JohnDs PI fuel pump development

[Nick Jones]

14 minutes ago, JohnD said:

Is your fuel cooler in the return from the Mu? 

EFI (in this version at least) is different in that the PRV is up front on the fuel rail, so the return flow is everything the PRV dumps. The good news is that this prevents (very) hot fuel in the fuel rail, but it does mean that the recirculating fuel pics up a fair bit of engine bay heat, so the whole fuel volume gets quite warm. I’ve known the Vitesse fuel tank get to between 40 and 45 degrees, which is not ideal. Pump gets a bit shouty.

[JohnD]

Ah! Thank you, Nick, I had forgotten that Roger has a modern system!

John

[JohnD]

Right!  No reply from Glencoe, and that rubber washer/O-ring looked so - wrong!   So I've reassembled the Facet without it. And it runs fine!

Replumbed the system as described above, and ran the Facet, then plus the main pump.  And the return from the swirl stays just the same! Not the dramatic drop in return that happened before.    Either the Facet repair or the replumb has sorted it!

Test tomorrow, and do a flow check.  And hope to hear from Glencoe next week.

PS Email last night from Glencoe (Terry Huxley) to say that the ring should be there.  It's "not available as a spare part" but they will find me one!   Does that mean that any O-ring of the right size will do?   J.

Edited October 9, 2022 by JohnD

[JohnD]

Glencoe may not have a spare, but I do!   A search of my spares shelves found, in my box of O-rings, a packet marked "Viton" of rings that are a perfect fit!    They may be a bit thin, 1.2mm.   The old ring was too deformed to show its' original thickness but it looked a bit more substantial than that.  Any road up, I've fitted one.

A flow test showed the Facet to have a slightly lesser flow rate than without the O-ring, 1.8L/min or 108L/hour.  This could be due to a slightly smaller stroke with the ring in place.    And, with both pumps running through the new plumbing, the flow was no less.     Thank you, Nick!

Then, I ran the engine while I measured the temp from the parts of the system.    Ran it for 20 minutes, until the coolant temp had stabilised at just less than halfway on the gauge.   Temps measured at the M/u return, the PRV return (to swirl pot), Swirl pot return 1 and 2, the supply to the main pump and the main pump itself.    Here are the results:

As you can see, the swirl pot returns and the Main supply do not vary significantly, and appear to stabilise at about one degree higher.   This must be due to full mixing in the pot, so I won't biother to measure all three in future, justb the main pump supply.     Meanwhile the M/u return had risen towards the same value, perhaps as warmer fuel reached it and gained heat from the engine, and was returned.

The PRV return rose much higher and more quickly to nearly five degrees higher than at the start but appeared to plateau at the same time.

The main pump itself warmed up even quicker.   It was still rising at the end of the test when it was nearly ten degrees warmer.   The slightly greater starting temp is due to the preliminary flow tests.

Conclusion.

The replumbing has, as predicted and with the new O-ring in the Facet, allowed a better filling of the swirl pot.   This should prevent the intermittent loss of main pump pressure that plagued me before!

BUT, it needs a road test, to see what temp the main pump will plateau at,  and a race test.   Another session at Three Sisters would be the formal answer, but I have entered the CSCC race at Cadwell on the 29th, so that will do.   I hope!

Thanks to all for your advice and experience.     Call it the University of Sideways or the Sideways Massive, you are invaluable!

John

 

 

 

Edited October 11, 2022 by JohnD

[Nick Jones]

Good work. Suspect you’ll be ok now that the swirl pot is properly and reliably full.

[flatter4]

Well done John for finding the root of the problem.

It's a fascinating read, and your fuel system an elegant (and complicated) design.

Some thoughts:

Both pumps will be more than sufficient for your needs.  They create flow, and only develop pressure as a result of resistance to that flow (don't worry that the facet doesn't match the Bosch).  The vent from the swirl pots needs to be restricted, it's a bypass that wastes flow.  Make sure you're not taking too much out....

An additional strainer on the inlet to the facet would improve long term reliability, but as Nick says this is the most sensitive part of the system, so nothing that adds too much restriction.  A fine mesh "rock stopper" will do.  20 to 40 um feels about right.  It's low to negative pressure, I like the transparent glass units for a visual on the crap you've trapped....

You're measuring pressure at the engine end I think.  The pressure from the Bosch to the regulator needs to be sufficiently above control pressure to give reliable regulation under full fuelling conditions.  Not sure what you'd want (@Nick....) but probably twice control pressure at idle?

I doubt you need a cooler, plenty of radiation from tank, pipes and hoses.  Adding 3 litres of fuel will vastly increase the cooling capability (so don't run long periods at idle with an empty tank)

Cheers, Will.

[JohnD]

Thank you, Flatter!

The swirl pot vent is part of the design - I presume you know this? The input is circumferential to make the fuel circulate the pot, 'swirl', and ensure that bubbles go to the vent in the centre of the lid and are excluded.     Some flow is necessary, but the hoses are 4mm ID.   

The Facet has a replaceable, internal filter at 40microns.  I took the opportunity of the replumbing to remount the Facet with the opening end upwards, to facilitate filter changes that will become part of annual maintenance.

I have bought some 'temp tabs', sticky labels that change colour permanently with temperature and will run with them on test and when I next race.   They only register between 37 and 100C, but that was the lowest range I could find.    If any of the system gets up to blood heat, I am in trouble!

I've had bad experience with the glass filters and now avoid them.    Giant metal jacketed fuel filter arrived yesterday - about 70x90mm!   I've built a bracket for it rather than just hang it on the hose as it must hold about half a pint of fuel, and will fit that today.

John

[JohnD]

And today, grabbing a gap in the showers, I took a test drive, 40 mins of fast country roads.   And the new plumbing seems to be behaving.  Afterwards, the temps were as follows, all in Centigrade:

M/u return - 23

PRV return - 27

Swirl pot - 19

Post-Facet filter (the new big one) - 16

Post swirl pot filter - 17

Main fuel pump - 37   This is another ten degrees warmer than it got to in my static test (above)

None of the temp tabs on PRV return, swirl pot or main pump had registered at all.

The ambient temp is 14C

The main pump is my only concern.    It's getting warm because it's working hard, but the fuel supplying it is not, and presumably keeping it cooler than it would have been.   I didn't measure the temp downstream of the main pump (Doh!) but would presume that the PRV return temp represents that in the pump outlet, so the fuel is gaining about ten degrees by passing through the pump.

As Nick advised, a cooler is probably unnecessary in the present weather conditions!     If I had been racing last summer, I wonder if one would have been useful, so I'll be repeating these measurements next year.   

If observations then seem to indicate a cooler should it be in the main pump outlet hose?   The fuel will be hottest there, so heat exchange should work best.

John

PS next weekend, 29th I'll be at Cadwell.    I have tickets if anyone would like to come along!

[Nick Jones]

How are you actually measuring the temperatures?

I'm surprised by the 10C positive difference between pump temperature  and PRV outlet temperature. As I’m sure you know the Bosch pump workings including the open frame motor run totally immersed in fuel, so you could reasonably the case to be no hotter than the fuel exit temperature, with the possible exception of the area where the actual roller cell pump is approximately 1/4 way along from the inlet end.  Fuel pump exit temperature will be no hotter (and probably a couple of degrees cooler) than the PRV exit temperature as the shear through that will add something.

If I were adding a cooler I’d put it in the PRV return as that is where the fuel will be hottest, but also at lower pressure, reducing leak risk.

[JohnD]

Thank you, Nick!

I use an infrared 'gun', placed on the item to be measured. 

Yes, I doubt if any commercially available fuel cooler would be designed for Lucas Pi fuel pressure, so the PRV return would be safer.     But in road use, the proportion of pump flow being diverted by the PRV would be much greater than on track, where full throttle up to 80% of the time is the norm.   We are in experimental territory!    Testing, of different layouts is indicated.   I am attracted by this webpage, where a Raspberry is used to monitor nine temperatures at once!  Multiple DS18B20 Temp sensors interfacing with Raspberry Pi - Hackster.io

In F1, the FIA expects fuel to be no cooler than 35C!  Else teams would aim to get more in the tank thanks thermal contraction.  So maybe I'm stressing with no need!

John

[Nick Jones]

1 hour ago, JohnD said:

But in road use, the proportion of pump flow being diverted by the PRV would be much greater than on track, where full throttle up to 80% of the time is the norm. 

Yes, but that’s fine as the fuel temp is only likely to rise significantly when the proportion of diversion is high.

IR guns are low precision devices, especially when comparing things of different shapes, colours and materials. As I’m sure you know, the measured area is rather larger than the laser dot…..

[JohnD]

Absolutely, Nick, which was why I held the 'gun' with its window in contact with the object to be measured.

And your first line has dispelled all my concerns!

Onwards and upwards!

John