"We can build that!" The Union Chronicles

[RolandMM]

"Oh sure, we can build that."

I was pretty sure Marvin did not have a mouse in his pocket so that meant the  "WE" in "WE can build that" meant Roland & Marvin.  The overworked, behind schedule, overbooked and exhausted team of Morrison & Marvin Engine Works now had a new project.

The "THAT" I understood perfectly. Of course Marvin was correct, we could build the engine, but years ago I had declined this incredibly interesting project because time & resources are finite and I already had enough projects lined up for 3 lifetimes.

We were in a building full of one of a kind California built engines.  A longtime California Engine Collector has amassed parts and pieces of countless engines and made the missing parts to restore many.  He is a good patternmaker and works well with George at the Silverton Foundry and years ago he decided there needed to be a special model engine.  One that showed the best features of a particular early manufacturer.

Features that showed unique design like curved spoke flywheels, horizontal governor, top exhaust and vertical side shaft.

Please understand I am over the age of 12 and therefore cannot think in Binary.  I grew up in America when it was a free country.  The cars were made of steel and had engines that roared. The cup holder set close to the driver and wore a miniskirt. I consider the computer and cell phone to be a problem not a solution.

I am barely able to write a draft in MS Word and put in pictures where appropriate but the pictures will not copy to MEM.  The way around this is more complex than I  care to deal with so I will try writing until there should be a picture, inserting the picture and continuing another day.

Here is a picture of the castings sitting on my stainless steel fab table.

Roland

[Alan Haisley]

Any way that tells the story works for me. 

Alan

[Dave Otto]

Hi Roland

You are doing fine; putting my thoughts into words is something that I always struggle with.

Looking forward to hearing more about this interesting one of a kind project; more pictures too!

Keep on posting.

Dave

[tvoght]

Roland,
  I enjoy your writing style and look forward to updates on what appears to be a very interesting project.

--Tim

[RolandMM]

Alan, Dave & Tim, thank you for your kind words.  All my engines have a story and it is an integral part of the engine. How can you appreciate something when you don't understand why it is there?

An early​ California Engine, the Regan Engine, was made in San Francisco, California in 1879 to power sewing machines.  What made his engine so great is Daniel Regan invented and patented, in 1884, the "Make and Break" ignitor which became the most popular type of engine ignition for many years. In 1889 The Regan Vapor Engine Company was incorporated to manufacturer Regan Vapor Engines and Vapor Engine Launches.

In 1889, Daly, Cavanagh & Barrett organized  the Pacific Gas or Gasoline Engine Company for the purpose of manufacturing Pacific Engines.   Regan sued because some of the features were also on the Regan Engines under his patents.

After many long court battles and appeals with both sides winning and losing, the companys were unable to continue so the parties involved agreed to join companies and build engines together.  The name of this company, founded on May 16, 1892 was UNION Gas Engine Company.

The WHAT was a 7/16 scale set of castings based on a 1901 Union but with features from earlier engines.

There were 3 sets of castings, one which had been partially machined.  No plans, no dimensions, and hardest of all, no actual engine to measure. In fact, as this was the combination of several engines, no pictures exist because this engine didn't exist.

No problem, impossible things just take a little longer.  We understand engines and of course we can build a running model from these castings.  But why just build an engine?  It is only twice as much work to make an exact replica which was the original intent.

Union engines are exceedingly rare, especially the early ones.  Most are quite far away in the foreign country of California. (If you disagree, explain why I must go through Customs and Immigration when I arrive from Oregon, Nevada or Arizona!)

Marvin and I had just returned from the "California Gold" Show (you must bring a California built engine for admittance) where there were prime examples of Union Engines.  However at that time WE did not know about THAT new project.  Had we known, we would have taken countless pictures and measurements.

We borrowed a 1905 Union catalog which shows a later model engine.  It showed some details we needed and we called Union owners for pictures from other angles. We were able to borrow some parts which were applicable to our model.  Nothing beats a "hard drawing".

I will attach the picture for each of you to study as we build the engine.

Roland

[RolandMM]

The first thing Marvin & I do when starting a new engine is to define the project.  This requires a parts list developed from various pictures and knowledge of how these old engines were built. Parts lists are available for many common engines which can be a starting point but usually the engines we build have very little literature still existing.

Then​ we studied the castings.  I quickly decided that Marvin should take the heads home with him.  We had been told the cores were different in each head and it turned out none of the coring worked.

Parts with long leadtimes were dealt with first.  I had no material the correct size so I made a split pattern for a billet we would need for the cylinder sleeves and took it to George at the Silverton Foundry to be cast in iron.  I did not take many pictures of the pattern so let me explain the fine points so when you need a billet it will be easy.

Greensand foundrys are happy with a matchplate that fits their flasks.  However unless there is a quantity to cast or it is a complex part, a split pattern will suffice.  Some foundrys will work with a solid pattern but for a simple billet a split pattern is easy to make.

Start with a piece of clear grained wood, plane two faces so they are flat. Then drill a blind dowel hole close to each end.  Make one 1/2" and the other 5/8" diameter.  Drill matching blind holes on the other piece.  Then glue dowels into one side and fit the other on it tightly.  The reason for different size dowels is to make it a challenge for foundry workers to put it together backwards.

Put a hose clamp around it and make a shoulder for machining the pattern and also when you machine the billet it is a good place to grab so you can machine the whole piece in one pass.  Note in the picture I protect my lathe from wood dust with papers.

Foundrys have stock sand cores but usually they are more a problem than just drilling the hole later.

I found a few pictures of the sleeve and rough machining it.  The lathe has a 4" spindle hole so it is easy to center drill both ends and true up the shoulder. Then chuck the billet on the shoulder, support with a center in the tailstock and take off the scale .

Then I chucked it and rough drilled the cylinder bore. Drilling with a 1-1/2" drill bit is quick because the tailstock attaches to the carriage for power feed.

Roland

[sshire]

Roland
Your process is absolutely fascinating. Looking forward to the rest of the adventure.

[metalmad]

Hi Roland
I'm looking forward to this one
Pete

[RolandMM]

Building a model engine requires a few deviations from a real engine.  The cooling water passages in the cylinder doesn't scale well.  New technology may be able to do it but for practical purposes we do not attempt to cast water jackets in place.

The practical way around this is to sleeve the cylinders after relieving an area for the water passage.

On the Union Engine the sleeve was so thick I decided to make the cooling water cavity in the sleeve.

After being drilled on a larger lathe I chucked the sleeve in a precision lathe to bore it and machine to size.

I purposely made the sleeve 1" too long.  This acted as a chucking lug for machining and assembly.

First I bored the sleeve to within .050 of finished diameter.  Then careful measurement of the cylinder bore was taken. The object is to get a tight fit between the cylinder and the sleeve.

Roland

[Dave Otto]

Hi Roland

I have been enjoying your Union story and the construction of this interesting early CA engine. The Regan, Pacific, Union history has probably got to be right up there as one of the most interesting in the US internal combustion engine development.

Please keep posting more progress photos as they become available; it will be wonderful to see this rusty pile of castings become a beautiful running example of our early engine pioneer's work.

Dave

[smfr]

This is shaping up to be a fascinating story, and I'll be following along, Roland.

Thanks for posting!
Simon

[RolandMM]

Simon & Dave, Thank you for your interest.  The world has no idea of the variety of engines developed and manufactured in California in the 19th century.  Besides the Regan, Pacific & Union there was the Garratt, the Oriental which later became the Corliss Gas Engine Company and the Hercules (not the common Hercules).  Then there was Adam Schilling & Sons which morphed into The Golden Gate Engine Company.  Even Syranus Standish built engines as did Daly & Cavanagh of the Fremont Iron Works.

Most California built engines are old and all of them are rare.

On with our Union model now;
Before assembling the sleeve into the cylinder there was some work to do on the casting.  I don't have pictures of the cylinder machining but I did find one showing how I drilled the passages needed.  The cooling water inlet is on the side of the cylinder and exits through the head.

Union engines had a pre-heater for the air.  This was typical of early California  engines with just slight differences, The Pacific had a diamond pattern of small holes for air inlet but the Union has a round hole cast opposite the boss for the mixer so the air traveled around the cylinder, in a passage similar to a water jacket.

All these holes go only through the cylinder wall so it made sense to drill them before inserting the sleeve.

I set up the cylinder on an angle plate squared with the table.  Height at this point didn't matter so I used some firewood to support the casting and held it in place with a couple bolts.  As the top of the casting was machined square with the bore, everything was true except with the mounting flange holes.

This was easily solved using an apprentice pipefitter technique of "two-holing".  Simply put 2 cylinder support columns in the top holes and level with a Starrett 98.

Then find center from the machined flange at the top of the cylinder and you know exactly where you are and everything is true.

To finish, I machined the mixer boss flat and drilled the holes.  Oiler holes in the bosses on the side need to go through the sleeve also so will be drilled later on a similar setup except a square will be put against the vertical mixer boss to make the holes at right angles.

Roland

[RolandMM]

Now that the cylinder casting has been machined as far as possible at this stage and the sleeve has been bored to within .050 of the finished bore (1.7500) it is time to consider assembly.

Careful measurements of the casting bore were made so the outside of the sleeve could be turned to fit.

After reading on this forum, I understand several inexperienced machinists are eager to learn what to do and especially what not to do. For them I will explain more than I know about fitting sleeves.

I know splendid model builders who use Loctite to hold a sleeve in a cylinder and it is quite successful.

I was an apprentice boy when a Loctite salesman showed up and told us how revolutionary his new product was.  I remember him saying someday it would even replace arc welding.  We all got a free sample and tried it and weren't impressed. It is probably much better now and I have used the hydraulic sealant with success but I am not one to put things together with hot glue and dryer lint.

I feel a sleeve in a cylinder should be an interference fit.

There are 2 kinds of interference fits: press and shrink. I like shrink fits because they are so simple and effective. The same interference with a shrink fit is 1-1/2 times stronger than a press fit.  Not as much chance of damaging a thin part and you don't need a press.

Machinery's Handbook gives more details than you want to know but a FN1 fit is good for model engines. Too tight and something will crack. Too loose and the water will leak.

Shrink fits are used in industry.  My largest shrink fit experience was on a liquid sodium pump coupling. The shaft was about 10" diameter and the coupling hung from a chainfall.  The interference was .0080 and 4 rosebuds heated the coupling for about an hour being very careful to equalize the heat.  When the tempstick showed the correct heat the coupling was lowered down to the stop clamped onto the shaft and left overnight.  The next day the coupling was connected and the shrink fit, without a key in the shaft, handled the 2500 horsepower from the motor.

Another time I watched a factory rep shrink a gear on a shaft.  It was a timing gear on a 2000 horsepower emergency diesel.  Because of the heat treating in both parts the gear was only heated in boiling water but the shaft end was cooled.  He used dry ice and alcohol which worked fine.  There was liquid nitrogen available in the plant but for some reason he chose not to use it. It is the difference in temperature that allows a shrink fit to be assembled.

Now back to the simpler job of shrinking a sleeve into a casting with a 2.6250 hole.

Those who know all about this can skip to the pictures;  the rest of you pay close attention because when you try this you only get one chance.

I wanted a .0030 interference but due to my lack of ability ended up with a .0035 interference.  Those of you who have tried to machine off .0005 know it will get away and pretty soon it will be a Loctite fit.

You can't test fit the pieces because they won't fit.  The best you can do is make sure there is no burr and break the edges so when you assemble the pieces they don't hang anywhere, because if they do, that is where it will stay.

I use the kitchen oven (stinks up the kitchen) and leave the piece in at least an hour to make sure it has expanded.  I used 450 degrees because it was a bit tighter than I planned.

During that time I fasten a stop to the sleeve so it could only go as far as it should.  This is a very important step!

Then I wrapped the casting in an old towel to keep it warm and hustled to the shop where the sleeve was still held in the chuck and slid it on.  Had some coffee and wrote a piece for Model Engine Maker bragging about what I had just accomplished while it cooled.

Roland

[90LX_Notch]

Nice work with the shrink fit Roland.  This continues to be a great build.

-Bob

[Dave Otto]

There is always a bit of a pucker factor when assembling shrink fits, but is sure is nice when everything slides into place (maybe forever!).

Thanks for the update Roland.

Dave

[RolandMM]

Bob and Dave, thank you for your encouragement and comments.  Now that the cylinder is permanently attached to the sleeve the bore can be finished.

It is common for the sleeve bore to have tight spots where the shrink area is located.  That is why I leave it undersize and finish it after the shrink has cooled.

Remember the sleeve is still in the chuck just as it was machined with the cylinder shrunk over it.  That eliminates the need to rechuck and assures the sleeve is exactly centered.

Wanting a 1.7500 bore to fit the piston ring, I used a boring bar to bring the hole to 1.7495 ending up with light cuts and slow feed for a good finish.  Then to get a true bore to the exact size I used a homemade lap with 350 grit abrasive.  I dreamed up this lap design years ago and probably got the idea from something I saw and used available materials.

Leaving the sleeve / cylinder still chucked up, I start small so the lap can adjust to the size and lots of oil on the lap and on the tailstock vee-ways.  Then add some grit and slide the tailstock so the lap goes the complete length of the bore.  Lock the crossslide so it allows the lap to only protrude halfway out the back of the cylinder.  If it extends out too far, this goes quickly from a fun easy project to a total disassembly, cleanup, and start all over nightmare. 

Tighten the lap until a drag is felt and by working the tight spot more, soon it will feel the same over the complete bore.  Only then take a measurement and determine how much farther you must go.  The first 0.00025 comes out quickly and as more high spots are removed, the bore gets a polish and removal gets slower.  Usually by the time all the tight spots are gone the size is close.  Finish up by moving the tailstock rapidly back and forth to leave a cross hatch so the rings will seat easily.

NOTE:  In my eagerness to take a picture I did it before I put down newspaper to protect the lathe ways.  Any newspaper will work but use several layers -and- if you happen to have tomorrow's newspaper you should read it first - (and then tell me how you got it).

The lap is simple to make and I have several sizes.  First make the stem, threaded on the end.  This stem works well with a shoulder and can be used with many lap heads.

Then cut a section of copper tube a couple inches long that is about the diameter of the bore.  Slit it lengthwise and if necessary, remove a piece so it will go inside the bore.

Then drill and tap a steel bar to fit your stem and slide inside the copper piece.  Drill and tap a hole, centered front to back, almost through the side of the bar and tap for a NF set screw.  I used 3/8 - 24 for this lap.

Cut the bar lengthwise through the center almost to the stem end and cut the half not tapped, to the slit so it is a separate piece.  Enlarge the tap drill hole in that piece to a body drill size so the set screw can fit inside it and hold it in place.

When you see the pictures it will be as clear as Carborundum grit.  The tapped hole piece fastened to the stem holds the setscrew which holds the loose piece in position and also acts as a crowd screw to enlarge the lap.

Now is time to silver solder the leading edge only of the copper shell to the tapped block so the turning direction will wrap the copper around the loose piece.  Remember to heat the whole copper piece to anneal it as the grit embeds easier.  Drill a hole when done to insert the set screw when assembling and you are ready to lap.

After lapping, I faced the end so it is perfectly square with the bore and broke the inside edge.

Now, and only now, is the time to release the chuck and continue on with the rest of the engine.  Sure nice to have my lathe back!

Roland

[marv in minn]

and now i know how you got such a nice fit on the piston 

[ths]

Thanks Roland, a very useful description. Cheers, Hugh.

[Dave Otto]

Hey Marv

Nice to see you made to MEM! Be careful you just might be asked to post an intro in the Introduce Yourself section.

Roland, thanks for the lesson on making laps; one of the other guys on the forum (Ramon) makes his laps much the same way. 

Good to see a nice progress report on the Union.

Dave

PS. the Regan book should be delivered tomorrow.

[RolandMM]

Marvin!  There you are.  I have been doing all this work and you are getting all the credit, just backwards of the way it should be.

Now that you are here I will tell what you have been doing and you add the details I leave out.

One of the parts with a long lead time was the nameplate.  We had borrowed a very early style and did a rubbing. I scanned it and sent to Marvin who patiently traced it so he could design toolpaths to cut a mold.

Now all the lettering needs to have 1 degree of draft so the wax can be lifted from the mold.  Since most nameplates are curved to fit a cylinder, we long ago decided to cast the wax flat and then curve the brass nameplate.  On larger pieces we have curved the wax before sending it to the foundry but usually it is easier to bend the brass.  Note Dave at the Idaho Division did that when he built his Pacific which is written up in another thread.

We commonly also make a buckle when we make a nameplate.  In this case Marvin made 2 backs for the mold.  One flat for the nameplate and the other rectangular for the buckle.  The picture will explain it well and you can see the design cavity will fit either back.

Since this was a very short run (3) I did not bother to cut vents and just used strips of tape to get the mold to fill.  The pins are the cores to make the mounting holes.  It is easy to locate holes when making the mold but difficult after the tag is made.

Roland

[marv in minn]

in order to make a tag or buckle mold, i start with a photo, rubbing or ideally an original tag .
then comes the process of finding a font that is close and then modifying it to match the original lettering.
this entails changing the shape, spacing and elements of the letters and adding flourishes that complete the design.
this process can take up to 40hrs like the Kansas City Lightning or Samson tags.  the Union and Pacific tags were fairly straight forward.
i have to make sure the letter size and spacing will be able to fill when Roland injects the wax and be sharp to read.

i use Surfcam 3 axis CAD_CAM program to draw and create the tool path.
for the engraving, i use a carbide single flute cutter ground on my Deckel cutter grinder with 2 to 5 deg draft.
i have to allow for the shrinkage of the wax (.017) plus the brass(.006) which comes to .023 per inch.
i also have to cut the mold in reverse so the end product will read correctly. (Dave, be quiet  )
i've heard that mold makers are backwards people, i've only been one for 44 years and still walk toes first. 

to cut the Union tag cavity, i used cutters with diameters of .187, .062, .012 and .007, the programs ran for about 6+ hrs.
the back plate has both locations for the plain tag and a buckle back where the tag cavity is the front of the buckle.

i now let Roland inject all of the waxes.
i tried it one time and wound up spraying his garage ceiling and my beard with green wax when a core pin blew out. 

[RolandMM]

Marvin,  Glad to get some details and maybe it is not as easy as you make it look.  Using a .007 cutter may make readers think this type of project can be done on a desktop mill.  Send a picture of the mill you use to cut delicate parts like nameplate molds.

The green wax story would almost make a new thread .......

Roland

[Dave Otto]

The green wax story would almost make a new thread .......

Or at least a little comic releif
The quality and detail of investment casting still amazes me.

Dave

[marv in minn]

Warning!! this is a working shop, not a show case

i use an older Milltronics Partner1 that was rebuilt to factory specs.
it uses Cat 40 tooling and can run up to 5000rpm.
it is one of the only full size CNC machines that can run on 220 single phase so i don't need a phase converter for it.
i also work on full size engines and machinery so this works best for me.

so far, i have not found a benchtop CNC that satisfys my needs

re: green wax,  at least everyone else thought it was funny   

[RolandMM]

Time to catch up on the Union Engine.

As previously mentioned, parts requiring castings were addressed first. One of these were the governor weights.  This was an easy mold and Marvin turned it right out.  It helped that we had been able to borrow a complete Union governor so had a good pattern.

Marvin cut the mold in an aluminum block and I injected wax and made several patterns.  Then off to the foundry, carefully wrapped in dollar bills for padding, and a few weeks later the bronze castings arrived looking good.

There was some machining to be done on them.  We had decided against cores because this was a short run.  A simple fixture with a hole to hold the ball and a clamp plate worked well.  It was fussy to get everything square and true and in addition to drilling holes, the dog had to be undercut so the spoolpiece didn't bind - but then I am getting ahead of myself.

Roland

[Alan Haisley]

It's amazing how much cleaner investment castings look than the cast iron, aluminum or brass sand castings. Ignoring material costs, how do the prices compare for the two processes?

Alan

[Dave Otto]

Hi Roland

Nice to see an update; looking forward to the next one.

Dave

[tvoght]

It's good to see an update Roland. I assume the castings were not received in the same packing material in which the patterns were sent...

--Tim

[RolandMM]

Alan, Dave & Tim,  glad to know someone is watching. The lack of comments makes me think no one cares.

Investment casting is expensive.  With green sand you use a pattern on a matchplate, usually made of wood and the foundry charges by the flask.  In other words whatever you can get on the matchplate is included.    I can get one (1) flywheel, bearing caps, a cylinder head, and other small parts on a matchplate.

With investment castings, there is the cost of the wax, now $8/pound + shipping, the postage of the patterns to the foundry and the postage back needs to be included.

Then there is the mold.  We make ours of aluminum because we can.  If you hired it made it would be thousands of dollars.

Now Alan,  to get to you question. My foundry charges $40/flask for cast iron.  The investment foundry charges $160 for one (1) flywheel.  Same cast iron, same size but there is a world of difference in detail and surface finish.

Morrison & Marvin Engine Works uses a lot of investment castings and our casting sets are exquisite but not cheap.  We strive to have the quality all other casting sets are compared against.

Here is a picture of the Union Governor mold with a wax pattern.  Also the finished Governor Balls showing all the machine work.

Roland

[Roger B]

Thank you for that interesting oversight of casting methods and costs    Quality costs remains true 

[ths]

You're right about the investment castings Roland, they are exquisite. Thanks for that insight. Cheers, Hugh.

[fumopuc]

I like investment castings also. Roland, thanks for showing it.

[RolandMM]

Roger, Hugh & Achim,  Thank you for your comments.

Now that the governor weights are made, there needs to be the holder and assorted pieces.

We borrowed a full size Union governor so had good details. It was a simple, fun exercise in machining.  If it looks complicated, maybe it is because we need three (3) so will make four (4), because it is easier to hold and surely will find a use for the extra one someday.

First was to find some bar stock that was suitable.  A piece of cold finished 1018 was on hand so a couple pieces were cut off.  It was decided to make two (2) sets of two (2) and save some steel.

After the blocks were squared up the center holes were drilled and reamed.  This step assured everything would be true with the bore.

Then the sides were faced to finished width, one side cut to finished length, and that side was grooved to fit the governor weights.

All this was very accurate because I knew exactly where the center hole was because I had just drilled it.

Then same operation on other side and repeat on the next block.  We are almost done.

Roland