Closed loop?

Kdw1403 · 08-30-2021, 10:05 PM

B/SA. Lt1. What should I set my correction limits to. Been running open loop in alpha N. Just kicking around idea of putting it in closed loop. Any advice. I Know nothing about closed loop.

KRatcliff · 08-30-2021, 10:29 PM

I wouldn't run closed loop on a bet. I prefer Alpha N/Open loop by far.

Glenn Briglio · 08-30-2021, 11:55 PM

Quote:

Originally Posted by KRatcliff

I wouldn't run closed loop on a bet. I prefer Alpha N/Open loop by far.

Why not?

KRatcliff · 08-31-2021, 06:37 AM

There are multiple reasons why I wouldn't run closed loop. The ECU can calculate very quickly, but the information coming into the ECU isn't reliable nor can the ECU read the information and make the necessary change to improve the performance within the short time of the pass in my opinion.

Our passes are only a few seconds and they are even shorter when you consider the gear changes. The ECU would have to make these adjustments reading off a single O2 sensor that isn't exactly in a sealed environment and it is only reading on bank of cylinders for an average. Plus the O2 sensors don't exactly like the fuel that we run.

When you consider these factors the ECU would be behind the curve in reading the data, calculating the change, and then effecting said change. I tune our cars in Alpha N/Open Loop in 10 different sections (for a 3 speed). The 2 Step (this would be chaos for closed loop) and 3 sections for each gear for a total of 10.

My fuel map leads in at each section similar to how we all use timing advance. There is also a cumulative affect that needs to be considered. None of this is possible with closed loop in our applications in my opinion. If closed loop was an advantage down track the Pro Stockers would run it. Closed loop is best suited for street cars.

Jeff Niceswanger · 08-31-2021, 08:45 AM

Quote:

Originally Posted by KRatcliff

There are multiple reasons why I wouldn't run closed loop. The ECU can calculate very quickly, but the information coming into the ECU isn't reliable nor can the ECU read the information and make the necessary change to improve the performance within the short time of the pass in my opinion.

Our passes are only a few seconds and they are even shorter when you consider the gear changes. The ECU would have to make these adjustments reading off a single O2 sensor that isn't exactly in a sealed environment and it is only reading on bank of cylinders for an average. Plus the O2 sensors don't exactly like the fuel that we run.

When you consider these factors the ECU would be behind the curve in reading the data, calculating the change, and then effecting said change. I tune our cars in Alpha N/Open Loop in 10 different sections (for a 3 speed). The 2 Step (this would be chaos for closed loop) and 3 sections for each gear for a total of 10.

My fuel map leads in at each section similar to how we all use timing advance. There is also a cumulative affect that needs to be considered. None of this is possible with closed loop in our applications in my opinion. If closed loop was an advantage down track the Pro Stockers would run it. Closed loop is best suited for street cars.

Hi Kyle !

Kyle and I disagree on this one. It's like Chevy and Fords. There are two distinct different theories. I do not tune in Alfa, as its old school. Very similar to a Carburetor. It repeats and does the same exact thing every run. And many people prefer that approach. When your running a carb, you have the availability of rods and jets to stay on target. Air goes from 1000 feet on Thursday to 3000 feet on Sunday, you can change rods. The Alfa on open loop is locked into what ever tune is in it.. It just repeats.
I use VE, usually with a Alfa idle circuit. All the low RPM #'s are in Alfa with a separate Alfa Map,. But as soon as you start getting into the higher RPM's it switches into VE. VE is just a glorified Speed Density Program. On some tunes I stay with Speed Density, as I already have pre made tunes that are really close to the customers combos (in SD). Its just easier and quicker to use one of them. I've seen no advantage ET wise in either of the two.
The trick is to get the tunes to where you can feel confident you need no Closed Loop Compensation. That includes burnout, and slightly after (milliseconds) you go full throttle after launch.
That way, you can set your Closed Loop Compensations very low, like 3 or 4 percent, and know that the tunes will stay dead nuts during the entire run. Even IF, the air changes. If you look at your Data after the run and your car was using more than that 3 or 4 percent of adjustments, then your tune is not correct. You will know that as your Learned Compensation is 4 percent in places and your Target will be slightly off. Reason being that 4 % was not enough and the CL "Could not get there" . There are dis advantages to having CL active. IF the 02 fails, and they do occasionally, the 02 always fails lean. Meaning the Holley will think it's in a lean condition and puts all what's available in the tune. In this case all 4 %. But if your up high, like 50%, the car will instantly go full rich, die or be impossible to keep running. If it does run, it will be horrible. Its not a bad idea to change 02's occasionally if your depending on your 02 ....
Hope this helps...

Robin Lawrence · 08-31-2021, 09:05 AM

Best answer is "it depends". I have many cars running in closed loop that are very consistent. It also depends on the ECU. We have been told that ours corrects faster than many others. Our hardware and software guys say it also depends on how the code it written in addition to the speed of the ECU.

Alpha N is great on applications where you have a very large cam. Where there is little vacuum at idle and no resolution. Speed Density allows for baro changes. If running Alpha N you should have a baro offset table.

I run my stocker with a Combo program. It is Alpha N up t0 2000 RPM and 20% Throttle Position. After that it is in speed density. I typically see less than 2% correction in closed loop.

To answer the OP's question, as far as setting limits in closed loop. I would suggest 10% additive or + and limit the negative to 5%. As 02 sensors degrade from race fuel the 02 will tend to show lean thus adding fuel. I recommend installing new 02 sensors annually.

The ability to correct with little latency will depend on your combination. On dyno we can see very fast response at sweep rates 300 to 400 rpm per second. Most sporstman cars engines will accelerate in that range in higher gears making closed loop function very well. In Pro Stock the engine acceleration rates can be upwards of 2000 RPM in the lower gears. I did not recommend that they use closed loop. I had several teams that enabled closed loop only in 4th and 5th gear.

I don't enable closed loop in my Mustang until 6000 RPM. With almost no collector after the 02 sensor it doesn't give good information at low RPM.

Of course this is my opinion. Results may vary.

Robin

KRatcliff · 08-31-2021, 09:46 AM

Quote:

Originally Posted by Jeff Niceswanger

Hi Kyle !

Kyle and I disagree on this one. It's like Chevy and Fords. There are two distinct different theories. I do not tune in Alfa, as its old school. Very similar to a Carburetor. It repeats and does the same exact thing every run. And many people prefer that approach. When your running a carb, you have the availability of rods and jets to stay on target. Air goes from 1000 feet on Thursday to 3000 feet on Sunday, you can change rods. The Alfa on open loop is locked into what ever tune is in it.. It just repeats.
I use VE, usually with a Alfa idle circuit. All the low RPM #'s are in Alfa with a separate Alfa Map,. But as soon as you start getting into the higher RPM's it switches into VE. VE is just a glorified Speed Density Program. On some tunes I stay with Speed Density, as I already have pre made tunes that are really close to the customers combos (in SD). Its just easier and quicker to use one of them. I've seen no advantage ET wise in either of the two.
The trick is to get the tunes to where you can feel confident you need no Closed Loop Compensation. That includes burnout, and slightly after (milliseconds) you go full throttle after launch.
That way, you can set your Closed Loop Compensations very low, like 3 or 4 percent, and know that the tunes will stay dead nuts during the entire run. Even IF, the air changes. If you look at your Data after the run and your car was using more than that 3 or 4 percent of adjustments, then your tune is not correct. You will know that as your Learned Compensation is 4 percent in places and your Target will be slightly off. Reason being that 4 % was not enough and the CL "Could not get there" . There are dis advantages to having CL active. IF the 02 fails, and they do occasionally, the 02 always fails lean. Meaning the Holley will think it's in a lean condition and puts all what's available in the tune. In this case all 4 %. But if your up high, like 50%, the car will instantly go full rich, die or be impossible to keep running. If it does run, it will be horrible. Its not a bad idea to change 02's occasionally if your depending on your 02 ....
Hope this helps...

Jeff,

Nice that you weighed in and your points are well taken and you bring up some things that reinforce why I tune ours in Alpha N/Open loop. Your comment about it being old school and somewhat like a carburetor is correct, but that doesn't make it necessarily a bad thing.

Electronics are quite fickle and hard to diagnose in many cases. Nothing is more electronic than an ECU relying on electronic sensors and prone to gremlins than in our environment in my opinion. Alpha N/Open loop is much like a mechanical gauge vs and electronic gauge with Speed Density/Closed loop. Blown cars have to use Speed Density, but I am referring to our NA combinations.

Our sensors are not always reliable like the throttle position sensor that is key to Alpha N for example. I block out our fuel map to have the same MS pulse width from about 75% throttle all the way to 100% incase I have a TPS that is being touchy. I know that the throttle is wide open going down track so it won't matter if the TPS is showing something like 81% or something. I don't want it to lean out on me by mistake.

My approach is somewhat like trust, but verify and count on some failure by sensors at some point. Another thing about my tuning approach is that you immediately know if it helped or hurt. I don't want the ECU acting like Hal on 2001 Space Odyssey saying "Kyle, we are not going to run fast today."

GUMP · 08-31-2021, 10:33 AM

Great points on both sides of this debate. I have cars running both.

My 396 port injected combination has always been in Alpha-N. It is very consistent, but needs regular tweeking.

The two cars that I am currently towing around are direct injected. I run them in closed loop. I aim for a 1% correction window. With the maximum being 10% in case I run into crazy air. That way I can see how much adjustment I need to make to fix the base tune.

I check my TPS on a regular schedule. I also verify that the throttle is opening 100%.

The O2 sensor gets changed at the start of each season and any time I have a major engine failure.

08-30-2021, 10:05 PM	#1
Kdw1403 Member Join Date: Nov 2016 Posts: 202 Likes: 47 Liked 68 Times in 22 Posts	Closed loop? B/SA. Lt1. What should I set my correction limits to. Been running open loop in alpha N. Just kicking around idea of putting it in closed loop. Any advice. I Know nothing about closed loop.

08-30-2021, 10:29 PM	#2
KRatcliff Senior Member Join Date: Sep 2011 Posts: 762 Likes: 75 Liked 522 Times in 140 Posts	Re: Closed loop? I wouldn't run closed loop on a bet. I prefer Alpha N/Open loop by far.

08-31-2021, 06:37 AM	#4
KRatcliff Senior Member Join Date: Sep 2011 Posts: 762 Likes: 75 Liked 522 Times in 140 Posts	Re: Closed loop? There are multiple reasons why I wouldn't run closed loop. The ECU can calculate very quickly, but the information coming into the ECU isn't reliable nor can the ECU read the information and make the necessary change to improve the performance within the short time of the pass in my opinion. Our passes are only a few seconds and they are even shorter when you consider the gear changes. The ECU would have to make these adjustments reading off a single O2 sensor that isn't exactly in a sealed environment and it is only reading on bank of cylinders for an average. Plus the O2 sensors don't exactly like the fuel that we run. When you consider these factors the ECU would be behind the curve in reading the data, calculating the change, and then effecting said change. I tune our cars in Alpha N/Open Loop in 10 different sections (for a 3 speed). The 2 Step (this would be chaos for closed loop) and 3 sections for each gear for a total of 10. My fuel map leads in at each section similar to how we all use timing advance. There is also a cumulative affect that needs to be considered. None of this is possible with closed loop in our applications in my opinion. If closed loop was an advantage down track the Pro Stockers would run it. Closed loop is best suited for street cars.

08-31-2021, 09:05 AM	#6
Robin Lawrence Sponsor Join Date: Jan 2002 Location: Galesburg IL. Posts: 173 Likes: 247 Liked 695 Times in 117 Posts	Re: Closed loop? Best answer is "it depends". I have many cars running in closed loop that are very consistent. It also depends on the ECU. We have been told that ours corrects faster than many others. Our hardware and software guys say it also depends on how the code it written in addition to the speed of the ECU. Alpha N is great on applications where you have a very large cam. Where there is little vacuum at idle and no resolution. Speed Density allows for baro changes. If running Alpha N you should have a baro offset table. I run my stocker with a Combo program. It is Alpha N up t0 2000 RPM and 20% Throttle Position. After that it is in speed density. I typically see less than 2% correction in closed loop. To answer the OP's question, as far as setting limits in closed loop. I would suggest 10% additive or + and limit the negative to 5%. As 02 sensors degrade from race fuel the 02 will tend to show lean thus adding fuel. I recommend installing new 02 sensors annually. The ability to correct with little latency will depend on your combination. On dyno we can see very fast response at sweep rates 300 to 400 rpm per second. Most sporstman cars engines will accelerate in that range in higher gears making closed loop function very well. In Pro Stock the engine acceleration rates can be upwards of 2000 RPM in the lower gears. I did not recommend that they use closed loop. I had several teams that enabled closed loop only in 4th and 5th gear. I don't enable closed loop in my Mustang until 6000 RPM. With almost no collector after the 02 sensor it doesn't give good information at low RPM. Of course this is my opinion. Results may vary. Robin __________________ STK 308 SS 3088

08-31-2021, 10:33 AM	#8
GUMP VIP Member Join Date: May 2002 Location: Shelby, NC Posts: 1,769 Likes: 2,045 Liked 2,167 Times in 504 Posts	Re: Closed loop? Great points on both sides of this debate. I have cars running both. My 396 port injected combination has always been in Alpha-N. It is very consistent, but needs regular tweeking. The two cars that I am currently towing around are direct injected. I run them in closed loop. I aim for a 1% correction window. With the maximum being 10% in case I run into crazy air. That way I can see how much adjustment I need to make to fix the base tune. I check my TPS on a regular schedule. I also verify that the throttle is opening 100%. The O2 sensor gets changed at the start of each season and any time I have a major engine failure.