ETC Vacuum Getting "Stuck"

[CCS86]

Hey guys,

I have a customer car with a very odd issue. It is a PD blown Coyote, which I have tuned quite a few. This once has a bit of bipolar disorder, where throttle control is great, and then periodically goes to hell.

I have narrow it down to what I believe is the main observable difference: [Manifold Vacuum ETC Model] getting "stuck" low. Every car I have tuned, when air load is low, calculated MAP is low, and ETC Vac is maxed around 13.4 in-Hg. When this car is running well, it does the same.

But something happens, usually after a blast of WOT, where ETC Vac stays low.

ETC Vac Problem2.jpg


Here is a side-by-side of otherwise almost matching engine conditions:


ETC Vac Problem.jpg



Here is a comparison view of log data showing my personal car which always runs well, to this car and it running well and poorly:


Image 11.jpg



Short log attached from the before after hit example above. When he found that his bypass valve was hitting the inside of the manifold and not opening all the way, I thought for sure that was it. But after fixing that issue, this one remains.

The issue has also repeated with a Ford twin 60mm TB, and a Twin 69mm. So, that's 2 sets of TB electronics and 2 TPS sensors.

I'm down to only two thoughts: some sort of electrical / wiring issue, or a PCM glitch that I have never seen before. I would have him switch strategies to test the latter, but would feel bad to drop that much non-refundable cash on a wild guess.

Any ideas?

[CCS86]

The car has now gotten a new bypass valve, new throttle body harness, and a new battery.

Issue still exists.

[smokeshow]

Hard to say without seeing the file. It's clearly making a correction via throttle feedback and then clipping the ETC model-based vacuum at what is expected to be the new sonic flow pressure threshold with that residual throttle angle error. If your toon has the knobs available, you'd be better off installing a MAP sensor wherever possible. You really can't ask it to accurately calculate MAP with no pressure sensors and an extra pressure drop in the intake. The logic is robust to the ETC PR calculation alone, but that extra PR across the blower rotors causes issues.

[CCS86]

Hard to say without seeing the file. It's clearly making a correction via throttle feedback and then clipping the ETC model-based vacuum at what is expected to be the new sonic flow pressure threshold with that residual throttle angle error. If your toon has the knobs available, you'd be better off installing a MAP sensor wherever possible. You really can't ask it to accurately calculate MAP with no pressure sensors and an extra pressure drop in the intake. The logic is robust to the ETC PR calculation alone, but that extra PR across the blower rotors causes issues.

I would love a way to integrate a real MAP sensor (into actual PCM calculations) for these boosted Coyotes, and I tried a number of ways. None of them have worked so far. I had Eric add [ECM] 44204 - MAP Sensor: Master enable/disable for MAP Sensor to my strategy but I couldn't get any reading on that channel when applying voltage to a few different PCM pins.

That said, it is absolutely possible for the PCM to calculate MAP accurately without one. I spent many hours with an external MAP sensor connected, locking out each individual mapped point, gathering data and correcting the speed density tables. Even with the blower in between the TB and the ports, I have accurate calculated MAP in all ranges of operation. I have tuned around 20 TVS Coyotes with my same base and they all drive beautifully. Hell, Lund and others don't even touch the speed density tables and get "driveable" cars.

In this specific case, looking at this "before/after" data:

- the ETC angle error change is pretty small. The reason it changes at all, IMO, is because the bogus ETC VAC value is causing lookup from the "wrong" part of the ETC tables.

- The calculated MAP in the "stuck/after" condition is lower, air load is lower, but the ETC VAC is lower when it should be higher (or clipped to maximum).

[RobCat030]

I'm still trying to 100% understand all this myself (so apologies if this is useless), but what I notice in addition to inferred MAP being lower is that desired load and airmass vs actual have a larger difference after the pull compared to before. What I would want to figure out is if the PR/ETC Vac is causing the issue or is it a symptom of it. On my car I also have a PID called "MAP from Desired Airmass," might be worth checking out if you have it.

How much higher RPM, low load data did you have when you calculated SD? Is the 4.7inHg correct against your measured MAP?

[smokeshow]

That said, it is absolutely possible for the PCM to calculate MAP accurately without one. I spent many hours with an external MAP sensor connected, locking out each individual mapped point, gathering data and correcting the speed density tables. Even with the blower in between the TB and the ports, I have accurate calculated MAP in all ranges of operation. I have tuned around 20 TVS Coyotes with my same base and they all drive beautifully. Hell, Lund and others don't even touch the speed density tables and get "driveable" cars.

You can without a secondary pressure delta in the intake tract. With throttle near closed, blower inlet pressure is near identical to MAP because the bypass is open...so one effective pressure delta, and the throttle is well defined enough to allow MAP calculation. At WOT, the pressure delta is across the rotors. Still just a single one, easy enough to calculate MAP with mass flow. But when you transition between these two modes of operation, you get that double pressure delta that the software can't account for. The bypass begins to close and the blower begins drawing in more air than the NA engine could alone. That's what makes the estimation pick up error. You're on the right track though. The ETC tables are a method to sort of 'hack' it into behaving. If you can bury the error from that transition area into it's corresponding vacuum rows, you can minimize the problem. Using a slight restriction in the bypass actuator reference line can also help by slowing the closure of the bypass.

[CCS86]

You can without a secondary pressure delta in the intake tract. With throttle near closed, blower inlet pressure is near identical to MAP because the bypass is open...so one effective pressure delta, and the throttle is well defined enough to allow MAP calculation. At WOT, the pressure delta is across the rotors. Still just a single one, easy enough to calculate MAP with mass flow. But when you transition between these two modes of operation, you get that double pressure delta that the software can't account for. The bypass begins to close and the blower begins drawing in more air than the NA engine could alone. That's what makes the estimation pick up error. You're on the right track though. The ETC tables are a method to sort of 'hack' it into behaving. If you can bury the error from that transition area into it's corresponding vacuum rows, you can minimize the problem. Using a slight restriction in the bypass actuator reference line can also help by slowing the closure of the bypass.

Definitely valid points. I have spent countless hours tuning this exact thing. I even have an adjustable bi-directional flow controller on my vacuum line to the bypass actuator.

It took creating an incredibly complex Excel sheet for calculations, but I have amazing predictable transitions between low and high load, with the datalogs to show it. I have ported that base tune into nearly 20 other cars and they all drive beautifully as well.

This car is doing something differently. After ETC Vac gets "stuck" low, it won't idle or cruise correctly. Which, as you mentioned, are relatively easy to tune.

[CCS86]

I'm still trying to 100% understand all this myself (so apologies if this is useless), but what I notice in addition to inferred MAP being lower is that desired load and airmass vs actual have a larger difference after the pull compared to before. What I would want to figure out is if the PR/ETC Vac is causing the issue or is it a symptom of it. On my car I also have a PID called "MAP from Desired Airmass," might be worth checking out if you have it.

How much higher RPM, low load data did you have when you calculated SD? Is the 4.7inHg correct against your measured MAP?

I remapped all areas of the SD tables. That isn't the issue here.

The reason for the desired vs actual difference is because the falsely low ETC Vac value is pushing the ETC table into the wrong range, so the TB is not being opened far enough for the requested torque.

[smokeshow]

Definitely valid points. I have spent countless hours tuning this exact thing. I even have an adjustable bi-directional flow controller on my vacuum line to the bypass actuator.

It took creating an incredibly complex Excel sheet for calculations, but I have amazing predictable transitions between low and high load, with the datalogs to show it. I have ported that base tune into nearly 20 other cars and they all drive beautifully as well.

This car is doing something differently. After ETC Vac gets "stuck" low, it won't idle or cruise correctly. Which, as you mentioned, are relatively easy to tune.

Damn. Well if you've given that approach its due diligence...it starts to smell like a rationality diagnostic. 'If your measurement is not believable, stop updating the variable so it doesn't affect anything else'...essentially. Can't say with any certainty though, especially not without a file. I've never actually seen you post an actual toon file

[CCS86]

Damn. Well if you've given that approach its due diligence...it starts to smell like a rationality diagnostic. 'If your measurement is not believable, stop updating the variable so it doesn't affect anything else'...essentially. Can't say with any certainty though, especially not without a file. I've never actually seen you post an actual toon file

I don't post my tunes because of how much time I have spent remapping things like the SD tables, inferred rail pressure, ETC tables, cam angles, etc. To hook up external gauges, gather that much data, calculate, iterate, rent dyno time... takes an incredible amount of time. To copy those tables takes 5 seconds.

I can't envision anything "in the tune", besides corruption, that could even cause this.

[smokeshow]

I don't post my tunes because of how much time I have spent remapping things like the SD tables, inferred rail pressure, ETC tables, cam angles, etc. To hook up external gauges, gather that much data, calculate, iterate, rent dyno time... takes an incredible amount of time. To copy those tables takes 5 seconds.

I can't envision anything "in the tune", besides corruption, that could even cause this.

Oh come on now. No new ground is broken lol. I give out all my stuff freely - if you can understand how I got there, you deserve to see it. If not, copy it all...still going to be riding the tooner struggle bus.