Monday, 29 December 2014

More on Oil

So today, we have something like 6 degrees Celsius outside. I've just completed about 3 miles around town and I have measured the oil temperature using a meat thermometer (not the one that was stuck in the Christmas turkey last week), and we have 73 degrees Celsius.

Please don't stick it in the Turkey
If you use one of these though, you will need to stop the engine first. The RF interference driven by the HT cables is phenomenal and stops the thermometer working at all. I guess the thermocouple wires pick up the RF.

Now, looking at the oil data sheets, we can see that the kinematic viscosities are similar at 100 degrees Celsius. Plotted they look like this:

This gives us a clue, but we can't get too excited because there are only two data points per grade. The temperature/viscosity relationship is not linear. Look at this, plotted on a logarithmic scale:

This gives us a better picture. Notice that the viscosity of the SAE 50 is about 20 cst at 100 degrees C, which agrees with the data in  the first chart, but that at 75 degrees C the viscosity is about 40 cst. That should give us about a bit more oil pressure for the same temperature than the 20/50W we are using now.

Friday, 26 December 2014

Oil & Oil Pressure

So this is one of those posts that we use to collect thoughts.

Amelia has done about 25 miles, and has had her head torqued down and the oil pressure relief valve spring replaced. She is running in on Tetrosyl 20w/50 multigrade oil, to API SE spec. When cold, the oil pressure gauge shows about 55-60 psi at idle speeds; after a 6 mile run around town, at speeds up to 25 mph or so, the oil pressure drops to almost zero at idle.

Darcy's Law tells us that flow through an orifice is proportional to the area of the orifice, the pressure drop across the orifice and the viscosity of the fluid.

The viscosity of the 20w/50 will be something like 18 centistokes at 100°C, rising to something like 135 cst at 40°C. After the run, the oil was visibly very free flowing and hot.

The relief valve may be controlling the pressure at idle - it may prevent the pressure exceeding 55-60 psi, but if this is the case it will have no effect at higher temperatures unless it is stuck open which seems unlikely given that the pressure displayed on the gauge falls gradually as the temperature rises.

Additionally, the effect is repeatable - the cold idle oil pressure is always 55-60 psi, and it always drops to near zero when it is hot, so I do not think we are seeing mechanical changes in the relief valve or the engine clearances - yet.

The Morgo is a gear pump which requires clearances between its gears to operate, and these clearances will allow more fluid to slip ineffectively around the gears when the viscosity is low. Are we right to fit gear pumps when the engine runs so hot?

And then there is the gauge. Ariel indicate that these engines should run at about 40-45 psi on SAE 50, so my gauge readings are not unreasonable, but we don't really know how accurate the gauge is or even if its accuracy is linear.

But one thing is clear - when the oil is hot (and we don't yet know how hot it really is) the loss in viscosity is such that the pump can push its full flow capacity (which itself may be compromised by low viscosity) around the engine without creating any pressure in the system.

And pressure is needed to prevent the bearing surfaces coming into contact with each other.

Wednesday, 3 December 2014

Oil... everywhere

So a few days later and I run the engine with the front main bearing cap off. I'm not going to show you a picture, because there would be oil everywhere, but you can see from this shot with the cap loose that there is oil there; with it off it gushes out:

And here we have a new spring just arrived from Draganfly, about 25% longer and somewhat stiffer than the old one:

And here we have an oil pressure gauge showing a previously unreached figure, 55 psig, at tickover - engine relatively cold (about 5 minutes running). Thats more than 50% up on last time.

So now we need a road test. Yippee!