BimmerFest BMW Forum banner

DIY: Thread dedicated to E39 SAS Delete

23K views 25 replies 7 participants last post by  eohagan 
#1 · (Edited)
I want to dedicate a thread for those who want to experiment with BMW ECUs to do the SAS Delete mod. The following is just a theory based on the Volvo, but I am hopeful that some BMW gurus will use this info to do a BMW SAS Delete.

So, if you use the info here and do a successful E39 (or other BMW models) SAS Delete, please post your success stories.

Remember that virtually all European cars, whether VW, Audi, Volvo, BMW, Mercedes, Porsche etc., use similar systems:
- ECU made by Bosch or Siemens.
- Air Valves (small electrical air valve, mechanical air valve, and air pump) made by Pierburg.
- Similar principles of operations of the SAS system.

---------------
For those of you who also own Volvo, you would be familiar with "SAS Delete". Remember that virtually all European cars (Volvo, VW, Audi, BMW, Mercedes, Porsche etc.) use the same setup for SAS, i.e., same Pierburg air pump, same air valve, similar ECU monitoring system.

Basically in the Volvo world, Radio Shack Diode 1N4003 (90-cent diode) bridged between FRONT O2 sensor and Aux Air Valve Solenoid circuit does the trick in fooling the ECU that the Air Pump is working. I have done this diode thingy in my Volvo and basically forget about the Air Pump. Zero CEL.

How does this work? The FRONT O2 sensor monitors exhaust condition for LEAN mixture during cold start. This happens b/c the Air Pump injects extra air during cold start. Let's say the FRONT O2 sensor sends a voltage of let's say 0.2V, the ECU knows it is lean and therefore "happy".
On the other hand, if the FRONT O2 sensor sends a voltage of let's say 0.8V, the ECU knows it is rich and therefore "unhappy", setting a SAS error code.

Just read a bit on how O2 sensor works on the web and you will see.

http://en.wikipedia.org/wiki/Oxygen_sensor

So if you know the Pinout of the E39 ECU, you can do this diode thing for 90 cents.

Below are photos of the 1998 Volvo S70 circuit for your use.

---> First photo shows the circuit diagram and how the diode is installed.
---> Second photo shows the actual mod...





--------
During cold start, the ECU sends a signal lasting about 100 seconds to 2 devices:

1. The Small Electric Air Valve (under intake manifold) to open its solenoid, allowing vacuum to feed the Mechanical Air Valve sitting on the exhaust manifold. This opens the port for the Air Pump to inject extra air into the Exhaust.

2. The Air Pump to run and inject extra air into the Exhaust.
During this time, if the SAS system works as designed, the O2 sensor will send different signal to the ECU:
- At first start, no extra air in Exhaust, i.e., rich condition, a ~ 0.8V[/] signal is sent to the ECU.
- Now the Air Pump injects extra air, the O2 sensor senses this, and sends a ~ 0.2V signal to the ECU, telling the ECU that the SAS system works as designed.
- In other words, the ECU is looking for voltage drop signal, in this case, about 0.6V or so.
- When your SAS system fails (either the small electrical air valve, mechanical air valve, or air pump malfunctions), the O2 sensor sends the same signal of 0.8V to the ECU, no voltage drop signal. The ECU will set an error code of SAS failure.
- The way to do a SAS Delete is to somehow fool the ECU into thinking there is a voltage drop signal.

I was helping my kids with AP Physics (College Physics), so here you go***8230;.Ohm's law states the following; where R = resistance (in Ohms) , V = voltage, and I = current (in Amperes).



This is for single resistor circuit, when you add a diode in series (as in the Volvo SAS Delete), you need to calculate voltage drop across the resistor (recall that the O2 sensor is essentially a resistor) with and without the diode.

I can go into complex calculation (Calculus, and differentiation etc.) but no need to do that. In a series circuit:
R (total) = R(1) + R(2)

And your college professor will tell you that when you add R(2) to the existing R(1) circuit, there is no change in current, but there is a change in voltage drop across R(1) because this is a series circuit. When you add a diode (a diode is basically a 1-way resistor) to a resistor, there should be a voltage drop across the resistor: I verified by a real experiment...

Think about the O2 sensor as a resistor, actually it is but its resistance varies with O2 in the exhaust.

1. Reference voltage is 8.92V. The diode is the 90-cent item from Radio Shack 1N4003.
2. With resistor only, Voltage = 8.88V.
3. With diode connector in series (note the direction of the diode), Voltage = 8.21V.
Or Voltage drop is 0.67V.

Now you know why when you install the diode into the Volvo ECU, for the first 100 seconds or so, it causes a Voltage Drop to pin #32, fooling the ECU!

That is it boys and girls, a multiple-choice exams will follow LOL.





 

Attachments

See less See more
12
#2 · (Edited)
I want to dedicate a thread for those who want to experiment with BMW ECUs to do the SAS Delete mod. The following is just a theory based on the Volvo, but I am hopeful that some BMW gurus will use this info to do a BMW SAS Delete.

Now you know why when you install the diode into the Volvo ECU, for the first 100 seconds or so, it causes a Voltage Drop to pin #32, fooling the ECU!

That is it boys and girls, a multiple-choice exams will follow LOL.
I have much respect for your contributions to this forum but I don't believe this approach will work with our cars. First, most O2 sensors used after 2000 are not variable resistors. The most common type of sensor is made with zirconia (zirconium dioxide) and platinum electrodes. It generates a voltage in response to the oxygen content in the exhaust gas. It is essentially a battery with variable output voltage. Some cars did use a type of O2 sensor that was a variable resistor. They are made with titanium dioxide and usually called Titania sensors. That is not what the E39 M54 and M52tu use. It is possible that the earlier M52 engines use Titania sensors. That is a 1990's technology. I don't know anything about what the M60 and M62 engines used and have to do a little homework there. For Zirconia sensors, I would worry about adding the impedance of a diode in series with it. I don't know output impedance of the Zirconia sensor but that may not play well with added external resistance.

Here is a very good link that describes how an O2 sensor works:

http://www.autoshop101.com/forms/h37.pdf

I don't understand how the Volvo diode approach works because I don't see a way to switch the diode in and out of the circuit.

For those of us with Zirconia sensors, the lean voltage output the DME looks for is in the range of 200-225 millivolts. The only type of SAS simualtor I have seen in the BMW world that is known to work uses relays and a voltage regulator to switch a regulated DC voltage between 200-225 millivolts on to the O2 sensor signal lines to the DME while the activation signal for the air pump is on. As soon as the O2 pump signal is turned off, the relays in the simulator switch the real O2 sensor signal back to the DME. This design is very popular in the E36 world and is used with M52/S52 engines. It has not been proven to work on the E39 yet but I believe it will. I built one of these designs about a month ago and started a thread to discuss it. I have been sidetracked by some very bad luck with my S52 swapped E34. The engine ate some valves for reasons i have not determined yet and is trashed. I have been prepping another engine to install in it and had to put my E39 SAS simulator on hold.

I had 3 small printed circuit boards made up and bought enough parts to build 3 SAS simulators for experimentation. (I have 2 bimmers with SAS.) I hope to return to this project in a week or two.
 
#4 ·
...I don't understand how the Volvo diode approach works because I don't see a way to switch the diode in and out of the circuit...
The Volvo terminal #37 feeds the small electrical air valve (sitting under the intake manifold), and it is activated for only 100 seconds or so. The BMW E39 does the same thing, the E39 ECU activates the small electrical air valve for about 100 seconds during cold start.

After the 100-second period, the small electrical air valve is off.

This is why whoever invented this diode thingy on the Volvo is very clever!
 
#3 · (Edited)
So others find this, I'll cross ref in the bestlinks.

Here's what I currently find by typing /sas F3 in the bestlinks, among others:
- How to maintain (1) (2) (3) & replace (1) & troubleshoot the BMW E39 SAS SAP valve secondary air pump system (1) (2) (3) (4) & an SAP valve group buy (1)

Updating that keyword-rich sentence to add this thread, I get something like:
- How to maintain the secondary air system (1) (2) (3) & how to troubleshoot the SAS (1) (2) (3) (4) & how to do an SAS delete (1) & how to replace the (SAP) secondary air pump (1) & an SAP valve group buy (1)

Cam:
I always try to put the right keyword set in the bestlinks.
The way I've written it above, they'll find it with "sas" or "sap" or "sas delete", in addition to a few others inherent in the sentence above.
Is that enough?
 
#7 ·
Circuit simply shunts 0.67 v to ground when SAS solenoid is energized. Pin 37 is grounded ONLY while SAS solenoid is energized, hence 0.67 v drop ONLY occurs during ~100 seconds when SAS solenoid is energized. O2 sensor voltage drop to DME does not occur when SAS solenoid is not energized. I agree, BMW's need two diodes, one for each bank.

Very clever, simple solution........provided the CEL/SES light stays off.
 
#8 · (Edited)
Look,

I can only propose to the smart BMW members.
I know this works for Volvo because I did it on my Volvo. Also, hundreds of Volvo owners do this (just google "Volvo SAS Delete").

Too bad my SAS works perfectly, otherwise, I'd be happy to experiment this.
Worst-case scenario is that this mod does not work, but I am very optimistic that it works.
I will wait for another E39 owner with SAS issues to do this experiment on behalf of forum members.
 
#10 ·
I did a little homework.

1. M52 engines use Titania sensors that are variable resistors. They are fed a fixed voltage from the DME and as the resistance of the sensor changes, the voltage returned to the DME changes. This type of sensor has best chance of working with diode modification. I am certain that the Volvo S70 also uses a Titania sensor. I believe the M52tu engine also used this type of sensor but I have not been able to find proof.

2. M54, M60, and M62 engines use Zirconia sensors that generate their own voltage in response to oxygen levels in the exhaust gas. These sensors may not work with diode mod and some experimentation is required.

Given that I already built one of the devices that uses relays and voltage regulator to fool the DME, I do not intend to experiment with my M54 with diode mod. If someone else wants to try it, they will need to still build a simple electronic circuit as opposed to soldering a diode between two pins on the DME. The line on E39 DME that turns on the air pump will need to be used to switch a double pole single throw relay. Each pole will need to have a diode connected to one side of the switch and a good engine ground to the other side of the switch. The other end of the diode will be connected to Bank 1 or Bank 2 pre cat O2 sensor. When activation pin on the DME for air pump control valve is switched to the on state, the diodes will be connected to vehicle ground and the voltage getting to the DME should be reduced by 0.67 volts. Like I said above, this has a fairly good chance of working with Titania sensors. Not as confident it can work with Zirconia sensors.
 
#11 ·
Compared to the volume of zirconia sensors, titania sensors seem to makeup a very small percentage.
Zirconia sensor output is ~0-1 volt. According to Wiki, Titania sensor output varies from 0-5 volt. If Wiki is correct, diode won't work for titania sensors.

At bottom of page in link is a loooong list of Bosch O2 sensors with descriptions. Use Ctrl-F to search for Titania and they will be highlighted.

http://www.rockauto.com/dbphp/x,mfrname,BOSCH,parttype,5132,d,BOSCH_Oxygen_Sensor.html

It will be interesting to see if Volvo guys who use this fix have Titania or Zirconia sensors.

This link says Titania sensors have RED or Yellow wire.
http://www.lambdasensor.com/main/bmwtitania.htm

This link shows BMW with Zirconia sensors. My M52TU and M54 pre-cat sensors are the same PN and have white & black wires.
http://www.lambdasensor.com/main/bmw.htm
 
#12 ·
Personally, I think it will work either way. This is because the diode is "downstream" from the O2 sensor, so the diode changes the voltage "leaving" the O2 sensor.

In other words, the diode does not care what the O2 sensor material is, all the diode does is to "divert" some of the voltage to another circuit point.

I guess we can debate back and forth between Titania vs Zirconia, the truth will lie in the real experiment!

So we need real-life experiment, anyone out there willing to try?
 
#13 ·
No doubt the diode will reduce titania O2 sensor output by .67 volt. The trouble is we don't know the threshold DME is looking for to detect proper SAS operation w/titania sensors. If titania sensor's output is 0-5 volts as Wicki says, perhaps shunting 0.67 volts won't be enough to prove SAS is working.
What's the 5 digit (13xxx,14xxx,15xxx) Bosch PN for your Volvo O2 sensor and is it titania or zirconia?
 
#16 ·
It is very likely that the Titania sensor gets a 5 volt DC supply from the engine ECU because that is the standard internal operating voltage for most ECUs built in the 1990's. Then, the return signal from a Titania sensor goes through a voltage divider circuit made up of an internal resistor in the ECU and the resistance of the Titania sensor. This voltage divider circuit reduces the voltage going into the analog to digital converters to a 0-1 Volt signal. I am certain that it either works that way or the voltage divider circuit is on the supply side of the 5 volts. Either way, you end up with a sensor signal that is scaled to be from 0-1 volt DC. If the voltage divider circuit is on the return signal from the Titania sensor, then the resistance on the return leg is the parallel combination of the diode and whatever internal resistance is inside the ECU (DME).

The easiest way to determine how the BMW DME does this, is to measure the excitation signal going from the DME to the Titania sensor on an M52 equipped car. If it is 5 volts, then the voltage divider is on the return leg. If it is 1 volt, the voltage divider is on the supply leg. For the purposes of the diode circuit mod, it would be better if the voltage divider is on the supply leg.

The diode mod as implemented on the Volve ECU is still not guaranteed to work on our cars. It depends on the input impedance (resistance) on the O2 sensor input pin. The total input impedance is the parallel combination of the diode and the ECU internal resistance. If the input impedance is a very large number in the 10's or 100's of thousands ohms, then the diode conducts most of the current and the mod will work as it does on the Volvo.

The ideal situation to explore this would be a test bench that can power up the DME and an O2 sensor. I don't have that capability or I would try some experiments. I am still reluctant to try this on my operating vehicles.
 
#14 · (Edited)
I did the Volvo SAS delete but to be honest wit you, I don't even know the Volvo O2 specs that well.

According to eeuroparts.com, the Volvo FRONT O2 Sensor is Bosch 15097:

http://www.eeuroparts.com/Parts/4298/Oxygen-Sensor-Front-15097/

PS: For those who want to try this on your M52 or M54 engine, rest assured that this mod, in the event that it does not work, is easily reversible.
If you don't want to mod, then simply remove the diode and you are done (well, back to square one).
 
#18 ·
Trying to do a little more research and it appears that Volvo use Titania or Zirconia O2 sensors on 98 S70 depending on which engine was installed. Some Canadian and UK market models had Titania. My initial info came from a UK web site. Which engine is installed in your S70?
 
#19 ·
Today, someone asked about the SAS delete, over here, so I bring the relevant references over, for one-stop shopping convenience ...

See also:
- How to maintain the secondary air system (1) (2) (3) & how to troubleshoot the SAS (1) (2) (3) (4) & how to do an SAS delete (1) & how to replace the (SAP) secondary air pump (1) & an SAP valve group buy (1)
 
#20 ·
I thought this was already solved, in testing at least, on this board. In fact, Im sure it was. Someone on the 3 series board came up with a solution for the I6. Someone else made it for his 528. He said he was going to follow up, but I havnt seen it.

I was waiting for the follow up but forgot about it. I have a failed SAS w CEL that I refuse to fix. Its a junk system.
 
#24 ·
This nice sas delete diy was posted today, for cross reference:
This is a DIY for a Secondary Air System (SAS) delete and installation of a SAS simulator.
The Secondary Air Pump (SAP) on my 528 went bad years ago. I didn't want to replace it because of several reasons: 1) Cost, $400 for the pump and a new valve, 2) The SAS has no performance value. It is for cold start emissions purposes only. 3) It's a very poor design. The only down side to this is living with a continuous CEL. My fix was to simply cover the CEL with black tape. Cheap, but (sort of) effective. I would do a code check whenever I was under the hood, just to make sure there were no legitimate codes. The "Secondary Air System flow too low" code for bank one and two were always there.



Last week I learned of a simulator for the E39 I6 (M52) that would fool the ECU into thinking the SAS was just fine. It meets the requirement of faking the expected output of the precat O2 sensors during open loop operations when the SAP is called on to operate. See this link for sourcing the SAS Sim: http://e36-electronic.com/ The parts and build quality all seem very good. Note that this sim is not available for all E39 engine types. Make sure you verify before you order.



So began my project for the afternoon- Install the SAS Simulator and remove the dead weight of the SAS. Both were pretty easy. You can leave the SAS in place and just install the SAS Sim, if you so desire. SAS removal requires a few additional parts. You need to make a valve block-off plate as well as install a vacuum line cap. I suppose instead of making the block-off plate you could leave the SAP valve in place so long as it is in the closed position. However, failed valves are usually the cause of failed pumps.
You have a choice when ordering the Simulator of a plug and play version or splicing the wires. The PnP version is naturally more expensive because it has OEM plugs for the O2 sensors and SAP power plug. So in the interest of time, I chose the PnP version.
I will discuss this as if it were two topics- SAS Sim install and SAS delete, for those interested.

SAS Simulator install:

The simulator installation could easily be done in less than an hour. But for me, I always get distracted and start looking at other things as Im working. Tools are minimal: 10mm socket and ratchet with 3" extension, small flat head screwdriver.



- Start by removing the fuel rail cover. Pop off the decorative bolt covers and remove the two bolts holding the cover on.



- Mount the Sim. The mounting flange on the sim is for mounting only, and does not need to be grounded. I used a stud for ground located under the passenger side cabin air filter that provides a perfect location for the sim mount. Remove the 10mm nut on the ground terminal and mount the sim. NOTE: I had to drill the hole larger on the flange to match the stud (1/4" or 6 mm). Route the wires for the O2 connections over to the Os plugs. Keep the wiring high and away from the exhaust pipes.







- Connect the plugs to the O2 sensors. Unplug the O2 sensors and plug them into the simulator harness. It doesn't matter which O2 sensor goes to which plugs on the harness, except don't cross the harness from one O2 sensor to the next. Easier to see it than explain it. Secure with zip-ties.



- Plug in the SAP lead. The Pnp Sim version comes with a plug that will mate with the power connector going into the Secondary Air Pump (SAP). If you don't have the PnP version you could chose to splice into the SAP wires from the engine compartment. However, I find that trying to unwrap the wire bundles is a pain. It is easier to just gain access to the SAP.
-Remove the front right wheel.
-Detach the outside air temp sensor on underside. Remove the forward wheel well plastic.
-Squeeze sides of cannon plug going into the SAP and pull. Plug the SAS Sim harness into the SAP power supply. Note: When I was at this step my cannon plug completely crumbled into dust. I ended up cutting the remains of the power plug off and splicing the wires into the Simulator harness.









-Reinstall wheel well plastic and temp sensor. Or, while you're in there, relocate the temp sensor to the brake cooling duct like it did. Much better temp reading and doesn't get ripped off from road hazards.
-Install and retorque your wheel.
-Install the fuel rail cover.
-Clear any engine codes. Start engine, and enjoy a no Check Engine Light day!

How to delete the dead weight of the Secondary Air System

-Remove the Pump. Remove/cut the two hoses going into the pump. The pump is held in with three 10mm bolts. You can get pretty good access with an extension and ratchet. One hose goes out to the engine compartment and valve. The other just goes to the pump intake filter visible next to the pump.
- Unplug the pump power supply. Splice or plug-in the SAS Sim harness. Brown is ground, black, red yellow stripe is Pos.
- Remove air hose. Pull the big ugly air hose out from the engine side. Detach it from the valve.
- The valve. At this point you need to either remove the valve and cover the port with a block-off plate, or ensure the valve is closed and leave it in place. If you leave it in place and it is not closed, you will hear the exhaust leak. In either case, detach the vac line and cap it off with a 5/32 vac cap. I made a crude block-off with some scrap 1/8" aluminum (you can get it at a hobby shop), using the valve flange as a template. Try and save the valve mounting gasket, or get a new one. My dealer wanted $25 for that gasket. Pelican was only $5, but had to be shipped. Fortunately, I was able to reuse my old one.







Give the 6-7 pounds of junk an appropriate spot in the trash can, and admire the leaner and meaner look of your engine compartment.



I confirmed the operation of the sim this morning when I had a cold start. I had my scanner hooked up, and while warming up, the system was in Open Loop and the SAS read "Upstream Air". After a minute or so the ECU went to Closed Loop and the SAS read "Closed" Im pretty sure that means its doing what it should. No codes in the pending pile. Im happy.



Legal Stuff: It is illegal to remove emissions equipment. That's why now I only drive my 528 on the farm to check the fences. Im pretty sure with this sim installed the car would pass an emissions test, but that's illegal too.

.
 
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top