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

[Lloyd-ss]

Aha!
Found the article I was looking for...!
k2

Inside the Cylinder of a Diesel Engine – by Harry Ricardo

Sir Harry Ricardo was one of the foremost engine designers and researchers of the internal combustion engine. In this article, he takes us through the first moments of combustion in a diesel engine…

oldmachinepress.com

Ah, some of those British engineers were brilliant, and prolific. Thanks for those references. And it seriously is amazing how prolific some of the best of the best were. One new idea after the last, with many completely changing the course of prior technology. Zora Duntov was one in the United States. He designed the first Small Block Chevy V8 engine in about 1953. Arguably, the first "modern OHV V8." It started out as a 265 cubic inch and eventually grew to 400 cuin. And the basic design is still in use today. There were so many millions of them produced that they became the engine of choice for hot rodders, for many years. The main reasons being that 1)there were so many of them in existence that they were cheap, 2)after-market manufacturers jumped on the band wagon and made all sorts of hop-up equipment that was affordable, and probably most importantly, 3)Duntov's design was so good that almost no changes needed to be made to the basic design, which led to, 4)a very broad family of engines with very amazing amount of parts interchangeability, which means, cheap aftermarket parts. Genius!

With that said, yes, I think I am more or less finished with the reinventing the wheel stage, but I will probably have a relapse.....soon. Given that the basic concept of the model I am building is the Detroit Diesel 2 stroke (another amazing family of engines), and now that I have the Roots blower and the Unit injector built, probably trying to scale the model to the basic Detroit Diesel design (including the combustion chamber design) will be a reasonable approach. Physically, the injector ended up being larger than I had hoped, so the basic scale of the blower might not be visually correct, but I hope it is.

And I also realize that if the model is, lets say, 1/4th scale, that the piston area is 1/16th, and the cylinder volume is 1/64th, and that the stiffness of the crank pin is 1/256th??? So, if the engine power is proportional to the cylinder volume, i.e., 1/64 the power, does that mean the crank pin will end up being only 1/4th as strong/stiff as it really needs to be? So the crank pin will really need to built to 1/2.8 scale to the original? That is, 2.8 being the 4th root of 64? So as to keep the stiffness of the crank pin in proportion to the power? Or, because the crank pin is also 1/4th as long, and stiffness changes with the cube of the length, will doing a simple 1/4th scale of the crank be ok. Or maybe I am overthinking it, but maybe not. Ha ha, I was just going to say that this is rhetorical question and I need to figure it out myself first, but there I go re-inventing the wheel again. But the journey is more fun than the destination.

I know there are some older threads about the complexities of scaling engines, so I better dive into those first.

Lloyd

 

[Roger B]

It is interesting to think in the other directions and compare the dimensions of a big ship diesel engine (conventional not crosshead) or railway traction diesel engine with a 2L automotive diesel. The proportions are much the same.

 

[Nerd1000]

For my diesel I put together a large and incredibly convoluted spreadsheet that estimates a lot of the critical values e.g. bearing loads, stresses on the conrod, cooling water flow demand, valve sizing etc. Sometimes I use FEM analysis in my CAD software to validate a part design, but I'm not an engineer (well I am a genetic one, but DNA is different from diesel engines) so I'm sure I miss things a first year engineering student would know about. So some trial and error is involved in the design process.

 

[Steamchick]

Hi Lloyd. Re scaling. You are absolutely correct! Scale the bore and stroke then se those to re design the crank (journals) to be strong enough and stiff enough to do the job.
My first lesson in this way a basal wood and tissue glider. Kit made at 27 in wingspan, I doubled all the sizes - when it had so much lift from the wing it would only fly extremely slowly (in still air), with 2 lbs of ballast to stabilise it against the huge lift from the "wrong" aerofoils...
Since then I have had to learn the Engineering of scale!
When you do the calcs for stress and stiffness, you'll be able to relax with your design.
Cheers,
K2

 

[Lloyd-ss]

For my diesel I put together a large and incredibly convoluted spreadsheet that estimates a lot of the critical values e.g. bearing loads, stresses on the conrod, cooling water flow demand, valve sizing etc. Sometimes I use FEM analysis in my CAD software to validate a part design, but I'm not an engineer (well I am a genetic one, but DNA is different from diesel engines) so I'm sure I miss things a first year engineering student would know about. So some trial and error is involved in the design process.

I love spreadsheets,too, and also use them to excess. But I have a couple of them that I use for high pressure tube calculations and internal ballistics for airguns that I have been tweaking for almost 10 years. Pathetic, nah.

I misinterpreted (I think) your statement about being a genetic engineer. I thought you meant that the "engineering gene" in your DNA was strong, and you could not resist the force. That is fine,too.

 

[Lloyd-ss]

Injector Volume Test

The flame test was pretty dramatic, especially at max output from the injector, but I had not tested the actual flow rate.
Here is the set-up I made for flow testing. It is driven with a toothed belt by a really slick sewing machine servo motor that I got on ebay for about $100, including the controller. You can set all sorts of things like rotation direction, top speed, ramp-up rate, etc. I have already made a prony brake for it to check the output torque and was amazed that it held its target RPM even as I increased the load on the prony brake. It didn't loose its RPM until it hit close to 1/2 HP.

But, back to the flow testing. Earlier in the thread we calculated about 5 or 6 cubic mm of fuel per stroke at max power. In the picture of the test set up you can see that injector squirts into a piece of clear vinyl tubing with 3/8" I.D.. That is a volume of 71 cubic mm's per mm of depth. The motor was set at 500RPM and I just ran it, with a stop watch going, till about 5 or 10mm depth of fuel had squirted into the tube. Then, calculated the actual number of strokes and figured out the actual volume of fuel per stroke in cubic mm's. Here are the results.

500 RPM volume test, with control lever adjusted each time.
Lever set to approx 10 degrees off no-flow = 3 cubic mm/stroke
Lever set to approx 30 degrees off no-flow = 6 cubic mm/stroke
Lever set to approx 70 degrees off no-flow = 35 cubic mm/stroke

I didn't test at a higher RPM because the volume of excess fuel was pretty obvious.

That's a serious over-kill at the 70 degree setting and explains the dramatic flames at max setting in the flame test video. Looks like I need to modify the ramp on the pump plunger in the injector. I already did that once, but now I see how it needs to be fixed. Probably also need to modify the cam, also. The starting point seems to be ok, but the overall stroke is much too long. The cam mod will be easy, too.
Lloyd

 

[Lloyd-ss]

I modified the helix (flattened the angle) on the pump plunger, modified the cam to shorten the length of the pump stroke, and re-made an internal spacer to eliminate some dead volume on the output side of the pump plunger.
The output of the injector at 500 RPM will now go from a minimum of 1.7 cubic mm per stroke to 5.7 cubic mm per stroke, as the pump shaft is rotated through 45 degrees by the control arm. I am more satisfied than I was, but I am not sure how to get the fuel to shut off totally, or if that is even possible with the style of pump I have.
1.7 cubic mm still seems like a lot of fuel for a 56 cc diesel to idle with. The minimum stroke volume of fuel was 3 cubic mm per stroke before the modification, so definitely an improvement.
Lloyd

 

[Roger B]

I couldn't wait any longer. Its a pretty day, perfect for a flame test.
No idea what it proves, but yes, it was fun.

View attachment 144911

My experience has been that if you can ignite the atomised fuel with a small flame it will ignite in the engine. If it is too much of a jet it will not ignite. This is an extreme example with just a 0.35mm hole. All it would do was cover my lighter in oil.

 

[Lloyd-ss]

There was a wonderful, long discussion thread about needle bearings and other aspects of this diesel model build, but I will try and bring the discussion back to this thread. Thank you.

This build is of a single cylinder "interpretation" of a Detroit Diesel 2 stroke, 1-71. I am having a hard time finding a manual specifically for the 1-71 (single cyl, 71 cubic inch), but have them for the 2-71, which is different from the 3-71 thru the 6-71.
Anyway, here are some pics of the 2-71 that hopefully will help clarify what this icon looks like.

A couple of things to note:
4.25" bore x 5" stroke
It uses a "unit injector", that is fuel pump, metering, and nozzle all in one unit.
The piston is over 5 " long because of how it needs to keep the inlet ports covered.
The cylinder liner is about 10-3/4" long to accommodate the full travel of the piston.
The conn rod is 10.2" long to be able to swing and miss the bottom of the cyl liner.
The camshaft and balance shaft are mounted lower in the engine than in the 3-71 and larger sizes.
The cam and balance shafts are counter weighted to minimize side-to side vibration.
The roots blower is geared to run overdrive at twice the engine speed.
Note that the blower rotors are similar in diameter to the cyl bore.
The 2-71 weighs 960 lbs, is 41" tall, makes 68HP at 2000rpm. That would make a Honda Civic sit funny.
The model will have an open crankcase and closed loop lubrication with supply and return lines.

Here are 3 thumbnails. Click on them to get up-close and personal with the 2-71. Sorry, no pics for the 1-71.