting topics related to this one that I would hope to discuss if this topic proves of interest to readers of the forum. One that comes to mind is the variability of lead as the reversing lever is notched up. Some valve gears provided for constant lead and in others the lead varied with the cut-off as the reversing lever was notched up, although I do not understand how this had any real practical value (or detriment) on a working engine. So many questions. The simplicity of these engines belies their true complexity!

BTW – I have the Accucraft GS-4 (butane) and have not modified it based in part on my analysis. It seems to be a good running model and I have 25 hours on it.

OK thanks for reading this far! I am interested in feedback and related “steam geek” topics!


Also reference http://www.mylargescale.com/Community/Forums/tabid/56/forumid/11/postid/60047/view/topic/Default.aspx
http://www.mail-archive.com/[email protected]/msg04353.html

Premium Member
3,378 Posts
Welcome to you and your inquiry....

In our initial work with Gordon Watson who developed the design and analysis he put forth an overview of stock vs. modified in regards to your bottom line: "It is this lack of steam lap that I believe renders the modification to a working combination lever – well – non-working. "" or in other words lacking improved efficiency.

Here are Gordon's cliff notes (version 2006) to us:

[2] Chassis and running gear.

A test steaming then took place, the Locomotive ran , but the exhaust was ragged/harsh. it used lots of water and fuel and seemed sluggish , in that a larger throttle opening than expected had to be used , the boiler made the steam needed, Just!

We then opened the valve chests and examined the valve events,,

Steam opening 35Deg after TDC. both ends.

Cut off of steam 5 after BDC..

Exhaust opening 35Deg after BDC,
compression period not sure there was one!

In linear terms [position of piston in cylinder]

Admission 13%[ late]

cut off 2% after BDC

Exhaust 12% after BDC[late]

In other words the steam was late in, admitting after BDC , working against the compression after BDC for 12% of the stroke. exhaust was opening at this point and then still open after TDC. no compression period..
This explained the ragged exhaust and huge steam use..

This paragraph tries to explain how Walschaerts Valve gear works, if you understand this , the cure is obvious!

Walshaerts Gear derives its motion from two movements.

{A] Expansion link , which rocks about acentre pin , this provides the reversing function and the cut off of steam admission..

combination lever /crosshead , motion.

This motion provides the eccentric advance function for the valve "Laps".

The valve gear fitted is a single eccentric type, this can be ,and has been used in simple models to give a reversible gear which ,when combined with a valve having zero or minuscule "Lap" will provide almost full stroke admission of steam , ideal for low speed narrow gauge garden models.
in many ways it mimics piston valve models in its full stroke admission and slow steady running characteristics.
However, they have fitted valves with 1mm of "Lap" but have no eccentric advance to drive it correctly, so all the valve events are occuring about 12% late in the steam cycle..thus the large steam usage and sluggish running. this probably tended to implicate the boiler as not up to the job.whereas with the flues balanced and the baffle the boiler is fine.

I then drew out the valve gear, and made new combination levers, these where fitted into the existing radius rods and valve guides, using pins retained by the valve guides sliders.
a very small amount of filing clearances had to done on the radius rods to clear the upper pins swings.
Once these where fitted, the valve events became:

Admission at TDC.

Cut off at 86% of stroke..[ alittle more than I would have liked , but to reduce would mean new valve blocks and a different combination lever.. I was looking for fairly simple fix!

Exhaust open at BDC

Compression 8% before TDC.

These settings gave easy starting, and strong performance
the exhaust is a quiet purr... and the water and fuel usage dropped to more normal levels..the test load was 60lbs[ bricks on flat cars] and I could hold the loco back and it still had power to spare.. the original cylinders are fine and it will be a good hauler..

To summarize:
-Valve gear was examined and determined that the valves (with built in 1mm of lap) were acceptable, and combination lever were developed.

-Baseline tests were preformed and determined that fuel and water utilized in anywhere from 12-25 min (alcohol- it was a struggle otherwise it would not have gone to GW) due to imbalance of the production of steam and inefficiency in the utilization of steam in particular at the other end ...this info was compiled before the combo lever retrofits. Post retrofits we are running 45 minutes-1 hour of run time for alcohol. The original models varied in performance with the butane having the better stock performance over the alcohol (which was the model Gordon based his work around).
There are many gauge one live steam engines with the capacity of cut-off, in particular Aster which can be used to compare apples to apples Aster version of GS4 which is 80.2% , the WM Shay, Mikado and Castle class are 75%, you can find this information @ Southern Steam trains web site: http://www.southernsteamtrains.com/roster.htm. May not be idea but every little bit helps and in the case of the Accucraft GS4 (alcohol) it now can run with the Aster instead of the Accucraft K27.
-Timing was set
-Levers were made and installed, with final cutoff ratio bring 85%** For many that run their steam engine with great enjoyment items such as cut off or superheater tubes make not difference to their experience. You seem to be very interest in the subject and it's possibilities. Would be interesting for you to explore this... We hope you have the opportunity to make a comparison between your stock model and modified GS4 for speed, power, and efficiency. A MLS neighbor nearest to you, Don Plasterer (dcplasterer is his member name) has a modified gas GS-4 that is stock except for combination levers, fuel and water efficiency increased substantially from the baseline test runs done on it.

Actually, please PM us your snail mail address, we will send you a set to tinker with, to one steam geek from others!

**Since the 1mm valve lap is .001" different (.039") when compared to proper .040" lap (1.016 mm). This accounts for the 10% difference in cutoff ratios. (Cut off of 86% is within the realm of many Asters and is perfectly acceptable given the cylinder displacement of 5/8".)
-And there was much rejoicing... at least for us, Gordon and others that have the combo lever set.

BTW- The work with the AC-11 will have newly designed larger valves with .040" lap.

Reference of basics: http://www.kesr-operating.org.uk/valves_and_pistons.htm
Valves and Pistons

A classic book available online(free for the referencing):
Steam engines

308 Posts
George , welcome..
The GS4 valves actually have "lap" exfactory.. thats why the events where able to be improved, with the single eccentric gear they have. the "lap" caused the late admission and the peculiar exhaust conditions...ideally after fitting the combination levers . new valves could be made up with increased lap, and by varying the ratios on the combination levers you could probably get down to a50% cutoff..

Lead in full size locos is used to admit steam to provide a cushion of steam in conjunction with the compression period as the reciprocating parts [piston, rods etc] approach TDC..helps to smooth out the reversal of large masses.


1 Posts
Discussion Starter #4
Hey guys - thanks for the replies - I get it - the Accucraft GS-4 has a good amount of lap to start with! Excellent.
Glad to see that you "measured twice and cut once" when undertaking this modification. Well done.
Also thanks much for the offer Charles for the "tinker toys" - I will likely take you up on that offer - I'll try to figure out how to send you an email off-board.

Still pondering - why Accucraft would have designed the valve with any lap considering their intention to not have a working combination lever?
No lap with non-functional combining lever would have provided better performance in stock form eh? Eliminated the late admission etc. Hmmmm.....
If they had included cross-over ports too.....but hey this give us an excuse to wrench on the little bugger right?

In case your interested - here are some links to the simulations I utilized to kick off this thread. FYI.

I had utilized the Valve Gear Simulation routines by Charles Dockstader as a tool to investigate this. Ref. http://www.tcsn.net/charlied
This is an excellent tool! I started with the Walschearts outside admission routine. This utilizes a working design and a working combination lever. It does indeed provide for variable cut-off and expansive working of the steam as you can see.
Be sure to turn on steam display. Notice the increasing expansive working (color change from red to purple) as the reversing lever is notched up.

I then modified this design with what I believe to be compromises similar to those of the Accucraft GS-4 model as delivered (or so I thought) i.e. little or no valve lap and a dummy combining lever. I did not however re-dimension the whole engine to reflect the Accucraft GS-4 as the functional results of my investigation should hold true in any scale. Specifically I set the “Combination Lever Upper” Frame Dimension = 0. This turns the combination lever into a “dummy” as in the Accucraft GS-4 as delivered. Next I set valve to center position and modified its dimensions so as to have no steam lap. The result behaves as one would expect. Notching up the reversing lever does not provide cut-off, variable or otherwise.
(Save this link to your computer and open it with the Wals_OUT.exe from above.)

Lastly I modified the above design by adding a working combining lever back into the mix, providing a simulation which I believe mirrors the modification made to the Accucraft GS-4 by Charles et al. As you can see by manipulating the reversing lever, no variable cut-off is produced because the valve has no steam lap.
(Save this to your computer and open it with the Wals_OUT.exe from above.)
Perhaps I will update my simulations with the info you provided. In my spare time. Happy steaming guys and thanks again! George

Premium Member
684 Posts

We'll make it easy on you. Send your address to tr3services[@]gmail[.]com (no brackets of course) and we'll send off the kit to you.

To answer your question, yes, it would have been easier and the locomotives would have run better in a stock form had they originally been produced with zero-lap valves and no combination levers. However, since the valves fitted had 1mm of lap, fitting a proper combination lever allows the loco to preform better than it ever would have with zero-lap valves and single eccentric valve gear.

Cross-porting can and has been done on the Accucraft GS's, one way to do it would be to create a crossover plate, perhaps when Gordon returns from Diamondhead he can enlighten us on his methods of making the radius rod look correct.

Easiest way I can figure is to plug the existing holes, machine ports, and drill new cross passages in the cylinder block.

Premium Member
31 Posts
As a relative newcomer to live steam but having always had an interest in things mechanical I have found this thread to be one of enormous interest. Thanks guys for the questions, answers and debate, please keep us all informed of you thoughts, ideas and developments.

1 - 6 of 6 Posts