Air Pump Removal

Last updated: February 10, 2002

Date: Fri, 24 Oct 1997 14:33:53 -0400
From: "Robinette, Maj Robbie D." (

A while back I asked if anyone could explain the theory behind the RX-7's air pump. I never got an answer from the list but I did run into a book on emission control and it described how it worked. The air pump injects air into the catalytic converter when the RPM is below about 2800 RPM (can't remember exactly what the RPM cutoff is, it's in the shop manual.) Above this RPM the magnetic clutch on the air pump pulley releases because the temperature in the cat is hot enough without the additional air. The injected air works like a bellows in a fire place. It feeds the fire (chemical reaction) and gives the nitrogen oxides (referred to as NO or NOx on emissions tests), hydrocarbons (HC--unburned fuel and oil) and carbon monoxide (CO) plenty of oxygen to combine with to transform the pollutants into carbon dioxide (plant food and greenhouse gas) and water. Without the air pump the cat is less effective at low RPM and excessive HC, NO and CO will be present in the exhaust.

While on the subject, has anyone removed/disabled their Exhaust Gas Recirculation (EGR) valve? I assume some of the racers have. Did you have to modify timing or the air/fuel ratio?


Date: Mon, 24 Apr 2000 01:58:35 -0700
From: Aarroneous (
Subject: Re: (rx7) Clatter/grinding sound from 93 air pump

> 1. Is this a known sign of the failure of the air pump?

It sounds more like the air pump's clutch engaging and disengaging. However, it's not supposed to be so loud so that one would take notice.

> 2. Can it be repaired?  Or is replacement the only option?

Unfortunately, no shop I've spoken with can rebuild the damn things. After taking mine apart, it looks like there are 3 fibreglass "fins" which force air into the cat. It's these parts which I believe are considered "unserviceable." The only way to get them is to buy a whole new air pump from Mazda. The rest of it looks like one might be able to repair it.

> 3. Is there some other way to diagnose it's failure other
> than by merely listening to it?  I did a visual inspection
> of the belts, hoses etc. and everything appeared to be normal.
> All the fluid and oil levels are at full or just below.

You can determine if it is indeed the air pump causing the noise by disconnecting the electrical connector. This will prevent its clutch from engaging. Drive around after disconnecting it. If the sound goes away, it is indeed the air pump.

> 4. Is there something underneath/near the air pump that's
> making the noise and it only appears to be coming from the pump?
> (I am 90% certain it is, but it's possible it's coming from
> elsewhere in vicinity of/under the pump).

See above.

> 5. Is it safe/non-damaging to drive the car the 3 miles to
> the Mazda dealership to have it looked at?  Would disconnecting
> the electrical connection on top of the pump be safe or advisable
> for that trip?  I am terrified of something seizing.

The air pump is only used for smog purposes, and will not hurt your car if it is disabled or removed. Some claim a lumpy or hunting idle when it's disconnected, but may others have no problem. I have been told that removal of the air pump will result in the early demise of your catalytic converter. I highly doubt that though, as above 3000RPM the air pump is not being used.

As an interesting side note, for those of you who wish to remove the parasitic drag (not much really) of the air pump, but want to keep a "stock looking" car for visual inspections, you can remove all the internal parts of the air pump, and leave the outside connected. Of course, simply disconnecting the the electrical connector would be much simpler, and just as effective.


Shiv Pathak sent the following questions in response to my queries about removing the airpump after installing a new cat of the type that does not require air to be injected via the airpump. I follow up with some questions of my own.

From: Shiv
Date: 12 Jan 00

>Does disconnecting the air pump really free any hp?  Isn't it clutch
>operated only at lower engine speeds?  I'd imagine that the only effect
>air pump removal has on the car is a lumpy stinky idle.  What am I
>missing here?

From: Steve

You are probably right that removing the air pump itself will not free up any horsepower. However, removing it will allow me to install my RP pulley kit, which should free up a few HP.

Could you explain the lumpy idle thing? I heard this before, but apparently wasn't paying much attention. Did it have something to do with the airpump injecting air upstream of the O2 sensor, in addition to injecting it into the main cat? Did the lack of air going in above the sensor mess up the idle?

As far as stinky idle, would the new cat that does not require the airpump take care of this? Or is it a question not of uncatalyzed gunk, but more of a fuel-rich idle due to the O2 sensor issue, and the stink was of unburned fuel?


Date: 6/5/00 4:50:51 PM Eastern Daylight Time
From: David Breslau (

Late 90's Corvette uses an electric airpump (smog pump). It's lighter than the Mazda pump, you can fit it anywhere, and it's what I'll be using when I go single turbo. Cost from a local (MA) dealer is about $180.

Not sure if anyone else has tried this yet, but with an appropriate relay to take the stock pump clutch signal and send it to the 'Vette pump, I'm pretty sure it will work. May not pass visual smog inspection, but what single turbo does...

Date: Tue, 7 Oct 97 08:55:46 -0500
From: Brooks Weisblat (

>I just removed my air pump, to fit some IC lines, and I had to use a 30" 4 
>rib belt, between the alternator and water pump.  Now with the reduced 
>contact on the water pump pully, doesn't the belt slip?  If so, what about 
>overheating on the highway or the track?
>-Tosh Desai

You'll extend the life of your waterpump with this modification by putting less strain on the water pump pully. You'll also open up space and allow for more air to flow near the turbos.

I was also worried about heating up. There are a few options for you at this point.

1) Find a 30" 5 rib belt (good luck, hard to find).

2) Add the pettit power pulley, and use a readily availiable 28" 5 rib belt.

3) Add the Greddy pulley system, which enlarges the waterpump and alternator pully, adding more tension and deflection to the belt. This system also comes with the proper 5 rib belt that will fit with your airpump being removed. Elite motorsports has them.

4) Add the Greddy and Pettit pulley system at the same time. This is what I did. You need a 29" 5 rib belt for this configuration. I ordered the belt from Discount Auto parts.

There is more info on my website along with part #'s and pictures of the setup.


Date: Sun, 24 Sep 2000 13:42:13 -0400
From: "Zach Keller" (

If you plan on doing the mid-pipe, you must defiantly install Pettit or M2's upgraded computer

Here is what I did:

A) Remove the air pump and ALL the hoses. You must cover the openings on the intake where the hoses connected. I made my own blocking plates.

B)Removed the air control valve under the plenum. You must keep the sensors plugged in or you will get an error code. Just unscrew them from the ACV and lay them under the plenum. I then made a block plate out of aluminum to put in place of the ACV. You can also remove the vacuum lines associated with the ACV. Leave the solenoids plugged in.

C)Removed the EGR valve, solenoid, and the vacuum lines associated with it. Also made a block plate for the EGR. The EGR removed does not hamper drivability or performance. You must remove the sensor on top of the valve and tuck it under the plenum so you do not get an error code.

D) Get the Greddy pulley kit so you can drive the car w/o the air pump in place.

This really cleans the under the hood up. Check out Rob Robinettes web site. That is what I used as a reference. There are also other things that he did which I did, i.e.. get rid of the double throttle, PCV valve, AST explosion device. Good luck, Zach


Date: Wed, 13 Aug 1997 21:13:55 -0400 (EDT)
From: Francois McKellar (

To omit the air pump with the stock pulley I would suggest a 5 ribbed belt of 30"-30.5". Unfortunately, they do not make it that size. I was able to barely fit (not too tight or the alternator bearing will fry) 29" belt.

I would assume the Greddy pulley to be larger than stock so that it will underdrive the systems by rotating more slowly. The 29" would be impossible on a larger pulley. The next smallest 5 ribbed belt I could find was 31.5" (there is also a 32.5") which I belive would work perfectly for your application (adjust the alternator height to make the best fit).


Date: Fri, 10 Apr 1998 00:17:28 -0500
From: (P Desai)

Use the 30.5" belt, and drill the slot on the alterator bracket to make it longer, so you can pull the alternator further and tighten the belt enough. Don't over tighten the belt, else the waterpump bearings will suffer.


Date: Wed, 09 Feb 2000 20:12:54 -0500
From: Wael El-Dasher (

I had several configurations run so I used several sized belts, I have them all but I can't tell which is which now. But I'll tell you how I buy my belts. I use a string, run it the path of the belt, and then measure it. I used Gates belts because that's what the local speedshop carries. The string trick is a very old but it works.


Date: Wed, 09 Feb 2000 23:24:55 -0500
From: Wael El-Dasher (

> i purchased a kit from pettit racing to do the task< it consisted of
> the alternator pulley, waterpump pulley, and Gates micro v belt part 
> # K050290,   i forgot the price of this kit, i think i paid between
> $100-$150 for it. 

This is the way Gates numbers their belts so you can speak Gatese (grin)

Part number K050290, the K is the series, 05 means five rib, 0290 means 29.0 inches. If you are planning to use a stock pulley then you will need a shorter one, perhaps just 1 inch, so try part # K050280 if it exists. The alternator gives you some reasonable adjustability, but do not over do it. The belt should have some slack.


Date: Mon, 06 Jul 98 07:24:08 -0500 From: "Linthicum, Sandy" ( >When removing the air pump on a 3G car, do any of the tubes
>need to be capped off? The one coming out of the air pump
>to the air valve would be a likely candidate. What about the
>air silencer assembly from my Pettit Intake? Can I loose that?
>Any help would be appreciated!

Cap off any holes left open with (if you can find them) 7/8 inch end caps. You can force 3/4 inch on and use them (I did) but it's a lot of work and 7/8 is the right size.


Date: Tue, 12 Oct 1999 22:47:14 EDT

>Has anyone put a resistor on the air pump clutch harness?
>if so what rating resistor did you use?

300-330 ohms.


Date: Tue, 12 Oct 1999 21:31:36 -0700
From: "Carlos A. Iglesias" (

>  300-330 ohms.
> Did this defeat the lopey, crummy idle and was idle
> mixture still rich? 

Unfortunately, the post-airpump-removal lumpiness has nothing to do directly with the airpump circuitry. Instead, it is caused by built-in biases to the Idle ECU's logic for maintaining Lean Best Idle.

I seem to recall alleviating the problem by sufficiently subtracting out fuel at ilde with the PMC. I recall the figure was in the -18 to -32 range, though I can't remember with which fuel injector setup that was.


Date: Thu, 20 Apr 2000 14:41:21 -0700 (PDT)
From: Jay (

Everyone talks about a lumpy idle when removing the airpump. I ended up removing mine when doing my non-seq conversion. In doing so I removed all the vacuum hoses connected to the solenoids.

My idle is quite smooth. Maybe a +/- 20RPM at the max. On the AVC-R the RPM usually just flucuates on the ones column.


Date: Thu, 20 Apr 2000 16:53:54 -0500
From: "Westbrook, Chuck E." (

The air pump is engaged at idle! That is why its removal causes the bad idle without ECU reprogramming. Without the added air, the engine is running too rich as the O2 sensor sees it, but in reality isn't rich. Then the O2 sensor tells the ECU to cut fuel. Then the whole control system tries to keep the engine running but in a terrible way.


Date: Thu, 15 Apr 1999 16:10:59 +0200
From: "Bernd Kluesener" (

> If my car is used strictly at the track, and I have already
> removed the air pump is there any reason to leave the ACV mechanism
> hooked up?

The ACV (Air Control Valve) consists of actuators (Relief, Switching) and solenoid valves (relief 2, port air bypass, split air bypass). It gets its air supply from the air pump and directs air - under ECU control - to the exhaust port, main catalyst or back to the atmosphere. Once the air pump is removed, the ACV becomes useless and you might as well remove it (you need to make simple blocking plates). If you still have the sequential setup, you will also need to make a mounting bracket for the Turbo Control 1 solenoid valve, which was attached to the ACV housing.

To avoid ECU error codes, either keep the solenoid valves attached to the wiring harness or replace them with appropriate resistors.


Date: Tue, 17 Aug 1999 11:21:04 -0500
From: "Westbrook, Chuck E." (

Removing the air pump causes the idle mixture to go very rich. The AAS/ABS is located under the curved plastic inlet pipe that connects to the throttle body. It is just above and to the left of the bottom right nut that holds the throttle body to the intake extension. Get a flashlight and small mirror and you can see the brass flat head screw inside the aluminum hole of the throttle body.


Removing the airpump is supposed to make the idle go a little lumpy. No one has really explained to my satisfaction why this is so. The airpump is supposed to inject air upstream of the O2 sensor, and some people have said that the lack of air causes the O2 sensor problems at idle.

This theory does not seem to hold water - the O2 sensor is not supposed to be consulted at idle - the engine computer is just working off maps, not readings from the O2 sensor.

Maybe it has something to do with the ACV that Bernd mentioned above. I personally don't have all the info to understand what the ACV does.



Date: Thu, 10 Feb 2000 08:40:37 -0500

... the search for the cause of rough idle. In looking at my shop manual, the air pump feeds "secondary" air to the air-control valve, which in turns routes it to one of three locations.

First of all, what do they mean by "secondary"? Second, of the three locations, two are understandable. If you remove the main cat, then you can remove the air pipe leading to it. I can't recall the second one off the top of my head (it was late and now I'm at work) but it was understandable as well. The third place was to the "switching" valve. Now this is where I think it gets hairy... because anytime a solenoid is involved... well, you know.

Anyway, I have seen people's cars where the airpump was simply removed and the pipes capped off. This doesn't make sense because as long as those components are in place, they should at least be fed ambient air... not sealed off.

I noticed that when I disconnected the electrical connection to the airpump after the car warmed up, the idle got a little rough. When I then removed the top hose, it returned to normal. I think the switching valve lets air into the exhaust ports or something. It seems the only way to totally eliminate the air pump is to remove all the related components except the switching valve.

Remove the vacuum hose and just stick one of the solenoids little filters on the end of it. then we should get smooth idle without all that emmissions crap. Some of the preceding is speculation as there are still a couple of questions, however, I have spent some time studying the shop manual and done some homework. I even came up with an idea of a one-way valve that can be welded onto the downpipe. Anyway, if anyone can offer some insightful input at this point, I would appreciate it. I am this >< close to getting this phenomonom licked short of replacing the ECU with a techII or something :-)

______________ From: "kevin kelleher" (

On the FD, the air pump NEVER pumps air into the intake system. It NEVER pumps air in to be used in the combustion process. It only pumps into the exhaust stream, or vents to the atmosphere, or is clutched out.

I previously noted a single, limited case where it INDIRECTLY effects mixture strength, at low speed / load (NOT AT IDLE), by diluting the exhaust before it reaches the O2 sensor, which is active at this specific condition.

I will repeat what I think may be a fix to the lumpy idle, based on my understanding of the system:

With pump disabled/removed, I would try resetting the idle speed per the manual ( the 'TEN' terminal of the diagnostic connector MUST be grounded during the adjustment ..... this was not mentioned in my manual ). Use an accurate tach. The adj't screw is under the TB, screw head faces the left fr wheel. No guarantee.

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