Discussion in 'Sand Casting' started by Mister ED, Nov 17, 2018.
Wow, those are some really nice gears !
As far as a concourse restoration ... They all look concourse 30-40 foot in the air, LMBO.
Time period on this mill is 1912 - 1923, certainly they had appropriate machine tooling (Flint & Walling was a pretty big company at that time). Brown (of Brown & Sharpe) came out with involute gear cutter sets in the 1860's or 1870's.
My best guess as far as how they made these originally ... oversized match plate and then machined. I wouldn't suspect that they even used a true indexing head. I would suspect they had two rather simple indexing "mechanisms" ... one with 15 indexes and one for the bigger gear which is 45, I think. The gears have a keyway (anyone want to cast that in, LOL) I suspect they threw a bunch on a keyed arbor and started cutting away.
Before Pat stepped in, my initial plan was to machine the gear from cast round (that would test my mill a bit). I think on the original pattern I sent Pat I included a machine allowance to allow me to go in and take a lite pass on each tooth, its hard to remember (that was late Nov). When this thread took off again, I had forgotten that. No worries though to those that have a pattern, I poured the babbitt thick ... even the std dimension gears I have printed are a bit too tight once everything is assembled.
I got your address, haven't had time to work on the pattern yet.
I tried my patterns together and they're fine with reasonable separation for backlash. Undrafted they fit well, expansion draft they have to be separated then they fit.
OK, I am back from the Windmiller's Trade Fair had a great time visiting with friends. Picked up a couple of pretty rare windmills to restore.
The first gear that OIF made, arrived when I was gone. Finished boring it out this morning and stuck it on the mill. I still need to clean up the parting line, but it does work. It it a bit on the small side and I expected that it would be, based on measurements that we were sharing.
Here is a pic of it on the mill:
Also while in Nebraska City we visited a small windmill factory (now museum) that was basically a time capsule when opened up and converted to museum 8-10 years ago. One particular pattern caught my eye.
Something to consider with cast gears is that the width of the space and the width of the tooth at the PCD are supposed to be the same, so in casting the tooth shrinks and the space grows. Add to that the fact that the tooth is smoothed so the pattern can be drawn and the difference in the two dimensions is exaggerated. Adding in shrink for the pattern only corrects radially.
The old timers might have known this, but I have never seen it mentioned in the old technical books.
Ed, do you consider that gear a keeper, or are you looking for a closer match?
That piece doesn't look like a complete gear. Do you suppose it could be a pre formed blank that could then be machined to finish? Nice pieces.
Interesting discussion. The width of the tooth and gap at the pitch circle need to be different however to allow for backlash. You mentioned the reasons the gear will be smaller than the pattern, and those are somewhat offset in sand casting by rapping the pattern before withdrawal.
Zero backlash is unacceptable except under very high tolerance design and manufacturing. Significant backlash has to be designed in for using uncut gears cast with draft. My initial pattern I deducted draft from Ed's design file to avoid interference. Ed thought I should have added draft instead of deducting it based on his knowledge of the gear fits and model size he provided. The current pattern I'm working on has draft added, provided by Ed.
You said adding shrink only corrects radially. I'm thinking on that one, since when you add shrink to an electronic model, everything gets larger. But even if it does add to tooth thickness 2% on top of a 1/4" wide tooth is only 0.005" and dressing down teeth as you mention easily exceeds 0.005".
We can guess that the windmill will operate satisfactorily with much more backlash than would an automotive gearbox.
Ed: They wouldn't loan you that pattern, would they?
ESC, I think you hit the nail on the head in both your first and last paragraph. Yes, I believe that pattern to be a blank to be machined, and think your first paragraph could be part of the reason for it (although its hard to get my head around). Standing there looking at the gear, a cast blank was the only logical explanation I could come up with.
It was very puzzling to see a gear pattern in the museum, this firm only made direct stroke windmills (no gear). As such, I couldn't compare the pattern to a part. I also know for certain that no pieces were brought into the museum for display, it was originally there. I think I have quite a few pattern pics, maybe I will post them in another thread on the factory/museum.
As far as the gear being a keeper ... in a pinch sure, no offense to Andy (he's the one that called it a paperweight, LOL). Although it would be interesting to see how the lost PLA comes out, I think a slight different approach is needed ... one with machine allowance. I had some in the pattern that I sent to Pat. With the machine allowance added, a few things become a little more controllable ... runout becomes a non issue, draft is removed in machining and also the left and right gear will be the same.
To Andy's point, I imagine these gears had quite a bit of backlash (by design) than a tranny gear. One of the best restorers I know will spray weld worn large gears (small too I suppose) and then use a grinder & file for final fit.
As far as borrowing that pattern, its for a different gear (although I do not know what its for). I was standing there thinking about how I could borrow all of their other patterns to make a brand new ELI windmill (LOL).
Oh, here is a windmill gear that I thought was interesting and took a pic to share with you folks.
Yes, it is one casting and the teeth are offset (tooth on one side, aligns to valley on the other). Diam is roughly 12".
Interesting they had a gear pattern but did not use gears. Suppose they took in custom foundry work to supplement their foundry business? Or did they not have a foundry?
I was joking about borrowing it. But they might work with you if you wanted to reproduce their original product.
I've learne da lot trying to make gears for you.
Very cool double gear. Two ideas come to mind:
1 You can minimize backlash if you have adjustable pinions to run on those gears.
2 You get the advantage of smoother running with a lot of backlash (gears look as-cast) because you effectively have twice as many teeth engaging but still have larger teeth. Kind of like a helix gear where one tooth is still engaged when the next tooth engages.
Just spitballing, but when the tooth shrinks the space between grows by double that value. The flanks on your gear casting look thin and if there is a way to correct that when Ed prints a new pattern it will easier done then.
Are we also sure about the 14.5 degree pressure angle?
It is unclear as to whether they did their own foundry work. There was a lean to out back (used as a shed in later decades) that it could have possibly been done there. Certainly not done in the main building.
I believe they were 14.5 PA, although not 100%. We did originally print and test both 14.5* and 20* PA gears. 14.5 appeared to work the best. Also 14.5* seamed to be the most prevalent PA for the very early 20th century. Its been quite a while since I was trying to figure that out, but if I remember correctly the first Machinerys Handbook only mentioned 14.5 PA (if not Machinerys Handbook, it was another early piece of machine tool literature).
For those interested, I did post some pics of the factory here:
So, Ed, is there any point in casting the pattern you sent me if it does not contain necessary machining allowances?
The last pattern Ed sent to me won't mesh fully even at minimum draft so I think they'll work. Have you tried meshing your patterns?
Yes, the don’t fully mesh. So, maybe with light cleanup they will be usable.
I did go out to the barn briefly and used some 2-week old slightly dry green sand and did a quick test mold. It did draw ok. The first test I rapped too much and broke off a couple sand teeth. Trial 2 I was very gentle with the rapping and it pulled. Whew!
So the clean up detailed in the other thread was adequate. If it pulled in green sand, I am confident it would also pull in Sod Silicate bound sand.
Denis, I think OIF is correct that the pattern you have should be fine. The two outside edges are at finished size, but with the draft we are basically adding a machine allowance from each outside to the center. If I remember correctly I have 1% shrink added.
I have a 5DP #7 involute cutter on the way (I was waiting to order it until I knew for sure if I needed it). I will take some clean up passes with that.
Andy, your package should be there Wed.
I finally got some good molds in green sand. My problems were 1 sand too dry, 2 sand not packed tight enough, 3 rapped too hard, a lot of very light rapping worked, 4 my inability to lift the pattern straight. I built a slide hammer puller to ease the pattern loose after rapping then used the stripper plate when pulling. Not sure the stripper plate was ultimately necessary but it seemed stupid to not use it since I had failed at least 6 times in green sand. This one looked good but had some fractured teeth which fell off in handling.
That was yesterday but instead of starting over I poured it in aluminum today to check the feed setup.
These were good.
I had to repair the bottom of the cope through a stupid mistake (cut the basin after splitting the flask) but the repair held.
Little furnace running all out on oil, melted 10 lbs of cast iron in 59 minutes. Not too impressive.
I pulled the casting out when it was orange and had to leave so it will get inspected tomorrow. I had the same problem as before, runner froze, and had to finish pouring through the riser. That's getting to be a bad habit. Maybe not hot enough. First glance casting looked good.
I'll be starting the shell on the first one tomorrow.... waiting to do the second to make sure the first one is right...
This is much more fun than going to the movies.
And the popcorn is cheaper too.
Its like a soap opera.
Will gear 1 marry/mate with gear 2?
Will gear 1 pull from the sand?
Will the shrinkage be right?
Will the part be machinable?
We should call this thread "As the Gear Turns", or "Days of our Gears".
Munch, munch, munch................................ pass me some more beer and popcorn, this is getting good.
You can't buy entertainment that is this good.
And I can really relate to what a pain in the ass it is to pull that pattern without damaging the sand, for those trying to do that, especially without any draft.
I am just glad I am watching and no longer struggling with it.
It was great fun, but tricky for sure, and ultimately my method failed.
What's the ring around the pouring cup? Is that shell-specific?
Looking forward to the results. I would worry the gate is not large enough to freeze last but you've got lots more experience than I do.
Separate names with a comma.