DES 8/1 Propane Conversion project

[BruceM]

Waiting on the machine shop for the hardened valve seats... meanwhile my new neighbor is camping in a small storage building and the generator/DC power shed until the shop building interior is completed, then I can set up the engine in it's permanent home.

[BruceM]

Got the head back from the machine shop who installed hardened valve seats for propane use.  We also had the head and cylinder planed.  Alas, though I didn't see it before, now you can clearly see a hairline crack near the web between COV and coolant passage.  It is not in a thin section, seems to be a casting (shrinkage) flaw.  It took some side lighting and fiddling to get it to show up in the photo.
 


I lack the expertise to know what to do about this.  Should we get a replacement or just run it and see if we get bubbles in the coolant or around the COV plug?  

[32 coupe]

OUCH ! The crack may not go all the way through. It may just be a casting flaw.

I would give it a try before I gave up on it.
It would take some work to pressure check it. A custom head plate or assmble engine
then pressure check.

Gary

[dieselgman]

I concur with testing rather than assuming that is indeed presenting a problem (especially after investing the machine work in that one). There are quite a number of fine scratches in that surface but it is impossible to tell from the photo exactly what is a surface mark and what might show an actual casting flaw.

We will replace it later, if it proves to be a defect.

dieselgman

[32 coupe]

Hey Bruce,
Have you had any luck yet ??
Gary

[BruceM]

There's been a long hiatus on this project, due to the generator shed being used for temporary living space while the shop building was being completed for heated winter quarters. We are now back to the fine tuning, with the engine now housed in the permanent generator shed and supplied via the home propane tank and exhausting into an leach field type "earth muffler".  We made a copy of my pneumatic exhaust valve lifter and have electronic push and toggle button control of all the necessary functions now. Remote starting will come next.

The engine was starting flawlessly and running well but either cold weather or the planing of the head and cylinder by the machine shop raised the compression too much and we had a pronounced knocking at load (air compressor).  We tore her down and added 0.075" under the cylinder. That improved the knock and power significantly but was still slightly present at load.  We added another 0.030 and suddenly it would not start, and we're smelling a lot a raw propane. We would get maybe one ignition, then nothing.   Spark was confirmed good visually and by timing light when cranking. Compression is fabulous, no leaking. The CA110 carb diaphram was checked and still looks fine.  Very odd to suddenly have a no start situation with lots of raw propane smell after such utterly reliable and unfussy starting.  I next checked the IMP52 regulator, and we seem to be getting gas leakage on it's output, which should require negative pressure, so I will call Century Fuel Products Monday and hope to get some advice.

Very frustrating to be so close to getting her finally "dialed in" and then to be back to a no start situation.

[BruceM]

We've got her starting and running quite nicely again. Our problems were my homebrew spark module, confirmed via spark timing light.  It still misses under heavy load after a bit of run time and also with the generator head, but this has been confirmed as glitches in my 555 based hall effect/dwell circuit, with secondary failure of the power transistor due to coil kick back.  Coil voltages climb as compression increases with load.

I am hoping some auto savy readers might be able to suggest a three or four wire COP (coil on plug) unit that I could readily use.  Ideally it would be a unit that has a short spark plug wire, not a direct to plug type.  I have seen photos of such things but can't seem to find a make and year so I can purchase one.  My spark plug is buried in the head with access too narrow for the direct to plug type.  I need make, model and year(s) for your suggestion.

 I think Jeff would be best off with a single commercial COP unit that replaces coil and ignition module in one for ease of swap out troubleshooting/maintenance.  Then only my hall effect/dwell timing circuit is custom, and it has a blinking LED to confirm it's operation.

 My plan B is to go to a 4 wire GM HEI ignition module and our existing coil, that will be time consuming as it's designed for a reluctor/coil type input and what I have is hall effect, so I have to fiddle with a conversion circuit.

[EdDee]

Hi Bruce,

What about a SCR in lieu of the transistor? Swing the gate to negative trigger voltage to break the latch...

Regds
Ed

[BruceM]

Thanks Ed. There are plenty of cheap high voltage igbt transistors; due to the low spark rate, this one doesn't work very hard and stays cool.   I just didn't realize the flyback voltage would go to 500v+ on an HEI spark test plug.  I will add a better snubber to lower that a bit.

A 3 or 4 wire ''Coil-Near-Plug'' ignition coil sure would be nice; coil and ignition module all in one.  I just need to find a popular one so it's affordable.

Meanwhile I'm separating the dwell timing from the ignition switching, and sending the timing pulse via plastic optical fiber, since I have plenty of those relatively cheap parts on hand.  I've got a 600v IGBT transistor on hand I can use for switching, and will mount it on perf board spaning  the coil terminals.  Hopefully that will let us go forward while considering other options for long term.

[EdDee]

Agreed Mike...

What about really old school.... mech point and condenser...simple, easy, bullet proof? ... don't need advance/retard either because of fixed speed.... or on the other hand, the Honda Gx340 has a built in ignition circuit in the magneto coil... or like the robin EC10, a single wire module instead of the points, fitted as an aftermarket replacement often...

Regds
Ed

[BruceM]

A wasted spark system with mechanical contacts would be doable, though a pain the implement mechanically. I'd rather go with a modern electronic spark system.  I already have a dual hall effect spark timing system that works well and was very easy to implement.  One hall sensor is sensing the IP cam follower (with magnet on top) , which enables a second which senses a magnet on the flywheel.  Under $2 including the ceramic magnets.

Both the honda gx340 and the robin ec10 seem to be quasi-magneto spark systems,  I'm much more interested in an automotive coil-near-plug with ignition module integrated into the coil.  If only I can find one, preferably one in high volume with aftermarket parts available.

If someone knows of a vehicle with a coil-near-plug (3 or 4 wire, integrated ignition module) I'd sure love to know the make, model and year(s).
 

[buickanddeere]

Just use the platoon sensor and transistor pack from a 80's vintage Dodge .

[BruceM]

B&D - Thanks, could you maybe suggest a specific model and year?  What the heck is a platoon sensor?  Ford used hall effect, GM used reluctor type inductive pickups, but I've never heard of a platoon. 

I revised my spark board to optically isolate the spark driver from the hall effect/dwell timing section, and added 9v linear regulation to the hall/dwell section to give it better noise immunity.  That and adding 0.2 uF of capacitive snubber across the coil seems to have done the trick.

I still have some knocking at full load and temperature- so more shimming of the cylinder is needed today.  I did my full load test yesterday- compressor unloaded, 1/2hp submersible well pump running 1140 watts, washing machine starting spin cycle.  It handed it OK. 

Some serious carb metering issues still with the CA110 propane carb.  I have to manually adjust the fuel mixture via inlet restrictor for each load.  It needs to be richer to start and no load, leaner proportional to load.  There is supposed to be a way to shim the fuel mixer to adjust this- but I haven't been able to find a ''how to'' article yet.   The fuel rich/lean adjustment on the carb has no effect at all.  The idle air bleed adjustment has only a minor effect.  I get missing (no combustion) and/or backfires (more like a little pop) when the fuel mixture is off.  Spark is confirmed during missing via inductive pickup timing light. 

I think we'll try the CA125 carb again and see if that behaves better.   

I'm going to look for a VERY long reach plug in hopes of changing spark placement within the pre-combustion chamber, and also will try increasing the spark gap in hopes of lessening sensitivity to fuel mixture.  I have already bench tested the spark system with a huge gap and the flyback voltage was still OK with the new 0.2 uF snubber.

[Tom]

A 2001 Subaru has a very long reach plug. The 2500cc 4 banger.

[BruceM]

What model Suburu, please, Tom?

After reading the Impco training manual I'm getting suspicious that the CA110 carb might be having trouble. It seems I'm going to learn a lot more about propane metering/carbs than I'd like to.      Shims under the metering valve can be used to lean the full power mixture, and that may be something I need to do. 

One concern I have is the abrupt lack of ignition we get when the fuel mixture is too rich or lean. That's not normal.  Sparking in the indirect injection chamber is less than ideal.  I'll proceed and hope I eventually get it figured out or, preferably, just get the lucky combination that works.  One thought I had is to have the cylinder head plug bored and tapped for the sparkplug, and then I could make a hollow cylinder that would fit in the injector hole.  This would help reduce compression (and our growing stack of shims) and allow more fuel mix to flow into/through  the combustion chamber.  Just rambling, I hope that isn't necessary.

I'm adding a neon spark indicator to the coil- powered by the flyback spike.  It should let me reduce the snubber capacitor (condenser in automotive speak), and having a visual spark indication is handy when trouble shooting rough or missing situations. It will let me have less wiring dangling around, as I've been using the timing light for this.