'Speedy Revisited' -valve dimensions & lead

[Gary L]

Hi all

I am reboring and re-valving a 5" Speedy which has the original imperfect LBSC geometry. I don't want to change the motion if I can avoid it (or at least, not yet), but I have studied Don Ashton's EiM article (Jan 09) hoping to ameliorate the faulty timing as far as easily possible.

According to Don, LBSC's lap of 0.157" gives a negative lead of 0.022" which isn't good, and the geometry seriously limits the starting cut-off, which is worse. Don proposes removing 0.022" from the inside edges of the valve bobbins which reduces lap to 0.131" and produces a 4% better starting cut-off. Since this shortening is by the same amount as the negative lead, am I right in thinking that this mod in itself produces the desirable condition of zero lead?

I gather that lead has little value in our small sizes, but I always assumed this is adjustable to some extent at the valve setting stage... so presumably the idea of this mod is to provide equal lead of zero in reverse as well as forward gear. But at this point I get completely out of my depth...

My question is, if shortening the bobbins as above improves the starting cut-off by 4%, would it help matters to shorten them a touch more?? And if so, by how much?

There was a thread a while ago here on this same subject where it was suggested (amongst other things) reducing Speedy's lap to 0.100" -a significantly bigger alteration than that proposed by Don. With an actual max valve travel of only 0.440" (and mine is also very close to that figure; it should be more by rights, if only LBSC had cut his bits of cardboard better!) it was said that this would give a port opening of 0.120" and an improved cut-off of 79%, which sounds very appealing. Presumably there is a down side... I suppose this will result in a positive lead? Which by adjustment can be reduced in forward gear, at the expense of astern? But even if I've got this right, is this going too far?

My Speedy is definitely challenged in the starting cut-off department (about as challenged as I am in comprehending valve events ). I can't resort to computer simulations to help BTW, because (1) I wouldn't understand what I was doing and (2) in any case, AFAIK, none of the simulation programs will run on my Mac.

So I'm hoping the assembled brains and experience can give me a pointer or two please!

Best regards

Gary

[Gary L]

Paul1979 said
"You could also have a look at the Don Young revision, which works really well, "

-Well yes, but Don Ashton is pretty dismissive of both Don Young's and KN Harris's improvements. I don't know enough about the subject to argue, but Don A's article is pretty convincing. More to the point, I really don't want to start rebuilding the valve gear as well, at least not before I've exhausted the possibilities of what can be done in the valve chest.

Gary

[steam4ian]

G'day All

Dr Alan Wallace of valve gear simulation fame has built a Speedy. It would be interesting to see what modifications Alan has made to the valve gear, his loco works well and is arguably his best puller.

His simulation file is available on his web site.

I understand Alan and Don do correspond.

Regards
Ian

[donashton]

Hi Gary,

Certainly get rid of the excessive negative lead and you will gain some cut off for starting, but this is only part of the story and will do nothing for the poor valve events. Reducing the bobbins further introduces POSITIVE lead and SHORTENS cut off. In the longer term, you cannot put things aright piecemeal - every alteration affects all the other components.
Paul suggest that DY's gear works well, but measured against what? I spent a lot of time trying to keep remaking down to a minimum without much success, probably because I'm rather fussy about what I recommend. Summarising: mod the valves and see how it then runs - it may satisfy you, but not me.
Tongue in cheek - why do you want a simulator - I've already done it for you.
Best of luck,

Don

[donashton]

Ian,
Allan Wallace (double L) built to KNH before I knew him. Having great expertise in programming, but admitting no special study of gears, he would back me up by accepting my judgment. Incidentally, he arrives to meet me on Sunday and we have a lot to discuss. Rain or shine he will be driving an engine or two that afternoon. He loves driving, is very practical, and is a very clever guy.
Don

[Gary L]

Hi Don

You wrote "Reducing the bobbins further introduces POSITIVE lead and SHORTENS cut off. In the longer term, you cannot put things aright piecemeal - every alteration affects all the other components."

Thanks; that's just what I wanted to know; so I shouldn't go down the road recommended in an earlier thread and reduce lap still further.

Very much appreciate you and others taking the time to reply. I have to go with the piecemeal approach for the time being... I realise it won't approach optimum, but all improvements are welcome, however small, and improving the motion might be a project for next winter. In the meantime, I won't have actually made the running any worse.

While you are on, could I ask about the valve setting advice at the end of your EiM article? ("Centralise the valve exactly over the ports"). I have had differing advice from various sources including (as you mention) equal leads and port opening. My question is, is your recommendation valid only for your modified geometry (which is how I read it)? Or, in the absence of the valve gear mods, is one of the other methods to be preferred?

Regards

Gary

[pondok]

Hi Gary, I haven't looked too closely at Speedy, and don't think there are any at my club, but would it be worth analysing the valve dimensions seperately from the motion imparted by the expansion link, and deal with each seperately, on the job, as it were?

I have had to tackle timing problems in the past and every time it turned out to be the eccentric rod + return crank combination making it impossible to get equal swing of the expansion link. This meant that the valve dimensions became largely academic, since no amount of adjusting would rectify bad valve events.

If you ascertain that the expansion link has unequal swing, then it might be worth the extra effort and possibly expense sorting that out first. It could be as simple as ajusting the position of the return crank on the end of the crank pin (assuming it's not pinned or square), but often it means resizing the eccentric rod, silver soldering to rejoin. The main join can be facing in, and the outer face of the rod can be very discreet so the join is almost invisible. Bruce Hope of the Harlington/ Guildford clubs did a great job on my loco.

It's a simple matter to determine the correct length of the rod, it involves removing the eccentric rod, and assuming the return crank isn't pinned (or the pin can be removed) or fixed to a square, one can use ajustable/digital calipers or dividers to find the correct rod length.

This is done by fixing the expansion link in mid-swing so it's arc forms the exact radius of the radius rod (check by moving the radius rod down and up - the valve crosshead should not move) and moving the return crank gradually so that the reading on the caliper/divider is identical between expan. link centre and return crank small end, when the wheels are both in front and rear dead centre.

Without this running true, the expansion link swing is unequal either side so the valve events will always be out. Resizing is relatively easy compared to the benefits of this being right.

Once the rods are back on, you can work on the position of the valve in the sleeve (hopefully you have sighting/doping holes) and valve dimensions (bobbin ends = port length + half the total valve movement when engine is in precise mid-gear). The "end to end" dimension of the valve bobbins would equal the distance between the outer edges of the ports in the sleeve.

This would give no lead to the valves, though, but this way should in theory mean the events would be as effective in forward as in reverse gear.

Hopefully I didn't get the completely wrong end of the stick Gary, and hopefully this info could be of some use to you.

andy

[donashton]

Hi Pondok,

Much of what you say is good practical sense but I beg to point out one or two issues.

Your exercise to find the eccentric rod length is nothing to do with equal expansion link swings - it is to ensure that the gear is capable of supporting equal leads so that the valves can be set so.

You cannot divorce the two harmonic functions - the action is too complex and involves both all of the time and this shows up forcibly on simulation.

Equal expansion link swings will not produce equal excursions of the radius rod or the valve. Of all the gears I've checked perhaps the nearest that came to equal swings was the BR 9F 2-10-0 (about 0.010 degrees out!) and the cut offs are far from equal. The Darjeeling 'B' class had no backset and unequal expansion link swings by around 2 degrees, but the equality of the valve events is very good.

"End to end" of valve is exactly as you say for exhaust line-in-line. Bobbins should each be port + lap full stop. Lead comes from an advance in the gear and is not a physical quantity that can be machined. Gary will see that removing excess negative lead by machining it off the valve bobbin is in fact changing the lap, and doing so alters all the valve events. That is why in the longer term he needs to correct the gear so that it matches the valve.

Gary - thanks to a few postings from friends you see the complexity of Walschaerts' gear and will learn from what everybody says. Our model designers did not have computer simulation to show up whether or not their efforts were secure. Mostly they didn't do such a bad job, really, but today we don't have to play Russian roulette blindfold with one arm behind our backs.

Do hope it's all helpful. Try my website, but it will be a bit hard going for beginners.

Don

[steam4ian]

Don

Give Allan my regards; he told me he was seeing you but I did not feel at liberty to make that public. Your gain will be our loss this Sunday. We are forcast for fine dry weather so I guess allan would have brought Juliet had he been with us.

Regards
Ian (one I ;D)

[steam4ian]

G'day all.

We have had engine/locomotive valve gear capable of being linked up for about 150 years. During the time of the development of the various gears there was no computer simulations, at best there would have been model.

Just how did our forebears get such reasonable outcomes?

I am not decrying the work of simulation programs but there must have been some "rules of thumb" they followed which led to an operating loco/engine as a minimum. What are the "rules of thumb"?

For instance with Stephenson's launch gear you set the forward eccentric and other linkages as you would for simple non reversing gear to get the valve events you want for say 75-85% cut off. The reverse eccentric should then follow this. The dickying around must come with the curve in the link and the suspension of the link and die block. Is this too simplistic? I agree it may not be optimum.

Regards
Ian

[Gary L]

Pondok- thanks for your comments, especially the suggested technique for reworking the rods. I hadn't thought of cut and shut with silver solder... I was thinking of (and dreading) a lot of intricate milling, and your idea sounds much more practical.

Don- thanks again for the input. No need to convince me about the intricacies of Walschaert's gear, just spelling it is a big enough challenge to get the head round! Quite agree about the computer simulations, though it is a bit tough on the people working from a set of plans to find later on that the plans could have been better; but that's the nature of progress. I can't help feeling that it behoves the retailers of said plans (which are seldom cheap) to keep abreast of subsequent improvements and at the very least point the builders in the right direction to locate them. (Not that /I've/ anything to complain about, but the original builder(s) of my Speedy must have been well hacked off when they found the engine didn't work terribly well, and their work had been superseded years before!).

That's one of the benefits of this very useful forum. I've learnt a great deal by trawling through the archive threads. Thanks to all the contributors!

-Gary

[donashton]

Gary, I didn't answer your question on setting which would probably use up a whole page. Normal procedure is to set leads equal at dead centres. The caveat is whether the gear design can support equal leads. Ignore port openings - they are unlikely to be equal. When the valve is exactly over the ports (twice in a rev) the piston is just short of full travel. If we have the simulator figure for this, setting to preserve all the simulated events is two minutes of child's play. You don't have this option at present. My sympathy if the head's aching - try hibernating for a week!
Ian, regards will be passed on with pleasure. If Allan can fly to Perth with Juliet he may have it in his pocket at Heathrow - just right for the metal detectors. Your remarks about Stephenson's gear are perfectly right - dealing with the last bit is easy enough once understood.

[havoc]

I am not decrying the work of simulation programs but there must have been some "rules of thumb" they followed which led to an operating loco/engine as a minimum. What are the "rules of thumb"?

A comptuer is a means to ease repetitive calculations. So doing the geometric calcultaions on paper you can design it as good as with simulation it only takes longer. And then they had those setups like there was a photo of in another thread here where they just build it and looked at it. They knew algebra and geometry.

The problem with simulation is that you really need to know what you are doing. It only helps up speeding the optimisation of the design, not the design phase itself. It is not a substitute for designing, more a tool for confirmation.

[Gary L]

Feb 6, 2010 10:28:08 GMT donashton said:

Gary, I didn't answer your question on setting which would probably use up a whole page. Normal procedure is to set leads equal at dead centres. The caveat is whether the gear design can support equal leads. Ignore port openings - they are unlikely to be equal. When the valve is exactly over the ports (twice in a rev) the piston is just short of full travel. If we have the simulator figure for this, setting to preserve all the simulated events is two minutes of child's play. You don't have this option at present. My sympathy if the head's aching - try hibernating for a week!

Thank you Don; I was going to email but you've saved me the trouble. Actually you've saved me a great deal more than trouble, so many thanks indeed. Aching head? Not as bad as it could have been, with your easy to follow instructions. I remember when I was a great deal younger I possessed a booklet on Walschaert's and I think I understood it as well, or at least convinced myself that I did. But that was a great many years ago, and both the booklet and the understanding have long evaporated! Now I've taken up ME again, it is a different world.

Best regards

Gary

[donashton]

Havoc,

Thanks for very sound advice about simulators. Early ones attempted to be design 'helpful' with suspect maths and did not allow verification of existing designs either.

This latter facility is important to me as I am trying to read the mind of the original designer to discover his skills. Beyond some basic design formulae little has been passed on from the experts, and sometimes they themselves seemed not too expert!

"build it and look at it"......yes, sounds like a cartoon and perhaps near the truth in many cases!

Don

[keith1500]

Gary,

I did have ago at improving the valve performance on my Speedy.
I must have read Don's book several time and each time a little more would sink in. Never-the-less I still made an error! Its not an easy subject particularly when working with an existing model. You will have several compromises along the way.

Still, to encourage you I measured the ports positions, I checked the travel of the valve rods and found I had to make adjustments to get equal travel on both side for both forward and reverse.

Based on the travel and ports positions I design what I thought would be a suitable valve. But as I said i didn't quite comprehend all in Dons book and made the valve 20 thou too long. Its/they are due to be shortened this year!

Having said that the loco performs well as I use it for shunting some quite heavy trains. but I know it could be better.

Heres my final evaluation that I have the valves set as best they could be.

www.youtube.com/watch?v=WN9PHtcx_-E

Keith

[donashton]

Keith - you hit the nail on the head. The model builder should not be required to know anything about gear design and only comes to a problem scratching his head over a valve setting. This puts nothing right, but tries to get the best of what's given. Making the valve fit the ports is step number one.
The real 1500 class was only required to shunt empty stock to and from Paddington and the previous 9400s could do this well enough. Maybe Swindon designed it to provide a handy-sized 5" loco for us. for we have built far more than the GWR did.
Before altering your valves you could measure the travel in mid gear and compare half this with however much your bobbins measure beyond port width. That's the lap/lead you have irrespective of the drawing.
Don't brag about making an error - I can beat you anytime!

Don

[keith1500]

Don,

I need to revisit my notes and sketches but from what I recall the valves are 20 though too long.

This is based on mid gear and getting the valve dead middle of the two ports. In my case both ports are open by 10thou.

As an experiment I counter-acted this by introducing 20thou lost of movement on the valve spindle, but obviously this then causes a delay in the movement of the valve and the opening to exhaust. plus a horrible tapping like a worn out old loco! But she runs very well, except for occasionally when shunting I'll find a dead spot caused by the 20thou slack in the valve movement.

Keith