jza80

iTrader: (32)

Awhile back I bought a new Supra fuel pump kit for my FD from one of the trusted Forum sellers. Later, I read up on how the fuel pump wiring on our cars is marginal and that the Supra pump while being quiet and decent flowing also pulled higher current than the OEM pump or some of the alternative replacement “upgrade” pumps and as a result some users rewired their fuel pump circuit to use a new relay triggered off the fuel pump power wire using a more direct feed from the battery and bypassing the low speed fuel pump voltage circuit. Other users have bypassed the ignition switch feed to the high speed circuit relay in favor supplying the circuit relay with a direct to battery feed to improve voltage, but then with the resulting loss of a potentially critical safety disconnect capability in case of a stuck closed relay. And then a few other users with the rewired pump have reported problems with the internal fuel tank bulkhead fitting burning out and having to modify that to allow larger in-tank wiring – with mixed results such as fuel wicking past the new connections.


All this has made me question my selection of pumps, and even if the Supra pump can really be considered an “upgrade” for a basically stock car that will never see more than 300WHP and maybe the original Mazda pump is a better choice, or perhaps a pump such as the Walbro that seems to draw less current.


For my case, wiring a constant battery feed via a relay energized by the fuel pump power is not really an option as I must maintain the ability to restore the OEM ECU every two years for California smog check. Although I could tune the PFC I use otherwise to account for the extra fuel at low speed, there is nothing I can do to the OEM ECU. And then there is the possibility of burning out the tank bulkhead fitting and having to cob up something that may or may not be safe and 100% fuel tight. What a mess, just trying to improve reliability by replacing a 22 year old fuel pump could wind up causing more damage.


So I thought, since there are a few known points where resistance could build up over time, and since surely the fuel pump circuit must have worked okay when the car was new why not try to attack those points and improve the system while maintaining the OEM design? So that’s what I did however the results were not as good as I hoped. Here’s what I tried:


First, I took a baseline of the battery voltage and fuel pump supply voltage (taken by back pinning the rectangular harness plug that resides behind the driver side trunk panel using a Fluke DVM).


Battery voltage at idle and at 3,000RPM was about 14.3v +/- 0.1v or so. With the AC on full blast, headlights on and rear window defroster on the voltage dropped to about 14.0 at idle and about 14.1-14.2 at 3,000 RPM. So supply voltage does not seem to be a problem.


Baseline voltage at the OEM fuel pump was 7.5-7.6v on the low speed circuit (idle, up to about 5,000RPM free rev), and about 10.8v at 7,000RPM wide open throttle on the road in second gear (I assumed this would be the worst case condition, max fuel flow – although I found that for all the tests, there did not seem to be a measurable difference any time the high speed circuit was active beyond maybe 0.1v be it low boost or max boost at redline).


For the next step, I installed new OEM fuel pump and fuel pump circuit relays (I was going to do this regardless, along with the main relay as preventative maintenance). This did not change the fuel pump voltage much, however it did improve about 0.1v to a more steady 7.6v at idle and high speed circuit seemed to shift to 10.8-10.9v.


Next, I tried a new ignition switch. I thought this might offer a bigger improvement since the fuel pump circuit voltage flows through this switch, and this assumption was correct but not by as much as I had hoped. With the new switch, the idle voltage moved up to 7.7-7.8v, and the high speed circuit increased to 11.1-11.2v. So some improvement but not that big a jump. What was interesting about this was that the original switch body was discolored brown around the contact area, clearly this switch does get warm as you might expect but it was still working generally pretty well albeit it was drawing the circuit voltage down about 0.3v.


Next, I installed the Supra pump. This caused a big change - in the wrong direction. With all the improvements still in place, idle circuit voltage dropped a full volt to 6.5-6.6v, and the high speed circuit dropped back to about 10.8v. This change seems to validate that the Supra pump draws a larger current than the stock pump. After the Supra pump install, I started tracing the wire harness that the fuel pump plugs into, and found the ground block at the rear of the car that supports this circuit and several others. I pulled this apart and although it looked perfectly good I re-cleaned the connection points, tightened the spade connectors and applied some dielectric grease to all the contacts (this was also done to all the connections for the relays and ignition switch, too). Doing this helped a lot, the idle circuit voltage increased to 7.0-7.1v, and the high speed circuit jumped up to 11.0-11.1v. Also, interestingly I had noticed that the fuel pump circuit voltage had consistently dropped almost 0.5v every time I engaged reverse to back out of my driveway for the WOT runs, and after the ground block cleaning this improved to only about 0.1v drop. The reverse lamps share this ground block, and operating them must have pulled up resistance some for the fuel pump. So cleaning the ground block really made an improvement overall. And I wonder what the circuit voltage might now be if I reinstalled my original pump back in...


So now, I am considering reinstalling the OEM pump back while I figure out what to do long term. Clearly the Supra pump is putting additional strain on a marginal wiring design, and my build will not require more than what the OEM pump supplies, as long as the pump is working correctly. The reduced low speed circuit voltage is particularly troubling, and my fear is that the circuit resistor could suffer an early failure from the clearly higher current draw. There was really nothing wrong with my OEM pump, other than that it was 22+ years old, so I think the original fuel pump may actually be a reliability upgrade over the brand new Supra pump. Or maybe a new Walbro might be a better choice. Or something else? Hard to decide what to do.

DaleClark

iTrader: (55)

The original writeup I did on ClubRX is gone, but this thread has most of the information you need -

https://www.rx7club.com/3rd-generati...615964/page10/

Basically you run fused 12v from the battery to the blue wire feeding the fuel pump relay under the hood. Doing that will keep the 2 speed operation, won't be a problem with emissions or codes, is minimal extra wiring, and will greatly bump the voltage to the pump up.

Dale

arghx

iTrader: (3)

uploading the pic from the thread for posterity's sake since it's linked from an outside website (don't need a photobucket disaster)...

jza80

iTrader: (32)

Dale and arghx,

Thanks for the feedback. Probably I will do some variation on the direct battery to circuit control relay mod, but I am really reluctant to eliminate the failsafe that disconnects power to the circuit from the ignition switch. With just the bypass, a stuck closed circuit relay means the fuel pump would never shut off. So I'm thinking maybe feed the battery power to a new relay that is controlled by the ignition switch circuit somehow. That would introduce a new failure path though. Maybe I'm overthinking it...

The direct battery to fuel pump via a new relay approach is definitely the best way from a voltage perspective, but then I would lose the low speed circuit which would seem to have a big impact on idle and low speed A/F. That would kill me when I needed to do a smog check with the OEM ECM. And I'm still wondering of the impact on the high current draw Supra pump on the wimpy bulkhead fitting and in-tank wiring (looks like 16GA at best if not smaller). Although higher voltage would reduce current draw a bit I think.

arghx

iTrader: (3)

Honestly I don't see how any car at stock ECU level of mods needs any fuel pump changes at all. The wiring is marginal if you are trying to run more boost, it's not marginal if you operate the engine within the limits Mazda intended. You can only do a few mild exhaust mods safely on stock ecu. You can't really turn up the boost much. If you have near stock boost levels on stock turbos you should be fine. That's why Mazda put a boost based fuel cut in the stock ECU. If you need a new fuel pump, get a new OEM FD fuel pump. Problem solved.

for what it's worth, the 3 rotor Cosmo had a 3 step fuel pump speed system and a higher flowing pump running at more amperage.

and the normal way to do the pump rewire is to use the factory wiring to trigger the relay. The factory pump wiring has safety control built into it because it is controlled by the ECU, not the ignition power. Anyone who is controlling their fuel pump by constant or ignition circuit is needlessly bypassing safety features. Look at the ECU pinout. One wire controls the fuel pump speed relay and one controls the pump itself.

if you're in California put a racing beat catback in there and keep the stock cats, or get a JDM catless OEM downpipe and only run the main cat until inspection time. Keep stock intakes and other stuff under the hood and you will pass emissions. You're not going to make 300whp. Welcome to emissions regulations. Keep it near stock.

jza80

iTrader: (32)

I agree 100%. My car is only basic BPU with JDM DP and stock main cat, PFC, Blitz SMIC, Apexi Noir catback, basically stock boost, stock airbox. I really don't need a bigger fuel pump with all the complexities that seem to go with it. Just want solid reliability.

Looks like new fuel pump assemblies complete with hanger are still available from Mazda, and and cost barely more money than a Supra pump. I put one on order this morning. I should have researched more before I bought the Supra pump upgrade kit, but its only money and not the first time I have made that mistake :-/

An almost new Supra pump will be coming to the classifieds soon, someone with more fuel needs than I will score a good deal...

adam c

iTrader: (2)

While you probably dont need a Supra pump, it isnt hurting anything. I have been running a Supra pump for almost 14 years. This is with no wiring upgrades. I have similar mods to yours with higher boost, and around 320 RWHP.

Many of us overthink things to the point of being ridiculous with these cars.

arghx

iTrader: (3)

The concern is that the higher flow rate will overwhelm the fuel pressure regulator at low load and raise fuel pressure, resulting in a richer mixture and worse emissions. That certainly applies if he raises the fuel pump voltage. I'm not sure if it applies if you keep the resister.

you can tune around that, but not on stock ECU.

jza80

iTrader: (32)

Adam,

That is a good data point, thank you for sharing. There is limited information on what people are running, what they have done as supporting mods, and what happened as a result. It's helpful to hear your results over a long period of time.

I am still struggling a bit about the voltage drop, the wire gauge used by Mazda is pretty large, at least 14GA if not 12GA and there should not be such a pull down unless it's just a combination of slight resistance at all the connection points. I wonder if the voltage at the pump was in the 11-12v range even when the cars were new and what users are seeing now is basically normal.

existanzrx7

iTrader: (8)

Thats a good point might be something worth looking at maybe we should all check that ground block. Where is that ground block actually located? Thanks

jza80

iTrader: (32)

looking forward from the rear of the car, it is just to the left of the hatch release mechanism, lower down. If you remove the rear plastic trim you can see it, just follow the wire harness that routes in from the drivers side.

Speed of light

iTrader: (8)

Here's another data point for you. I put a Supra pump in quite a few years ago and ran a regime of voltage drop and fuel pressure tests. I concluded that all that I really needed to do [to the wiring] to support 400-ish whp FD was to implement the mod that Dale and others have mentioned above. The majority of voltage drop was occurring in the ignition switch circuits. Rewiring the relay to take power from the main buss (through a fuse) worked beautifully and removing the fuel pump from the ignition circuit also improved the voltage available to the coils and other accessories that share that circuit.

My 93 FD has factory #10 gauge wiring extending from the firewall mounted electrical distribution block, through the body and to the pump connector. No evidence of any issues with the wiring or connectors. Measured resistance from the firewall at the block (just to the left of the steering column, iirc) to the pump assembly, including all connectors was less than 1/10 of an Ohm. Suffice to say, I deemed the wiring to be adequate.

Different mods suit different applications and usage, but for a general purpose street car that sees considerable use, I would keep the multi-speed pump arrangement in place. Reducing the voltage to the pump when not in boost substantially reduces the load on the pump--with corresponding benefits--when full output is not required. Power is a square law function, and by reducing the pump voltage by about 5 volts (as the factory does), one reduces the pump power input by about a factor of 3 (~50 watts vs. >150 watts full load), based on my measurements.

Hope this helps.

jza80

iTrader: (32)

To finish up this project I installed the new Mazda OEM fuel pump that I spoke about earlier in the thread and ran the voltage tests using all the other improvements (new switch, relays etc). The results were better than my expectation.

Idle voltage was 8.2v, with the AC on it dropped to 8.0v

On-boost voltage was now 12.2v, with the AC on it dropped to approximately 12v. This is about a 1.2 volt improvement compared to a new Supra pump! I also ran a test with measuring voltage with the pump plug disconnected so no load. With the key on, the pump is powered for a few seconds by the ECM, in this mode I saw 12.4v (this was with a battery tender attached, battery voltage is about 13.5). So the difference between free voltage and load voltage is only about 0.2v, making me suspect that the basic design was for around 12v running voltage from new.

At this point, I think the system is working exactly as designed. Almost all the points of resistance have been optimized (new switch, relays, ground point), all that is left is the bulkhead connector but the negligible change from the relay replacements leads me to think that not much improvement will be found there.

It is interesting that the new Mazda OEM pump, visually, was a near twin to the Supra pump. The housing color is slightly different (goldish color for the Mazda pump versus silver for the Supra pump), there is a plastic cover plate around the electrical posts that is not present and there is a plastic shield that wraps over the pump outlet pipe on the top that is not present on the Supra pump, also the strainer sock is about twice the size. But the pump housing physical size itself is exactly the same as far as I can tell. The part number prefix (195130-) is the same, but the last digits are different (Mazda OEM is 195130-0782 5U17, the Supra replacement is either 195130-1270, -1276 or -1279, I could not tell).

All that said they are clearly not the same internally. The Mazda OEM Denso pump creates considerably less current drop and allows a significantly higher operating voltage. I don’t know if this also means that the pump flow output is commensurately lower as well. But for my build, a properly functioning OEM pump is all I need.

Anyway, this wraps up this project for me. Hopefully this information helps someone planning a fuel pump change.