Fuel coolers and FI supply (Split from “Her name is Bella”)

[RedRooster]

On 11/18/2024 at 7:43 PM, Nick Jones said:

And oil cooler and fuel cooler. Oil cooler is a good shout on a 1500.

The fuel coolers got me interested, how much does it cool? 

[rogerguzzi]

28 minutes ago, RedRooster said:

The fuel coolers got me interested, how much does it cool? 

Hello RedRooster

                             I have no Idea I fitted it as most moderns have them I think and Nick said the fuel gets warm when the tank is low as it is going around the system and collecting heat when under the bonnet!

If I remember right it is fitted in the return line! but then again it maybe in the flow line I will look (if I remember)

I think it is a Bosch one as fitted to Diesels and cheap!

Roger

[RedRooster]

Interested in that, as I've got a return to tank line fitted for the triple DHLAs, not sure if I really need it now as the initial conking out problem when hot I traced to the solenoid, anyway I've kept it as it does get warm here in the summer.

[Escadrille Ecosse]

Curious about fuel return lines and whether or not they're needed.

I realise diesels generally do, along with associated fuel coolers but petrol engines are much more of a mixed bag. Carb and EFI.

Triumph never fitted them to 'our' cars on SUs or Strombergs but the Weber DGAS on the Essex in the Scimitar had a return to the tank. When I fitted the Rover V8 and Weber/Edelbrock 4 barrel I ditched the return line as the 4 barrel Weber has no facility for one. The RV8 EFI sysyem incorporates a return from the fuel rail but this is combined with a regulator sensing manifold pressure for mechanical fueling regulation without ECU involvement, rather than a MAP sensor feeding electronic signals to the ECU as you would have now.

I didn't fit a fuel return to the DCOEs on the Spitfire. This seems to be the usual arrangement for DCOE installations at least.

I will ultimately fit injection to the Spitfire and I have been wondering whether or not a fuel return is actually needed for this.

I realise most DIY EFI installations have a return from the fuel rail. But is this because it's been proven to be needed or because 'everybody else does it'?

Modern OEM systems frequently don't have a return from the fuel rail, thinking here of Ford and VW. The Fiesta doesn't and nor do at least some of the smaller VWs.

These deal with pressure/flow relief at the tank, either through a short loop or more often as part of the in-tank pump assembly with just a pressure regulator at the fuel rail inlet.

So does pumping a lot of excess fuel (itself adding energy) through a hot engine bay and round a hotter engine actually make things worse?

 

Edited November 19 by Escadrille Ecosse

[rogerguzzi]

6 minutes ago, Escadrille Ecosse said:

I will ultimately fit injection to the Spitfire and I have been wondering whether or not a fuel return is actually needed for this.

I realise most DIY EFI installations have a return from the fuel rail. But is this because it's been proven to be needed or because 'everybody else does it'?

Hello All

               My understanding is it runs at a constant pressure but adjusted by a pressure regulator that is vacuum controlled!

Then  the injectors just take what they want at any point depending on the programming!

Roger

[RedRooster]

26 minutes ago, Escadrille Ecosse said:

 

So does pumping a lot of excess fuel (itself adding energy) through a hot engine bay and round a hotter engine actually make things worse?

 

My concern as well, I did it to keep the fuel moving and not overheating, to be honest it seems to be OK, my main gripe is I can hear it going back to tank, I need to make the return line into the tank much lower. 

[rogerguzzi]

Hello All

              On Spitty I used the original feed pipe for the return and it points towards the fuel pump that I built into the tank (Range Rover type) plus it discharges under the fuel

I bought an original one not after market and touch wood it has been fine for about 40,000 miles maybe more?

RANGE ROVER CLASSIC V8 EFI IN-TANK FUEL PUMP UPTO 1990 - PRC8318 - NEW PUMP | eBay

Roger

 

 

[egret]

15 hours ago, rogerguzzi said:

Hello All

               My understanding is it runs at a constant pressure but adjusted by a pressure regulator that is vacuum controlled!

Then  the injectors just take what they want at any point depending on the programming!

Roger

Apologies for the essay and thread drift, maybe we split this into a separate topic?  I've been looking at fuel injection stuff recently and I believe you're correct, but happy to be corrected.

Because the injectors are essentially just clever valves (my understanding corrected following a conversation with John at Brands where I think I was confused with diesel injectors).  They need a constant pressure drop across them so when they open for a given amount of time they always deliver a set amount of fuel.  I assume there's complexities with certain pressures delivering optimum spray patterns too.  By doing this you're reducing a variable in the system making control simpler.  As the manifold pressure will vary, the fuel pressure needs to vary too, so the pump will need to provide the maximum required pressure for any eventuality at all times and then the pressure regulator will reduce this to the required pressure for the specific conditions.  The excess fuel pressure is bled off back in the return fuel line the swirl pot/fuel tank.  If you have lots of fuel flow around this high pressure loop then it's a fairly small volume of fuel having to deal with the heat gains of this system.

Understanding the manifold pressure varying and needing an excess of fuel pressure is probably easiest to understand for turbo/supercharged cars.  The manifold pressure increases with increasing boost levels and so the fuel pressure will need to increase in line with this to keep consistent fuel flow.  If it doesn't track the boost pressure then increasing boost will reduce the differential pressure across the injector and reduce fuel flow.

To throw some made up numbers at this, fuel pressure might be 50psi, and at idle with no boost the inlet will be an atmospheric 14.7psi meaning you have 35.3psi across the injector driving fuel flow.  If you add 7psi of boost then the inlet will be 21.7psi so you only have 28.3psi across the injector.  So a manifold pressure connection to the fuel pressure regulator allows the fuel pressure regulator to increase the 50psi of fuel pressure to 57psi and maintain 35.3psi of fuel pressure across the injector.

For cars without boost it should drop the fuel pressure in line with atmospheric conditions and help give consistent fuelling.  Not knowing how the lucas injection system works I'm only guessing here but if it doesn't reduce pressure under lower atmospheric pressures then it might be part part of the issue of over fuelling on the alpine passes.

I think modern cars have pumps which can vary the pressure they deliver fuel and so there is no need for a bypass or return pipe.

[rogerguzzi]

Hello Egret

                   That is about how understand it and you are probably right about modern cars and variable pump pressures I have no idea and not interested as to how they work

It takes me all my time to understand Spitty but at least I have chance!!

Roger

ps the hardest bit is to stop tinkering with the settings

[Escadrille Ecosse]

Apologies for thread drift mostly my fault.

Regards fuel rail/injector pressure, the Lucas system adjusts the injector rail pressure by having the pressure regulator operated by manifold pressure directly removing this variable from ECU control which otherwise relies only on engine speed and mass air flow.

https://www.gomog.com/allmorgan/flapperEFI.pdf

I don't know about variable pressure pumps on moderns, however on 'our' cars with aftermarket ECU fuel rail pressure is fixed and MAP compensation is managed by the ECU in the same way as all other load/speed variables.

In other words Megajolt/Megasquirt, Emerald, etc all use a fixed rail pressure controlled by a dumb regulator and so the question is where does that regulator go. At the engine, with a long return line to the tank or can it go at the tank with only a short loop?

Assuming the fuel pump is suitably sized so as not to give an excessive flow/pressure then heat buildup from pumping losses should be relatively low.

So would this work adequately?

[Nick Jones]

On 11/20/2024 at 12:53 PM, egret said:

following a conversation with John at Brands where I think I was confused with diesel injectors

If talking about PI injectors, then they really are pretty much the same as old school diesel injectors. Main difference is they “crack” at a mere 50psi rather 90 Bar upwards for diesels. They are opened by pressure generated by an upstream injection pump.  EFI injectors (and common rail diesel) are electrically triggered.

As regards fuel supply and cooling:

Cooling is probably only necessary if you can’t /won’t run the fuel lines away from significant heat sources as the main heat source is “soak” from adjacent hot objects.  
For example, because I’m lazy and cheap, I used the original carburettor fuel lines including the block-hugging copper former pump to carbs section for the return line. This really picks up heat and fuel tank temperatures can reach 45C+. I suspect I could fix most of this by changing that section of pipe. I guess if you can’t avoid picking up heat, you do still have to consider using a cooler.

In high pressure systems (common rail diesels @2,500 Bar and direct injection petrols, much higher than conventional port injections @45psi or so), wasted energy from pressurising a constant, large flow of fuel and forcing it through a relief valve can cause unwanted temperature rise. Generally these are vehicles that have the coolers used here.

These high pressure systems gain real advantage from pressure control by flow control, sidestepping the losses caused by full flow/relief valve systems. 

As regards pressure control for conventional port injection at nominal 3 Bar / 43psi, as already mentioned, the traditional method is to pump to the fuel rail, have regulator at the far end of the fuel rail, usually referenced to manifold pressure to maintain a near constant differential across the injectors. This reduces the variables needed to be accounted for by the ECU. Modern ECUs can manage without this help and many moderns have the relief valve back at the pump and do without a return line. Some monitor fuel rail pressure and control the pump speed to maintain it. Whether the low-end aftermarket stuff we’re typically using qualify as modern is debatable!

 

[mpbarrett]

As I understand it if you want to control the flow rate by timing the pulse width then you need a constant differential pressure across the injector. Then, assuming a  constant nozzle orifice  factor and an very fast rise and fall rate on the injector, controlling the pulse width will give you a known fuel flow rate into the engine. (Google "flow rate through an orifice formula" and it comes up with a quite good AI explanation of the formula  ).

On my setup I have the regulator near the injectors in the engine bay and then a return pipe to the tank (original fuel pipe). 
The regulator does not know about manifold pressure although I think it does have a extra port on it that could be connected to the manifold.
Although the Megasquirt  does know the set system pressure (but does not measure it) and the manifold pressure I dont think it does any corrections for changes in injector differential pressure but I might be wrong!
I think modern system do measure the system pressure.

mike

 

[egret]

I think my talk about manifold pressure compensation is really only relevant in forced induction applications. For most of what we do (speaking as a resident of the Fens) there's not really any point in altering fuel pressure once it's set for the injector preference, and it's just another mechanical item to go wrong.

At 2000m atmospheric pressure drops to 11.5, so the pressure drop across the injector increases by around 10% (50-14.7=35.3 and 50-11.5=38.5) which is probably within the ECU's control authority and I assume a wideband O2 would pick it up and compensate.

[Escadrille Ecosse]

For clarification, the Lucas sytem I was talking about was the Hotwire/GEMS and not Pi. Which is what the link relates to.

All my comments and queries here relate to EFI and system installation on the Spitfire using aftermarket ECU and throttle bodies as I am wanting to understand what I need to do when it comes to swapping the Webers for EFI down the line.

As noted in the posts above to provide a precise quantity of fuel from the injector (talking about EFI here) then for a given pulse time the actual quantity of fuel delivered is dependent on the ratio of rail pressure to manifold pressure. So higher manifold pressure as measured by the ECU via the MAP sensor (or the independent manifold regulator on the Lucas Hotwire/GEMS) should ideally require a higher rail pressure.

However this can only be achieved if the pressure regulator is capable of varying based on a signal from the ECU (or manifold regulater as GEMS).

This is true regardless of whether the regulator is sited at the rail, engine compartment or tank (I've seen all locations used). Assuming of course that the delivery pipe is adequately sized and the rail has sufficient volume to cope with instantaneous variation. Or an accumulator is located near the injectors.

However....

Unless I'm missing something, none of the EFI systems I have seen here, Nick, Roger, Mike, etc appear to have variable pressure regulation.

So we are back to my original question. Does it matter where the regulator goes?

My gut feel is that for a normally aspirated engine, particularly one of ours, then the actual variation in manifold pressure isn't actually significant enough to cause driveability issues. So variable rail pressure regulation is not that critical. You just want it fairly constant so help the poor old ECU.

Different for a boosted engine and perhaps if you are @rogerguzzi wanting to improve on your already galactic mileage .

Or am I missing something entirely??

This is all new to me hence the questions

Edited November 23 by Escadrille Ecosse

[RichardB]

My EFI installation has the fuel pressure regulator on the engine bay bulkhead, and a return line that passes back to the fuel tank return like Roger's. I don't have a problem with 'heat soak' per se, but even with the fixed regulator, when the engine, manifold etc all get nice and hot, it throws the tune off - despite the engine temperature being constant at 88 degrees centigrade. The difference between driving at full engine temp for after it has been warm for 5 minutes, versus driving at full engine temp after you've been doing 70mph for 20 minutes is really noticeable - AFR 14.7 at a tuned point in the table becomes 13.7 or so, it is that sort of magnitude. Moving the IAT sensor from the manifold to the front of the bonnet doesn't change this.

It's been off the road for over a decade now so my memory is a bit hazy, but I think fuel and oil temperature were next on the list of things to measure to understand why this happens. I did wonder whether a fuel cooler would keep things more stable. It probably doesn't help that the fuel rail is made from aluminium billet.

Edited November 23 by RichardB
added photo

[egret]

I'm in a similar position, wondering what I need for efi, how I can do it economically, ideally using OEM/pattern parts, and how I do it in stages so I can spread the cost and trouble shooting out!

I think a low pressure fuel pump and speeduino ECU are a good starting point. I can then add additional temp and O2 sensors in the future, before making the jump with manifold, ITBs, swirl pot, high pressure fuel pump and return system with pressure regulator. Also all the other things I've forgotten about!

[Nick Jones]

3 hours ago, Escadrille Ecosse said:

Unless I'm missing something, none of the EFI systems I have seen here, Nick, Roger, Mike, etc appear to have variable pressure regulation.

Mine is variable to the extent that the regulator diaphragm is referenced to manifold vacuum.  It’s mounted on the end of the fuel rail. 
 

As my injectors are batch-fired I’ve got them arranged in two banks (odds & evens) which fire alternately to even the load on the fuel rail.

[Escadrille Ecosse]

4 hours ago, Nick Jones said:

Mine is variable to the extent that the regulator diaphragm is referenced to manifold vacuum.  It’s mounted on the end of the fuel rail.

Ahh. Didn't realise that the regulator on the fuel rail was variable.

Batch firing vs sequential is just something else to look into come the time

[rogerguzzi]

8 hours ago, Nick Jones said:

Mine is variable to the extent that the regulator diaphragm is referenced to manifold vacuum.  It’s mounted on the end of the fuel rail. 
 

As my injectors are batch-fired I’ve got them arranged in two banks (odds & evens) which fire alternately to even the load on the fuel rail.

Hello All 

                This how Spitty is set up and seems to work well enough for me?(I do need to ask Nick and forgot his name((Drink)) as to the high up bits they used )

Because I have not got it right yet but other than that it it works well(The AFR gauge worry meter is a problem and leads to Tinkering!!)

But that's what we do and it's looked very lean at times but is still running with no engine problems after about 35,000 miles at about 45MPG 

I admit not driven hard that often if the Memsahib is on board but has given the TR6,s a surprise now and then!! 

Now the driving season is over for Spitty the bloody gearbox has to come out again to replace the CRAP after market clutch cover(Was ok for 3000 miles that's years and years for posers??((7350miles this years so She is used not polished!!)

But I have bought a NOS Borg and Beck NOS from Fleabay for less than the Crap repro one!!!!!!!!!!!! and even a NOS thrust bearings RHP made for about the same money as the Cr**P cover !!

I would just like to fined a NOS friction plate as I was offered a new modern one for FREE and the centre boss is shorter and thinner(looked Crap to me!!) so the old plate went back in as I Still think it was better after about  50000+ miles????

Roger

ps Spain Ferry Booked for next June again?(Daft old buggers but as my old mother used to sat once you give something  up you will Never Start it again!) 

And she was right I have 3 motorcycles I have not ridden for about 5 years now?(Must sort them out and sell them to buy more wine and Adventures in Spitty before the Grim reaper comes knocking!!)

[mpbarrett]

Thats bad luck with the clutch plate. Just helped a friend with his gearbox in a 1500 spit, what a pain refitting it, sure it was much easier 10 years ago when I was more flexible!. I need to take the gearbox out of the herald, again, next year and am tempted to take the bonnet off and take the whole lot out and change it in the garage..... 

Back to injection.....
I think the advantage of having the regulator near the injector is that they is no chance of pressure drop along the pipe to the injectors if the pressure is regulated near the tank. Also a lot of modern (1980's) had the regulator on the fuel rail to the injectors, I assume for the same reason.
No fuel pressure control on my system, just fixed pressure regulator and no reference to manifold pressure (but I might change that).

mike

[Nick Jones]

6 hours ago, mpbarrett said:

No fuel pressure control on my system, just fixed pressure regulator and no reference to manifold pressure (but I might change that).

If it’s working fine as it is I doubt you’ll see any real gains. The vacuum signal from ITBs is pretty weak compared to plenum and single TB.

[Escadrille Ecosse]

6 hours ago, mpbarrett said:

No fuel pressure control on my system, just fixed pressure regulator and no reference to manifold pressure (but I might change that).

Presumably if using an aftermarket ECU the tuning process to generate the fuel map will by default include the small rail pressure variation due to load. Simply because it happens as you vary the load generating the map. So while nice it's probably not absolutely vital and why just having a fixed regulator works OK.

6 hours ago, mpbarrett said:

Back to injection.....
I think the advantage of having the regulator near the injector is that they is no chance of pressure drop along the pipe to the injectors if the pressure is regulated near the tank. Also a lot of modern (1980's) had the regulator on the fuel rail to the injectors, I assume for the same reason.

Primarily though I think that if you are going to have a physical link for the vacuum signal between the regulator and manifold then you must have the regulator at or very close to the manifold otherwise the response will be inadequate, too delayed and 'flatened'.

Doing this electronically via a MAP sensor and the ECU to either adjust the regulator or simply trim the injector opening time eliminates that particular constraint.

I accept that regulation at the tank might have an effect on actual pressure at the rail depending on load. I suspect though that this will be fairly predictable, greater load, more demand lower pressure but I imagine that can be dialed out by a combination of adequately sizing the delivery pipework/rail and through injector open time trimming via a table in the ECU. Probably easier to do with smaller capacity engines than monsters.

I don't know if aftermarket ECUs have the ability to do this.

I suspect this is also used for controlling the variable speed pump on moderns, although pretty sure that will be more to do with minimising pump/bypass valve losses and fuel heating rather than rail pressure, because it's a bit too coarse as control for that.

 

Edited November 24 by Escadrille Ecosse

[mpbarrett]

5 hours ago, Nick Jones said:

If it’s working fine as it is I doubt you’ll see any real gains. The vacuum signal from ITBs is pretty weak compared to plenum and single TB.

but there is always the temptation to fiddle even if its working well....  

[mpbarrett]

5 hours ago, Escadrille Ecosse said:

Presumably if using an aftermarket ECU the tuning process to generate the fuel map will by default include the small rail pressure variation due to load. Simply because it happens as you vary the load generating the map. So while nice it's probably not absolutely vital and why just having a fixed regulator works OK.

 

 

unless it has a model to determine what happening, and to do that it need pump details, pipes sizes, injector characteristics etc, it has no measurement to make any correction.
So I don't believe the Megasquirt does this....

Mike 

[mpbarrett]

5 hours ago, Escadrille Ecosse said:

......

Primarily though I think that if you are going to have a physical link for the vacuum signal between the regulator and manifold then you must have the regulator at or very close to the manifold otherwise the response will be inadequate, too delayed and 'flatened'.

Doing this electronically via a MAP sensor and the ECU to either adjust the regulator or simply trim the injector opening time eliminates that particular constraint.

I accept that regulation at the tank might have an effect on actual pressure at the rail depending on load. I suspect though that this will be fairly predictable, greater load, more demand lower pressure but I imagine that can be dialed out by a combination of adequately sizing the delivery pipework/rail and through injector open time trimming via a table in the ECU. Probably easier to do with smaller capacity engines than monsters.

I don't know if aftermarket ECUs have the ability to do this.

I suspect this is also used for controlling the variable speed pump on moderns, although pretty sure that will be more to do with minimising pump/bypass valve losses and fuel heating rather than rail pressure, because it's a bit too coarse as control for that.

 

pipe directly from the inlet manifold to the regulator (in my case about 100mm away) simple mechanical/analogue control.
I think we can overthink this problem and MOST of the time it works well enough. I agree on large engines with very high injector flow rate this might need to considered. I am sure modern systems will consider this issue and with more sensors can be corrected for change in injector differential pressure in the software.
If I was developing it I would model the whole flow system, collect data and then develop a model and from that create an algorithm to correct for variations in the whole system. Easy to  do in a modern ECU and enough time to develop and test the code!
Also modern system are concerned with cost and emissions which raise all sort of other problems we dont have  an old Triumph engine designed in mid 1950s that we are playing with!
but it is all good fun!

Mike