Where to look next? Feedback/advice?

[johnb]

I know we're probably past this but for the record: from a quick look at the code, the threshold to activate decel fuel cut off looks like 360 above the threshold to turn it back on, which is 1200 for normal operating temps. So fuel should only be cut above 1560, and re-enabled at 1200. At 39C and below, the threshold is 4000rpm for re-enabling (so 4360 to activate cutoff). This logic is located at 0x0708.

[spacecad3t]

No progress, no time these last few days. Also getting colder/darker.

I dont follow you two here. Im aware of decel fuel cut off, the car seems to do that. If I am cruising and I let off the gas and engine speed starts slowing, around 1400-1600 you can hear the engine speed and tone change. From that git repo, It seems like this is that "flare" period, intended to dampen returning to idle?

Either way- maybe Im wrong here... What I do not follow is "fuel cut off with a warm engine". When Im not decelerating, the engine is static and idling, why would there be a fuel cut off?

Im not sure I have seen any bouncing off the cut off. The original 1200 warm and now 1400 all temps is steady, not surging or bouncing.

And just to recap. The car is an NA not a Turbo. And with the ISV out of the loop, car is idling correctly, just cold starts rough, as expected.

Next steps: Testing the DME plug for Temp and TPS signals. Smoke testing again. I also want to test restoring the venturi and y-pipe to the ISV loop.

I must have misinterpreted something you said above about surging. When the TPS idle contact is closed, the DME cuts fuel to the engine somewhere in the 1200 to 1600 rpm range. (The book and the maps say 1200-1300 when warm, but actual experience suggests it's a bit higher.) So, if there is a vacuum leak that is allowing the motor to idle at, say, 1800, with the throttle closed, what happens is the rpms start to climb up until the DME cuts fuel, at which point the rpms drop, then when the rpms are low enough again, the fuel comes back on and the rpms climb again, until the fuel shuts off again, and so on and so on. That's when you see the rpms 'bouncing' back and forth between idle and 1500rpms or so. If that happens, it means the throttle idle contact is working. But it sounds like that's not your situation, so I'd double check the idle contact at the 35-pin dme connector.

I think your next steps are good ones. I can't think of anything that would cause the DME to intentionally command a 1400rpm. A bad temp sensor might cause add a few hundred rpm to the idle (if the DME thinks the motor is at sub-zero temps), but not 1400. You either have a leak somewhere, or your ISV is not working correctly (ISV/wiring to it/DME). The motor is getting extra air somewhere. My hunch is that the ISV will always let a certain amount of air in, and that blocking it off really just cuts off that expected airflow, bring the total amount available to the motor down to idle speeds -- in other words, blocking off that airflow is just masking the problem and not revealing one...

Ok, thats helpful, thanks Tom. Yeah, the idle doesn't surge, the idle is steady, just high. Im hoping to find a few minutes to try and smoke it again and test those pins this weekend.

[spacecad3t]

I know we're probably past this but for the record: from a quick look at the code, the threshold to activate decel fuel cut off looks like 360 above the threshold to turn it back on, which is 1200 for normal operating temps. So fuel should only be cut above 1560, and re-enabled at 1200. At 39C and below, the threshold is 4000rpm for re-enabling (so 4360 to activate cutoff). This logic is located at 0x0708.

Just so I can put experience to this. When I stripped the bolts on the TB, I didn't realize until I tried to start the car and it seemed like the engine was surging like crazy, but what was happening was the vacuum leak was driving the RPM up to fuel cut off and then falling back to 1200rpm and "looping" back and forth?

If I have a vacuum leak still that is causing the idle to reach 1200-1400, it only appears steady because it isn't triggering the cut off?

[johnb]

I know we're probably past this but for the record: from a quick look at the code, the threshold to activate decel fuel cut off looks like 360 above the threshold to turn it back on, which is 1200 for normal operating temps. So fuel should only be cut above 1560, and re-enabled at 1200. At 39C and below, the threshold is 4000rpm for re-enabling (so 4360 to activate cutoff). This logic is located at 0x0708.

Just so I can put experience to this. When I stripped the bolts on the TB, I didn't realize until I tried to start the car and it seemed like the engine was surging like crazy, but what was happening was the vacuum leak was driving the RPM up to fuel cut off and then falling back to 1200rpm and "looping" back and forth?

If I have a vacuum leak still that is causing the idle to reach 1200-1400, it only appears steady because it isn't triggering the cut off?

Yes that is what I was thinking of - I brought all this up the other day because I didn't realize you had that surging, so I was wondering why your fuel isn't being cut at 1400....but it looks like you have to go higher for it to activate.

Reviewing what you wrote, it's not clear if you had the TPS connected or not when you saw the surging? If it was connected, and you saw surging just from opening up the bypass screw, then that definitely sounds like what I'm talking about. So that proves the idle contact is working.

[spacecad3t]

OK, I performed all my tests this morning. Here are videos too.

First test, which I didn't record, was just restoring the venturi, y-pipe, and complete ISV loop with ISV connected. No change, idled at "1400", I used quoted here because its 1400 on the tach, but I used my timing gun for all of these tests and saw 1300 rpm on that. Take away here is no change... So to simplify it all, I took the venturi/y-pipe/ISV all back out of the loop and plugged the jboot connections, etc.

Next test was smoking this configuring. I did not find any leaks. Here is a video trying to show that smoke is exiting form the airbox intake, but I did not see even the smallest amount of smoke anywhere else in the system! In addition, to that, I pulled the jboot when I was done and smoke came bellowing out (2nd video)... This says to me, not only do I not have an obvious vac leak, the system is holding smoke! alot! Please tell me I'm wrong...

Smoke test video:


Smoke bellowing video:


OK so moving on. I pulled the DME plug and performed the TPS and DME Temp sensor tests. The TPS is working/signaling to the DME. Here are videos:

Idle test video:


WOT test video:


Finally, the DME Temp sensor at the DME Plug... Warm-ish engine(~100F) and 0.75-0.85k ohms... This seems wrong? However, it would indicate not running rich as the resistance is so low compared to "cold start resistance" which should be 3.3k ohms, per clarks.

EDIT: Maybe 0.75k Ohms is correct at 100F given 86F ~1.4 and 176F is ~0.28?

At 59 °F approximately 3.3 k-ohms
At 86 °F approximately 1.46 k-ohms
80 °C (176 °F) 280-360 ohms

DME Temp Sensor test video:


Original: This would say "hot" engine no? Thus leaning it way out? Could this be cause the issue with the ISV?

EDIT: Or is too cold still and causing enrichment?

Another test that I tried adhoc, was reconnecting the altitude sensor. Per F9Tech website, the MAF provides this and I was instructed to disconnect it.

3. Disconnect the factory altitude switch (this is no longer needed).

No change. So I left it disconnected.

This is where Im at right now:
1. ISV is unplugged and capped off.
2. Venturi deleted, jboot capped.
3. 1.75 turns out on the idle bypass screw.
4. Car starts (40F ambient), idles at 750-800.
5. Car warms up, hitting 880rpm just before "warm".
6. Car idles at 900-1000 warm to hot, stopping around 1050rpm "hot".
7. If I restore the ISV, and leave the idle bypass at 1.75 turns, I get "fuel cutoff" surging between 1200-1900rpm.
8. If I restore the ISV, and close the idle bypass completely, I get 1000rpm idle cold and 1300pm idle warm/hot.

[Tom]

Did you ever try the other throttle body? That butterfly plate looks like it might be non-factory, and you can see the screws stick out too far in the shaft. It's getting extra air somewhere... Did it ever idle correctly?

[spacecad3t]

Did you ever try the other throttle body? That butterfly plate looks like it might be non-factory, and you can see the screws stick out too far in the shaft. It's getting extra air somewhere... Did it ever idle correctly?

Interesting observation. The bolts holding this throttle body down had the same yellow paint I think Porsche used to mark the fasteners, as I see it on other parts of the car. I did try another throttle body, the one I mentioned here.

However, that one had a very sticky throttle plate/shaft and was causing huge vacuum leaks (cutoff surging) when installed on my manifold. So I went back to using my original one, with a replacement TPS and new o-ring. I did buy this last night... going to try it when it arrives.

Almost forgot... "Did it ever idle correctly?", I don't think so. When I acquired the car it was surging, minor, but when hot it would float between 1100 and target idle. That was when I started replacing vacuum lines. During which I observed that the idle bypass screw was very closed up, less than a half a turn open. That the AFM mixture was played with and that someone had pried open the AFM housing to what I assume was a "cog spring tension" adjustment.

Correct me if Im wrong, but that makes sense if a prior owner was covering up vacuum leaks? Decrease air bypass and then force richer through the AFM? The car is exempt from smog, so they wouldn't care... but also why... given vacuum leaks are typically easy to fix.

After the vac lines, sensors, DME, MAF, the idle was correctly hitting target during cold start and warm up but reaching 1100rpm when hot. But no longer surging and the car overall drives excellent, I drove it 1600 miles... My original post in this thread contains more details around what I replaced trying to reach a consistent target idle.

[spacecad3t]

I asked GPT-5... It knows the car/work already.

I am testing my DME Temp sensor at the DME plug and found: 0.75k Ohms at 100F
References for testing say these are correct values:
At 59 °F approximately 3.3 k-ohms
At 86 °F approximately 1.46 k-ohms
80 °C (176 °F) 280-360 ohms

Is my sensor reporting a correct temp value to the DME?

Short answer: it’s reading too low (too “hot”) for 100°F, so it’s not right on the factory curve. It’s close enough that the car will run, but it could absolutely contribute to lean/rough cold running.

Comparing your reading to the spec
Factory reference values you quoted (these match the usual Porsche/Bosch data):
59°F (15°C) ≈ 3.3 kΩ
86°F (30°C) ≈ 1.46 kΩ
176°F (80°C) ≈ 280–360 Ω

You measured at the DME plug:
0.75 kΩ (750 Ω) at ~100°F

Using the 86°F and 176°F points as a rough curve:
At 86°F → 1.46 kΩ
At 176°F → ~320 Ω

A typical NTC thermistor will be around 1.0–1.2 kΩ at ~100°F (≈38°C), not 0.75 kΩ. Your 750 Ω reading is more like what you’d expect around 120°F (≈50°C), i.e. it’s telling the DME the engine is significantly warmer than it really is.

So:
Expected at 100°F: ~1.1 kΩ (ballpark)
Measured at 100°F: 0.75 kΩ
Error: about 30–35% low, which is well outside a normal ±5–10% sensor tolerance.

[Tom]

I think the temp sensor is 'close enough' since the coolant is likely hotter than the outside of the motor where you are pointing the gun. I see 850 ohms which roughly equates to maybe 140F. Even if the reading were actually off by 40F, I really doubt that could cause your issue.

Kind of a hack suggestion, but I might be inclined to seal off that butterfly-to-bore gap (packing tape?) to see if that's where the air is getting by at idle. The plate looks replaced and there is some curious scratching near the bypass port -- all of which makes me suspicious of that throttle body...

[spacecad3t]

I think the temp sensor is 'close enough' since the coolant is likely hotter than the outside of the motor where you are pointing the gun. I see 850 ohms which roughly equates to maybe 140F. Even if the reading were actually off by 40F, I really doubt that could cause your issue.

Kind of a hack suggestion, but I might be inclined to seal off that butterfly-to-bore gap (packing tape?) to see if that's where the air is getting by at idle. The plate looks replaced and there is some curious scratching near the bypass port -- all of which makes me suspicious of that throttle body...

Thanks Tom. Im not sure what would work ok there either... masking tape might be hard to get right. I'll wait for the other TB I have coming and try that first. It looks to have a factory plate, they should be plated yeah? Mine isn't plated looking...