Fuel Pump Upgrade

Last updated: January 30, 2003

Date: Sun, 09 May 1999 18:24:04 -0500
From: "Kevin T. Wyum" (aspi@winternet.com)

Default (fuel pressure) is 35 or so at idle.

Two speed pump, fuel pump resistor does it, but once the throttle is hit it bumps to a base of about 40. Remember the fuel pressure regulator will reduce pressure with more vaccum as well.

So base at 0 pressue/vac is 40 with throttle. +1 per psi of boost.


Date: Wed, 25 Mar 1998 09:51:09 -0500
From: rotary@mediaone.net (Carlos Iglesias)

Don't need to upgrade the pump until fuel pressure drops as a result of exceeding the stock pumps capabilities. In-turn, ways of exceeding the stock pump require either:

1) increasing injector sizes
2) maxing out fuel pressure at around 50-55 psi (base pressure) + 15-18 psi of manifold referenced boost

As for testing the fuel under max load, apply the Cameron Worth/Carlos Iglesias fuel pressure gauge mod. Buy a fuel line adapter with an 1/8 NPT gauge outlet, plump in a -04 braided line, and at the other end of the line, attach a 100 psi fuel pressure gauge. Disconnect the wiper CB, and attach the gauge to the wiper arm on the driver side. Now take fuel reading at MAX RPM/boost. Pressure should rise approximately linearly with boost.

Uncharacteristically, I disavow all responsibility for the stupidity of the mechanically inept in performing this mod/test. At a min, carry a fire extinguisher for the first few days.


Date: Wed, 22 Oct 1997 23:43:03 -0400
From: "Carlos A. Iglesias" (rotary@mediaone.net)

> MSD offers a fuel pump that has a capacity of 43 GPH.

GPH -> Gallons per Hour.

I've been living and breathing fuel system components for the last week trying to find out what really is the best. As it turns out, all of Porsche and Mazda race teams, most of the fuel injections specialist, and the majority of WSC cars run Bosch pumps and regulators. A little pricey, but the pump that I'll probably buy (from a Porsche 962) flows almost 60 GPH conservatively (at 40 psi). And unlike the high flowing Paxton or SX pumps, it maintains its flow rate up until about 110 psi. And like the Bosch products tend to be, they retain an acclaimed degree of reliability.

For comparison sake, the stock pump flows about 36 GPH and the Cosmo pump flows about 42 GPH (both according to Mazda Comp).

Now I've just got to take the fuel rails off, and modify them so that I can use use an adjustable rising rate pressure regulator... and that of course means replumbing the entire fuel system from the pump, to the injectors, and back to the pump in -8 and -6 braided line. Isn't it funny how simple ideas always turn more complicated and complex before final implementation? ;-)


Date: Thu, 23 Oct 1997 16:39:08 -0400
From: "Carlos A. Iglesias" (rotary@mediaone.net)

>Do you mind sharing with me what you have done with your fuel
>system and what your objectives are. What type of rotary are you
>doing this to? I am interested in modifying my fuel system and
>would like to know what you know.

I've got a '93 that I've decided to upgrade the fuel system, in preparation for larger injectors/more boost. The stock injectors/fuel system starts to max out around 15-16 lbs.. as evident from the injectors (aggregate) duty cycle above 85 percent at WOT and O2 sliding towards less than .80 volts.

So this last week, I set out with my usual badger-like tenacity to find the best way of upgrading the stock fuel delivery system. Several benchmarks came to mind immediately for me:

  1. I want to be able to decrease as well as increase fuel pressure, in order to be able to runner leaner for emissions, and richer for racing. This meant that I'd have to remove the stock fuel press reg, since it is non-adjustable, and placing an adjustable regulator downstream of the stock one would only allow raising the pressure, but not lower it to lean out the mixture.
  2. For ease of installation, and to keep the noise down, the pump would have to fit in-tank (with some modification to the retaining bracket), pump significantly more than a Cosmo pump to justify the addition installation requirement, and be "bullet proof", since rotary have a "Zero Tolerance" of lean conditions.
  3. The pressure reg has to be a rising rate regulator, with two inlet to provide for future plumbing upgrades.
  4. I didn't want to mess with the stock evaporative system.
  5. At least the pressure side of the fuel system, and preferably the return side would have to be replumb with larger lines to maximize the high output.
  6. That sorry excuse for a fuel filter definitely would be replace with a more efficient (which isn't saying much), easier to access filter that has a replaceable element.
  7. Most importantly, the components would have to be race proven, and reliable.

With all of this on the plate, I started to call around. All three fuel injection specialist that I know of (Kinsler, RC, and Marren), 8 different Porsche specialist, 3 race component and 2 general car catalogs (Pegasus, Racer Wholesale, True Choice, Jegs, Summit), so here goes...

I started out thinking that I'd go with either the Paxton or SX product. The prices were reasonable for their very high numbers. However, I was dissuaded by several of the sources that I contacted (including two that sell Paxton) because they were of the opinion that the products were less than reliable, and especially when they are used towards the edges of their operating envelope. So I started to get opinions.

The first thing to settle on was that it would be a Bosch pump. If it's good enough for Porsche's and Mazda's race teams, it's good enough for me. I call five different Porsche tuners to ask what was the largest stock pump. I got four different answers. But three of them told me that the 962 pump were the way to go if money wasn't a big issue. It flows 65 gal/hr (about 400 lbs../hr), at 40 psi, but unlike the Paxton/SX pumps which flow over a third less at 70 psi, the Bosch still flows 325 at 70 psi. So that settled the pump.

As for the pressure regulator, I stuck with Bosch for the same reasons. The same press reg that's on the WSC cars: dual 14 mm (converted to AN -6) inlet, an 14 mm outlet, with a boost input for rising rate regulation.

Haven't decided on a filter yet, but will probably go with a Canton Mecca Racing inline fuel filter, since I've been so impressed with my Canton remote oil filter. Excellent construction, and it filters down to 8 microns, as opposed to the 10 micron filtration that is the norm. For the hard core, they also have 4 or 2 micron (can't remember) filters available.

The only tricky part is going to be converting the stock fuel rail ends so that I can put AN fittings on them. I think that since the stock press. reg and pulsation dampers are integral to the stock rails, I may be able to remove them, and tap the end so that so that the fitting can be screwed on. Worst case, I'll have the fittings TIG welded on.

Add about 20 feet of Aeroquip -6 braided line, and about $120 in fitting.

Finish it off with a 100 psi VDO fuel gauge, have the injectors cleaned and balanced, and forget about giving Christmas presents this year.

All said and told, I expect about $900-$1000 buck for the project. But it's relatively cheap insurance against fuel system and therefore engine degradation. And I'll be all set to spend the big bucks for INDY injectors and an engine management computer.

Hope this answers both of your questions. Nuff said.


Date: Wed, 4 Feb 1998 21:04:41 -0500
From: rotary@mediaone.net

It'll be a piece of cake. Just don't forget:

1) Depressurize the system by running the car with the fuel pump CB pulled.

2) Disconnect battery

3) Lots for rags/towels

4) Preferably a low tank of gag (less messy)

5) Prime and leak check the system when you're done

Date: Wed, 25 Mar 98 09:28:40 -0500
From: "Linthicum, Sandy" (linthias@sandy-ntws.usps.gov)

I am going to but am going to use the Bosch pump. The Cosmo is a drop in but costs more. The Bosch is overkill (you will never push its limits) but requires a little work by your shop (ie. AN fittings) to install. You can get the Boshe from Mostly Mazda for about $290 or from Pettit (do not know Pettits price)


Date: Tue, 30 Jan 2001 11:40:30 -0500
From: "Lanham, Wade A." (wlanham@alleghenypower.com)

The M2 or Kinsler big Bosch pump and Walbro GSS-341 flow practically the same amount of fuel. The Cosmo/nippondenso pump flows much less. The Bosch appears to be higher quality than the Walbro, which seems flimsy in comparison.

I have information on my web page about fuel pumps.


Date: Thu, 30 May 2002 22:42:01 -0700
From: "Jim LaBreck" (jimlab@earthlink.net)

Was your Bosch pump ordered from M2 Performance? Was it the model with fittings at both ends intended for in-line installation? I've seen instructions for making a "castle" nut out of the bottom nut with a grinder to improve fuel flow. I think Carlos Iglesias may have been the one who came up with that mod. :)

There are three versions of that Bosch pump. The correct model has a screened inlet which covers nearly the entire bottom of the pump. Here's the correct Bosch pump compared to the Walbro.

I got the Bosch pump from Doug Rippie Motorsports (www.dougrippied.com in the fuel section) for $180. The price has since been raised to $200.

Date: Thu, 14 May 1998 18:20:21 -1000 (HST)
From: F8LDZZ (f8ldzz@lava.net)

> I asked about this before and got no response...has anyone checked the
> voltage at the pump in their FD? In my Talon I had a problem with a lean
> condition at the top end. (fuel pressure drop). So, I checked the voltage
> RIGHT AT THE PUMP. It was like 11.5 to 12V. So, I took a 10 gauge wire
> DIRECTLY from the battery to the pump and voila! 14-14.5 volts which causes
> the pump to run faster and *SIGNIFICANTLY* increases flow....problem
> solved. I used a relay and fuse also in the project...to turn the pump on
> and protect the whole system respectively. I assume this is what the Kenne
> Belle deal does.

Nope, not exactly. The Kenne Bell Boost-A-pump is a fancy voltage regulator. The BAP can be set to be a strict regulator from ~12.5VDC to 16VDC by inserting a jumper to one of it's leads. The BAP has a dedicated pressure sensor that can also raise voltage levels to the fuel pump by a fixed ratio...so higher the boost, the higher the voltage goes to the fuel pump. An adjustable knob sets base voltage levels remotely.

Date: Fri, 15 May 1998 11:21:22 -0700
From: Spencer Hutchings (hutch@spnetworking.com)

Did anybody see the latest Jegs catalog? It has a high volume, 255ltr, in tank fuel pump upgrade for the RX7! It's $199.00 but does not list the brand name.

Date: September 24, 1998
From: "Carlos A. Iglesias" (rotary@mediaone.net)

Let me begin by saying that I think that the Walbro pumpl is a very good pump, at an excellent deal. However, I do question some of the flow rates for stated in the VFAQ website. You may want to call around to Kinsle, TWM or RC Engineering to confirm the Walbro numbers. The numbers I provide below are from Kinsler Fuel Injection, and confirmed by Marren Motorsports.

My pump setup is not exactly drop-in, however, it should only take about 2-2.5 hour to complete. I do have a picture of the setup on my web page (http://julio.simplenet.com/rx-7_37.htm), if you go to the bottom and click on "Fuel System Upgrade," and then click on the fuel pump diagram. Subsequent to that picture, I have revised the design so that the pump is connected not by two straight fitting, but by a single double swivel fitting. This allows the pump to move up along the stock bracket about 2", which in-turn allows drawing fuel from the rear most part of the tank via a hose connected to the pump inlet. This resolved the fuel sloshing starvation problem that I originally experience under heavy acceleration.

As for the Bosch pumps performance, here are the numbers (at 13.2 v.):

PSI        LPH        Amps
0             400        5.5
20           370        6.9
50           325        8.7
70           300        10.0
100         260        11.7
120         200        13.0

I ran a fuel pressure gauge for about the first two months after puttin in the new fuel system, and once I was satisfied, I removed it and plugged the port. There are fuel pressure adapters for both AN and standard size fuel hoses. All of the performance catalogs (ie Jegs, Summit, et.) carry them.

From: Carlos Iglesias
Date: Thursday, December 03, 1998 8:00 PM

>What is the equation to calculate the change in fuel flow in
>relation to a change in fuel pressure?

And the answer is:

SQR [SysP1/SysP2] * StaticFlow1 = StaticFlow2

    Where: SQR         is Square Root
           SysP1       is original specific fuel pressure (psi)
           SysP2       is proposed new fuel pressure (psi)
           StaticFlow1 is the original static (non-pulsed) flow rate (cc/min)
           StaticFlow2 is the proposed new static (non-pulsed) flow rate (cc/min)

Sourced from: Fuel Injection: Installation, Performance Tuning, Modification by Jeff Hartmann.

Also check out RC Engineering for more F.I. goodies.


Date: Sun, 19 Sep 1999 21:48:42 -0700
From: "Carlos A. Iglesias" (rotary@san.rr.com)

Subject: (rx7) [All] Spic Racing Factoid #09199901


  1. Fuel Flow Change = (New Pressure/Original Pressure) ^ .5 (i.e. change in fuel flow is equal to the square root of the percentage change in pressure.)

  2. Fuel pumps flow LESS as the outlet pressure increase.

  3. Common wisdom indicates that injector should not be run above 70-80 psi.

  4. A benefit of increasing fuel pressure is better atomization of the fuel by injectors.

  5. Most injectors are flow rate at the 30-40 psi range.


Date: Tue, 17 Oct 2000 10:22:37 -0400
From: "Lanham, Wade A." (wlanham@alleghenypower.com)

Doug Henry (from the Supra list) tested each of these pumps through the stock Supra fuel system (so be warned, these numbers will appear lower than the published flow rates for the pumps because of the pressure drop of the fuel system). I cannot back up his tests, but I see no reason for him to skew the results. I don't have the time to test them, unfortunately.

The Walbro was the 340/341, but not the "hot"/high pressure version. The high pressure version flows much more on the high pressure side. I have graphs of the flow of the Walbro (directly from Walbro). I've been using the high pressure pump for months with a baseline fuel pressure of 50psi (peak of about 65psi) and I haven't had any problems maintaining pressure. If I set my base higher than 50psi though, I'll get fuel pressure bounce at idle, which I believe may be caused by my idle getting too rich, my injectors not working correctly at such low duty cycles, or some other reason. The only way to know if this pump is good for 70+ psi (as Wael runs his Bosch) would be to use it with a rising rate regulator, which I don't have. You're welcome to try it. This pump can be purchased for about $100. That's not a typo. My only complaint with the pump is I'd rather have an AN attachment. It has a plastic nipple that I'm not crazy about. Please don't blame me if you have a problem with this pump, I'm only relaying my experiences. If I were selling this pump marked up to $200 though, I might give some support. But I'm not. :)

This graph suggests that anyone thinking of going with a Cosmo might as well get the Supra MKIV pump. It can be purchased straight from dealers for < $200. I'm pretty sure this is a direct drop in. You can see from the graphs that the flow is almost identical to the Cosmo, especially in the high pressure area (where it counts). We were told by one of the Supra dealers that one of the RX-7 tuners buys their pump from them. I won't name any names, I don't want to spread rumors, but I will say that it's probably not who you think.

The numbers for the Bosch that Carlos posted on Cirian's site are even higher than Walbro's published numbers. If it is the same pump as M2 sells (and Wael uses), then it should perform even better than the Walbro. I don't know if M2's pump is the same as Carlos described... but M2 should be able to provide the flow rates for their pump. If not, buy one from Kinsler Fuel injection or another place that knows.

There's no sense in buying a pump rated at 500 gallons a minute at 43 psi if i drips at 60psi! You'd be better off with the stock pump.

    PSI     Walbro          Cosmo           MKIV       MKIII
    - -------------------------------------------------------------
    26       60.0 gph       61.6 gph        60.4 gph     40.1 gph
    30       58.1 gph       57.5 gph        55.4 gph     35.6 gph
    40       51.0 gph       48.4 gph        47.2 gph     26.3 gph
    50       43.9 gph       38.0 gph        37.3 gph     16.7 gph
    60       37.3 gph       28.0 gph        28.7 gph      8.0 gph
    70       32.1 gph       16.6 gph        16.5 gph      no test
    80       20.9 gph       dribbling       dripping      no test
    85       dribbling      no test         no test       no test


Date: Mon, 23 Oct 2000 10:38:57 -0400
From: "Lanham, Wade A." (wlanham@alleghenypower.com)

I've done a little research on the Bosch pumps. If I keep this up, none of the tuners will sell me parts any more. :)

The information for the Bosch pump that Carlos Iglesias describes on Cirian's site is incorrect (the numbers he lists as LPH are actually LBPR - *big difference!*) It is also the same one that Wael got from M2.

The bottom line is the Bosch flows almost exactly the same as the GSS-341 Walbro pump in the "useful" pressure ranges. The difference is the Bosch has a much more linear fuel flow curve, while the Walbro flows more at lower pressure and barely flows at very high pressures (100+ psi).

BUT, the Bosch costs twice as much, and my opinion is that the Bosch appears to be a higher quality pump. I don't think that anyone who has seen both pumps would argue with that. So, go cheaper (Walbro) or more expensive (Bosch), either one will flow more than you'll ever need. I'd probably buy the Bosch if I knew then what I know now. However, the Walbro is working fine for me, so I'll use it until it fails, if it ever fails.

** Published flow rates (provided by manufacturers):

** 13.2v for the Bosch and 13.5v for the Walbro (so the Bosch numbers are slightly higher than this)

BOSCH (10208,10209,10210) at 13.2v
Pressure GPH Amps
0	66	5.5
20	62	6.9
50	54	8.7
70	50	10.0
100	43	11.7
120	33	13.0

WALBRO GSS-341 at 13.5v
Pressure GPH Amps
0	86	5.0
20	79	6.1
50	65	8.5
70	57	10.2
100	34	13.2
120	12	15.5

The performance of both of these pumps is excellent, and according to David Henry's tests, they destroy the Cosmo pump. The only question in many of your minds is probably quality. If you are one of these people, just buy the Bosch and forget about it.

Also, the flows I quoted from David Henry (also, I have his testing method if anyone wants to know more):

PSI     Walbro          Cosmo           MKIV       MKIII
- -------------------------------------------------------------
26       60.0 gph       61.6 gph        60.4 gph     40.1 gph
30       58.1 gph       57.5 gph        55.4 gph     35.6 gph
40       51.0 gph       48.4 gph        47.2 gph     26.3 gph
50       43.9 gph       38.0 gph        37.3 gph     16.7 gph
60       37.3 gph       28.0 gph        28.7 gph      8.0 gph
70       32.1 gph       16.6 gph        16.5 gph      no test
80       20.9 gph       dribbling         dripping       no test
85       dribbling         no test           no test        no test

The flow shown for the Walbro pump is the GSS-315. The GSS-315 is the same as the GSS-341, except the 341 flows more at high pressure. So, the numbers shown above for the Walbro are actually on the LOW side. It should be obvious that the Cosmo should only be used as a very mild upgrade. It is also noteworthy that the MKIV Supra pump could be used in place of the Cosmo pump, for those wanting to take that option. Chris Davis found a dealer selling the Supra pump for $190 IIRC.

Many of you are probably wondering why the measured flow for the Walbro pump is so much lower than the published numbers. The pressures measured for these tests were at the pressure regulator, not at the pump outlet. This suggests that the 3rd gen Supra fuel system has approximately a 20psi pressure drop through the lines, fuel filter, etc. at working pressure. If the pressure had been measured at the pump outlet, the fuel pressure might have been 10psi or 20psi higher. My point is that measuring the flow of a pump working against a fuel system will always give lower flow numbers than the published numbers.

In reality, it is the flow at the injector rails that matters (which is usually just before the pressure regulator), so the numbers measured by David Henry would actually be better numbers to use when planning your fuel system. If a pump will supply just enough pressure for your application based on its published flow rates, then it probably won't be up to the task in reality.

From: Carlos A. Iglesias (carlos@the-rotary.net)
Date: 10/25/2002 08:59 PM

Though you might want to add to my legacy of Fuel Pump Geekdom:

Here's the INSIDE track on THE Bosch pumps:

Bosch p/n	Kinsler p/n	Porsche p/n		Inlet		Application
0 580 254 984	10208		911.608.102.00		.475" barb	'79 911T front pump
0 580 254 979	10210		???			14 mm		'79 911T rear pump

I've consistently found the best price for both of these at www.thepartsbin.com , plus ground shipping is free.

Date: Mon, 29 Jan 2001 06:29:05 -0800 (PST)
From: e s (rx712psi@yahoo.com)

> Hi, I've been digging into the my engine bay and I noticed a small silver
> module under the brake booster that I can't identify.  It looks kind of
> like the igniter in that it has heat sink fins on it but it's smaller
> (and not mounted on the inside of the left fender like the igniter).

If you are talking about the thing that is bolted to the top of the frame rail then that is the fuel pump resistor. there is a relay that bypasses it at WOT or 3KRPM to increase the supply voltage to the fuel pump.

Its main purpose is to keep fuel pressure low at idle.

[Mail me] [To Lightning home] [To my home page] [Copyright Notice]