MAN Diesel engine from 1907 "DM 2 * 100"

[gus]

You can't beat the old MK1 eyeball Gus

Hi Skyline,

Way back 2001 when I went around Central China to audit foundry suppliers.I sure hated myself for putting a foundry owner and his "expert" into a tight spot. Asked owner how do they read cupola furnace melt temperature and has he got a pyrometer. The reply was pyrometers no good,not accurate,expensive and require calibration.My assistant has 40 years experience and he can read within plus/minus 50 degees C. by eyeball!!!! And every now and then we borrow neighbour's pyrometer to calibrate him.Ha Ha.We took off the "Certified Suppliers" List.

We had to go back to him as he supplies best Nodular Iron for our con-rods.
Suppliers with pyrometers that we qualified did not.We had to rewrite our ISO 9002 Procedures and Work Instructions to qualify/certify him as Suppier.
His Nodular Con Rods rang like a bell.Others sound like dead C.I.

So much about eyeball temperature reading.

 

[MuellerNick]

Please keep the thread on topic.

I'll post new pictures at Sunday. Tomorrow will be casting day with two friends. And I'll do "production casting" (no new patterns). But I do have a lot left to show.


Nick

 

[MuellerNick]

I milled a core box today. So I couldn't resist to make a test-cast with it.
It is one of the small core sand moulds that are vacuum assist cast.

Machining time was about 6 hours. As the box is quite deep, I had bought a new mill. Diameter 1.5 mm, 16 mm long (plus the shaft). Full carbide, costs 40 € a piece. They will make you cry if you break one. They are so sensible, that they do have an RPM limit of 30000. I milled with 20000 RPM.

This is casting #8, an arm for the collector brushes of the generator.


The part, sawn off.


Other view.


This mould casts two parts (I need more of them). The block marked red gets sawn off. But first, it helps clamping and machining the part.


Core box and cores. One core box makes two halves. The core sand mould lacks a gate. I drill that later (that's more efficient). You can see the hole in the upper mould half.


An assembled mould.

Unfortunately, both test castings are a bit shifted at the splitting plane in the same direction. Have to find the reason for that.
With these core sand moulds (their boxes), you have almost no chance to repair them. They either work, or you throw away at least a day's work.

There are some areas with which I'm not completely happy with, but I didn't have a release agent in the box yet. And, the boxes get a bit better with some usage (and experience, how to rattle that specific core out)..


Nick

 

[MuellerNick]

A few more castings ...

Casting #9, post for rocker-arm shaft.


A simple piece cast in sand. You might ask why I don't make that on the lathe. Well, only castings look like castings.


The split pattern.

Casting #10. The housing for the injector pump. With this size, I don't think it will work. But in the original, it was low pressure, as the injection system was with blast air. They didn't get a high pressure diesel pump working. They had enough accidents around that.


The casting (sand) ...


... and the pattern.

Castings #10 … #12. Rocker arms.
I'm a bit lost now wich one is witch. But one is for intake and exhaust, the one with the 90° bend for the blast air and one is for the fuel.


All three of them, cast in sand.


One of them from the side.


And the patterns.
Initially, I had a set of patterns, that looked quite similar, but didn't work, or with a 90% reject. The problem was the gate, as I didn't had any good place for that. The other problem was, that they kept falling out from the sand. They were to shallow. So I added the cylindrical part (no draft). That is a machining surface, so the extra material doesn't hurt. Now the cylindrical part plays two roles: Keep the pattern in the sand and act as a gate. They always work now.


Nick

 

[MuellerNick]

Casting #12, cylinder head.

The cylinder head of the original is water cooled, so the head of the model has to be water cooled too. I had no information of how it looked at the inside. I only found a picture of that time of a cylinder head repair that allowed a look inside. But if you look at the features of the visible outside, consider where the valves, injector and compressed air starter go, it isn't too complicate to construct a water jacked from these constrains. At least, the water jacket is reasonable, but certainly off in some details.
The biggest challenge here was to find a way how to integrate the water jacket into the pattern, get it into the mold (without falling out), finding a way that it doesn't float during the pour and finally get the core sand out again.
All that are a few days staring at the drawing and scratching my head over and over again.

But here we go …

The cylinder head as cast.


Other view (of a reject) with bottom and top milled flat. The original had those faces milled too.


Other view (reject). Those two "nibbles" required a core just to get them. I found no other way to solve that detail.


Other view (reject) into the inlet/exhaust port. The heads are symmetrical, there is no distinction between inlet and exhaust port. On the two-cylinder version, both inlet ports are face to face.
Left from the port is the inlet for the cooling water. After drilling, it will have access to the water jacket. The little feature right and below of the port is a passage to the cylinder (A-frame) for cooling water.


A reject that has some part milled off to show the inside and the port.


The pattern for the cylinder head.


A view that shows where all the cores go to.

TBC …

 

[MuellerNick]

Now to the cores …

The water jacket.
As I described above, getting that thing into the casting was a bit tricky. My solution after some experiments was to "nail" it into the mould's sand. Initially, I glued the "nails" into the oil bound sand with CA. Sounds like that doesn't work, but it does. But then, I found that it is good enough to push the core (with the nails) into the sand and be done.
To avoid the core to swim to the top (the core is fixed in the cope), it has four "nipples" at the "nails" that have contact to the cope. So the core for the water jacket hangs down from the ceiling, so to speak.


Seen from the top, with the four nails (Al filler rod) and the four nipples at the root of the nails.


From below.

The core for the water jacket is made out of three parts that get glued together.


The top part's core box. It has two loose inserts. This core box requires an additional fixture for the nails to guide and stick out from the bottom of the core box (not shown).


The core box for the lower two parts of the water jacket.
These boxes are from my early days where I painted them. I no longer do it this way. The binder actually sticks to the paint. Now, I just mill the boxes and they get a wash of release agent (some PTFE I guess in an evaporating agent).


Left and right cores for the port features and the coolant inlet. Port glued in place.


Core box for the core from above. One assembled, the other one disassembled.


Core box for the port (inlet/exhaust).


Jig to glue the two cores together properly.


How the ports and the water jacket cores are located in respect to each other.


Different view.

TBC …

 

[MuellerNick]

... and the core and core box for the "nibble" feature on the outside:

Core


and core box. It is open on two sides, so I lay it on a flat surface, ram the core sand, activate the binder and split the box.

All that is quite logical. Problem is to find the logic.

Here is a video about the cylinder head:
[ame="http://www.youtube.com/watch?v=_M0DxxlIK3I"]YouTube[/ame]

Enjoy!
Nick

 

[CF-EKM]

Fantastic work!
Thanks for sharing your ongoing project!
The cylinder head is very nice, it must require an enormous work and time to earn these results.

Christian

 

[MuellerNick]

Just came to my mind:
My melting furnace is oil fired. It is "a bit" oversized for these parts. It has about 200 kW at full throttle. But I don't let it run full open, because most of the time I use an iron crucible (that won't make it much longer). Capacity is about 10 kg. The advantage is, that I can make moulds until I run out of boxes or sand, and pour them the next day with about two melts. I also have a pattern (spindle mount) that needs almost 5 kg of melt.

I measure the temperature of the molten aluminium with a thermocouple directly in the melt. Pouring temperature is between 660 °C and 680 °C, depending on what I cast. The above mentioned spindle mount works best at 640 °C. With these temperatures, I get less shrinkage and a better surface finish. Initially, I have cast with up to 720 °C, but that is nonsense.

For repeatable and predictable results, you need precise and repeatable temperatures.

BTW thermocouple:
The pyrometers you can get for a few dollars are absolute crap! They work in just one spectrum and rely on the emission rate of the material. For aluminium (and all metals) the emission is unpredictable. If the measuring spot (in case you know its correct location and size) is on a thin layer of oxide (you always do have that), the reading is off to a useless value. A real and working pyrometer costs well above 1000 $.

Nick

 

[kvom]

After watching the video, I gather that the molten metal goes down the hole from the top, through the filter, and then through the "tunnel" to go up into the mold. what is the reason for doing this vs. going directly from the top into the mold cavity?

Also, how do you get all the core material out afterwards?

 

[MuellerNick]

It is better to fill the cavity from bottom to top. If you fill it the other way round, the melt splashes, picks up sand, the oxide layer tears apart, etc, etc.
Sometimes that is impossible or would make more work like in the engine's base. The engine's base in an example of bad design from a foundry POV. Huge sprue, huge gate, cast from the top. I cast the base with 660 °C, but next time I'll try it with 640 °C and I'm sure it will work too and give a better surface finish.

The cores are baked out in an electric furnace at 480 °C … 500 °C for about 4 hours. I can save some time when I hammer out/off most of the cores. But with the cylinder head's water jacket, the only solution is to wait. And then shake out the sand through 4 small holes.


Nick

 

[billmac]

Nick -

In cast iron blocks and cylinder heads we used to use chaplets to stabilise cores and to prevent them floating out of position (maybe still done this way?). The chaplets became part of the casting itself, the size being chosen so that they retained just enough strength to hold the core but form a a properly fused part of the cast iron.

I'm not sure if you could use chaplets in Aluminium castings; they would have to be made of Aluminium of course and getting the right size would need some trials.

I'm not sure how 'chaplet' translates in German, so if not clear let me know.

 

[MuellerNick]

Chaplets are called "Kernstützen" in German.
But with the dimensions I work at, they seem to be of little help for me.
For the A-frame, I made a set of special boxes (for cope and $forgotIt) where I screw in the core. It is a bit odd, but works perfect.


Nick

 

[MuellerNick]

This vacuum assisted casting is going to blow my mind!

When I started this project, I didn't know how I could cast certain parts. They were out of the boundaries of sand casting. When I switched to core sand moulds, it opened new ways. The lid (above) worked at about 50%. When I switched to vacuum assistance, the lid jumped up to 100%. I couldn't ask for more.

But there was one part, that was behind my imagination. It was too thin-walled and a year ago, impossible for me to make.
I kicked my own ass and said to me: If you don't try, you'll never know. So I milled the core box, made a bunch of cores (including other ones) and gave it a try.

Here is my first try. Casting #13, bearing block for speed lever axle:

Without wash, without release agent, without tuning parameters, just the plain first try.
The wall thickness of the ribs is 1.4 mm.


One mould makes two of them, mirrored. The runner also serves for workholding and referencing.


Core box and core sand mold.

I need a beer now!
Nick

 

[AussieJimG]

Thanks for that beautiful video. I was getting a bit lost in the detail of that very complex head.

Jim

 

[gunna]

This is just amazing! I find it very interesting as I am presently involved with the restoration of a 12" scale engine of almost identical design to the MAN. If anyone would like to see pictures of the full size components that Nick is modelling from, perhaps I could start a new thread in the Break Room? I guess that would be the best place.

Ian.

 

[MuellerNick]

If anyone would like to see pictures of the full size components that Nick is modelling from, perhaps I could start a new thread in the Break Room?

Sure I would!
Similar design? If it is very similar, that would be a Langen & Cie from Italy. They have stolen the MAN design and ran into legal fights with MAN. Or a license build from USA (have no names in my head, just in Cummings book).


Nick

 

[gunna]

Hi Nick, the engine I am working on is a 3 cylinder Willans made in England in 1927! However, it appears they designed it in 1911 and never upgraded the design. As all diesels being manufactured at this time (1911) had to be made to Herr Diesels licence or not at all, they all tend to be very similar. It will take me a few days to get suitable pictures, but I will try to get the same image direction as the pieces you have made as that will make for an interesting comparison. This engine is 15"/380mm bore and 22"/560mm stroke. That seems to be the same as "your" MAN. Do you want me to continue in this thread or should I start a new one as I originally suggested?
Ian,
in Melbourne, Australia.

 

[MuellerNick]

Do you want me to continue in this thread or should I start a new one as I originally suggested?

I'd prefer if you start a new thread.
But you sure can set several links to yours (as progress goes on on both sides) here and back from your side. And if you want, you can take as many pictures from this thread as you want.

This sounds like a very interesting and promising parallel-thread!

Nick

 

[gunna]

Ok, I have started over in the Break Room with my pictures.
http://www.homemodelenginemachinist.com/f16/willans-blast-injection-engine-19174/