Why did Ford go returnless?

YtnGT - your description of the system is pretty darn accurate. It's not as precise as you might imagine, because the PCM can compensate for variations in differential pressure "on the fly", and recalculate every individual injector pulse.

Regarding the black relay in trinityGT's post, it's a standard Bosch automotive relay capable of switching up to a 30A lamp load. These cost Ford less than a buck in quantity. A relay that can reliably switch an inductive (fuel pump) load and operate during cold cranking is a little more expensive - like $1.20. The diagnostics are EPA mandated and must be OBD2 compliant; that adds a 50 cent serial transceiver IC, plus an 18 cent power conditioning IC. Add to that a housing, bracket, and connector, and the whole deal represents less than $2.50 in Ford (actually Visteon, who builds the Ford modules) cost.
In comparison, the PWM fuel pump driver IC with integral OBD2 port costs $1.38 and the driver MOSFET 17 cents (my company makes both). Add in the PC board, housing, bracket, and connector, and the total easily meets or beats the relay solution. What finally tips the scale is warranty labor costs . . . the mechanical relay has 2X the failure rate of the solid state module over the emissions warranty period.
 
  • Sponsors (?)


i see some of the engineers chirped in on this one... one of the reasons ford went too the returnless system is it provided a more controlable fuel system for emmisons purpose.. also with the ability too change the fuel presure they can increase the presure at high load conditions... this increases the dynamic range of the injectors by alowing for a lower injector pulse width..

i do know this... ford dont care how easy any off there engine controls are going to be for modifications.. in fact when i went too a svt class the instructor spent most of his time telling us how too spot different mods and what part of the warrenty it will void...
 
n0v8or said:
YtnGT - your description of the system is pretty darn accurate. It's not as precise as you might imagine, because the PCM can compensate for variations in differential pressure "on the fly", and recalculate every individual injector pulse.

Yes I was aware of that strategy (was trying not to bore anyone). It is a reall nifty thing that ford thought of. It really damps the effect of FP variations on individual cylinder fuel delivery. BUT in my case with wet nitrous, the fuel pressure control loop variations causes more or less fuel to be delivered with the N20. FYI returnless control systems really dont like the step change in fuel demand when the fuel solenoid opens. So my N2O a/f moves around with fuel pressure.
 
n0v8or said:
Regarding the black relay in trinityGT's post, it's a standard Bosch automotive relay capable of switching up to a 30A lamp load. These cost Ford less than a buck in quantity. A relay that can reliably switch an inductive (fuel pump) load and operate during cold cranking is a little more expensive - like $1.20. The diagnostics are EPA mandated and must be OBD2 compliant; that adds a 50 cent serial transceiver IC, plus an 18 cent power conditioning IC. Add to that a housing, bracket, and connector, and the whole deal represents less than $2.50 in Ford (actually Visteon, who builds the Ford modules) cost.
In comparison, the PWM fuel pump driver IC with integral OBD2 port costs $1.38 and the driver MOSFET 17 cents (my company makes both). Add in the PC board, housing, bracket, and connector, and the total easily meets or beats the relay solution.

So $1.20 for an inductive-load capable, completed relay in quantity versus $1.38 for just the ASIC, let alone the interconnects, other semis, the PWB, assembly and test labour and capital, and of course the NRE costs up front. Sorry...I don't buy that the FPDM comes out to cost substanitally the same as a simple Bosch relay.

If I was to walk into a Ford dealer and ask for a Bosch relay and a FPDM for a 2002 Mustang GT, can you tell me, roughly, what they'd charge me for each? I don't really feel like doing so but a call to my local dealer will prove which one is the more expensive part. I'd expect Ford to charge $500 or more for an FPDM and less than $15 for a relay. Assuming this is correct, why the vast disparity for two parts that you say are substantially equivalent in terms of cost?

The issue was one of going from a return-system that utilizes a fixed-speed pump, relay and FPR to a returnless that trades the FPR for a pressure sensor and the relay with a FPDM. I'm going to assume that a fuel pump happy with being PWMd at 9500Hz costs about the same as a fuel pump motor designed solely for a DC voltage in the 5-24 volt range. I think that the main driving force at work here are federally mandated emissions requirements under OBD-II but I don't buy that the move to a returnless system saved Ford a bunch of money. Indeed, I rather think the move to a returnless system in fact costs Ford more per vehicle on the assembly line and that we, the customer are eating that cost.

What finally tips the scale is warranty labor costs . . . the mechanical relay has 2X the failure rate of the solid state module over the emissions warranty period.

:shrug: There are thousands (millions?) of cars on the roads today utilizing the KISS-philosophy relay and have zero reliability issues with them.

Anyway, very cool discussion n0v8or. Nice to have gearheads and geeks here involved in the design of components on our cars. :hail2:

Now if only someone from Tremec would show up so we could pummel him over the 3650 synchros :banana:
 
YtnGT said:
http://www.kennebell.net/techinfo/general-info/jimbells-supercharged-perf-guide.pdf

Read the article on the right hand side, that should explain the decision to go returnless.

That's a very interesting article.

However your car should not ping when you open the throttle up fully on any car. If it does and your running the correct grade of fuel (i.e. manufacture stated) then it's a poorly utilized application for that vehicle. What ever part of the fuel system is not up to the job.

Don't get me wrong, I am not saying it a bad system just not very good for this application, as it is standard.