Engine Instrumentation Project - Page 8 - PeachParts Mercedes-Benz Forum

dieselbenz1 · #**106** 11-23-2015, 10:41 AM

The system I'm focused on building, uses a Holley carburetor float bowl to eliminate the return line to the tank. Return from the injectors/IP flows into the float bowl and fuel actually burned in the engine is made up through a low pressure boost pump through the float valve. This allows only one flow meter to be used. I've identified an eccentric gear flow meter which allegedly reads in the range required using a Hall sensor.

This is interesting and will likely be very accurate depending on the flow meter as you indicated. Just food for thought and I haven't put a whole lot of thought into it but the lift pump on the IP now is a constant volume pump. For each pump stroke a fixed volume is delivered you could easily infer rate and volume theoretically based on rpm. Still would require the return flow measurement and by paddle wheel I interpret that as a volumetric meter for a fixed volume you get a pulse? You maybe should also add a variable to each meter as a calibration aid in your software.

mach4 · #**107** 11-23-2015, 10:59 AM

While it's true that the lift pump is a constant volume pump, it provides excess fuel to the IP. This excess is returned through the overflow valve which also maintains constant pressure within the IP. Additional overflow comes from the injectors. Calibration is handled through constants in the code.

dieselbenz1 · #**108** 11-23-2015, 11:28 AM

Correct and that's why the return meter would be required. The holley bowl would not be needed but that's a good use for the bowl I always had better luck tuning quadrajets and eldlbrocks. The code becomes a tad more complicated though.

funola · #**109** 11-23-2015, 12:20 PM

Quote:

Originally Posted by mach4

..........The system I'm focused on building, uses a Holley carburetor float bowl to eliminate the return line to the tank. Return from the injectors/IP flows into the float bowl and fuel actually burned in the engine is made up through a low pressure boost pump through the float valve. .............

I must be missing something. I have a difficult time visualizing how this is possible. The lift pump outputs approx 1/2 a gallon a minute to the IP at idle (much more at higher rpm). Fuel returned to tank vs fuel burned I think is 80 to 1 ratio. How big is the float bowl?

mach4 · #**110** 11-23-2015, 12:40 PM

Quote:

Originally Posted by funola

I must be missing something. I have a difficult time visualizing how this is possible?

OK, imagine that you're doing a diesel purge. You've got a quart bottle with the purge in it and your pickup and return are both plumbed to the bottle. The engine happily runs until the purge is exhausted. Now imagine that you desire to keep the volume in the bottle at exactly the same level regardless of how much is used, so you have another little bottle of fuel and everytime you notice that the level dropped by a millimeter you poured in a couple cc. This makeup fuel is what was burned by the engine. Now imagine that you automate this process by introducing a float so that everytime a bit of fuel was used by the engine a corresponding bit was added. This added fuel is what is being measured by the flow sensor every 100ms or so. The size of the float bowl is irrelevant other than to say the smaller the better within reason. A float bowl the size of a toilet bowl wouldn't work very well.

mach4 · #**111** 11-23-2015, 01:06 PM

Here is a Holley float bowl. It will mount to an aluminum backing plate. The line to the lift pump will be in the backing plate.

funola · #**112** 11-23-2015, 05:11 PM

Quote:

Originally Posted by mach4

OK, imagine that you're doing a diesel purge. You've got a quart bottle with the purge in it and your pickup and return are both plumbed to the bottle. The engine happily runs until the purge is exhausted. Now imagine that you desire to keep the volume in the bottle at exactly the same level regardless of how much is used, so you have another little bottle of fuel and everytime you notice that the level dropped by a millimeter you poured in a couple cc. This makeup fuel is what was burned by the engine. Now imagine that you automate this process by introducing a float so that everytime a bit of fuel was used by the engine a corresponding bit was added. This added fuel is what is being measured by the flow sensor every 100ms or so. The size of the float bowl is irrelevant other than to say the smaller the better within reason. A float bowl the size of a toilet bowl wouldn't work very well.

I have done purge out of a gallon jug where the return fuel is back into the jug. In what you described, you are drawing from the tank and not returning to the tank. You are at the mercy of the lift pump, which you have no control over. Where is all that unburned fuel going to go?

mach4 · #**113** 11-23-2015, 05:30 PM

Quote:

Originally Posted by funola

Where is all that unburned fuel going to go?

It returns to the float bowl. A low pressure boost pump is what provides the pressure to the float valve. Think of your toilet bowl - as the level in the tank goes down the float valve opens and returns the level to preset.

funola · #**114** 11-23-2015, 05:44 PM

Quote:

Originally Posted by mach4

It returns to the float bowl. A low pressure boost pump is what provides the pressure to the float valve. Think of your toilet bowl - as the level in the tank goes down the float valve opens and returns the level to preset.

Are you bypassing the engine driven positive displacement lift pump?

mach4 · #**115** 11-23-2015, 07:47 PM

Quote:

Originally Posted by funola

Are you bypassing the engine driven positive displacement lift pump?

No, it's what pumps fuel from the float bowl to the IP.

Here's a block diagram...

<10,137>

dieselbenz1 · #**116** 11-24-2015, 09:13 AM

If the eccentric gear flow meter is a volumetric meter it should work well. The issue with measuring flow especially using a different type of meter is that when flow is interrupted it takes a while once flow is established to stabilize a flow meter therefore accuracy is compromised. If you can setup the float to generally be cracked open at idle that would be ideal, (proportional acting). If the float operated more as an open closed valve that may result in more flow than is actually being measured as the meter would see no flow then flow.

funola · #**117** 11-24-2015, 10:11 AM

Quote:

Originally Posted by mach4

It returns to the float bowl. A low pressure boost pump is what provides the pressure to the float valve. Think of your toilet bowl - as the level in the tank goes down the float valve opens and returns the level to preset.

Thanks for the diagram, it was sorely needed. I see you have rerouted the engine driven lift pump but that really does not change anything. You are still at the mercy of the lift pump. Do you know what dead heading means? That is what you are doing practically to the lift pump. Results will not be good. Problems that will arise may include: fuel leaks from blown hoses and gaskets. Overheating of injector nozzles and injection pump elements, damaged lift pump, from insufficient cooling due to lack of fuel flow.

Other problems I can see (from looking at your diagram), aside from the more serious problems mentioned above, is the boost pump must be capable of higher pressure than the lift pump, otherwise the float bowl will not get any fuel. The paddle wheel will be pulsing at the rate of the lift pump, going back and forth. Signal processing will not be simple.

ROLLGUY · #**118** 11-24-2015, 10:57 AM

Quote:

Originally Posted by funola

Thanks for the diagram, it was sorely needed. I see you have rerouted the engine driven lift pump but that really does not change anything. You are still at the mercy of the lift pump. Do you know what dead heading means? That is what you are doing practically to the lift pump. Results will not be good. Problems that will arise may include: fuel leaks from blown hoses and gaskets. Overheating of injector nozzles and injection pump elements, damaged lift pump, from insufficient cooling due to lack of fuel flow.

I disagree. The float bowl is acting exactly like the fuel tank. The fuel is taken from the bottom, and unused fuel is returned to the top. No extra pressure is in this system that would not otherwise be in the stock system, so no chance of blowing hoses. The fuel used is replenished by the stock fuel tank and added pump.

mach4 · #**119** 11-24-2015, 11:29 AM

Quote:

Originally Posted by funola

Problems that will arise may include ... overheating of injector nozzles and injection pump elements, damaged lift pump, from insufficient cooling due to lack of fuel flow.

I referenced this possibility in an earlier post, which is why I stated at that time that I'll be adding a temperature probe at the float bowl to monitor this in real time. And if it is a problem I'll add a fuel cooler to the return line very similar to the power steering fluid cooler.

Actually there is one concern that I have with the system and that is the capacity of the flow sensor and the potential pressure drop across it which could allow the engine to drain the float bowl and draw air at the very limits of the operating envelope...I'll share my calcs on that later.

Quote:

Originally Posted by dieselbenz1

If the eccentric gear flow meter is a volumetric meter it should work well. The issue with measuring flow especially using a different type of meter is that when flow is interrupted it takes a while once flow is established to stabilize a flow meter therefore accuracy is compromised. If you can setup the float to generally be cracked open at idle that would be ideal, (proportional acting). If the float operated more as an open closed valve that may result in more flow than is actually being measured as the meter would see no flow then flow.

Your theory is exactly correct - the problem here is not with the design of the positive displacement eccentric gear meters but the manufacture of this particular cheap Chinese one with tolerances and precision so poor it takes a huge flow just to start the gears turning and then sometimes they stick. The aircraft flow meters are almost all the vane type, which actually reduces the possibility of jamming due to contaminants. The floats are extremely sensitive so are "generally cracked open" at idle anyway. The k-factor in flow meters is surprisingly linear except at very low flow rates. The trick is to get one where the idle flow is in the linear range without exceeding the capacity on the top end. I'll share the calcs later.

mach4 · #**120** 11-24-2015, 07:25 PM

The calculations I used for sizing the fuel flow sensor are based on the very informative BMEP chart for the 617 engine. This particular one has a road load line for the 300SD. The map would be equally applicable for any turbocharged 617 engine. The road load is essentially the sum of aerodynamic drag and rolling drag on level ground at steady state at sea level.

You can calculate almost any fuel performance number from this chart. For example to get fuel usage in gallons per hour for a point on the chart locate the power in kW from the scale on the right and the g/kWh line and multiply them together to get grams/hour. Divide grams by 454 grams per pound to get pounds per hour and that number by 6.2 pounds per gallon (the density of diesel) to get gallons per hour. Now, if you know your rear end ratio you can find speed at that rpm and then convert gallons per hour to miles per gallon.

Here's an example - assume level ground, steady state, sea level - at 60mph I'm turning about 2300 rpm (I've got a 2.47 rear end) - that takes about 18 kw of power (24hp) and burns 300 g/kWh = 5400 g/hr / 454 g/lb = 11.9 lbs/hr / 6.2 lbs/gal = 1.9 gal/hr. 60mph / 1.9 g/h = 31.5 mpg which happens to be very close to what I get on long road trips.

So relating all this to the mpg computer, I was concerned about the possibility of exceeding the published top end of the sensor (2 to 30 liters/hour). A little work showed that while it is theoretically possible to exceed that the danger zone lies at the furthest reach of the envelope. At under essentially 3500 rpm I can't exceed the published specs and at max power at 4000 I'm only able to exceed the limit by 10% or so. What happens at 30 liters per hour I'm not sure - maybe the accuracy drops or possibly the pressure drop exceeds some threshold, or there is cavitation - but for a few seconds it probably won't matter. I'm not even sure I've ever been over 4000rpm with the engine