william_b_noble

I had thought of DTL - a diode OR to be specific, but we need an AND function - both the motor and the SAI must be 12V to activate - maybe I'm getting rusty but I don't see an easy way to do that with two diodes and a resistor - am I missing something? (I suppose if you rely on the impedance of the source pulling the signal down to zero when not energized, that would do the trick, reverse the diodes and pull up to 12V with the resistor.....)

Mike J

Look at the engine rebuilding thread on my web site, specifically http://www.pcarworkshop.com/index.ph...e_SAI_passages.

That might help you understand the SAI passages and where they go.

Cheers,

Mike

ltc

The easiest way (for a quick hack with mostly existing parts) would probably be to just add a simple transistor inverter to go from NAND to AND.
(NAND/AND ... just a matter of taking the output from the emitter or collector I suppose)

You could also just use a dual optocoupler or use a single CMOS AND/NAND gate as well.

Is it really as simple as just having both motor and SAI valve at 12VDC to turn on the 4 output FETs?
if that is the case, then a solution for 97-98 993's is very easy.
in fact, the new circuit could be used for both 96 as well as 97/98 by just tying the 2 inputs together for the 96 case.

ltc

A simple RTL AND gate, as you mentioned, would work:

http://www.transtutors.com/universit...AND%20Gate.JPG

However, it would require bringing in a 12V lead for bias.

If I have some spare time, I would like to think about doing it without the need for a bias supply, similar to the current 96 implementation.

william_b_noble

Thanks, that is most excellent, and it explains why a snake run in through the little port over the exhaust valve hits a solid object that it won't get past - what a crummy design. In the past I've been able to get the light to go off by cleaning the ports with solvent, I was so hoping that they could be cleaned mechanically without removing the cam towers. I guess it's back to solvents, or attaching a bypass tube to inject air into the manifolds elsewhere than through the head.

ltc

OK, if I understand the requirements for the 97 and 98...both the SAI pump AND the newly added SAI valve signals must be 12V in order for the O2 sensors to be shorted by the MOSFETs (please correct me if I'm wrong in this assumption based on the previous posts)... then I believe I have an updated design.

(there are moments when being an insomniac with OCD does have its advantages )

The updated circuit should work for both 96 (which has been tested successfully with the current circuit's single input... SAI PUMP only) and the 97/98 (assuming they are both the same and they follow the assumptions at the top of this post).

I can assemble a modified (Frankenstein) circuit using 2 of the 6 spare boards from the first build. If someone can test this on a 97/98, please let me know.

Lorenfb

"OK, if I understand the requirements for the 97 and 98...both the SAI pump AND the newly added SAI valve signals must be 12V in order for the O2 sensors to be shorted by the MOSFETs (please correct me if I'm wrong in this assumption based on the previous posts)."

Right, which has been the case all along, i.e. both the valve and the pump are
wired together thus both getting power at the same time. So the present workaround
when powering the SAI relay powers both the valve and the pump.

Please check the Porsche '97/'98 993 diagrams WKD 483121 per Sheet 5B
and the '96 diagram Sheet 5B, i.e. both the valve & pump are wired the same,
before designing workaround circuits

ltc

As I've said all along, I don't have Porsche wiring diagrams. I don't even change my own oil.
I was simply trying to donate my electronic skills to help fellow members.
If someone can define the function necessary, I can implement the hardware.

william_b_noble

I'll look at the Porsche schematic later, but if Loren is right, and I have no reason to believe otherwise, then the speculation about the SAI valve being monitored in the later 993s is not correct and the reason the circuit doesn't work in the later cars is still unknown.

william_b_noble

OK, I looked at the schematic for the 97, specifically diagram 97-131, look in zone M-150 on the diagram and it shows the secondary air pump and the SAI valve - and as Loren said, they are wired to the same +12 excitation. The exictation comes to the motor and valve via connector X60 pin 3. The exictation is via a relay that is driven from pin 37 of the computer.

So, I don't see any value in adding another input. I have attached the relevant part of the schematic to this post

william_b_noble

here is what the 993 OBD manual, 1995 has to say:
P0410
Secondary Air Injection Signal Implausible (cylinders 1 - 3)
P1411
Secondary Air Injection Signal Implausible (cylinders 4 - 6)
Function
To improve emissions and to faster heat the three way catalytic converter (TWC), the secondary air injection pump is activated after the engine is started and when the engine is cold, thus blowing secondary air behind the discharge valves. If the engine temperature at start-up is less than - 10° C, the secondary air injection pump is disabled until the engine temperature has reached more than 60° C.
Diagnosis runs while the secondary air injection system is active. The voltage of the oxygen sensors is checked by the ECM during this time.

there is additoinal information on page C-12 and 13 of the manual of the manual.

Secondary air diagnostic
Through the secondary air system additional oxygen is introduced behind the exhaust valves to reduce harmful emissions. This occurs after starting the engine within a defined engine temperature range and for a predetermined time.
While the secondary air injection is active, the diagnosis of the system is being performed. During the diagnosis of the secondary air system, additional air behind the exhaust valves creates a surplus of
oxygen at the oxygen sensors causing a reduction of the voltage at the oxygen sensors ahead of the Three
Way Catalytic Converter

The diagnosis checks in a predetermined rpm/load and temperature segment whether the sensor voltage is too low or too high because of missing secondary air.

There is a graph on page 13 that seems to suggest that the sensor voltage should be around 100 mv instead of above 450 mv. A forward biased silicon diode is about 700 mv - I wonder if there is a way to introduce a 100 mv offset in the ground for our magic circuit - something simple???

william_b_noble

Loren (or anyone else) - can you please confirm that on a car without the CEL the sensors both before and after the catalytic converter read about 100 mv? when the weather meets my fancy, I'll try adding some adjustment to the circuit and see if I can get it to read whatever the right value is, maybe that will work.

Lorenfb

"can you please confirm that on a car without the CEL the sensors both before and after the catalytic converter read about 100 mv?"

Will check for the actual values the next time a readiness test is done.
But from what can be remembered, they never reach zero. The voltage
is in the order of 10s of millivolts but less than 100 mv.

ltc

This is what I believe we know to date:

1. The circuit, as currently implemented, has been successfully tested on several 96 993's.

2. I did receive the following feedback from a 98 993 owner who tested the circuit:

Just wanted to let you know that the SAI bypass works for me - my car is a 1998 C2S....no issues...

and then, I changed the wires myself from blue to black at a later time. My guess... just from my own observations...is that the wires that are spliced in to the porsche terminals are not seated properly by the installer, so not making a good connection.

(when I changed the wires from blue to black - within a day I had the CEL come back.... but then I re-seated the wires.... and it disappeared)

Also note, that when we did the install and reset the errors.... I came back a week later to check on his OBD about the readiness codes.... everything was fully ready ! (we were both surprised) - all I did in the interim week was one bmw-warm up drive cycle. (and not a very good one at that)

So, based on a small sample size, and based on the information above that the SAI Pump and SAI Valve are driven by the same single line, the current implementation appears to be work, or at least be worthy of additional user feedback.

Lorenfb

"based on a small sample size"

One.

And how many boards were sold, i.e. '96 - ?, '97/'98 - ?.