Using a steam indicator

[k4kfh]

I took a look at those pictures and my recommendation stands, take the 4 screws out and make a dummy cylinder head that the pressure transducer screws into.

Sure it will look goofy but who cares, using some very tiny tubing is going to introduce a host of new variables and problems that you don't want and could create all sorts of issues.

From reading your postings my impression of you is that you are a young engineering student who has picked an interesting learning project to test your new found skills in measurement, instrumentation and programming. I think this is pretty cool, should be commended, you you will learn a lot and get to put something neat on your resume to talk to at an interview for an internship but if you are going down this path why not go the whole way and do it right?

I get called all the time by my customers when a machine isn't working correctly to come out and do all sorts of crazy measurements. My recommendation if you want to teach yourself some engineering skills is to go the whole way and think of this as a customer's machine that you are called out to investigate and have 2 options;

1. You show up at their facility with a temporary cylinder head with a pressure transducer screwed right in that you made ahead of time in your company's machine shop. You unscrew the cylinder head set up your instrumentation run the tests and prove your ideas, and show them it will work. When you are done at the end of the day you put the machine back together the way you found it go back to the office and produce a nice report showing them the performance of their machine.

2. You show up, with all sorts of tools, perform major surgery on a customer's machine, drilling a hole or two that you can not undo that physically alters their machine to install a setup that will impart a whole host of issues and uncertainty that could taint your results. Then you put your completely unproven software and DAQ system to the test only to find out that nothing is working and you are forced to walk away (or walked out of their plant) with results that are inconclusive at best.

Which path sounds better to you?

In my experience in engineering always pick the simplest, easiest, lowest risk route first that will not impart a whole host of variables beyond your control over one that looks pretty! Even if the requirements are to eventually make it look pretty and elegant and over complicate it, almost everyone will give you the option/funds to try the the quick and dirty, reversible cheap proof of concept route first!

Your impression of me is dead-on, the only trick is I am an electrical engineering student, so it will take some serious linguistic gymnastics to put this on my resume.

I think you're right on, I had not considered how easy it would be just to swap the cylinder head and put the modified version on. The only remaining question is whether the pressure sensor I had in mind will hold up to steam temperatures, but for $15 I guess that's worth the gamble.

[k4kfh]

I haven't worked with electronic transducers, but my impression is that they take little volume and movement of the fluid to make them register, compared to an indicator which uses a piston which has to be displaced. If that's true, a small tube would work just fine. With minimal volume and flow, there will be minimal friction and inertia of the fluid in the tube to affect the measurement. Ensuring that the tube is full of water will avoid lost motion due to compression of any air in the line to the transducer. Because there will be no exhaust from the tube between the cylinder and the transducer, it's unlikely that you would get live steam at the transducer. If would all condense in the tube.

This was the other reason I suggested the tube, a small tube would almost certainly fill with water.

Assuming the proof of concept works, perhaps I could put the cylinder cocks on the end of the small tubes using some sweat unions, rather than directly underneath the cylinder? Of course all of that will depend on how much it will cost me to get another machined cylinder from PMR, which will probably be a pretty hefty sum considering the machining complexity of that part.

[RET]

Hi k4kfh,

Both apm & Dick Morris are offering you very good advice. One way would be to mount the pressure sensor on an extra cylinder head, do an internet search for "Arduino miniature linear position sensor" and you should come up with the DFRobot website which has a potentiometer position indicator for $5.90. You can make a temporary connection to the crosshead on your steam engine with it. Although it doesn't say so, I assume that price is for multiple pieces, but you might be able to wheedle a sample from them.

As I said previously, use both of these together with an arduino board to do the whole project via a temporary connection to your little engine. By doing it this way you will learn a lot (mechanical, thermodynamics, electrical & electronics, programming, etc) and this is also good practice for the "real world" when you graduate. If you aren't familiar with the arduino system, look it up. While I haven't had occasion to use it myself yet, I know it is very versatile and relatively inexpensive.

When you are finished, remove all the extra "stuff" and then you will be back to a nice working model, but you will also have all the data you collected while the engine was "instrumented." Unfortunately from your pictures, I think your engine only has slip eccentric valve gear which means you can't do any form of "notching up." You can still do an indicator diagram, you just can't see the result of changes in valve cut off. Charlie Dockstader's valve gear software is well worth your attention. You will learn a lot from it and it even includes the slip eccentric valve gear.

What Dick Morris says is also true in that you could use 1/16" copper tubing to the cylinder head as a permanent connection to the pressure sensor together with another rod to the crosshead that drives the position sensor out of sight somewhere.

Finally, if you are an engineering student, you should have access to a machine shop facility somewhere on the university campus. If you supply them with a drawing, they might even make the part(s) you needed.

Hope this helps.

Richard Trounce.

[apm]

Dick,

Yes you are right about pressure transducers, we do a ton of hydraulic measurement and use this rubber hose that maybe has about 0.025"dia and it works great. I always marvel at how well 30ft of this hose still gives me good measurements. I have seen over 4ft delays maybe on the order of 2-4miliseconds in response time delays but this is all with an incompressible fluid which steam is not. So yes, tiny diameter tube would work but I think the trick is going to be getting it fully bled out so you are sure it is 100% full of water not some steam water, cylinder oil, and air combination that will wreak havoc on the measurements.

Then notice how k4kfh is talking about sweat soldering the thin tube, what is the chances that he wicks up a small blob of solder into the tip of the tube? Been there done that, and now the thing definitely won't measure worth a damn!

All this leads me back to teaching the young lad the correct and professional approach which will go a long way towards talking about in an interview, start small, and simple and work your way up to complexity. In other words;

Step 1 does my position measurement work,
Step 2 does my pressure measurement work, (take your transducer hooked up to an air regulator and see that it all registers on your computer before even hooking it up to the engine
Step 3 does his logger plotter work when it is run on air with the quick and dirty setup. (No need to screw with running a boiler, debug software code, and worry about steam burning out your $15 transducer all at the same time.)
Step 4 does the thing work when run on steam?
Step 5 hook up copper tubing and try the elegant permanent setup on air does it still work?
Step 6 hook up copper tubing and try the elegant permanent setup on steam does it still work?

In engineering until you have confidence in one thing it is critical to avoid dumping too many variables into each of your processes. I have seen all too often in my career and fallen victim to it too myself where engineers inadvertently bite off more than they can chew and wind up spending more time spinning their wheels then they ever would have done if they started simple and built their way up to the complexity. So yea while I am not saying copper tubing won't work I am saying don't do it right away. To k4kfh, the common sense things I speak of above isn't really taught in engineering school so don't feel so bad if you are not there yet but is crucial to becoming a good engineer in my opinion.

Going back to something else k4kfh wrote, don't worry this will look great on your resume. Today's engineering market demands more than ever multidisiplined engineers than ever before. Where I work our electical engineers are expected to know software programming and have a good enough idea of mechanical engineering principals and measurements. Our mechanical engineers are also expected to have a good working knowledge of electronics and software. Go back to that article I posted on Harold Crouch instrumenting the steam locomotive. The day and age of 10 engineers out to preform an experiment like that are long over. Today you would have 1 people out there on that locomotive maybe 2 expected to do the work of those 10 guys you see in the photos, and they would need to be well versed in the mechanics of the situation, software, and electronics all in one.

So yes I know in school you are an EE but in the real world for most engineering jobs you are far more valuable if you can cross over into the mechanical and software disciplines too!

[NP317]

When I was instructing at the Univ. of Washington Mechanical Engineering College, all of our undergraduates were required to take several electrical engineering/electronics classes. As stated above, this makes better engineers.
I used to consult on their projects, and often found the need to help them simplify their goals.
Interesting to read the same/similar info above, from apm.
Some things do not change.
~RN

[B&OBob]

Hi -- In reply to your question of 10/24 on what I meant by an earlier comment about "meaningful measurements," I was referring to unavoidable small influences that deteriorate measurement accuracy, sometimes called "parasitic effects." They might be caused by clearance volume, piston or valve blowby, lost motion in valve drive, condensation in the cylinder and steam passageways, or the many other small, but non-trivial losses that can occur in a steam engine. The influence of parasitic effects is generally greater on small engines than larger ones so, in the case of your PMR 3 with a displacement of only 0.147 cu.in., they will have considerably more detrimental effect on measurement accuracy than on an ST 10H/10V with 0.33 cu.in, or Hall's 1.27" x 2" of 2.55 cu.in test engine. with 17 times as much displacement.

A simple GHoogle search for testing miniature steam engines yielded the URL for the paper mentioned in my PLS article. It can be found at: greenloco.com/hall01.pdf. Another useful reference is: greenloco.com/resources.htm where the schematic is shown for Hall's Digital Indicator, a very simple circuit using a few gates and an Op-Amp. Note that I have not used the Chaski designation for direct access to URLs, as it does not support the addresses listed; they must be entered independently in your browser as listed here.

In his detailed description of his tests, Hall shows electronic indicator cards he obtained that are so close to theory it's scarey. Admittedly, he made a special engine for his measurements, nevertheless, he has produced cards of such exquisite quality they will serve as the benchmark for future experiments involving miniature steam engine indicator cards. Good luck with yours!

B&OBob

[k4kfh]

Thank you all for the advice! Looks like I have plenty of reading to do.

Regarding my access to a machine shop, I'm working on that, but currently I do not have access to a shop until I take the intro level machining/manufacturing class. Further, I have no experience in CAD, so I'd have to get better at that before I could produce any meaningful drawings. But this is on my to-do list (I am only a freshman so I have time, I'm going to try to get into that class next semester).

Many people have pointed out that I could use a linear potentiometer or encoder, and this is true. However, I think it will be much simpler to use a rotary encoder connected to the crankshaft with an off-the-shelf shaft coupler. This engine's crankshaft is unsupported on one end, so a rotary encoder could be disguised as a bearing by simply making a bracket for it with some hardware store bar stock. This also has the advantage of needing no direct modification to any of the engine's parts, as apm suggested, it is simply a "bolt on" solution. And it works with any engine, in theory, because all you'd have to do is set what position corresponds to top/bottom dead center and let the microcontroller do the math to figure out where the piston is in its stroke. So that's my rationale for not using a linear potentiometer. I am quite familiar with the Arduino system and I'm in the middle of an intro-level C programming class anyway, so the programming for this should be trivial.

apm, I had not even considered the possibility of blocking up the copper tubing with a drop of solder. Yikes.

To everyone who has suggested the valve gear simulation software, WOW, you were right. I started playing with it yesterday evening and it is extremely educational.

Again, thank you to everybody for the advice and reading, I really appreciate it. I have access to a lathe/mill at home through a friend, so around Christmas I may be able to fabricate a cylinder head and crankshaft encoder mount, which would give me something I can toy with in my apartment. I will definitely keep you all posted on the progress, this is shaping up to be a highly educational project.

[B&OBob]

The commment about having to learn CAD before a meaningful drawing could be made was truly astounding. It ignores the reality that every prototype locomotive was designed and produced without the "benefit" of CAD! That every airplane in WW-II (whenever that was) was designed and built without CAD! The same goes for every achievement in engineering introduced before invention of the transistor and development of electronic computing. Pencils and paper (and, perish the thought, ink on linen) served us rather well for designing live steam locomotives and making those designs available to other builders. CAD is now essential for a meaningful drawing? What have we come to?

Maximum accuracy is achieved when measurements are made as close as possible to the elements being measured. Earlier advice to use a linear transducer on the crosshead was well conceived. By attempting to measure piston displcement indirectly with a rotary encoder on the crankshaft will introduce errors from clearances in the croshead guide and the many intervening bearings. Additionally, as acknowledged, there is the further need for complex calculations of crank angular position relative to piston displacement. OK, IknowIIknow, "there's a program for that," but multiple physical dimensions of the engine's drive mechanism are required for that. Are all those dimensions accurately known? And most important, why go through an overly complicated exercise at all when croshead guide position provides direct, losslesss data of piston position?

Good engineers achieve required results in the simplest way.

B&OBob

[JJG Koopmans]

Having done this exercise with a real steam locomotive, I might add some comments after reading the posts.
As for the mechanics, make a new cylinderhead. Hole it, put the copper tube in, expand on the inside and solder it tight on the outside. My experience is that the condensed water in the tube protects the sensor.
As the water is incompressible the pressure measurement is instantaneous.
As for the piston displacement I used a linear system which worked fine.
The real problems I found were with the sampling rate and the sequential reading of the sensors. Pressure and piston location were not sampled at exactly the same time by the Omega software I used and the hardware prevented it also. So whatever electronics you use make sure your samples/data match with time/each other.
A few years later I learned from some chaps in the UK that they had used a computer hardware add-on called Picoscope.
It displays a nice indicator card on screen and with printscreen (PrtSc) a lot can saved for future use. I am sure a modern version of Picoscope works fine as will Arduino's or Raspberry PI computers.
Kind regards
Jos Koopmans

[B&OBob]

As for the piston displacement I used a linear system which worked fine.
The real problems I found were with the sampling rate and the sequential reading of the sensors. Pressure and piston location were not sampled at exactly the same time by the Omega software I used and the hardware prevented it also.

I smiled when I read of you experiences. They mirrored mine of over 70 years ago when I attempted to get a decent card for a 4"x5" vertical stationary engine using a traditional indicator with string coupling to the crosshead. For every "good" card, there were ten worthless ones caused by random manual opening of the indicator's valve in mid-stroke, or some other issue associated with those antiquated methods. All the more credit to those valiant engineers who rode inside a temporary shack on the pilot of a locomotive to successfully take cards at 60 mph!

We have come a long way from that -- all the more reason to celebrate the success of W.B. Hall with his landmark experiments on a model engine.

B&OBob

[k4kfh]

The commment about having to learn CAD before a meaningful drawing could be made was truly astounding. It ignores the reality that every prototype locomotive was designed and produced without the "benefit" of CAD! That every airplane in WW-II (whenever that was) was designed and built without CAD! The same goes for every achievement in engineering introduced before invention of the transistor and development of electronic computing. Pencils and paper (and, perish the thought, ink on linen) served us rather well for designing live steam locomotives and making those designs available to other builders. CAD is now essential for a meaningful drawing? What have we come to?

Maximum accuracy is achieved when measurements are made as close as possible to the elements being measured. Earlier advice to use a linear transducer on the crosshead was well conceived. By attempting to measure piston displcement indirectly with a rotary encoder on the crankshaft will introduce errors from clearances in the croshead guide and the many intervening bearings. Additionally, as acknowledged, there is the further need for complex calculations of crank angular position relative to piston displacement. OK, IknowIIknow, "there's a program for that," but multiple physical dimensions of the engine's drive mechanism are required for that. Are all those dimensions accurately known? And most important, why go through an overly complicated exercise at all when croshead guide position provides direct, losslesss data of piston position?

Good engineers achieve required results in the simplest way.

B&OBob

Re: CAD
This was poor wording on my part. I meant that I needed to learn CAD to be able to do much in the university machine shop, because from what I understand, it is almost exclusively CNC. I may find out that I'm wrong, but I believe this is the case. I also have access to 3D printers, so CAD makes simple prototyping of non-temperature-sensitive parts (brackets, etc) easy. I need to learn CAD anyway. I understand CAD is not strictly necessary to do this kind of work. Thank you for clarifying though. Either way I need to learn how to either use CAD or make proper shop drawings by hand, or both.

[apm]

On the subject of CAD, the drawing you are looking for is a 2minute sketch that can be done by hand, and is probably takes no more than 15mins to make start to finish on both the lathe and mill. No need for CNC here, in fact CNC most likely would slow you down trying to use it for one part. Since this is all about teaching you real world skills that every engineer should have I guess it is probably time to add one more to your bucket list of skills to learn. That is how to work with tradesmen. To a good engineer the value of having great relationships with a skilled machinists, welders, electrical assemblers, electrical lab technician, fabricator etc... is something almost impossible to describe. The skilled tradesmen are the guys who make the ideas inside of your head come to life in the real world and if you know how to listen they will contribute immensely to the success of your project, from time to time these are the guys who will come by and rescue you when you get in deep over your head or mess up before anyone else notices. At the same time develop a reputation of being someone who can't listen, who is thin skinned, etc... and these are the same guys who will hand you the noose and help you wrap it around your neck and let you jump (figuratively speaking that is).

I think it is time you take a walk down to a few of the machine shops on campus and start learning the art of talking to and asking machinists for help. Bring your steam engine pictures of it and ask for help, if they say they can't ask for the shop tour and start sweet talking them along the way. Show some interest most of them like it, maybe they can even offer you an on campus job.

See what it would take to sneak a little "Government job" into the shop. (Government jobs are the ones that are top secret #1 priority, and no one including the boss can ever know what the parts are for, "G-Jobs" are basically slang for home projects that sneak under the radar often on the clock sometimes after work usually for the simple price of a case of beer, breakfast sandwich etc... ) If you are going to be in the live steam hobby this is a skill every live steamer should be lucky to be able to fully employ. Believe me, it comes in real handy should you ever find yourself working for a place with CNC punch presses, press breaks, CNC machining centers etc.... If you are good at talking to the guys in the machine shops at your school you should find one eager to help you with your project. University shops helping the students learn is always accepted and encouraged.

One other thought rotary positioner vs. linear, I am on the fence with this one. I think given the questions, your resources and abilities keep things simple and build in the complexity one step at a time as you go along the way. This project isn't too hard for someone with real shop knowledge and experience both in measurement and instrumentation, as well as basic fabrication techniques but it sounds like there is a bit of a learning curve. At the same time B&OBob's comments remind me that there is quite a bit of movement that can take place in a steam engine that becomes "lost" so the math of converting rotary to linear may not be the greatest. I am reminded about experiences timing a locomotive where you try to find dead centers. https://www.discoverlivesteam.com/magaz ... 08/08.html. There is a lot of lost movement there. Am I on the right track to think that this would really mess with a good indicator diagram trying to convert rotary to linear?

At the same time the fact that k4kfh isn't sure he can make a simple bracket to secure a linear transducer without learning solidworks and having a 3d printer has me thinking that rotary should get him somewhere first after which, once the whole setup is working right he can decide if he wants to go further in this learning process and put a linear measurement device in the loop? If you really want to learn some cool electronics make your own LVDT measurement circuit and figure out how to read position with one of them. Those things are pretty cool and nothing seems to beat them for speed, accuracy, and reliability.