william_b_noble

so far, from what I'm reading, NOBODY has reported success at clearing a CEL using this little circuit - I have reported FAILURE using my hand wired equivalent - it does make the car run poorly for the first minute, that's good evidence that it is doing something. So, still waiting for anyone to confirm that it doesn't work, or to report that it does. My car is a 97, so if you do report success, please include the year of your car.

Lorenfb

"So, still waiting for anyone to confirm that it doesn't work, or to report that it does."

- The Thoughts of Many -

jimbo3

^^^See post #1 and #79.

william_b_noble

yes, you are right - there are reports that it works on a 96. sorry, I missed that.

so, I am reporting that it DOES NOT work on a 97. So, has anyone else tried the circuit on a 97 or a 98, and if so, please report results

ltc

Loren,
I am somewhat confused and am hoping you could help me.

Flying_low started this thread based on a design he came across and also having tested it on his car.
I then offered to implement his circuit in a low cost assembled (or kit) version at cost.
Neither of us are commercial ventures nor have experience with Porsche ECU's.

I assume (based on your website in your profile) that you are a commercial enterprise and do have experience/knowledge of Porsche ECU's.
If so, perhaps you could offer some additional input/data on the differences between a 96 and 97/98 993 ECU?
With this information, it would be possible to come up with a modified version of the 96 ECU based SAI circuit, which would benefit a large number of RL members.

Jimmy586

And the purpose of the circuit is?
__________________

william_b_noble

well, the purpose of the circuit is to trick the computer into thinking that the air pump is putting extra air into the exhaust by tweaking the O2 sensor outputs to make them read "lean" - very lean. It does this by shorting the O2 sensors to ground, so the computer sees zero volts from each. This is apparently sufficient to trick the 96, but not the 97 (and presumably not hte 98). I have the ECU manual for my 97, but it doesn't really say what the thresholds are - it offers some graphs. I have been tempted to modify the circuit so that instead of shorting to ground, it just reduces the voltage by a preset fraction and see if that works. Alternatively, there is always the "clean out the passages with injector cleaner" method, which does work ..... for a while. I agree with LTC that any additional data on the tests performed by the 97 and 98 cars would be most helpful.

Lorenfb

"If so, perhaps you could offer some additional input/data on the differences between a 96 and 97/98 993 ECU?
With this information, it would be possible to come up with a modified version of the 96 ECU based SAI circuit, which would benefit a large number of RL members."

Would be more than willing to help, as I have access to many Porsche
independents and dealers that I could use for evaluations. But first it
would be very helpful for someone with a later car ('97/'98) to provide
some form of feedback. In all likelihood, only a minor change may be
necessary or just tweaking the way it's wired into the car. Let's not
assume that there's a real problem without some reports.

william_b_noble

ok, what feedback would you like? I built the circuit (there is a photo some where back in the thread of my hand built circuit, I didn't use the circuit board others used) - I installed it, verified via OBD2 that all 4 O2 sensors were zero volts during the air injection motor running period and normal thereafter, observed that the engine ran much worse during the air injection motor run period, left the circuit in place for about 20 to 30 drive cycles, enough for the CEL to reset if it was going to, also verified that P401 persisted, so SAI fault was still detected. Disconnected the circuit (I made it with little connectors to make disconnecting easy) and the engine ran better during the air injection motor run period, still have CEL. What more would be helpful to know?

Oh, the CEL has been on for most of a year, I am sure it is due to clogged ports in the head, I've had the exhaust manifold off a few times to fix it (it's a pain), once enough ports are free flowing (aprox 4 or 5 out of 6) the CEL goes away by itself after two drive cycles.

ltc

It is possible/probable that the 97/98 ECU does not want to see 0.0V, which for all practical purposes is what a fully enhanced MOSFET provides in the current implementation.

It is likely that the ECU wants to see either a small voltage (or small resistance, depending on whether the circuit is voltage or current biased)

There are 2 quick ways to test this that I can think of.

1. Replace the 2N7000 MOSFETs with NPN bipolar transistors (pick one). In saturation, this will provide a small saturation voltage which may be interperted by the ECU as a correct lean condition (again, no idea what voltage it is looking for)

2. Keep the circuit as is, but introduce a variable resistor (small resistor wired in series at the terminal block, a variable resistor/potentiometer, etc). Different values will be needed to determine what resistance emulates the lean condition.

If access to the specifications to the lambda (O2) sensor and schematics of the ECU were available, it would greatly expedite the process ...

Lorenfb

"observed that the engine ran much worse during the air injection motor run period"

This maybe a clue. If the '96 cars don't have this problem, then the early
ECM ignores the O2s from a running standpoint but just monitors the
voltage. To make any kind of a valid assumption, one needs more than
one sample. That's why more purchasers of the circuit need to provide
feedback.

"Replace the 2N7000 MOSFETs with NPN bipolar transistors (pick one). In saturation, this will provide a small saturation voltage which may be interperted by the ECU as a correct lean condition (again, no idea what voltage it is looking for"


That'll result in greater than 150 - 250 mv which is way too high.

"Keep the circuit as is, but introduce a variable resistor (small resistor wired in series at the terminal block, a variable resistor/potentiometer, etc). Different values will be needed to determine what resistance emulates the lean condition."

The O2 sensor is an high impedence output voltage source, i.e. any resistance
less than 50K - 100K ohms will result in zero O2 volts.

"If access to the specifications to the lambda (O2) sensor and schematics of the ECU were available, it would greatly expedite the process ..."

Good luck on finding the 993 DME ECM schematics. You thought the 964
and the '95 DME ECMs were complicated, take a look inside of the late
993 DME ECM.

Bottom line: The proper evaluation of the circuit requires the use of a Porsche factory tester
that can force the SAI test and then monitor what happens. Guessing not allowing!

william_b_noble

I am inclined to suspect that the later cars don't expect zero volts, and also that it would look for a ratio of the pre and post Cat sensors. It would be easy to add a 1 meg trimpot in series with each input and adjust for a proper voltage, but the graphs in the manual do not have voltages on them so I don't know what to adjust for. I am deeply suspicious that once the circuit is trimmed to force the proper voltages then the fault will go away.

Clearly, during warmup, the 97 at least (or at least MY car) is still closed loop. If it is a really smart piece of software, then instead of the transistor to short to ground, we would need an op amp to subtract the proper voltage that would represent the added O2 from the air pump.

If I (or it seems several other posters) knew what was required, the circuit is probably not all that hard. There is, of course, a mechanical solution - drill a hole and inject the air from the air pump in a location that is easier to clean - drilling a hole or two into the connection from the CAT to the manifold and then MIG welding in a fitting is not hard - that might almost be a permanent solution - I wonder, is it worth it?

ilko

I installed the circuit today on my '96. It is beautifully done and very easy to put together. Took a little over an hour to piece everything together and run the wires.

After I was done I ran VAG-COM in OBDII mode during start up and got the same values as flying_low. No CEL, or any soft codes.

In the interest of full disclosure, I've had P1411 pop up twice during my ownership of the car. Both times the CEL showed up a good half an hour after I started the car and while driving. Is this normal? My thinking is that if he system detects a clogged port (voltage doesn't drop to 0) at start up then it should then throw a code. Or am I mistaken?

Thanks again to flying_low for the schematic and paving the way for this. And a special thank you to Lewis for building these for us!

william_b_noble

so we have independent confirmatio nthat the circuit works on a 96 and we have my statement only that it doesn't work on later cars, right? Has anyone else tried the circuit on a 97 or 98? alternatively, can someone with a working SAI and no CEL tell me what the four sensors read during SAI pump run and immediately after? thanks

jimbo3

I got a P1411 code a couple of times, too. Same as your's- car thoroughly warmed up, driving along minding my own bidness, then it popped. Cleared it with a scanner and that was the end of it.

Did you notice any uneven idling or hunting after start-up as some of the others with later cars mentioned?