Injected Diesel 56cc 2 Stroke, Will it ever work?"

[Lloyd-ss]

Pressure and Volume Test Results from Roots Blower.

The Roots Blower that I made has a pair of 3-lobed rotors 44mm in dia and 27.8 mm thick. The p.d. and the C to C distance is 29mm. The inlet and outlet ports ar 20mm x 16mm. Each lobe has 3 "empty spaces" of about 5.9cc. Each revolution pumps 6 empty spaces from the inlet port to the outlet port, so the pump through-put at 100% efficiency could be 35.4 cc/rev . As a side note, because the Roots is symmetrical inside, if the direction of rotation of the rotors is reversed, the in and out directions are swapped.


Because the 2-stroke diesel that I have in mind is about 56.4, even if the Roots operated at 100% eff, the rotors would need to run at 1.6 to 1 (60% overdrive) to keep up with the piston movement. 100% eff, no way!

Here are the test results fro the various rotor speeds:

1,040 rpm
Maximum developed pressure (with outlet blocked)(in inches of water column):
7.12" W.C.
Blower chamber pressure measured, and volume calculated, with air exiting the plenum through a
.182" hole:
2.37" W.C.
19,820 cc/min
Blower chamber pressure measured, and volume calculated, with air exiting the plenum through a
.375" hole:
0.22" W.C.
25,770 cc/min
Theoretical swept volume of the Roots blower at 1040 rpm (at 35.4 cc/ blower rev):
36,820 cc/min, max theoretical

1,510 rpm
Maximum developed pressure (with outlet blocked)(in inches of water column):
12.75" W.C.
Blower chamber pressure measured, and volume calculated, with air exiting the plenum through a
.182" hole:
4.63" W.C.
27,690 cc/min
Blower chamber pressure measured, and volume calculated, with air exiting the plenum through a
.375" hole:
.62" W.C.
42,760 cc/min
Theoretical swept volume of the Roots blower at 1,510 rpm (at 35.4 cc/ blower rev):
53,454 cc/min, max theoretical

2,150 rpm
Maximum developed pressure (with outlet blocked)(in inches of water column):
21.75" W.C.
Blower chamber pressure measured, and volume calculated, with air exiting the plenum through a
.182" hole:
8.25" W.C.
37,090 cc/min
Blower chamber pressure measured, and volume calculated, with air exiting the plenum through a
.375" hole:
1.25" W.C.
60,880 cc/min
Theoretical swept volume of the Roots blower at 2,150 rpm (at 35.4 cc/ blower rev):
76,110 cc/min, max theoretical


How does the scaling to Detroit Diesel 1-71 look?
As I stated previously, this little Roots blower is approximately 1/5 th linear scale to 1-71 Roots.
The swept volume of the DD roots blower, per rev, is equal to about one cylinder's worth of air per engine rev. In other words, 71 cuin of air per rev, or 1,163 cc of scavenge air per revolution, or 1,163,000 cc of air per minute at 1,000 engine rpm. (Note- edit made to this paragraph to correct math errors)

The blower I am working on is supposed to be 33mm bore x 66mm stroke, for 56.4cc/rev. Doing the simple math:
56.4 x 500 rpm = 28,200 cc/min (for 1 cyl fill of air per rev)
56.4 x 1,000 rpm = 56,040 cc/min (for 1 cyl fill of air per rev)
56.4 x 1,500 rpm = 84,600 cc/min (for 1 cyl fill of air per rev)
56.4 x 2,000 rpm = 112,800 cc/min (for 1 cyl fill of air per rev)

Looking at the performance per rev chart, above, it looks like approx 1,510 rotor rpm might be a good starting point for 1,000 rpm engine speed, which would require at least a 1.5 to 1 overdrive for the 2 rotors.
There might be some mistakes in my math or logic, so please let me know if you spot anything.

Lloyd

 

[Richard Hed]

Pressure and Volume Test Results from Roots Blower.

The Roots Blower that I made has a pair of 3-lobed rotors 44mm in dia and 27.8 mm thick. The p.d. and the C to C distance is 29mm. The inlet and outlet ports ar 20mm x 16mm. Each lobe has 3 "empty spaces" of about 5.9cc. Each revolution pumps 6 empty spaces from the inlet port to the outlet port, so the pump through-put at 100% efficiency could be 35.4 cc/rev . As a side note, because the Roots is symmetrical inside, if the direction of rotation of the rotors is reversed, the in and out directions are swapped.
View attachment 128952

Because the 2-stroke diesel that I have in mind is about 56.4, even if the Roots operated at 100% eff, the rotors would need to run at 1.6 to 1 (60% overdrive) to keep up with the piston movement. 100% eff, no way!

Here are the test results fro the various rotor speeds:

1,040 rpm
Maximum developed pressure (with outlet blocked)(in inches of water column):
7.12" W.C.
Blower chamber pressure measured, and volume calculated, with air exiting the plenum through a
.182" hole:
2.37" W.C.
19,820 cc/min
Blower chamber pressure measured, and volume calculated, with air exiting the plenum through a
.375" hole:
0.22" W.C.
25,770 cc/min

1,510 rpm
Maximum developed pressure (with outlet blocked)(in inches of water column):
12.75" W.C.
Blower chamber pressure measured, and volume calculated, with air exiting the plenum through a
.182" hole:
4.63" W.C.
27,690 cc/min
Blower chamber pressure measured, and volume calculated, with air exiting the plenum through a
.375" hole:
.62" W.C.
42,760 cc/min

2,150 rpm
Maximum developed pressure (with outlet blocked)(in inches of water column):
21.75" W.C.
Blower chamber pressure measured, and volume calculated, with air exiting the plenum through a
.182" hole:
8.25" W.C.
37,090 cc/min
Blower chamber pressure measured, and volume calculated, with air exiting the plenum through a
.375" hole:
1.25" W.C.
60,880 cc/min


How does the scaling to Detroit Diesel 1-71 look?
As I stated previously, this little Roots blower is approximately 1/5 th linear scale to 1-71 Roots.
The swept volume of the blower per rev, is equal to about one cylinder's worth of air per rev. In other words, 71 cuin of air per rev, or 1,163 cc/rev, or 69,780 cc/rev. That works out to 4,186,800 cc/minute of air.

The blower I am working on is supposed to be 33mm bore x 66mm stroke, for 56.4cc/rev. Doing the simple math:
56.4 x 500 rpm = 28,200 cc/min (for 1 cyl fill of air per rev)
56.4 x 1,000 rpm = 56,040 cc/min (for 1 cyl fill of air per rev)
56.4 x 1,500 rpm = 84,600 cc/min (for 1 cyl fill of air per rev)
56.4 x 2,000 rpm = 112,800 cc/min (for 1 cyl fill of air per rev)

Looking at the performance per rev chart, above, it looks like approx 1,510 rotor rpm might be a good starting point for 1,000 rpm engine speed, which would require at least a 1.5 to 1 overdrive for the 2 rotors.
Lloyd

Do you have the plans for this? I was working on a two lobed roots blower but got nowhere as I didn't have a quality enough mill--that is I didn't have ANY mill.

 

[Peter Twissell]

Nice work Lloyd!
Good to see that your test results correlate well with calculated flowrates at higher rpm. This is to be expected, as the leakage in the blower is constant through the rpm range.

 

[Lloyd-ss]

Do you have the plans for this? I was working on a two lobed roots blower but got nowhere as I didn't have a quality enough mill--that is I didn't have ANY mill.

Richard, yes I do have some drawings of the rotors and housing. They are just 2d autocadd drawings, but the mesh of the rotors seems to be good throughout 360 degrees of rotation. The rotors are not true cycloids, but close approximations of them. All features on the rotors are portions of a circle, as shown in the post about the actual machining of the rotors. Start by getting the gears first, so that you know what the exact center distance between the rotors needs to be. The size of the rotors and housing are fully scalable, so you could make whatever size you want. I will post the drawings in a few days after I get them prettied up with adequate dimensions. I think I can post both pdf's and dxf's or dwg's of the blower parts.

Just like standard involute gears, where the tooth faces roll against each other with no rubbing, the cycloid shape is supposed to do the same thing, all rolling without any rubbing.
Lloyd

 

[Lloyd-ss]

Nice work Lloyd!
Good to see that your test results correlate well with calculated flowrates at higher rpm. This is to be expected, as the leakage in the blower is constant through the rpm range.

Thank you Pete! After reading your reply, I went back and did a few edits in post #61 to correct some fuzzy, late night, math and descriptions. I also added a "theoretical maximum" airflow at each test rpm speed, based on swept volume of the Roots rotors of 35.4 cc/revolution. That helps shed some light on the leakage in the blower as you pointed out.

I am a data junkie and love making spreadsheets and graphs to display the data in various ways. You never know what sort of unexpected trend might pop out at you, or how consistent and preditable the results might be. I might collect more data at the 2 remaining speeds on the little drill press, 585 and 3,000 rpm. There might be the makings of some actual pump curves, or at least my interpretation of a pump curve. The science of this adds a whole 'nother dimension of enjoyment of this hobby for me. But the cool thing is, you only need to get into the science as much as you want to, or even not at all. Great fun!

Also, I want to give a big thank you to ALL of the members of the forum who display and share their projects and ideas and comments. The challenges undertaken, and successfully completed, are true inspirations for a new novice member.

Lloyd

 

[Richard Hed]

Richard, yes I do have some drawings of the rotors and housing. They are just 2d autocadd drawings, but the mesh of the rotors seems to be good throughout 360 degrees of rotation. The rotors are not true cycloids, but close approximations of them. All features on the rotors are portions of a circle, as shown in the post about the actual machining of the rotors. Start by getting the gears first, so that you know what the exact center distance between the rotors needs to be. The size of the rotors and housing are fully scalable, so you could make whatever size you want. I will post the drawings in a few days after I get them prettied up with adequate dimensions. I think I can post both pdf's and dxf's or dwg's of the blower parts.

Just like standard involute gears, where the tooth faces roll against each other with no rubbing, the cycloid shape is supposed to do the same thing, all rolling without any rubbing.
Lloyd

When I was working on this, nearly 30 years ago, an engineer friend told me .012 clearance was good, but I thimpfk that is too much. I woujld thimpfk that a thou would be fine if one could achieve it. Is there any reason, other that making sure there is no actual contact, for making such huge clearances?

 

[Lloyd-ss]

When I was working on this, nearly 30 years ago, an engineer friend told me .012 clearance was good, but I thimpfk that is too much. I woujld thimpfk that a thou would be fine if one could achieve it. Is there any reason, other that making sure there is no actual contact, for making such huge clearances?

Richard, Here are 2 pages from the DD 71 series manual and it does indeed call for a .012" clearance between the rotor tips and the housing. But it only calls for a .002" clearance between the interference points between the rotors. During operation, all of the air flow is along the outside wall of the housing (air does not pass in a straight line thru the rotors. It calls for inspection every 100k miles, so i guess they figure that is where the clearances are needed. The rotors are huge and maybe they get hotter and expand more than the housing. Funny that only 2 of the clearances have max tolerances specified.

My rotors started out too tight to easily roll all the way around. Sandpaper, files, and lapping compound finally got it to rotate with some contact, but no binding. The cycloid features (like an involute gear shape) should roll past each other, not slide past each other.

The cross section picture really shows how huge the blower is.
Lloyd

 

[Steamchick]

OK Joe, let me start by saying that I do have a smile on my face
I used the qualifier "essentially" and I guess I could have said that the backlash should be "essentially" zero, but I didn't. Didn't think it was necessary given the context of the conversation.
But I will give you 2 examples where the backlash can be zero. If you roll a newly ground steel gear against a master gear to inspect center distance variation and find that the max is 10 millionths of an inch, I would say that that test fixture operates at zero backlash. The other is when I was making the aluminum gears to try on the blower in this thread. As I was creeping up on the proper tooth thickness/backlash, I installed the gears on the assy and got it where I could barely rotate by hand. I could feel the bump-bump from tooth to tooth as the clearances were all used up and the soft aluminum of the teeth yielded enough to rotate while maintaining point to point (line to line) contact.
Whether we call it zero backlash, or almost zero backlash, doesn't matter to me. I want it to feel, with my fingertips, like there is no backlash, and, if you checked with a dial test indicator, (one .0001" per graduation), you probably couldn't say for sure if there was or was not any backlash. I am still smiling!
Enjoy your weekend!
Lloyd

Just an observation from a real amateur - when it comes to gears, thrust faces, contact points, sliding surfaces, etc.... but Doctor Engineer in Tribology once told me that you always want sufficient clearance for the "molecules" of oil to roll between the peaks of the metal crystalline structure. This is what he called zero clearance, I.E. where the oil lubricated at "minimum film", without being "sheared" and rapidly broken down. He was talking of a minimum of 40 microns (4 x 10E-5) clearance in main and big end plain bearings in car engines. (Actually hugely greater than true oil molecule sizes. ~10 E-10m. A C60 molecule is about 0.7nm, =7 x 10E-10.!). He also advised that "minimum clearance" is naturally broken within gears, which is why oils for gear boxes have to have special oils (additives), to resist this mechanical breakdown. I think the additives are large molecule mineral compounds, dissolved in the oils, but never got to work on gear oils so have no real knowledge.
Not knowing, I wonder if running the gears so tight is actually going to run-in the gears, or simply distort and damage the ball races? Or will a very soft and fine abrasive (toothpaste?, Brass polish?) assist in running-in the high spots of the gears? I am not sure what you are trying to achieve by running gears "tight"? Are you simply marking high spots to be dressed with a tool or stone slip? Or are you trying to work harden the surfaces, or wear them away metal to metal?
I do think buying proprietary tooling to make the gears will be one step more accurate than grinding your own tools. So you have done the right thing there IMHO.
Enjoy, and keep posting on this interesting subject.
K2
K2

 

[Lloyd-ss]

Not knowing, I wonder if running the gears so tight is actually going to run-in the gears, or simply distort and damage the ball races? Or will a very soft and fine abrasive (toothpaste?, Brass polish?) assist in running-in the high spots of the gears? I am not sure what you are trying to achieve by running gears "tight"? Are you simply marking high spots to be dressed with a tool or stone slip? Or are you trying to work harden the surfaces, or wear them away metal to metal?

Steam, Thanks for your detailed response and questions. I might have been a bit snarky in response #59 to Joe, and if it came across that way, I apologize. That certainly was my intent.

Steam, your questions got me to thinking, and my general thought processes. Most of previous small size machining has been lathe work for PCP air rifles working in the 3,500 psi range. A lot of precision lathe work, but basically no mill work that precison center distances. Now, this roots blower is basically 2 sets of gears, involute and cycloid, turning on the same shafts and requiring rotor clearances of a few thousandths of an inch inside the blower housing. Basically, my shop equipment isn't capable of achieving without some tricks. Again, I am still learning at a basic level, in blissful ignorance, determined to make it work. The reason I want "minimal" backlash between the 2 involute gears is because all accuracies throughout the blower must be held to a high degree so that the cycloid gears of the blower will rotate smoothly without interference.

If you look at the tolerance stackup, a .001" change in center distance will cause about a .0014 change in backlash. This backlash affects the clocking of the cycloid gears. Add in a .001" circular runout in a couple of places and there could possibly be another .003" backlash. Now, one bad thing about cycloid gears is that with a pair 3 tooth gears, one cannot drive the other because they will lock up every 1/3 rotation. So, as the backlashes and tooth thicknesses, and runouts, all accumulate, they all end up being transfered to the cycloid gears. Then, add on top of that, that each time the scavenging ports are opened by the piston travel, the Drive, and the Driven loading on the 2 cycloid gears might reverse. If the accumulated backlash is too great the cycloid gears could jam.

The blower in the pictures is really more of a proof-of-concept device than anything else, although it DOES work, but just not well. I now have on hand some purchased steel mod1 gears (which match the design C to C distance), some flanged bronze bushings, some sealed ball bearings, and some 3/8" 1144 stress proof rod for the shafts. Basically, I will be keeping the rotors, endplates, and housing, and adding new gears, shafts, and bearings.

I can see why some of you guys make your own gears so that you can control the tooth thickness and therefore, the backlash. so much to learn, but that is what it is all about!
Lloyd

 

[Steamchick]

Lloyd. Don't apologise. Your explanation sounds good. This is a very "technical" build, much more than the "average machinist" may attempt - knowingly. In my experience, there are many superb machinists, but only a fraction of them can discuss the engineering and explain "what's what" as you do.
I look at it that "Toolmakers" and "Engineers" are the angels of different religions. Each may think they (and their "reilgion") are the most important, but each needs the other to get the best from their combined abilities.
Thanks,
I am enjoying the technicalities of the thread. (Though I know my desires and limitations, and will never be making a roots blower! Even though I know my bike would be a much better performer with one.).
K2

 

[Lloyd-ss]

I did run a pump volume test (the Roots Blower) using the set up shown in the video in post #60. Basically, I attached a sealed and flattened 33 gallon plastic trashbag to the pump to see how fast it would fill the bag, and how much pressure it would max out at when a "leak" hole was cut into it. At 2.150 blower rpm, it filled the 33 gallon bag (125,000 cc) in about 100 seconds. The leak hole was to simulate the scavenging ports, kinda. After some number crunching it looks like the blower might barely keep up with an engine running at 1,200 rpm with the pump being drive at 40% overdrive. A crude test, but at least it shows the blower is working.
Lloyd

 

[Lloyd-ss]

I look at it that "Toolmakers" and "Engineers" are the angels of different religions. Each may think they (and their "reilgion") are the most important, but each needs the other to get the best from their combined abilities.

Very well said. I could use a good toolmaker angel.....please.
Lloyd

 

[Lloyd-ss]

ROOTS BLOWER - THE ON-GOING SAGA

It has been a few months since I was last here, so first I want to say thank you to the true, hardcore members of the forum who are on here almost every day. You folks are the real life-blood, and I appreciate that. I feel a little cheesy just dropping in and out occasionally, but I do think about this a lot and work on it whenever I can.

To make a long story short, this was darn difficult, and very humbling. A CNC probably would have helped, but the clearances as the rotors rotate is soooo critical, that just a tiny discrepancy allows leakage from the positive pressure side to the negative, and the airflow basically stops.

The 2-lobe was the worst. Even though it has a huge swept volume, if the rotors don't stay in mesh 100% of the time, the flow drops to zero. The 3-lobe was only marginally better.

Then I started giving it some more serious thought, from the perspective of not having a cnc. The more lobes the rotor has, the closer the profiles can be approximated with circular arcs, so that they are always in contact. The 2 lobe is the worst, with only a short section of circular arc at the root and the tip. So all the rest of the cheesy cycloid approximation is just wide open leakage paths.

The final (maybe, LOL) 6 lobe is made from PTFE, which has good temperature qualities and can be run with almost zero clearance. I also used flanged ball bearing this time that can be preloaded with the lock nuts. I only ran it for a minute at 1500 RPM, but the thumb-test revealed some decent positive pressure.

Even though the swept volume of the 6 lobe is only one half that of the 2 lobe, a single point of failure does not make it loose ALL of its pressure. It will still pump some air, rather than NO air at all.

Next will be to set it up for testing to see how much flow and pressure it can achieve, and try and figure what what "adequate flow", might be.

Thanks to everyone who has shared their failures on the forum and showed that if you want it to work, you just keep trying. It made me realize that I am not the only one.
Lloyd

 

[ajoeiam]

ROOTS BLOWER - THE ON-GOING SAGA

It has been a few months since I was last here, so first I want to say thank you to the true, hardcore members of the forum who are on here almost every day. You folks are the real life-blood, and I appreciate that. I feel a little cheesy just dropping in and out occasionally, but I do think about this a lot and work on it whenever I can.

View attachment 135095

To make a long story short, this was darn difficult, and very humbling. A CNC probably would have helped, but the clearances as the rotors rotate is soooo critical, that just a tiny discrepancy allows leakage from the positive pressure side to the negative, and the airflow basically stops.

The 2-lobe was the worst. Even though it has a huge swept volume, if the rotors don't stay in mesh 100% of the time, the flow drops to zero. The 3-lobe was only marginally better.

Then I started giving it some more serious thought, from the perspective of not having a cnc. The more lobes the rotor has, the closer the profiles can be approximated with circular arcs, so that they are always in contact. The 2 lobe is the worst, with only a short section of circular arc at the root and the tip. So all the rest of the cheesy cycloid approximation is just wide open leakage paths.

The final (maybe, LOL) 6 lobe is made from PTFE, which has good temperature qualities and can be run with almost zero clearance. I also used flanged ball bearing this time that can be preloaded with the lock nuts. I only ran it for a minute at 1500 RPM, but the thumb-test revealed some decent positive pressure.



Even though the swept volume of the 6 lobe is only one half that of the 2 lobe, a single point of failure does not make it loose ALL of its pressure. It will still pump some air, rather than NO air at all.



Next will be to set it up for testing to see how much flow and pressure it can achieve, and try and figure what what "adequate flow", might be.

Thanks to everyone who has shared their failures on the forum and showed that if you want it to work, you just keep trying. It made me realize that I am not the only one.
Lloyd

Nice work.

A wondering - - - - these blowers are NOT a new thing - - - - wondering how they were fabricated 'back in the day' say the 1950s.
Anyone out there have any ideas?

 

[Lloyd-ss]

A wondering - - - - these blowers are NOT a new thing - - - - wondering how they were fabricated 'back in the day' say the 1950s.
Anyone out there have any ideas?

Great question! I have not researched, and do not know. But I am always amazed at the ingenuity and artistry of machinists from the past. Someone here must know.

There is a question that has been driving me a little crazy that I am sure a lot of the more experienced model builders have the answer to. It has to do with scaling of the models and how it changes the functioning of the actual model.

My roots blower is maybe one fifth scale. Each rotor is 1.62" long and 1.66" in diameter. If this were a 2 lobe rotor, each "space" would have a volume of approx 1.92 cubic inches. And for arguments sake say that each lobe space has linear seal/leakage lines of 3 lengthwise and 2 on the ends. That would be approx 8.1 linear inches.

The full scale rotor (5 times as big) would have a volume of 5 cubed x 1.92 cubic inches (EDIT, 5 cubed, not 5 squared) =5x5x1.92=48 cubic inches. =5x5x5x1.92= 240 cubic inches NEW ANSWER But its linear gaps would only be 5 times bigger, 5x8.1=40.5 linear inches. Correct??

So for the full sized 2-lobe blower you have EDIT .84 .17 NEW ANSWER linear inches of leakage path per cubic inch, and for the one fifth scale you have 4.2 linear inches of leakage path per cubic inch.

So to me, it looks like the one fifth scale model has EDIT 5 times NEW ANSWER 24.7 times the leakage potential of the full size. Of course, the scale model should have closer tolerances and therefore the leakage potential will be less than that. I am quite confused as to what the practical real-world implications of the scaling are. I am also wondering if this is beyond my skill level? I do want to get it to work, though. Thanks for any thoughts.

EDIT-Frankly, I am shocked that my math shows 24.7 times as much leakage potential for the one fifth scale model. (Probably should be 25 due to rounding errors). Is that correct ??? Thanks! Lloyd

 

[Steamchick]

Cam-following milling machines were used long before CNC was developed. The Cam being a shaped lump of metal, not Computer-Aided-Machining. WW2 did a lot to improve machining development in the USA, Germany, Japan and UK.
Even in the 1980s~1990s I saw machine shops making pistons (and lots of other things) using machines that "followed shaped cams" - e.g. for piston skirts that were NOT circular. But these were being superceded by machines that followed a "computer" shape with servo driven tooling.
K2

 

[Steamchick]

Hi Lloyd,
That sounds like you do some good calculations. Anything (like gases and leaks) that have a combination of square and cube factors cannot scale directly. linearly. e.g. model aircraft have a thicker wing than real full sized aircraft - simply to make them work. All about ratios like Reynolds numbers, etc... that are a simplification of the squares, cubes, etc. of nature.
I do a bit of work on steam boilers. Nothing is scale, except perhaps the external size and shape. Even then there are usually compromises. e.g. Water in water gauges doesn't follow scale due to the meniscus following the "water" characteristic, due to surface tension, of climbing up the glass. Below a certain size of tube, the surface tension dominates pressure differences of the water level. And steam bubbles can block the whole device anyway, when the bubble diameter is large enough to touch the tube all around, when surface tension over-rules the pressure differences that would otherwise cause the bubble to float through the liquid water.
Anyone who has attempted to make a small single cylinder ICE engine will know all about leakage when their compression is very low! - Just as you are learning. It is a part of the reason we had to develop better materials and machining to make smaller and smaller ICE engines than the early huge cylindered engines. Same applies to hydraulics, as well as pneumatics.
Cheers!
And good luck! - I am sure you'll succeed with some perseverance.
K2

 

[Lloyd-ss]

Cam-following milling machines were used long before CNC was developed...........

Steam, ahhh, that is the direction I was thinking. Thank you so much for your experience and input about when there were a lot of dedicated machine set-ups and the tool crib was filled with lots of big hulking specialty fixtures. A lot of ingenuity.
I was around for the middle of the manual-machine-to-cnc evolution as a YPE (young punk engineer) and just sucking it all in being amazed at how all these beautiful metal parts were made. I was in heaven!
There was a problem that had come up with a 10 to 1 template dresser on a pump-gear finish grinding machine. It ground the individual tooth spaces between the gear teeth. They could not get a proper profile on the large template and when the finished gears were rolled against a master gear, the ink paper trace showed an ugly bump-bump-bump as the gears rolled together. I had had just enough Fortran programming to think that I might be able to help. They gave me a chance and I worked with the math and geometry guru there to learn how the profiles were mathmatically generated. My first attempt was disheartening because it was not much better than what they had. But the guru was very understanding and we finally realized that my understanding of how the tooth space was measured (perpencicular instead of circular) was incorrect. I fixed that in the program (using cnc to generate the manual 10 to 1 template) and it worked. I can remember being so excited as I saw a few random out-of-charcter smiles popping up. I was on my way to loosing the YPE name tag. A wonderful time.
Lloyd

 

[Lloyd-ss]

.........Nothing is scale, except perhaps the external size and shape. Even then there are usually compromises.......

Steam, you have no idea how much anxiety that comment has relieved. Thank you so much. There are so many gorgeous scale models that I always assumed were scaled all the way down to every nut and bolt. But now I see that that is not the case. The tribal knowledge and tricks take over to make it look scale, but still be able to function. As I build this thing in my head while riding around on the lawn mower, I have actually been thinking about compression ratios and how to design the head for the engine so that the compression ratio could be adjusted up or down after it doesn't work the first time. Sometimes it is all about knowing the tricks. Thank you for the insight.
Lloyd

(Sorry for the rambling posts, but sometimes I get on a roll and can't stop.)

 

[ajoeiam]

Steam, ahhh, that is the direction I was thinking. Thank you so much for your experience and input about when there were a lot of dedicated machine set-ups and the tool crib was filled with lots of big hulking specialty fixtures. A lot of ingenuity.
I was around for the middle of the manual-machine-to-cnc evolution as a YPE (young punk engineer) and just sucking it all in being amazed at how all these beautiful metal parts were made. I was in heaven!
There was a problem that had come up with a 10 to 1 template dresser on a pump-gear finish grinding machine. It ground the individual tooth spaces between the gear teeth. They could not get a proper profile on the large template and when the finished gears were rolled against a master gear, the ink paper trace showed an ugly bump-bump-bump as the gears rolled together. I had had just enough Fortran programming to think that I might be able to help. They gave me a chance and I worked with the math and geometry guru there to learn how the profiles were mathmatically generated. My first attempt was disheartening because it was not much better than what they had. But the guru was very understanding and we finally realized that my understanding of how the tooth space was measured (perpencicular instead of circular) was incorrect. I fixed that in the program (using cnc to generate the manual 10 to 1 template) and it worked. I can remember being so excited as I saw a few random out-of-charcter smiles popping up. I was on my way to loosing the YPE name tag. A wonderful time.
Lloyd

I have been digging for info.
Have found a number of academic papers.
Are you interested in wading through them?
(I'm not guaranteeing that they're totally useful - - - just that they 'might' be useful.
I'd do some reading myself just that I've got a lot going right now and thought you might want to start soonish and and and . . . )