Upshur Horizontal Farm Engine - Mike's Version

[mikehinz]

Ron, I just measured the various piston positions on my engine.  See the following:

- The skirt of the piston is about .18" below the bottom of the cylinder when at BDC.
- The top of the piston is about .31" below the top of the cylinder at TDC.
- The top of the piston is about 1.31" below the top of the cylinder at BDC.

So, my initial 'just looking at it' observation was incorrect.  And shockingly enough, the stroke appears to be right at 1"!! 

So given that we're seeing about the same thing on both of our engines, it's a good bet that they're correct.  At least that's my sincere hope! 


FYI.

Mike

[mikehinz]

Ron, one more bit of followup regarding the piston position with respect to the cylinder.

I finally did what I should have done originally and that is, go back to my CAD model and look at it!  I had made it so that it rotates correctly and so I checked the piston distances at TDC and BDC.

What I noted in the model as follows:

At TDC the piston is .313" below the top of the cylinder.
At BDC is the piston is 1.313" below the top of the cylinder and protrudes by .188" from the bottom of the cylinder. 

So that's another good data point that what we're both seeing is correct.

FYI.

Mike

[RReid]

Thanks Mike, that is a good data point.

But since my measured number at TDC is 0.001" off, I guess I'll have to turn myself in to the Machining Accuracy Police in the morning. Or should it be the Measurement Precision Police?

[mikehinz]

Thanks Mike, that is a good data point.

But since my measured number at TDC is 0.001" off, I guess I'll have to turn myself in to the Machining Accuracy Police in the morning. Or should it be the Measurement Precision Police?

Lol!  You'll just have to have all your measuring kit sent off to calibration lab and certified at tremendous expense!!  The problem has to be with the instruments!   

Good work!

Mike

[mikehinz]

And here's my next post on several small parts that I made.

First is the valve rod guide block that's unique to the hit-n-miss feature. 

I had a bit of brass that I milled to dimension and then drilled two holes as indicated on the drawing.  Shown is the first drilling operation on the mill after locating the edges of the stock.  Nothing terribly exciting here.



And the finished block on the print after a bit of cleanup with some 400 grit paper laid on a flat surface.



And here's the guide block shown in place on the partially assembled engine. 



I also made a small brass block to act as the catch stop on the valve rod.  I designed this to be square as I thought it might offer a better edge to the lever edge to engage, but that's really just a guess on my part.  I won't know until I'm able to try it out.

This part turned out to be a lot of trouble as i didn't have any small square rod stock.  I first tried to make it from a larger piece of flat brass stock, but I ended up with no good way to hold it and managed to crush it out of shape in the mill vise.  So I started over with a piece of round brass stock and made the end square in a collet block held in the mill vise.



Then I stood up the collet block in the mill vise, made sure it's square and drilled a clearance hole through it, offset to the edge, for the valve rod.



Next I turned the collet block back on its side and drilled and tapped for a 4-40 setscrew to secure the block to the valve rod and to make it adjustable.  Upshur's plans call for Loctite'ing a block in place, but I thought it would be more wise to have it adjustable as I'm sure it's going to be tricky to setup the hit-n-miss mechanism and I wanted to give myself every chance for success.



To finish the part, I simply sawed it off the rod using a slitting saw.



Here's the troublesome little block laid on top of a dime for comparison.



And the block test fitted to the valve rod on the partially assembled engine.



And here's the head with everything assembled that I'd built so far.  This picture is important and tragedy will soon follow.  Note the small flush Al plug that I used to blank off the top of the thru-hole that forms the path to the spark plug.  That a 20P British coin for size comparison.  I noted that the Queen was not smiling as she knew what was about to happen.



I'll post a few more pix before i explain the tragedy.

Here are the valve retainer caps.  No pix of making them as it was just straightforward lathe work on a bit of brass rod stock.  These are slightly different than Upshur's design as I made a small recess that holds the retaining pin in place when installed.  The Queen actually smiled and approved of these bits. 



On to the valves.  For my Webster engine I made the valves in one piece, but I wanted to try 2 piece valves on this engine as turning down a .25" rod to .0938" is extremely slow and a bit painful to me.  So these are just 2 bits of .250 drill rod, drilled and reamed to 3/32" and parts off just a bit over .10".



I drilled .040" holes .10" from the end of each to-be valve stem.  I find it easiest to hold the stock in a collet block, find center of it, find the end of the rod and then spot drill and drill thru with a #60 drill with the Bridgeport running as fast as it can go.  Here's the 20P coin for comparison as otherwise, it's hard to see.



Here's one of the valves ready for silver brazing.  I used Harris Safety-Silv 56 along with Harris flux for the brazing.  I cleaned the parts with carb cleaner and then fluxed the end of the rod and the ID of the head.  I cut a very small piece of the brazing material off the roll and placed it on the head of the valve and applied heat with a plumber's torch.  To hold the head of the valve in place, I very slightly dimpled the top of it with a punch.



One of the valves after brazing but before cleaning it up and turning the head in the lathe.



Here's a pix of 3 valves after being cleaned up bur prior to turning the heads.  I actually made qty 4 valves and ended up with 2 good ones.



Here's a pix of what happened to one of the valves when I started trying to turn the 45 degree angle on it.  I was using the wrong tool and tried too large of a cut with the valve held a bit too far out in the collet.  I managed to ruin 2 valves in a similar fashion before I figured out how to do it successfully.



This is the setup for turning the backside of the valve.  This turned out to work well.  I used my 2mm groove/turn/part tool engaging just the corner of it with the work.  I fed it in using the compound and very gently advanced the cut using the carriage.  Using this tool allowed the valve stem to be held to the maximum length possible in the collet.  And it cut very well.



And here's 2 valves successfully completed!  Note that the silver braze completely penetrated the valve head and it formed a nice fillet on this side. 



And a pix of the top of the valves.  You can see that the brazing material completely surrounds the inserted rod, so I'm pretty comfortable of their strength.



I also made the valve springs using my home made guide tool in the toolpost and a mandrel for forming the spring held in a collet.  I used whatever The Machinery Handbook specified for the mandrel OD for each wire size and the springs came out OK.  I wound the springs by setting up for threading at 14 tpi and holding a bit of tension on the music wire and simply running the lathe at the slowest speed possible for just a few seconds. 



And here are the two springs after making them and cutting them to length.  One was from .014" and the other was from .022" music wire exactly as per Upshur's spec.  I also made the head gasket on my Silhouette cutter.  I used .015" Fel-Pro Karropak gasket material.  I had tried sheet teflon on my Webster but I found that under the heat and pressure of running, it tended to soften and extrude and the seal would fail.  The Fel-Pro material seemed to work much better for me as the heat doesn't seem to bother it.  I've previously documented how I made gaskets in a separate thread so please refer to that for more details on my process.



This post is probably long enough so I'll conclude this now.

Enjoy!

Mike

[90LX_Notch]

Mike-  Did you indicate the valve stems in the collet to make sure they were running true?  That would be my big worry making two piece valves.   Parts don’t always run true in collets. 


-Bob

[mikehinz]

Mike-  Did you indicate the valve stems in the collet to make sure they were running true?  That would be my big worry making two piece valves.   Parts don’t always run true in collets. 


-Bob

Bob, indeed I did indicate them.  I'm pretty lucky as my collets are all pretty good quality and even though I bought my lathe and related bits used, the spindle runs very, very true, at least as best as I can tell with a DTI.  But I also think the process I followed, that is after silver brazing the head to the stem then cleaning them up and turning the OD, face, and 45 degree angle on the back side, the head of the valve pretty much has to be true to the stem in any case.  I've not yet shown the pix, but I did lap the valves in and pressure test them and they hold pressure on my test rig.

I have done one piece valves also and they also work, but I wanted to try the silver brazing method as turned down a .25" piece of drill rod to .093" is a bit slow and painful and I manage to ruin several in the process typically, although I'm getting better at that method.

Thanks for taking a look at this thread!

MIke

[RReid]

Hi Mike. I like your method of turning the angled faces on your valves. I made mine pretty much the same way, but my technique at that point was not as good and shows my inexperience with IC engine builds. I just set the cutter as best I could at the required angle, knowing it wouldn't be reliably repeatable but hoping for the best. Most importantly, the seat cutter is probably not at quite the same angle as either valve. It never occurred to me ( ) to dig out and set-up the compound slide, which is not permanently mounted on a Taig. I haven't tested them yet, but if the valves don't seal after lapping in, I'll go to your method for the next pair, and the seat cutter as well.

[mikehinz]

Today's post will cover disappointment, heartache and eventually success! 

So, at this point I've made the head and all the bits for it, but based on my Webster experience, I definitely wanted to test it before assembly to the cylinder.  So, I needed a test fixture.  Here's the first pix of making one.  I had a scrap bit of 1.75" OD Al rod, so I chucked that up in my 3J and turned a bit of it 1.50" to match the head diameter and then drilled thru 7/16", tapping size for 1/4" NPT thread, and here I show boring out a .75" ID relief to allow clearance for the valves.



Then over to the mill where I drilled thru 2 holes for some 6-32 studs to secure the fixture to the head and tapped 1/4 NPT as shown here.



And here's the test fixture and a air fitting 1/4" NPT x 1/8" barb and the Queen looking on in approval.  The seal side of the fixture was lapped briefly with some 400 grit sandpaper on top of a small granite plate. 



The test fixture assembled ready to go to work.  I added a bit of Loctite 567 thread sealing to the threads prior to assembly.



And here's the entire test setup ready to go.  Shop air at 120 psi is coming from the right side via the large air hose and then enters a small regulator then the lower pressure regulated air flows the regulator to the test fixture via the 2 barb fittings and some 1/4" OD clear air hose. 



I perform the test by submerging the head and test figure in a small container of water and watch for any bubbles.  I was quite saddened at this point, as the valves appeared to hold, but the little Al blanking plug used to seal the top of the thru hole for the spark plug passage leaked.  I'd pressed it in place and used Loctite 638 on it, but evidently that wasn't good enough.  In another thread I asked for advise regarding how to repair this and I received some very good advice.  So thanks to all that responded on that other thread! 

In the end, I managed to press out the existing plug by making a small, close fitting punch and pressing it out.  I then used JB Weld and created a much rougher surface by making a new small plug and made some threads on it using a 10-40 die and a ran a 10-40 tap thru the existing hole.  This in no way created a good threaded connection but was an excellent way to create a rough surface for both parts.  I also made a small head on the new plug just to more easily locate it.  Here you can see the head and plug prior to assembly.



And here it is after the plug was in place with a liberal coating of JB Weld prior to assembly.



I waited 24 hours for the JB Weld to cure and then pressure tested it again.  This time the plug held, but my intake valve showed a slight leak at 20 psi.  So I tore it all apart and relapped the intake only using just a bit of Clover 600 grit paste.  Then after assembling it one more time, I was able to get a successful pressure test!!  Huzzah!!   

I think pressure testing the head as a separate assembly was a very good thing.  I suspect that if I had assembled it without pressure testing it, I'd have gone down a dark, depressing road trying to get the engine to run as I suspect that small leak would have been impossible to spot when attempting to run the engine. 

Looking back this adventure with the head, I'd like to offer some advise to any other Upshur engine builders.  First, don't drill thru the head to connect the spark plug passage.  Just drill from the bottom side of the head and stop when your intersect the hole from the spark plug.  Doing it this way totally eliminates the need for a plug.  Also, I'd suggest NOT drilling all the way thru the head for the rocker arm post.  Drill only down to about .40" and shorten the rocker arm post accordingly.  This eliminates another possible leakage point.

I had decided that if I couldn't get this repair done successfully, I was going to make a new head and would relocate the spark plug to the top of the head vs the side.  Upshur in fact has an alternative head design that shows this very idea.  But the way he designed it, the spark plug hole is only about 50% exposed to the cylinder.  The other portion lies outside of the .75" cylinder bore.  He did it this way as otherwise the spark plug is hard to get a wrench on.  I made CAD model of an alternative design where the spark plug hole was moved a bit and then drilled and tapped at a 15 degree angle to the top surface.  This gives sufficient clearance at the top for the plug and moves to the entire spark plug hole diameter to within the cylinder bore area.  If anyone is interested in a print of this design, or the CAD model I'd be glad to post it or send it to whoever might be interested.

So, that's it for today! 

Enjoy!

MIke

[Admiral_dk]

Hi Mike

Sorry about the problems, but at least it feels better knowing that you have solved them 

As I'm sure there are several who would like to see and probably use the revised drawing you did of the head - I wonder if it would be OK to upload it to the drawing section ...?

Best wishes   
 
Per

[RReid]

Hi Mike.
The test fixture I made the other day is almost the same as yours, differing only in minor dimensions. When I get the chance I will follow your lead and do the dunk test. The only unnecessary hole in my head (cylinder head, that is) is the one for the rocker post. The post does have a shoulder, which will help a bit, and I'll goop it up well.

I'll be following your build even more while I can't work on my own!

[mikehinz]

Hi Mike

Sorry about the problems, but at least it feels better knowing that you have solved them 

As I'm sure there are several who would like to see and probably use the revised drawing you did of the head - I wonder if it would be OK to upload it to the drawing section ...?

Best wishes   
 
Per

Per, I don't think I'm comfortable with uploading my revised drawing.  Upshur was careful to mark and date his drawings with copyright information and his estate still sells his drawings.  I can't see that he actually registered his copyright(s) but I think his intent is clear.  I'm not exactly sure of the law in this case, but I certainly want to respect the spirit of his wishes.

I think I can provide enough info for others to replicate what I did with respect to the sparkplug hole with the following info:

- First change:  Don't drill thru the head for the rocker arm mounting post.  Drill only .40" and shorten the rocker post accordingly.

- Second change:  Relocate the spark plug hole to the front using the following information:
     - When viewing the 'top' of the head, locate and mark a point at the following x,y coordinates:  x=.1875:, y=.3248".  Zero is at the head center.
     - This will create a starting point 60 degrees from horizonal and .750" from the center of the head.
     - Change your setup to locate this point and to be able to create an angled hole at 15 degrees from vertical with the angle toward the center of the head. 
     - Spot face with a .375" end mill to .25" depth.  This creates a flat surface to start drilling from and creates a seal surface for the spark plug.
     - Drill thru .221" (#2 drill)
     - Tap 1/4-32 to .5" depth (or deep enough to completely accept your chosen spark plug)

And that's it!  Hopefully the above is understandable.

FYI.

Mike

[mikehinz]

Hi Mike.
The test fixture I made the other day is almost the same as yours, differing only in minor dimensions. When I get the chance I will follow your lead and do the dunk test. The only unnecessary hole in my head (cylinder head, that is) is the one for the rocker post. The post does have a shoulder, which will help a bit, and I'll goop it up well.

I'll be following your build even more while I can't work on my own!

Yes, IMO a test fixture is a necessity.  And I very much like submerging the test part in water to check for leaks.  Just like with tires and inner tubes, submerging in water will find leaks that are otherwise almost impossible to spot.

On my head, I drilled thru for the rocker post, but it sealed using Loctite, so I was lucky in that respect.  But I'd not build another head with sort of potential leak point. 

Safe travels and I hope you enjoy your holiday!  I'll keep working and I think I'm actually getting fairly close to attempting a startup.

Mike

[mikehinz]

Today's post will cover 3 smaller parts for the engine.

First up is the starter hub.  This is exactly as per Brian Rupnow's design.  I used it on my first I/C engine and it worked fine so I decided to do it again!

I started with a piece of 1.375" round stock, I think it's 12L14 as it cuts so well.  I turned the smaller OD to 1" and the larger to 1.25" then drilled and reamed thru for the 5/16" crankshaft.  Here I'm boring the larger recess in the end that will accept the starter spud.



You can see why I left the 1" OD a bit long.  I mounted it a collet block, located it and drilled the holes for the 1/8" dowel pins and drilled and tapped 8-32 for the setscrews.  The collet block makes indexing the locations easy and setting a stop makes it repeatable in X. 



Then back over to the lathe to part off the starter hub to length.  I used my favorite 2mm carbide parting/grooving tool.



Broaching the 1/8" keyway on the internal bore of the hub.  I aligned the keyway to be inline with the 8-32 set screw hole, just by eye but that's plenty close.  For the 1/8" broach, two passes are required, one with just the broach and the other with a shim inserted behind the broach.  I oiled the broach liberally with some Tap Magic EP-Extra cutting fluid.



Here's the finished hub on the print with the 1/8" pins pressed in and the key laid on top of a 20P coin for comparison.



And finally here's a pix of the starter hub installed on the crankshaft showing the keyways matching and the pins installed.



I'm not showing any of the steps for making this, but this is the ignition timing disk.  It's a piece of 1" OD Delrin rod, parted off 3/16" thick, drilled thru 5/16" for the crankshaft with a 4-40 drilled and tapped hole from the rim to the center.  I drilled a 1/8" hole near the rim and pressed in a 3/16" long magnet to act as the trigger for the hall effect sensor.  Shown is test fitting to the crankshaft.



And of course there has to be some way to secure the hall effect sensor ot the engine.  I'm using one of Roy Scholl's S/S CDi systems and he supplies a hall effect sensor with a very nice insulated housing that prevents all sort of problems.  I very much like to use that setup if there's room for it.  Shown is a small bracket made from a bit of scrap Al that will act as the holder for the sensor.



And here's how it looks mounted on the engine with the ignition timing disk installed.



And that's it for today. 

Enjoy!

Mike

[mikehinz]

The next part(s) up are the flywheels.  I tend to like to use cast flywheels and I purchased qty 2 from Martin Model and Pattern from Scappoose, Oregon.  I chose the 4 1/4" six spoke castings that I think Gary Martin calls a 'heavy pattern'.  These are a bit larger than Upshur calls for, but I figured a bit heavier wouldn't hurt anything.   

I'll only show the work on one flywheel as they're identical.  One thing about these flywheels castings is that they are extremely high quality with plenty of extra material on them and with almost no noticeable inclusions.  Plus they turned to to machine very nicely!

Here's the start of the machining.  I mounted the casting by the rim and tried to get the interior surfaces to run as true as possible.  I gripped the casting so that I could get to just a little more than 1/2 of the outer rim.  Here's the casting with 1/2 the OD just about finished.  I used a standard ccmt insert and that worked very well.



And here I turned the casting around in order to machine the other half the OD.  I used a DTI to reach the machined surface as best I could.  I also made sure that just machine side was running as parallel as possible by spacing it off the vise jaws with a piece of .250 tool steel.



Here's the start of machining on the 2nd half of the OD.



And here's the casting showing the various features on the side being finished.  I turned down the OD of the hub and flattened the end of the hub also.  I then machined the 'side' of the flywheel just enough to remove any casting roughness.  For this I used a 3mm carbide tool as I wanted to make sure I left a small radius at the various corners.



Then I changed to the 3 jaw chuck, just since it's faster and I was dealing with 2 castings, and drilled and reamed the center hole to 5/16".



And here are the 2 flywheels test fitting on the engine.  I ended up taking down the OD to 4 1/8" just to make sure that I don't have any interference with the wood base this engine will be mounted on at some point.  Note that the very, very small difference on the OD can be seen, but I'll soon take care of that.  But the flywheels both fit well on the crankshaft and seemed to be quite true.



I made a couple of 5/16" temporary shafts to mount the flywheels for their final touch up.  These are just 2 pieces of 5/16" drill rod with center holes drilled on each end.  I secured the temporary shaft with a bit of Loctite 638 and let them cure overnight.  You can clearly see the OD differences on this pix.



I mounted each flywheel in the lathe, using a collet on the drive side and a live center in the other side.  Each of the flywheels ran very true.  I took a very, very small skim cut just on the OD to eliminate the very small difference between the two halves.   They sides were running completely true so I didn't touch them.



Here's the pix after one flywheel was touched up on the lathe as compared to the one that hasn't yet been finished off.



I wanted to add a key and setscrew to secure the flywheels to the crankshaft.  I ended up purchasing a 8-32 extended (pulley) tap and a #29 aircraft drill from McMaster-Carr.  I ended up tilting the flywheel by 5 degrees in order to place the hole as close as possible to the middle of the exposed hub area.  The best way I found to do this was to leave the temporary shaft in a collet block and then place the block in the mill vise with a 5 degree block under the collet block.  Shown here is spotting the hole with an extended center drill.  I pecked away slowly at it and the drill did stay on center. 


 
Here's a view of the overall setup.  I quicked the mill out of gear turned the tap while held in the chuck.  That worked well especially since the cast iron cuts so freely.



I don't have any pix of broaching the keyways, but I used a standard 1/8" broach driven by my Dake press.  The 1/8" broach needs 2 passes, one without a shim and the 2nd pass with the shim included with the broach.  Shown a pix of both flywheels after the temporary shafts were removed and the broaching operation was finished.  You can see the slight surface discoloring from heating the flywheels to break the Loctite bond in order to remove the temporary shafts.  And you can see where I drilled into the temporary shafts so that the hole through the flywheel hub was full size all the way through it. 



As a bonus bit of information, I'll offer the following which probably everyone other myself already knows!!  When making keyways, the keyway portion on the shaft needs to be actually milled slightly deeper than I originally thought.  When I made the crankshaft, I found the top of the shaft and then plunged in .062, essentially 1/2 of the key size.  It turns out that the cut is measured from the 'full chord' point that's equal to the size of the key, so in this case, a chord of .125"  and this is known as the "M" value.  There are complete formulas to calculate this number as well as approximations and of course The Machinery Handbook discussed this but doesn't have a table that extends to very small shaft. 

The approximate value can easily be calculated by squaring the key width and dividing that by 4 times the shaft diameter.  That gives a value of 0125" vs .013" for the exact formula. 

So, I ended up filing down my keys by .013" in order to get them to fit.  What I SHOULD have done was to touch off the top and then plunge in Z by .063 + .013 = .076".    Or do as others might:  Start plunging in until the edges of the cutter fully engage the material and then from that point, plunge in 1/2 the size of the key.

That's it for today.

Enjoy!

Mike