One of these tasks is to analyse and reset the valve timing of the engine to improve efficiency and performance.
This post focusses on this work (thus far) and finally solves the questions over the LP port dimensions and why they are shown with two different sizes in the drawings (as discussed here) .
The LP port size problem
The dimension problem was finally resolved yesterday, when I realised that a small drawing in the Model Engineer articles labelled “LP Chest Back Face” was incorrectly dimensioned, and this diagram had been copied into the BOOK, and I had used this drawing when machining the ports in the LP valve chest and therefore machined the ports to 3/8″ high instead of the required 5/16″ high <sigh>….
Why we are doing this
As you may have also read, we have built a Digital Engine Indicator, this is a computerised version of the mechanical engine indicators that were used throughout the Industrial revolution and time that steam was used on the railways (and they are still used in very large marine diesels). These instruments are designed to produce Pressure/Volume diagrams to indicate the power being generated by an engine, and to analyse the valve events.
The data we captured, indicated (see what I did there!) that all was not well with the timing on the engine.
So, having got the engine on the bench, and got most of the slop out of the valve gear, we set about remeasuring the valve events to allow us to understand what we had, and how it might be improved.
I am not a qualified engineer, and did none of the apprenticeship that a chief engineer in charge of steam plant might have gone through. This meant that when I tried reading up on the subject I could make little or no sense of the literature or explanations of how things should be or should be measured.
So, when I first built the engine, back in May 2016, I decided to measure with the aid of a digital angle gauge and listening to air being blown through the ports to determine when they opened and closed. (you can read about that here) . (I understand that this apparently unconventional approach is also recommended by no lesser man than Mr Roger Mallinson – so maybe I’m OK.)
What became clear in that process is there are a LOT of data points that needed to be recorded, and even printing a template to record the data in 2016 was a far-from-effective solution.
For the record you are going to record the following;
- inlet (admission) valve opening and closing crank angles
- exhaust valve opening and closing crank angles
- you need to record these data for
- the HP cylinder
- top of cylinder
- bottom of cylinder
- the LP cylinder
- top of cylinder
- bottom of cylinder
- the HP cylinder
- and then repeat all of the above measurements for one or two notched up positions to understand if the gear is producing appropriate changes to lead and cut-off.
That turns into about 100 individual measurements to be taken!
So, this time I went the whole hog and developed a spreadsheet to allow me to record the data and then draw timing diagrams directly from the data (see below). Below is a screenshot of the spreadsheet. (remember click on it for a full size view)
Sharing this spreadsheet is not so easy, but if you want to use it please just email or PM me with your email address so I can email the sheet directly to you…. Sorry this is a pain, but I can’t think of a better way at the minute…. (I could also send the VB program that draws the diagrams if you like),
Doing the measurement
The general game plan for measuring admission timing is to remove the drain cocks/relief valves from the cylinders, and then apply some low pressure air to the inlets (about 5psi) then manually bar the engine over listening for the port to open (letting the air into the cylinder) and close (air stops being heard entering the cylinder, and noting down the position of the crank as measured by a digital angle gauge stuck to the flywheel.
For the exhaust events, the opposite approach is used, connect the compressed air to one of the drain/relief ports, and listen for the air emerging from the exhaust port (as the exhaust valve opens) and the cessation of the air flow as it closes.
Doing all the measurements took a few hours, and some of the measurements were repeated two or three times, as I could not believe or make sense of the answers I was getting….
The LP events were reasonably easy to measure.
The HP events, much less clear, as it became clear that there is a LOT of leakage in the piston valve, despite it being fitted with cast-iron rings. I am pretty sure this is to a large extent being caused by leaks between the valve liner and valve chest. This caused by the old Harrison and A&S mill I was using (mid-1940s) with a lot of wear machining leaking tapered components.
A Word on Angularity
There is a deal of criticism for engines with short rods, as this results in increased “angularity”.
This is an asymmetric effect, being exhibited around BDC and not at TDC. What happens is the piston does not follow the expected Simple Harmonic Motion, and instead “lingers” around BDC. I must confess that from a position of significant ignorance, I think the effect appears pretty limited. Although I know from racing Chevrolet V8 engines, that the tuners in America play with rod length to allow more time for valve events!
So, as a man who spent a deal of time working in Computer Graphics, I am a sucker for a picture. So I created a spreadsheet to draw the effect.
The two diagrams below show how the piston movement is affected by the length of the connecting rod. Showing the piston position in blue and the simple harmonic motion (sine curve) in red. The first diagram is for the Leak’s 7-inch rods, the other is for a rod of 3-inches (equal to the stroke). Remember to click them for a bigger image. The faint pink line on the Leak diagram is me converting a 70% cut-off line to crankshaft degrees… (about 110-degrees)
Personally the effect seems pretty minor to me – perhaps you can persuade me otherwise???
The results in Pictures
The following 20 pictures show the results I got from the data we recorded – I wrote a Visual Basic program to read the data captured in the spreadsheet, and produce the diagrams and dat below.
- The “pie” chart shows the amount of time (crank angle) that the valve is open (red for admission and blue for exhaust).
- The data in the boxes shows the actual angles (all measured from TDC). (remember to click on the pictures to see a full size image) ..remember that in Astern degrees are measured anti-clockwise, as that’s the direction the crank is turning in!
- The caption shows the setting of the gear and the data point being recorded…
- The black diagram reminds us the direction of rotation of the crank/pictures in ahead/astern.
So, for example the first picture shows the timing when running Ahead, gear set into Full Gear, data from the LP cylinder, measured from the Top of cylinder.
So, let’s focus on that first image. The Ahead, LP top end, data in full gear: ..What does it say?
- The admission (inlet opening) occurs after TDC or BDC (at 8-degrees in this case. So we have a negative LEAD (it should occur before the start of the power stroke) . 😦
- The inlet closes very late (158-degrees in the first picture). One would expect the cut-off to occur at around 65% to 70% of the stroke, so this should occur at about 100 to 110-degrees.
- The exhaust also opens before the end of the power stroke (wasting steam) which could still be doing useful work. So, early again.
- The exhaust stays open for 176-degrees, the idea is that it should close before the end of the stroke to provide some cushioning. So, we might argue that this one event timing is perhaps OK.
We can see this is very similar when we look at the bottom end of the cylinder (as we should expect) but the diagram is 180-degrees shifted.
When we look at the same data for the notched up diagrams we see:
- There appears to be little reduction in the admission duration at the 20mm notched up position (movement of the reversing lever). So we are not getting the earlier cut-off we might expect. However, at 40mm there is some sign of the admission duration reducing.
- The bad news is that the admission timing appears even later (more negative LEAD)…
- The exhaust timing is similar, OKish
So, overall, we can see that the reduced LAP caused by the oversize ports is really getting in the way – John M has calculated that we really need a LAP of around a quarter of an inch (even more than the drawings would provide if they had not been incorrectly dimensioned) so a new slide valve is on the to-do list. <sigh>
These pictures paint a really bleak story. It doesn’t even warrant complete discussion, the leaking piston valve and iffy geometry of the design (and my machining) appears to result in occasions where the inlet and exhaust ports are open at the same time.
For example, look at that first HP graph (Ahead, Full Gear, HP Top) (1st in row 2), you can see that at the beginning and end of the stroke the inlet and exhaust overlap for 34-degrees… this starts to explain why the boiler can’t keep up, and why the engine reverses so easily….
So, again we are going to have to remake the liner and bobbin of the piston valve to attempt to bring the events into some sort of order. <sigh>
This mechanical mayhem perhaps also explains why we need such a big condenser, as quite a lot of the time we have the boiler almost directly connected to the exhaust!!!
I’ll leave you to browse the pictures and comment if you think my analysis is wrong, and lets hope at some point in the new year we will have a set of new components which will improve things, and then we can show those wake-board maniacs what 3-tonnes of steam launch at speed feels like!