There are also 'desired airmass' switches currently set like this. I'm not sure if the first one being disabled overrides the idle switch or not:
Idle-airmass.png
There are also 'desired airmass' switches currently set like this. I'm not sure if the first one being disabled overrides the idle switch or not:
Idle-airmass.png
So something I noticed while comparing your throttle body tables to stock tables, the areas I've highlighted match and correspond to one another, and those areas of your tables do not match like this. They look like this in the table areas that are pretty much only at idle, maybe if you can get yours too look something like this it may reduce the error? I could be way off base here but its just an observation I made.Originally Posted by CCS86
My tables are slightly modified versions of the OEM tables for the Twin 65mm Ford Racing throttle body.
I have tried significant changes to both TB tables, and changes to only Effective Area (not Predicted Angle). Neither have shown any change to ETC Angle Error.
That's because you are doing the math wrong!! . Get ready, I'm about to explain throttle error.
In reference to this screen shot you posted.
You have a Throttle Angle of .93* with an ETC Area of .054. Your ETC area needs to be decreased because an ETC Area of .054 is a perfect match for your Throttle Body Desired of .97.
You actually need a throttle area of .0525, but your calibration is incorrect because .0525 is calling for .45*.
Throttle Angle .93 - Throttle Angle Error .48 = .45 degrees
And yes, driver demand can cause torque errors.
That calibration is for a throttle body with a perfect circle. We're using twin 67mm oval so the flow at low angles will vary because of the wideness. It is possible to calibrate one using the same technique where the axis match the cells but you have to be flowing around 2000CFM hitting an ETC area over 50^2
Yes you're correct, I overlooked that very important detailWe're using twin 67mm oval so
I'm a little lost in your explanation, so correct me if I'm following you wrong, you're saying because the Throttle Desired Angle and ETC Effective Throttle Area in the log match up to his ETC Predicted Throttle Angle table, that the error is therefore in his ETC Effective Area table. Now in his ETC Effective Area table at 14.1 inHg and 1.0 Throttle Angle I am seeing the number 0.057, are you saying that this number needs to be decreased? And how do you quantify the change needed?
I think you are misunderstanding the data. You seem to be keying in on the difference between desired vs actual ETC angle (.93* vs .97*), thinking that the tables are aligned for the desired of .97*, when they should be aligned to the actual of .93*. I think this is wrong for two reasons:
-You are looking at a difference in angle of 0.04*, which is an order of magnitude less than the reported ETC angle error of 0.48*
-The two PIDs refresh at different times. You can't look at a specific time and assume that different values have conclusive meaning. They could actually be overlapping curves, that are refreshing at different times, making them appear out of step.
You claim I am doing the math wrong. Please, tell me what is wrong in my math below. This is straightforward bilinear interpolation:
TB-Angle-calcs.png
On the driver demand claim, prove it. Post a log during a certain run condition, with no IPC Torque errors, then change only DD and recreate the log.
Here is the difference table of my DD vs stock. My IPC Torque errors are 0.0 everywhere.
DD-diff.png
Last edited by CCS86; 08-01-2018 at 07:06 AM.
The Twin 67mm is not an oval. It's just like the Twin 65mm I am using: two round bores.
The reason the tables have values the decrease in predicated angle, and increase in effective area as ETC vacuum goes up, is more likely because of some non linearity in flow rate as pressure drop across the TB changes.
Murfie can you link that patent?
I did another test today, where I made some big changes to the TQ and Inverse tables of MP0.
I created a math parameter to subtract ETC TQ from Brake TQ and that seemed to trend very closely with ETC Angle Error.
After this change, my TQ DIFF quantity flipped in sign and was reduced, but the ETC Error stayed positive and similar:
ETC-TQ-change.png
I tuned on my 14 GT500 almost all day yesterday, got it driving pretty good, Whipple big elliptical tb, and lumpy cams, 4.5L Whipple blower, at one time, I adjusted only the DD table, had lotsa IPC errors, I would have never thought it, but it did. I can't see how this can happen, maybe there's some code to where these must match up close, changing them makes the errors come.
On tb tuning, the way I was taught is different than stock tb tables. On these, you put the same predicted values from top to bottom, at all vacuum levels, then take the values from the rows and make them the column values in the effective area table. I kept tuning tb tables, along side torque tables, & MAF trims & lotsa errors went away, drivability got way better. These three things seem to make the most difference in tuning these type of ecu's than anything. I am attaching a screenshot of my tb tables.
Clipboard01.jpg
Do you have back to back logs where you drove a custom DD table, then reverted to stock?
Checking in to see if you ever got this figured out, really interested to hear your findings if you have or when you do.
That's because Lund is a "tooner" himself. NOT a calibrator. You can not calibrate these cars on a Dyno-Jet. You cannot calibrate these cars without the proper instrumentation of the vehicle, an AC motored dyno and a solid understanding of the way a torque based system operates. At the OE level we don't drive on the street taking data logs. Ford is using so much inferred logic that without an actual measurement of the MAP, converter temps, and input shaft torque it's going to have issues.
Anybody who has claimed to have an understanding of ETC Angle Error, I have put the claims to the test and failed to produce results.
Yup. I found this out firsthand.
I am actually wondering about some of the big guns too. Both Roush and FRPP supercharged calibrations have a lot of unmodified GT calibration left behind.
- FRPP has completely stock GT speed density tables, and Roush 575 only has a few modifications to MP9 and OP.
- FRPP has heavily reworked torque tables, but leaves the inverse tables untouched
- Roush 575 uses essentially stock torque tables, only changing the last two rows to allow for higher load
Why aren't companies with the resources to do so, actually calibrating their kits?
For Whipple its a lack of money. development of this kind takes several million dollars and that's if you own the equipment. If not you're going to pay for the time. Typically this time isn't on an AVL dyno. I'm not convinced that we always have access to everything we need with HP.
Maybe that's the case for a full OEM level calibration, but that isn't what is required here.
I am one guy, new to tuning Copperhead, using only logged PIDs and mechanical boost/vac, and I have been able to retune all the speed density tables, for all MPs (that I use). Now my calculated MAP matches my gauge very closely all the time on my PD blower combo. It was WAY off before.
Why are FRPP and Roush too lazy to do anything about this? Calculated MAP being off throws off fueling, then that gets "fixed" in the MAF curve or injector settings, which is really bad practice.
Tuning to pass emissions may very well take several mil, but 99.9% of HPT customers are not concerned with that at all. Tuning for power/drivability is way easier, & cheaper, if you know how.
I'm digging this thread up from the dead.
It's really a shame that more isn't understood about the Ford ETC system, so hopefully more people dig in to test more theories and post their findings.
I went a different direction in how I tune the TB model, but there is still room to improve it and minimize these reported calculation errors. There are still ghosts in the machine. I have twin 60mm and twin 65mm models dialed in for my personal car. But when I use them on another car with a nearly identical build, they always need fine tuning. Some cars behave very differently with the same hardware and same calibration.
Here is an example from a car I am remote tuning right now that shines some light on one of the major sources of variability: ETC Vac. Being a primary driver of both TB tables, this is a very important value, that we can't directly control (to my knowledge). Here are two instants from the same log, where at essentially the same RPM, airload, ETC angle, and even cam positions; l I am seeing huge differences in reported ETC Vac.
You can see that even though 'calculated MAP' is slightly higher, it is reporting more than double the 'ETC Vac'. This variation makes throttle body tuning very difficult, since ETC Vac is a primary driver.
ETC vac.jpg
Reply With Quote