Model Diesel: 32mm bore, 38mm stroke, indirect injection

[minh-thanh]

Hi Nerd1000 .
My way is : make the engine run and then change it the way I want
When the engine is running it will tell me : fuel pump : ok , ball valve : ok , injector : ok , fuel quantity : ok , injection time , compression ratio ..... ok , And then I change the type of injector , the way it injects, change the fuel pump, change springs....and change ..... ,
I didn't change the type of combustion chamber so I don't know
Everyone has a different way of doing things and a different way of thinking, I'm sure that type injector is fine with my engine so go ahead with your project - I just hate it.
Make sure the injector spring doesn't change with the high temperature of the injector - with valve injectors it doesn't matter much.

 

[Nerd1000]

Hi Nerd1000 .
My way is : make the engine run and then change it the way I want
When the engine is running it will tell me : fuel pump : ok , ball valve : ok , injector : ok , fuel quantity : ok , injection time , compression ratio ..... ok , And then I change the type of injector , the way it injects, change the fuel pump, change springs....and change ..... ,
I didn't change the type of combustion chamber so I don't know
Everyone has a different way of doing things and a different way of thinking, I'm sure that type injector is fine with my engine so go ahead with your project - I just hate it.
Make sure the injector spring doesn't change with the high temperature of the injector - with valve injectors it doesn't matter much.

I use stainless steel springs and the injector is actively cooled by fuel flow (some leaks past the needle seal ring and will flow back to the tank via a return line). So hopefully the injector won't get too hot. Heat is also a big problem for coking of the nozzle tip so certainty something to avoid.

 

[Lloyd-ss]

edit for brevity ............

What is clear from my research is that my style of injector is totally unsuitable for a direct injected engine like yours. It's usable only for indirect injection designs, or possibly also engines using the MAN 'M-system' of a small swirl bowl in the piston crown. ......................................

Nerd 1000,
I am really enjoying your posts and appreciate your sharing, warts and all. Your progress is uplifting. You seem to do a lot of research and planning and that is a good thing, but we all certainly have our own philosophical approaches to such a project.. which is really a total learning experience that enriches our lives.

It looks like you are using a pump-line-injector system, with the injector being a differential pressure type. I am working on one also but haven't posted in a while. I downloaded the shop manual for the Yanmar "L-series" single cylinder portable engines that run at 3600 rpm and have displacements from .19 to .35 liter. The manual is 130 pages and is very complete with lots of theory, diagrams and specs. They also use the P-L-I system, with a differential injector, and inject directly into a chamber in the piston crown like you mentioned. I bought one of their injectors and pumps (pretty cheap, actually) to look at their construction. Impressive in their robustness considering the small size of the engine. They are obviously built to last. Just for reference, their firing pressure in a test set-up is 19.6MPa, and I tested the one I bought and found that to be true. Seems pretty extreme and it seems like the leakage problems of a home-built unit increase with the square of the operating pressure, LOL. I just thought I'd throw this info into the pot.

Keep up the excellent work that you are doing. I will get back to sharing the progress on my thread very soon.
Lloyd

 

[Roger B]

Some thoughts and suggestions on small diesels. I have got as far as a runnable 20cc, 25mm bore 40mm stroke, 4 stroke diesel. This runs on standard pump diesel. The video is a warm start, opening the fuel rack as I start to crank.



The atomisation from the injector is not very good yet. At least half the fuel is ejected unburnt from the exhaust. Compression pressure is around 35 bar and the injection pressure is over 100 bar.

The cylinder head is flat with a hemispherical combustion chamber in the piston. The injector is set at an angle aiming at the centre of the chamber.

The injection pump is 2mm bore and 2mm working stroke, the first 1mm closes the inlet port. The fuel quantity is controlled via a spill helix so the timing of the start of injection is constant. The injector is a poppet or mushroom type based on the GM 5.7L design. I have experimented with needle type injectors but getting a good seating down the end of a long bore was very difficult. With a poppet style the seating is visible. Find Hansen and Don Comstock both use this type.

The fuel pump plungers are made from commercial hardened and lapped pin gauges. The metering helix is ground with a small cutting disc. The pump body is made from hardened silver steel lapped to size using an Acro needle lap and 1 micron diamond paste.

The injector needles are turned from silver steel and just the tips are hardened. The bodies are stainless steel. The seating is lapped with 1 micron diamond paste. I have tried several different angles for the tip, currently 10° included seems to work. If the angle is too shallow it will tend to jam.

I built a test pump for the injectors so I can apply a known load to the plunger with a spring balance. 50N on a 2mm plunger gives around 150 bar opening pressure. Don’t get it anywhere near your skin. The spray pattern is still not very good with a mixture of well atomized fuel and wet drops.

 

[Nerd1000]

Very cool Roger, you've done what I am trying to do!

Regarding your incomplete combustion issue, I have a few thoughts...

There is some good public domain info on diesel combustion systems on the web if you look hard enough. What's clear from my reading is that you need two factors to get good combustion in a diesel: adequate atomisation, and sufficient mixing of the air and fuel during the burn. The swirl chamber I'm planning to use is one solution to this; the strong swirl in the chamber ensures intimate mixing of the injected fuel with the air in the chamber (about 1/2 of total). The remainder of the mixing process happens as the hot partly burnt mixture rushes from the swirl chamber into the cylinder, creating lots of turbulence that mixes the remaining air with the still burning fuel.

In a DI engine getting the mixing to work well is much harder. You need to provide sufficient air motion in the cylinder itself to mix the fuel spray with the air. This is usually done through careful design of the inlet ports to create a strong swirl in the cylinder. Having more fuel jets aimed around the cylinder reduces the required motion, hence most full sized DI diesels using multi hole nozzles. On our scale, perhaps you would want to look at adding a 'lip' to the injector needle to try to spread the spray pattern out more.

 

[Lloyd-ss]

Roger, your set-up with the adjustable injector test pressure is clever. Iike it.

 

[Roger B]

Thank you both

I have a couple of sources of diesel information, some older versions of the ‘The Modern Diesel’ and T D Walshaw’s ‘Diesel Engine Design’. (Washaw published model engineering things as Tubal Cain).

Diesel Engine Design by Walshaw T D - AbeBooks

The Modern Diesel - Fifth Edition

Turbulence appears less of a problem in a small engine as there is less distance for the fuel to travel to mix with the air. I have designed moderate turbulence with the air from the squish band being forced into the combustion chamber as TDC is reached. Too much turbulence can cause excessive heat loss.

I still thank most of my problem is atomisation. One of my tests is to ignite the fuel spray with a small flame. Experience shows that if I can’t light the fuel spray this way it won’t ignite in the engine. Whenever I make this test there is always of layer of fuel on the workbench underneath where the heavier droplets have fallen. The ‘ink blot’ test shows the same.

 

[Nerd1000]

Thank you both

I have a couple of sources of diesel information, some older versions of the ‘The Modern Diesel’ and T D Walshaw’s ‘Diesel Engine Design’. (Washaw published model engineering things as Tubal Cain).

Diesel Engine Design by Walshaw T D - AbeBooks

The Modern Diesel - Fifth Edition

Turbulence appears less of a problem in a small engine as there is less distance for the fuel to travel to mix with the air. I have designed moderate turbulence with the air from the squish band being forced into the combustion chamber as TDC is reached. Too much turbulence can cause excessive heat loss.

I still thank most of my problem is atomisation. One of my tests is to ignite the fuel spray with a small flame. Experience shows that if I can’t light the fuel spray this way it won’t ignite in the engine. Whenever I make this test there is always of layer of fuel on the workbench underneath where the heavier droplets have fallen. The ‘ink blot’ test shows the same.

It's curious to me that your injectors if anything seem to make a narrower spray than mine, despite my design being a pintle type that really should be expected to spread less than the poppet type you're using. My injector does seem to lack consistency, so maybe flaws in the finish of the tip are making the spray wander around the place and appear to be more spread out than it actually is.

Anyway, I've certainly noticed that I get a distribution of different droplet sizes like you do. I guess higher injection pressure would help- the only other thing I can think of for your style is to add a second conical surface to the poppet so it looks like this:

The sealing surface would still have a 10 degree included angle, but there's another cone below with a larger included angle. I'm thinking that impingement of the spray on the lower cone would help break it up into smaller drops, kind of like some garden hose nozzles.

For my design this isn't possible, maybe I could experiment with different pintle shapes, or go to a poppet type nozzle. Testing will be needed to see how well my style performs.

 

[Nerd1000]

The design of my engine is more or less done, I still have some odds and ends (governor, fuel tank, exhaust pipe) but we've reached the stage where I can start making major components.

I'm planning to make a lot of the larger parts (block, lower crankcase, etc) as castings. Here are the patterns and core box for the main block, I did have a

go at casting this part on the weekend, but had a misrun due to inadequate venting of the mould, and couldn't re-try due to having only made one core. Next try will be after Christmas, I'm fairly confident it will work.

 

[Arild]

Hi! I've made a lot of castings in both aluminium and bronze. Remember to have some overshoot at the surface you need to mill off, and add a 2% to it all, (in aluminium)

 

[Roger B]

It’s an interesting problem to know what is going on when the fuel is injected into high pressure (and hence high density) air. My initial experiments with needle type injectors with various orifices from 0.2mm – 0.5 mm seemed to produce good atomisation but were not successful when installed in the 2 stroke diesel. The airflow/swirl was obviously different to the current four stroke.



My needle injector design will need significant modification to fit the four stroke cylinder head.

 

[Nerd1000]

It’s an interesting problem to know what is going on when the fuel is injected into high pressure (and hence high density) air. My initial experiments with needle type injectors with various orifices from 0.2mm – 0.5 mm seemed to produce good atomisation but were not successful when installed in the 2 stroke diesel. The airflow/swirl was obviously different to the current four stroke.



My needle injector design will need significant modification to fit the four stroke cylinder head.

View attachment 143138

There are a couple of things we can know for sure: in the cylinder the atomisation will likely be worse (less pressure drop across the nozzle, though perhaps the poppet type nozzle isn't affected by this because the pressure helps the spring hold it closed?) and the air is much denser, so the droplets will not travel so far.

 

[Lloyd-ss]

Nerd 1000,
With your tenacity, success is the only possible outcome.

I hope you are not too tired of being bombarded with slightly relevant information (LOL, no offense intended), because here is some more. Yanmar makes a LA series portable engine in a range of 0.2 to 0.4 liter, and up to 3600 rpm. It is a direct injection. Notice that the combustion chamber is off center, but remains in line with the injector nozzle. Whether this was strictly because of space limitations around the valves, or whether it is off center strictly by design, I do not know.
I know that you are at the point of no return on many of the design aspects of the engine, but just for reference here are a few detail drawings from the Yanmar service manual. The pump plunger is .293" in diameter, the full srtoke is approx 5mm with a quick rise and then a 180 deg dwell on the plateau. The helix ramp on the plunger has a slope of approx 0.17" and controls the end point of the injection. When I measured the parts, I was rather surprised at how large the dia and stroke were for that displacement of an engine. The opening pressure of the injector nozzle is set at 200kg/cm^2 (2,800 psi). Even if the compression pressure in the engine is as high as 1,000 psi, that still means the relative injection pressure is 1,800 psi. Again, that pressure surprised me.
I hope you find something useful in this info.
Lloyd

 

[peterl95124]

Lloyd, thanks for posting the drawings for the Yanmar LA,
FYI, the plunger length is about 5 X its diameter, a good rule-of-thumb,
one thing that still puzzles me is the inlet hole to the pump chamber
is blocked by the plunger most of the time, so it seems that fuel
can't get back in, yes its open when the plunger is fully retracted, and
when the cut-off groove matches up with the inlet, but between those
two extremes it seems there would be "vapor lock" so to speak

 

[Lloyd-ss]

Peter, yes, I had already given the vapor lock situation some thought and do have a reasonable guess, but there are a couple of more oddities.

The first thing, which really isn't odd, is that the cut-off groove is connected to the tip of the plunger via a drilled hole. You can't see that drilled-hole feature in the drawings I uploaded. So when the cut-off groove passes over the inlet hole, all of the residual pressure in the pump chamber shoots back into the inlet line. The part I find a bit odd is that the supply line and the return line are the same thing. So there must be a LOT of turbulence in that fuel line.

What may be mitigating the vapor lock phenomenon is the fact that the plunger is pumping upward, and that the fuel tank is directly over the fuel inlet line, so there is strong gravity flow into the pump body, but also with all that turbulence. Somehow, that combo seems to work.

The second odd thing is that the cam lobe for the fuel pump (you cannot see the lobe in these drawings), holds the plunger at the maximum pressure portion of its stroke for a full 180 degrees of the camshaft revolution, which would be for a full 360 degrees of crankshaft rotation. Maybe holding the plunger all the way in, after it has spilled all of its residual fuel and pressure back into the inlet line, helps any vapor lock to dissipate.
Just my thoughts.
Lloyd

 

[minh-thanh]

Here are 2 pictures

 

[Lloyd-ss]

Again, interesting that the high pressure fuel pump spends 2/3 of its time in the full compression position, and only briefly blips up to expose the inlet port. The low pressure pump is synced with the inlet timing of the high pressure pump, which certainly makes sense.

 

[Nerd1000]

Again, interesting that the high pressure fuel pump spends 2/3 of its time in the full compression position, and only briefly blips up to expose the inlet port. The low pressure pump is synced with the inlet timing of the high pressure pump, which certainly makes sense.

This feature is likely intended to prevent the engine running backwards (with the exhaust valve serving as the inlet valve) if it 'kicks back' while being started. This is a legitimate problem for diesels with jerk pumps whether 4-stroke or 2-stroke, many older Mercedes cars have a one way valve in the inlet so the engine chokes on its back pressure if it tries, and you can find many stories of Detroit Diesel 2-strokes doing this resulting in black smoke coming out of the air filter.

Edit: an addendum regarding the 'vapor lock' issue on the pump. My injection tester pump plunger has no helix or drilling, therefore there is no way for fuel to enter the pump on the return stroke until the spill port is uncovered by the top of the plunger. Presumably the result is the formation of a partial vacuum (and some boiling of the residual fuel in the pump) during the return stroke, but this doesn't seem to cause any problems so long as the return spring is strong enough.

 

[Nerd1000]

I've been working on the crankshaft for this engine. In my typical spirit of 'doing it wrong' the starting point was not steel but a bar of grade 450 ductile iron. This material has strength similar to mild steel and a surface hardness of about 180 Brinell, sufficient for operating against soft bronze bearings. I began by roughing out the profile, then set it up on a pair of offset centres to turn the crankpin. The setup was assembled on the mill table for alignment and locked in using a pair of setscrews and loctite retaining compound, then set up on the lathe.
After a loooong time offset turning at more or less minimum backgear speed to avoid chatter I was able to finish the crankpin with 1mm radii at the sides and polish the surface with emery. I was careful to polish the surface in both forward and reverse directions, this helps mitigate an issue with iron shafts where microscopic burrs are left by polishing, such burrs may later act like file teeth on the bearings.

Being rectangular and accurately sized, the offset centres then provided a nice way of fixturing the crank to form the webs and counterweights on the mill. I forgot to photograph this setup sadly. The majority of the material was removed using an 8mm '"corn cob" roughing mill. The radii were then cut with a 10mm ball end mill. I then removed the offset centres by gently heating until the loctite let go, here is the result:

I then inserted a piece of scrap as a gag piece and glued it in place with super glue for turning of the main journals and other features:

Overall I'm OK with how it's gone so far. I did have some issues with my parting blade wandering while roughing which led to the crankpin journal being wider than intended. This compelled me to take a little material off the webs to keep the overall width within tolerance. Also I goofed while chamfering the webs and nicked the crankpin journal. I'm confident this won't cause problems in such a low performance engine; it is less disruption than you'd get from an oil drilling and has a generous radius so there is little chance of cracking. But it upsets my inner perfectionist. I am however very happy with the ductile iron material, after all this unbalanced machining it shows no evidence of having warped!

 

[Steamchick]

Hi Nerd1000. Some of us talk the talk, others get on and do the job! - Excellent stuff! Don't worry about that odd nick from the tool, just make it the location for an oil drilling anyway! How fine did you linish the journals? I'm sure when crank grinding as a teenager, we were using finer than 600 grit, - I think the next may be 1200 grit emery? It was quite expensive stuff as I recall being told! - It came as a big coil of 1" wide... We had a special tool that had a fine adjustment then locking lever arrangement - it used interchangeable jaws that were usually the journal diameter plus a bit, so the linishing emery was clamped "just so" with not a lot of pressure, rotated a dozen or so turns slowly (100~ 200 rpm-ish?) - so on models I use Mole grips for the same task. But I have not made special jaws (Yet) as the "best of my grips" has good flat faces. But we NEVER ran 2 directions for linishing journals. I'm sure the crank grinding was always "clockwise" looking at the timing end (the direction of rotation of the crank in service), and we never had cranks returned (even from Haulage high mileage customers, or car racing guys) for Scuffed journals. We washed the journals (and rinsed oil-ways through) with paraffin before oiling when finished: the clean Paraffin always washed the "smokey" abrasive residue from the finished journal.
When I worked on engines for a major car maker, the Glacier company senior engineer explained the composition of aluminium-tin for the white- "Glacier-metal" shells was selected to have enough silicon (nodules) in the alloy to polish cast iron, without causing undue wear. But I wonder what bearing material you will use? Of course, you are right about the "grinding directionality" of the micro-structure of cast iron shafts. That was a key point about using the aluminium tin bearing shells instead of Phos-bronze metals on cast-iron cranks. - Or so metalurgists and tribologists at both Glacier and Welworthy - and my head office expert - taught me!

Well done on a proper job.
K2