Bump valve uniflow engine

[astroud]

Hello
I am planning to build a simple bump valve uniflow engine to run on air. I have a 3/4" bore and a 1" stroke.
From reading up the inlet and exhaust valve opening and closing seems to be pretty flexible, I am planning the inlet to open c. 1/16" before tdc and the exhaust port to start opening half way down the stroke, does this sound OK ?

Is there any wisdom on how much dead space should be above the piston at tdc, I am thinking more than usual to allow for air compression.

thanks,  Andrew 

[PaulR]

Following along... 

[MJM460]

Hi Andrew, I find the best way to understand the effect of valve events and clearance volume on these (or any other) engines is to construct an indicator diagram like the attached sketch.  I have sketched the cylinder below the diagram so you can see how the corners of the diagram line up with the physical events as the engine moves.

An indicator diagram is a graphical representation of cylinder pressure and cylinder volume.  The area enclosed by the loop the operating pressure traces out is the work produced by the engine each stroke. The operating point moves clockwise around the loop in the direction of the arrows.  No need for exact pressures, just good enough to demonstrate some critical points.  Anything that increases the area enclosed by the loop also increases the work produced during each stroke, while anything that decreases the area enclosed decreases the work output.

I don’t pretend that this diagram is to scale, though the volume axis should be, as all the interesting points come directly from physical measurements of your engine.  However, the pressure scale is, well shall we say somewhat elastic?  But adequate to show the principle.

There is still a wealth of information contained in this simple looking diagram.
 
First draw the cylinder cover, piston top and bottom dead centres, and exhaust outlet to scale on the volume axis.  When the piston starts pushing the valve open and then allows it to fully close, changes the curve at points 1 and 5, while the exhaust port starting to uncover through to finally covered again changes the form of the curve at points 2 and 4.

  Starting with the piston just finally closing the exhaust port on the upstroke, I have assumed  for this purpose that the cylinder pressure exhausts to atmospheric pressure by the time the exhaust ports are closed (at point 4).  Realistic for the very large ports you have proposed, perhaps not so realistic for the smaller ports I have drawn.  (I have no particular reason for the size I drew, I had overlooked your proposed port diameter when I drew the sketch.)

The piston compresses the gas remaining in the cylinder through the remainder of the upstroke towards point 5.  The shape of the curve is approximately determined by adiabatic compression, and is generally the shape I have drawn. 

If the clearance volume is large enough, the cylinder pressure at point 5 will be lower than your supply pressure as I have shown. 

On the other hand, if the clearance volume is very small, the compression stroke might reach your supply pressure before the piston starts bumping the valve open, and results in an unusual indicator diagram for the remaining part of the compression stroke, but I am sure you can see the enclosed part of the loop and hence the work output will be smaller.

On the other hand, if the clearance volume is larger than I have drawn, the pressure reached at point 5 will be lower, and the area enclosed by the loop is larger, hence more work output.

The clearance volume also affects the power stroke.  When the piston moves towards the crankshaft, the gas enclosed by the cylinder expands, producing the work output. 

I have assumed in the diagram that by the time the valve closes at point 1, the cylinder pressure is about equal to your supply pressure.  The expansion of the gas in the cylinder as the piston moves to point 2 is determined by the volume ratio from point 1 to point 2.  If the clearance volume (plus the volume change during valve closure) is very small, point 2 will be lower than I have assumed, reducing the area enclosed by the curve, while if the clearance volume is larger than I have assumed, point 2 will be higher, so more area enclosed by the curve, and more work output.

You can see that more clearance volume increases the work output on both the expansion stroke and on the compression stroke, so increases the work output of the engine. 

This extra work is obtained at a cost however.  It will require more steam or air to get the pressure at point 2 up to the supply pressure, hence more running cost to get that amount of work output.  The cycle efficiency is lower.

Of course, I doubt if anyone reading this knows or cares how much fuel cost is involved in running a model engine, but on a large machine in industry, running cost becomes a significant issue.

I hope this answers your question, but it probably raises more questions than it answers.  Please ask if anything is unclear.  There is a lot more information that this diagram can tell us about these engines, but I have rambled on long enough.

MJM460

[Charles Lamont]

MJM40 has given an excellent explanation. There is one point, relevant to Andrew's question, that I don't think quite right. When the valve closes at point 1, there a certain quantity of air in the cylinder. At the bottom end of the stroke some of that air is let out. On the compression stroke at point 5 we have the same cylinder volume as at point 1, but less air, so the pressure cannot exceed the supply pressure.
The point remains that a smaller clearance volume will give greater efficiency, and a large one will give more power but use disproportionately more air.

There is, however, another consideration. The exhaust is going to atmosphere. Assuming the exhaust port closes with the pressure at 1 bar. A 7:1 compression ratio would mean that the engine will not run with a supply pressure below 7 bar.     

[Jasonb]

There is, however, another consideration. The exhaust is going to atmosphere. Assuming the exhaust port closes with the pressure at 1 bar. A 7:1 compression ratio would mean that the engine will not run with a supply pressure below 7 bar.   

The couple of uniflow engines I recently got running had very little head clearance so the compression ratio would have been quite high and probably more than 10:1 but they run on 2 to 3 bar with no problem.

You can see how little clearance and get an idea of piston movement from the crankshaft

https://www.modelenginemaker.com/index.php?action=dlattach;topic=12854.0;attach=184323;image

[astroud]

Hi MJM
Thank you for your detailed analysis and explanation, it certainly helps me understand what I need to think about.
You mentioned my large port sizes but I have not set these yet. Others have mentioned the exhaust should be a number of small ports around the circumference but I cannot see why, surely the same as one single large port.
The inlet timing can be changed by swapping out the plunger on the piston to a different length, the dead volume can be changed by altering the conrod length and I plan a 3 piece conrod to make this possible. However the exhaust timing is not so easy to alter, only by plugging and re drilling.
Time to start cutting some metal!

Andrew

[Jasonb]

One large diameter hole is uncovered for a longer part of the stroke than a series of smaller ones though I have seen engines done both ways.

[Charles Lamont]

There is, however, another consideration. The exhaust is going to atmosphere. Assuming the exhaust port closes with the pressure at 1 bar. A 7:1 compression ratio would mean that the engine will not run with a supply pressure below 7 bar.   

The couple of uniflow engines I recently got running had very little head clearance so the compression ratio would have been quite high and probably more than 10:1 but they run on 2 to 3 bar with no problem.

You can see how little clearance and get an idea of piston movement from the crankshaft

https://www.modelenginemaker.com/index.php?action=dlattach;topic=12854.0;attach=184323;image

Hmm ... scratching head ATM. That illustration does show a huge flywheel. And the poppet valve timing is probably different to what you would get with a bash valve.

[MJM460]

Hi Charles, thank you for some well thought out comments.  You are quite right that for any given compression ratio, there is a supply pressure below which the engine does not produce any work on the power stroke, so the engine will not run.

I must admit that it was not obvious to me, though I should have known, as Benson and Rayman mentioned it in their book on Experimental Flash Steam, which I bought back when it was first published many years ago, and have read it many times.  Their explanation was simply that it was “obvious”.  So back to the indicator diagram to work out what is going on.

Step one is to draw the compression curve for a higher compression ratio, one that gives a pressure above the supply pressure before the valve starts opening.  5B in the diagram below.

When the valve starts opening, the cylinder is no longer a closed space, so the curve changes as the gas, now at higher pressure than the supply,  can now flow into the supply system, and the operating point moves towards point 6 at top dead centre, and approximately constant pressure towards point 1 when the valve closes.

Now the issue finally is obvious, even to me.  The power stroke starts with lower pressure than the compression stroke pressure at the point where the valve is just closed!  At the higher compression ratio, the operating point would follow the dotted line as the piston moves down.  Clearly the power stroke does not produce as much work as absorbed by the compression stroke, so the engine cannot run.  In practice, the upper limit to compression ratio is a bit lower than necessary to just meet the supply pressure at  5, as there must be enough work produced to overcome friction and other inefficiencies in order to run.

We have already concluded that a higher compression ratio produces less output, we can add that the compression ratio must be lower than necessary to equal the supply pressure at the end of the stroke.

Benson and Rayman do mention another result from their experimentation, the power output does seem to be quite sensitive to compression ratio, and a significant 100% increase in power came from optimising the compression ratio on their setup, but that is another study.  I suspect the extra steam consumption at lower compression ratio was taxing their steam supply.

Andrew, I am glad you have found this helpful.   We can certainly conclude that designs that allow adjustment of the compression ratio are worthwhile.  As to the issue of exhaust hole sizes, assuming a similar total flow area, a larger hole requires more piston travel to completely cover and uncover, while multiple small holes can uncover the same flow area with less travel, leaving more travel for expansion (and compression)as Jason described.  The larger hole also gives more open time to let the pressure to get to atmospheric pressure before the hole is finally closed, a further advantage.  If you do the diagram assuming very small holes, then again with very large holes, you will see the difference.  You probably will see it best by exaggerating the difference in hole size, as the inaccuracy of the curve from points 2 to 4 in yesterday’s drawing may hide small differences.  In general, it is better to have the exhaust hole area too large than too small.  If you are interested in experimenting it may be worth making a cylinder with small holes and in addition, a larger hole that has provision to be blocked off with a screw.  You would be able then test the effect, but probably not worth doing unless you are a very keen experimenter.

Looking forward to seeing the engine progress,

MJM460

[Jasonb]

Hi Charles, thank you for some well thought out comments.  You are quite right that for any given compression ratio, there is a supply pressure below which the engine does not produce any work on the power stroke, so the engine will not run.

Surely the engine does work on the power stroke as the piston will move to BDC. However if you don't get enough energy into the flywheel then it will not return and pass TDC .

When the guy who owned those two engines first asked me about them I did say they would need a reasonable pressure and to try 50psi just over 3 bar and to give the flywheel a good spin. as you won't get them to start on a few psi like a slide valve engine but can wind down the wick once they are running.

You might want to look at the similar thread I did on ME about the restoration as there is a post which has copies of  the article on the "Don" engine it was said to be based on complete with indicator diagram and what pressures were showing.

The cams look to be lined up with the crankshaft throw so no lead which would be similar to the timing you would get with a bump valve

https://www.model-engineer.co.uk/forums/topic/me-vol-48-no-1136-of-1923/

[Charles Lamont]

As it happens, my observation about a minimum working pressure for a given compression ratio also came from a glance at Benson and Rayman, where they discuss Jim Bamford's bash-valve engine for hydroplane racing.

Another thought for the hypothetical indicator diagram; we don't have to wait until point 5 for the valve to open - it will do so as soon as the back pressure reaches the inlet pressure. Actually it will probably need a bit more, as the valve seat has width, so the area below is less than above. And assuming the valve is not (unnecessarily) spring loaded.

[MJM460]

Hi Jason, you are quite right, work is done during the expansion stroke as you have described, but the point is that it takes even more to drive the compression stroke, so that in each revolution, more energy is absorbed from the flywheel than is produced on the power stroke, there is no net work output to overcome friction etc let alone drive an external load.

Even if you give the engine a good start with a cord as for I.c. engines it will slow quite rapidly and stop, rather than continuing to run when the critical compression ratio is reached or exceeded.

The G1 and similar engines which have a pushrod operated valve for inlet have different valve events which have to be suitably represented at the appropriate points on the indicator diagram.  These are not really the subject of Andrew’s original post, and after all this is still his build thread. 

I suggest that the drawing in the ad in your first reply is probably not intended to show the eccentric timing, but shown with the valve in the same plane as the crank for drawing convenience and clarity.  I would rely much more on the timing  you ended up with on the engine you had running.

There is a huge amount of information in the thread you linked to, most interesting.  The point about the difficulty of attaching an indicator to a model is still mostly valid, though with modern high speed electronic pressure transmitters I suspect it could be done by someone suitably equipped.  But the drawings in that thread look very much like a standard slide valve diagram for an engine with some lead from the usual 90 degree position.  We should start another thread if you would like to discuss how a conceptual indicator diagram could help us understand those engines.

Apologies for not responding to your replies earlier, they are always appreciated.

MJM460

[Charles Lamont]

Hi Jason, you are quite right, work is done during the expansion stroke as you have described, but the point is that it takes even more to drive the compression stroke, so that in each revolution, more energy is absorbed from the flywheel than is produced on the power stroke, there is no net work output to overcome friction etc let alone drive an external load.
MJM460

Yup, it's a compressor.

[MJM460]

Hi Charles, And not a very efficient compressor at that.

Still due credit to you for picking up the comment in Benson and Rayman’s book.  I should have refreshed my memory of the book earlier.

I did however consider including of how the pressure would contribute to to the valve opening, but decided it was a matter of degree, and my post was already too long without going into the detail of valve area exposed to the cylinder pressure and spring loading.  But thanks for introducing it.

MJM460

[astroud]

Thanks for the discussion, no problem to carry on in this thread.

I decided on a horizontal engine and so have a spring loaded ball valve, spring from a biro pen i think. It could be left out or the preload changed just to add another variable.

Here is progress so far, the odd shaped frame plates are due to using up some offcuts.

[MJM460]

Thanks Andrew.  I hope the discussion has answered your original question and helped you understand how the various dimensions work together to make a working engine.

Those off cuts look like they were made for the part.  A great start to your engine.

MJM460

[astroud]

Progress has been made and I am nearly ready for the first run.
The crankshaft is 3/8"  diameter in anticipation of having to overcome quite high compression ratios.
The conrod is 3 piece Loctite'd together, the centre section being a 6mm tube from a radio aerial. The ball in the inlet valve has a biro spring to keep it against the seat as this is a horizontal runner. The piston has no rings, just a couple of shallow oil grooves.
I decided there is no exact science for the timing, the single exhaust port is 1/4" dia and starts to open when the piston is half down its stroke and the piston bumper starts to open the ball valve at approximately 1/8" BTDC.

[PaulR]

What a little beast, I guess it's going to be pretty loud...

[Sanjay F]

Looking forward to seeing this one running 

[astroud]

After a bit of fettling she is a runner ! I needed to increase the valve opening from 1/16" to  1/8" and increase the inlet pressure to about 50 psi. After building engines to run on the lowest pressure it was quite a gulp moment to wind the pressure up and I had supply pipes blowing off their barbs until I wired them on. Midway through the runs the engine would suddenly start to race away making me step back and the first time it happened I feared for the integrity of the engine.
It must hit a resonance spot, a rapid pressure reduction causes it to eventually slow. Unless I give it a good flick to start it will just oscillate backwards and forwards.
<a href="https://www.youtube.com/watch?v=wxMrnPKdZP0" target="_blank">http://www.youtube.com/watch?v=wxMrnPKdZP0</a>

[Sanjay F]

Nice! Very purposeful sound 

[PaulR]

Great stuff Andrew, really nice. These uniflow engines do make some interesting random sounds that's for sure but generally it reminds me of a taxiing WWII aircraft.

50 psi is pretty scary though!

[MJM460]

Hi Andy, good to see it running successfully.  Congratulations on a runner.

It might be worth a little experimenting to see if you can modify something simple to get it running at lower pressure.  If you have a look at Paul’s Bump valve engine thread you will see he had his running at much lower pressure, so the required pressure is more a matter of design details than a limit of the bump valve principle.

We talked earlier about the effect of compression ratio, and I suspect the main difference between your engine and Paul’s is the compression ratio, though I don’t know enough of the critical dimensions of either design to do a more precise analysis. 

Is it possible that your compression ratio is relatively high?  Could you reduce the compression ratio by playing with your conrod length?  Or introducing more clearance volume by modifying the head design?

It migh also be worth playing with the exhaust opening.  Starting to open at half the stroke seems a little early to me, but it may be worth increasing the exhaust by adding a second port, perhaps a little nearer the bottom dead centre, in order to exhaust more gas so less work is required to recompress on the return stroke.

Alternatively, on to the next project!

A runner is always a good result.  Further experimenting only if it is your interest.

MJM460

[astroud]

Hi MJM
Thanks for your thoughts. I originally thought I would build the engine and then experiment with the settings as you suggest principally because it was my first uniflow and I doubted I could get it to run. Now it does I am having second thoughts about changing anything.

More playing has shown it likes 50 psi to start but can then be turned down to about 30 psi and it will run nicely there. If I could get a fast enough flick speed on starting it would probably start with a lower pressure, if not high enough it just bounces backwards and forwards which is quite entertaining.

I had thought about a variable exhaust port position but could not come up with any easy design. The compression ratio is high, the calculated clearance at tdc is 0.080". I could shave off from the head locating spigot which is about 0.125" deep  but no so easy to put it back if needed....I want to swap out the 1/4" ball bearing valve for a smaller one but having trouble finding one on the workshop.

I will do some more simple playing, then go onwards and upwards to something else.

Andrew

[MJM460]

Hi Andrew, I can understand moving on.

If you ever had the inclination to experiment after a rest on this project, you could experiment with increasing the clearance volume by making a spacer to install between the head and cylinder.  Remember the extra volume only has to be volume, it does not have to be the same diameter as the bore.  You would need to extend the valve pin to suit.

Rather than simply giving it with your fingers, you could try starting it with a leather lace, the same technique as used for diesel engines in a model boat.  (You will need to clamp the base down )

What is the next project?

MJM460

[Jasonb]

String wrapped around a flywheel will also get it going as will a rubber faced drum in a cordless drill held against a flywheel rim

A good way to allow for some experimentation is to use a separate wrist/gudgeon pin holder then you can just use shims to alter the position of the piston and also add larger shims to the crown to alter compression ratio.

[astroud]

Hi Andrew, I can understand moving on.

If you ever had the inclination to experiment after a rest on this project, you could experiment with increasing the clearance volume by making a spacer to install between the head and cylinder.  Remember the extra volume only has to be volume, it does not have to be the same diameter as the bore.  You would need to extend the valve pin to suit.

Rather than simply giving it with your fingers, you could try starting it with a leather lace, the same technique as used for diesel engines in a model boat.  (You will need to clamp the base down )

What is the next project?

Not sure about the next project, I have a half finished IC engine which I should get back to, I baulked at making the cams and valves as being beyond my skills but should have a go. Otherwise buoyed by getting the uniflow to run a multi cylinder uniflow with possibly valves according to Chuck Fellows' thoughts and designs which were published here several years ago.

Andrew

[astroud]

String wrapped around a flywheel will also get it going as will a rubber faced drum in a cordless drill held against a flywheel rim

A good way to allow for some experimentation is to use a separate wrist/gudgeon pin holder then you can just use shims to alter the position of the piston and also add larger shims to the crown to alter compression ratio.

Thanks Jason, something to remember for the future. My conrod is a 6mm tube with little and big ends stuck on so I can unstick it, make a new shorter central section and stick it back together again.

Andrew

[astroud]

After playing with the engine I noticed if I unscrew the air valve the engine would slow down so I decided to build another air valve with a variable position valve meaning I would be able to move the ball closer or further away from the piston activating prong. I achieved about 0.15 movement which I had seen from my testing was enough.
The photo below shows the design, basically a moveable spool with the ball and spring within the housing. I found out after testing the air pressure within the spool can push the cap on the opposite end to the ball seat off its seating resulting in a loud clacking sound when running. The only way I see to fix this is either solder or Loctite or a much stronger return spring but it is not a problem if I keep the pressure below 40 psi so I will leave as is for the moment.
I have some smaller diameter ball bearings coming to see what impact they might have.

If you can watch the video to the end it does its party trick and starts oscillating backwards and forwards.

Andrew

<a href="https://www.youtube.com/watch?v=Tlg3QlicZzg" target="_blank">http://www.youtube.com/watch?v=Tlg3QlicZzg</a>