Soler 107mm TB idle tuning

[smokeshow]

More kudos to being publicly involved in the product improvement.

Mike, correct me if I am wrong, but it sounds like your approach here is to create a throttle body bore and blade that closely matches the behavior of the smaller production GM units at low throttle area where the Cd is most critical. I'd say that's a pretty cool way to solve the problem, and it sounds like you've succeeded, based on your internal testing. Could the outlier you found and the throttles that your customers here have worked with be just production variability? If you can take a poor performing throttle back as a return from a customer and send them one in its place without changing the product itself, it isn't the product that has an issue...rather variability in the manufacturing. Food for thought I suppose.

[Mike@SolerEngr]

More kudos to being publicly involved in the product improvement.

Mike, correct me if I am wrong, but it sounds like your approach here is to create a throttle body bore and blade that closely matches the behavior of the smaller production GM units at low throttle area where the Cd is most critical. I'd say that's a pretty cool way to solve the problem, and it sounds like you've succeeded, based on your internal testing. Could the outlier you found and the throttles that your customers here have worked with be just production variability? If you can take a poor performing throttle back as a return from a customer and send them one in its place without changing the product itself, it isn't the product that has an issue...rather variability in the manufacturing. Food for thought I suppose.

Exactly, that is the approach. Variability is also one of the lines of troubleshooting we are pursuing, so far they don't show in the company car. But heavily built cars idle very differently. More than anything we need logs so we can see the results, all we know is that there's negative idle timing, but nothing we can anchor to, like a fixed setup, with a fixed tune, where the only variable is the TB.

Update 10-01-2021, see Post #62. Do the on road idle learn procedure before tuning.

[Mike@SolerEngr]

Hi Mike

Good to see you jumped in this chat, feel free to continue speaking to me via our existing email thread or DM here or even in this thread.

I was actually just discussing with the shop guys in Texas if we should ask you about changing the gear by a tooth or two under the cover. The same tune has resulted different timing results with all 4 different throttle bodies so the issue isn't in the tuning. The one that I have running in the positive has a strange dip when coming to a stop that I am still working through to see if it is a TB thing or a Virtual Torque thing. The guys I work with are top notch mechanically and I can instruct them if there are any physical changes you want to make, or we can just send a unit back for replacement and see what it takes to get a positive idle without frankensteining the torque tables.

My one client really wants to keep the 107. Let me know what you can do to help us on this car. The #4 TB was the best one that would idle at -5 where the others would idle from -12 to -20 on the same tune. We are willing to keep working with you if you want to get involved. I was literally going to email you today about the next steps. Unfortunately, 2 of the 3 clients ran out of patience and do not want to continue. But I still have the one hardcore guy that is ready to make it work.

Thanks,
Jason

That's great, Jason. I'll get back to you by email. I have something here for you guys to try.

Quick question for all. At the most basic level; what is the root command the ECM gives the throttle body when idling? I.e. desired TP%, or desired MAF, or perhaps desired ETC motor ducty cycle, IDK. There are different ways to get the same result, but one of them should be the driver and the others depend on it; correct?

Here's why we ask:

So far, we know that the idle target rpm's are being achieved, but under different conditions, with higher airflow resulting in negative timing. It is as if the ECM strategy is fixing TP% and if MAF is not on target it then uses timing to control rpm's.

We have a log on a Z06 M7 with cams and 2300SC, our TB1 idled at ~20 g/s @ 20%TP with negative timing, TB2 idled at ~12g/s @ 12%TP positive timing. Same TB geometries; same hot idle target rpms, what could make it desire such different airflow and TP%'s?

[TriPinTaZ]

That's great, Jason. I'll get back to you by email. I have something here for you guys to try.

Quick question for all. At the most basic level; what is the root command the ECM gives the throttle body when idling? I.e. desired TP%, or desired MAF, or perhaps desired ETC motor ducty cycle, IDK. There are different ways to get the same result, but one of them should be the driver and the others depend on it; correct?

Here's why we ask:

So far, we know that the idle target rpm's are being achieved, but under different conditions, with higher airflow resulting in negative timing. It is as if the ECM strategy is fixing TP% and if MAF is not on target it then uses timing to control rpm's.

We have a log on a Z06 M7 with cams and 2300SC, our TB1 idled at ~20 g/s @ 20%TP with negative timing, TB2 idled at ~12g/s @ 12%TP positive timing. Same TB geometries; same hot idle target rpms, what could make it desire such different airflow and TP%'s?

In a very basic description, the ECU uses predicted torque and immediate torque to determine what proportion of throttle and timing should be used to maintain the desired idle torque. Predicted torque refers to Throttle Blade Movement to meet a torque value. Immediate torque refers to spark advance to meet a torque value. In the case of idle, the ECU uses Torque Management Advance to pull spark away from the value in the spark tables to meet the demand. As for the torque calculation, this is determined by the ECU in a complex mathematical equation but we can influence this directly with airflow (MAF or VVE) or by manipulating the Virtual Torque and Coefficients to tell the ECU it is making more or less torque at a given Airmass or MAP vs RPM value.

Where the negative idle comes into play is when the ECU cannot seem to reasonably close the throttle blade enough to idle in a low enough airmass where the torque values line up to play nice. Or the torque model is way off causing the throttle blade to be open more than it should. This alone causes crazy negative idle spark because the ECU says the TB is open 12% and the airmass is .500, well this is too much torque because the VT tables say the car is making 125lb of torque but the idle mode is asking for 15lb of torque. Then is starts pulling timing to lower the torque. Why does the throttle stay open too much? I don't know. That is not always in the tune. It can be, but not always the case. Without changing the tune on one of the ZL1's, the different 107's all idled with different TB % opening. There is either some tables we cannot see in HPTuners or something else is at play.

In the past I've actually taken the gears and TPS motor assembly from the original cars throttle body and put it on a ported TB to fix a negative timing issue. This was probably 4-5 years ago when ported 87's were the only thing out. I called up the well known porter and said hey I have a problem, he suggested to swap the parts over and try again and if that didn't fix it he would exchange. I swapped the parts from the original TB over and it worked great after that. Still no rhyme or reason. I sort of suspect that the motor and gears and TPS electronics are synchronized to the actual physical TB for proper operation and if they don't match up to some level of tolerance, everything just goes wrong. But I have no proof or data to demonstrate this suspicion.

[smokeshow]

Quick question for all. At the most basic level; what is the root command the ECM gives the throttle body when idling? I.e. desired TP%, or desired MAF, or perhaps desired ETC motor ducty cycle, IDK. There are different ways to get the same result, but one of them should be the driver and the others depend on it; correct?

Yes. lol

The answer is all of those things. In short, engine indicated torque is summed with desired idle torque reserve and through a lot of nasty equations is resolved into a desired throttle area. The effective area and the throttle airflow are linearly proportional so long as the pressure ratio is below 0.528, at which point the flow is choked. This of course requires the pressure after the throttle to be accurate - a supercharged application without an SCIAP will not be able to accurately calculate the pressure ratio, so that is a must.

So long as the effective area of your throttle closely matches the stock one from full closed/locked position to the position at the highest possible idle torque request, then in theory there should be no difference in the idle behavior. Curious, how have you been validating that behavior so far? Flow bench, etc?

I can break any of that down in further detail if needed.

[Mike@SolerEngr]

Yes. lol

The answer is all of those things. In short, engine indicated torque is summed with desired idle torque reserve and through a lot of nasty equations is resolved into a desired throttle area. The effective area and the throttle airflow are linearly proportional so long as the pressure ratio is below 0.528, at which point the flow is choked. This of course requires the pressure after the throttle to be accurate - a supercharged application without an SCIAP will not be able to accurately calculate the pressure ratio, so that is a must.

So long as the effective area of your throttle closely matches the stock one from full closed/locked position to the position at the highest possible idle torque request, then in theory there should be no difference in the idle behavior. Curious, how have you been validating that behavior so far? Flow bench, etc?

I can break any of that down in further detail if needed.

Yes, on the bench we've got the same slope of the Area vs. TP% curve (linear at low throttle). We can adjust the Cd so we get the stock airflow line at .528 pressure ratio. Our next shot will be reducing Cd and see what happens, although we cannot recreate a test vacuum high enough to choke it. We'll have to tell from a much different pressure ratio. The engine will have the last word.

So if the output from the tq model is area, then it would be looking for a TP% for which it thinks it will get that area; correct?

Thanks,

[smokeshow]

Yes, on the bench we've got the same slope of the Area vs. TP% curve (linear at low throttle). We can adjust the Cd so we get the stock airflow line at .528 pressure ratio. Our next shot will be reducing Cd and see what happens, although we cannot recreate a test vacuum high enough to choke it. We'll have to tell from a much different pressure ratio. The engine will have the last word.

So if the output from the tq model is area, then it would be looking for a TP% for which it thinks it will get that area; correct?

Thanks,

Not sure I follow. What do you have on the bench? Without test equipment to create sonic flow conditions, how are you validating the Cd changes to the bore? CFD goes a long way but it would still need to be proved out in real testing. And anything above 0.528 starts to become nonlinear, so it isn't really viable as a test condition.

Once desired throttle area is calculated, the lookup tables determine what 0-100% position will yield that area. The measured airflow is then back-calculated via the airmass torque model to estimate the effective throttle area for short/long corrections and diagnostics.

[Mike@SolerEngr]

Not sure I follow. What do you have on the bench? Without test equipment to create sonic flow conditions, how are you validating the Cd changes to the bore? CFD goes a long way but it would still need to be proved out in real testing. And anything above 0.528 starts to become nonlinear, so it isn't really viable as a test condition.

Once desired throttle area is calculated, the lookup tables determine what 0-100% position will yield that area. The measured airflow is then back-calculated via the airmass torque model to estimate the effective throttle area for short/long corrections and diagnostics.

28 Inch H2O max vacuum. Respectfully disagree about the non-linearity. Flow vs. Area is still linear above .528, the Cd doesn't depend on pressure ratio, only on geometry. Or at the very least, without the need of digging any deeper, it is theoretically linear enough for the purpose of determining the Cd.

But then again, as you well said, real test in the actual engine showing the symptoms will have the last word.

[smokeshow]

28 Inch H2O max vacuum. Respectfully disagree about the non-linearity. Flow vs. Area is still linear above .528, the Cd doesn't depend on pressure ratio, only on geometry. Or at the very least, without the need of digging any deeper, it is theoretically linear enough for the purpose of determining the Cd.

But then again, as you well said, real test in the actual engine showing the symptoms will have the last word.

Unfortunately that one isn't subject to opinion. It is most certainly nonlinear as your MAP rises and pressure ratio exceeds that critical 0.528 value. This is the core physics of how all manufacturers control airflow in throttled engines. There's a figure in the tech section of your website that shows this deviation from linearity as throttle position rises. If I were keen to dial this in, I'd want to see the flow vs throttle area between the two match exactly until just beyond the maximum expected idle throttle position...then let her rip

[Mike@SolerEngr]

Unfortunately that one isn't subject to opinion. It is most certainly nonlinear as your MAP rises and pressure ratio exceeds that critical 0.528 value. This is the core physics of how all manufacturers control airflow in throttled engines. There's a figure in the tech section of your website that shows this deviation from linearity as throttle position rises. If I were keen to dial this in, I'd want to see the flow vs throttle area between the two match exactly until just beyond the maximum expected idle throttle position...then let her rip

It is not a matter of opinion, the Cd is independent of the pressure ratio, which is what allows us to test above critical conditions. You said that the test would not be accurate because of that. The valve flow characteristic curve is a totally different subject and yes, it is non-linear as shown in our site.

What we are saying is that for any given (fixed) throttle position it doesn't matter what the pressure ratio is, you can test above critical conditions and still get the right Cd, because at that position (that pressure ratio) MAF and Area are directly proportional. So we can find the effective area and say that will also be the same effective area when under critical conditions or under.

What we cannot do, is make an airflow curve like the one you could get on the engine. But the Cd's, yes, those would be fine. Something has to be clear, we do not recreate engine conditions at the bench, we test under other conditions and compare OEM's and ours under those same conditions and they have to match, if they match at the bench, they should match in the engine. That is the idea.

The test method is still valid, even above .528; makes sense?

Oh, very important. The Cd in the idle region is much lower than the Cd as it opens. We are talking about the idle Cd only, not the spectrum of Cd's that happen after that, we have no issues there. This is about negative timing at idle.

[smokeshow]

The valve flow characteristic curve is a totally different subject and yes, it is non-linear as shown in our site.

This is the subject I am talking about lol. The linear or near-linear region is where a stock engine idles (there is a correction factor in the algorithm to attempt to correct for slightly higher than 0.528 ratios). In order for the flow to be accurate, the curves would need to match at idle throttle areas. But this data cannot be gathered at 28in H2O / 6.9kpa. The automakers don't even attempt closed loop control of the throttle with such a large pressure ratio. Idle needs predictable control though, so that's why the linearity at low position is necessary.

[Mike@SolerEngr]

This is the subject I am talking about lol. The linear or near-linear region is where a stock engine idles (there is a correction factor in the algorithm to attempt to correct for slightly higher than 0.528 ratios). In order for the flow to be accurate, the curves would need to match at idle throttle areas. But this data cannot be gathered at 28in H2O / 6.9kpa. The automakers don't even attempt closed loop control of the throttle with such a large pressure ratio. Idle needs predictable control though, so that's why the linearity at low position is necessary.

Agreed, now we are talking about the same thing...and "this data" the airflow cannot be gathered at low vacuum and be accurate, now, "this data" the flow area and Cd (effective flow area) can be gathered at low vacuum and made the same as the OEM, so when in an engine under the same operational vacuum the flows are the same. Agreed?

[smokeshow]

Agreed, now we are talking about the same thing...and "this data" the airflow cannot be gathered at low vacuum and be accurate, now, "this data" the flow area and Cd (effective flow area) can be gathered at low vacuum and made the same as the OEM, so when in an engine under the same operational vacuum the flows are the same. Agreed?

Agreed. We're on the same page now. I will be interested to see the results of the reduction in Cd at low throttle positions. It will certainly hurt your advertised mass flow rate transition of the throttle, but I don't think that's on you to fix. Lots of calibration definitions which can adjust that are not defined in HPT.

[Mike@SolerEngr]

Agreed. We're on the same page now. I will be interested to see the results of the reduction in Cd at low throttle positions. It will certainly hurt your advertised mass flow rate transition of the throttle, but I don't think that's on you to fix. Lots of calibration definitions which can adjust that are not defined in HPT.

It will hurt the transition, although we curved the transition area a bit to help, the knee of the curve wouldn't be as wide as we'd like but not as sharp as the stock one, we think it is a good compromise.

Thanks,

[Mike@SolerEngr]

Are the Soler's hand ported or is this a CNC operation?

Fully CNC machined, no cast surfaces, no manual grinding. This 103 is not ported, it is made from scratch. The 87's and 95's we port are also CNC machined.

[sevinn]

Agreed. We're on the same page now. I will be interested to see the results of the reduction in Cd at low throttle positions. It will certainly hurt your advertised mass flow rate transition of the throttle, but I don't think that's on you to fix. Lots of calibration definitions which can adjust that are not defined in HPT.

I hate that there's so much useful stuff undefined.

[lt1z350]

Hey Mike been a while I wish you told me going from the 100mm setups to a 107mm. I ran Mikes stright bore ported 95mm/100mm tb on my a10 zl1 kinda a one off I beleive and worked great on my car both with stock cam and when I did the heads and cam on it. I had only swapped to a granetelli 103mm as they had a huge sale and I got it shipped for around 400 bucks. I wanted to see if giving up performance to the 100mm over all probably not enough to bother spending more money really but too close to say.
To me compared to the typical Katech it does come with a cut shaft and its more of a straight bore style cut over Katech or MW 103mm units. I have not seen one of Mikes new 107mm to compare to that but the granetelli does by looks flow more then other 103mm. When I went from Mikes 100mm stright bore to the 103mm I had to change nothing on my tune to keep the same 16-20 degrees of timing I had at idle this is a heads cam car with an R&D ported supercharger from Jokerz. I think a lot of it is how the tune is also setup on how much changes when you change the size of the opening for air to enter the engine.
Why I say this is have any of you tuned for a 112mm throttle body yet? I was told by others running it that it is a nightmare vrs a 103mm. A car Vengance tuned I was told even had shifting issues on the a10 when they swapped over to the larger throttle body. I had yet to experience it but was told they can cause all kinds of issues. So one of the cars I tune she told me putting one on her 2650 I was freaked...lol The buddy doing all the work called me after he did it to tell me zero changed on the car. Idle didnt change any for timing and car drives great still. So like my own with a pretty radical change on the opening to intake nothing changed in the logs to show a mechanical change was made to either car.
I will say both cars run two totally different cams hers is on a very old design with a ton of overlap for a PD blower car vrs mine with 0 degrees overlap. Both idle at and around 0 to 30 lbs of torque on the log and both idle with around 20 degrees of timing. I am using very similar tunes in both of these cars and tweak by the logs I get.
So how much of this is tune related vrs a throttle body issue?? Im not saying anyone is doing anything wrong just that the way it is setup on the current tune its more sensitive to a change in inlet flow. I look at how much is the blade open vrs how much timing and how much torque it idles with. I do not change a lot of things other tuners are changing in idle tables. I actually go down on the two that most go way up on them. My cars idle excellent and are not affected by change it seems. My way of thinking is I like to see almost no blade opening so to achieve idle it has to add a lot of timing. My cars with cams idle at 675-725 I also dont put 800 or 850 in them like I have seen a lot of other tunes I have gone in and fixed. If you can get a car to idle near 0 delivered torque it has no choice but to add alot of timing to make the car idle correctly. Also how are the fuel trims on these cars?? VVe tuned well? I tune my vve at the same time using maf using a blended vve histogram. So tune the maf really good I like to see -3 to -5 I never like to see plus or minus 5 I dont like it adding fuel. Then I use the table to tune the vve all in real time and my values on all my logs are with in 5 percent of what it tells me to put into it vrs what Im using when I re check things later on down the road even if not tuning anything just looking to see whats changed. Seems the closer the tune is to being correct the more the ecm can adjust correctly to certain changes just my thought on some of this as once I figured out the low engine timing mess so much of it started to stay in line with other mods and airflow changes and less and less I had to tune anything with adding changes to my car.
I am on a very radical stock ported blower that currently in Tennessee at 3800 da is faster then my last track day at -100 in Orlando Florida on a cool morning at Cadillac attack. i went 9.30@150 that morning on a very soft second gear launch and current dragy times are crushing what I did there so my setup is pretty radical for a stock ported blower and drives better then stock and gets better mpgs actually. Just about everything I learned in tuning came from all the guys on this forum lots of reading and then trial and error. Its helped me continue to be one of the fastest stock blower cars . Sorry a little off topic Im just saying work on the basics and more comes together when a change is made to the car and not to have negative affects to it.
About the only thing you cannot tune around is a major mechanical issue IE vacuum leak or bad sensor reading due to a leak and I have found more and more a lot of shops working on cars I tune cant even tighten a clamp correctly or hook up a catch can the right way and not send fuel trims to the moon. So alot of this might not even be your tune or how you do something but a vacuum leak you dont even know is there and you are beating your head on a wall to try to fix something I guarantee you will not as I have tried to tune around my own vacuum leaks and until I did a smoke test did 50 tunes trying to get a good idle that never happened. Not until I fixed a very small leak I only found due to smoke testing it. So keep that in mind here too guys if your not hands on with these cars.
Mike cannot adjust his throttle body to a car that has issues. I will say if someone wanted to send me one to bolt on to see how it affects my own im all in to prove this is probably not a throttle body issue but other problem.
Mike if want to see some pictures of this granetelli for reference shoot me an email and ill send some pictures. I am always about more and if a 112 could even fit a stock blower Id put it on the car...lol That would be a lot of welding though.

[TriPinTaZ]

It is not a tune issue or a vacuum leak. I've demonstrated the variance between multiple 107's on the same car using the same tune. Negative timing at idle CAN BE a tuning issue, but that is not the issue here as clearly documented in my posts. I explained this in explicit detail.

[TriPinTaZ]

Update: We got one idling in the positive without needing to mangle the Virtual Torque tables. This 107 is idling just fine on the same tune as the ported 95mm was. Stays around +4 to +9 degrees of timing which is perfectly fine for this setup. The best part is that in drive, foot on brake, 0mph it is staying around 8 degrees. The other TB's had this area the worst and would always creep negative even with a Frankenstein virtual torque table.

[radz28]

Was this a new TB from Soler or did you get one of the existing ones cooperating?