Notices
928 Forum 1978-1995
Sponsored by:
Sponsored by: 928 Specialists

Twin Turbo 928 fixed and back out there terrorizing the streets!

Thread Tools
 
Search this Thread
 
Old 10-15-2020, 10:58 PM
  #2521  
jcorenman
Rennlist Member
 
jcorenman's Avatar
 
Join Date: Jan 2008
Location: Friday Harbor, WA
Posts: 4,061
Received 316 Likes on 152 Posts
Default

Cool, that has to be the ultimate sleeper!

Originally Posted by ptuomov
Idle hunts a little when cold, and when cold it smells a little rich. It needs a bit of tuning for cold startup which John obviously couldn’t do in Kentucky summer....
Yep, you can only tune for the environment you've got. But I forget... do you have a Sharktuner yet?

Last edited by jcorenman; 10-15-2020 at 10:59 PM.
Old 10-17-2020, 01:05 PM
  #2522  
ptuomov
Nordschleife Master
Thread Starter
 
ptuomov's Avatar
 
Join Date: Jan 2009
Location: MA
Posts: 5,610
Received 81 Likes on 64 Posts
Default

Originally Posted by jcorenman
Cool, that has to be the ultimate sleeper! Yep, you can only tune for the environment you've got. But I forget... do you have a Sharktuner yet?
Yes, I'm an early adopter of ST2.

I'm just going to drive it for a while and write down some impressions before even looking at the logs or maps.
Old 10-20-2020, 08:50 PM
  #2523  
ptuomov
Nordschleife Master
Thread Starter
 
ptuomov's Avatar
 
Join Date: Jan 2009
Location: MA
Posts: 5,610
Received 81 Likes on 64 Posts
Default

Solved the idle problem today with Jim Corenman's help. I had to get a new notebook computer and set it up, which was probably the most time consuming part of the project. Once I got the Sharktuner 2 hooked up, I did some experiments by myself and fixed some basic issues with the idle ignition map. Call this the zeroth step: make sure your idle ignition map is V shaped or U shaped with a downward slope of ignition advance as a function of the rpm in the idle range of rpm. With that fixed, it was time to give Jim a call and screen share to get the JDS idle controller calibrated. The controller is very logical if you've ever used a proportional-integral-derivative/differential PID controller (and it remains logical even if you don't understand it!). The first step was to decide the target idle rpms for different operating conditions and record them in one open-loop map that produces the target rpm value. The second step was to calibrate the open-loop base ISV% for each target rpm at each operating condition. At this point, the car would likely idle correctly without any PID control (or those parameters set to neuter the PID control), just based on the V-shaped idle ignition map and the open-loop idle calibration. This open loop controller provides the feed forward control to the PID controller by varying set point. The third and final step was to ad hoc calibrate the PID controller parameters to known good values and deviate a little from those based on the special circumstances of my car's engine. Thanks, Jim!

Last edited by ptuomov; 10-22-2020 at 09:15 AM.
Old 10-20-2020, 10:14 PM
  #2524  
Tony
Addict
Lifetime Rennlist
Member
 
Tony's Avatar
 
Join Date: May 2001
Location: Las Vegas
Posts: 14,676
Received 584 Likes on 305 Posts
Default

how do you switch between open and closed loop for the tuning...
Old 10-20-2020, 10:17 PM
  #2525  
ptuomov
Nordschleife Master
Thread Starter
 
ptuomov's Avatar
 
Join Date: Jan 2009
Location: MA
Posts: 5,610
Received 81 Likes on 64 Posts
Default

Originally Posted by Tony
how do you switch between open and closed loop for the tuning...
It's built into the JDS idle controller. Basically, it's always an open-loop map for the target rpm and then an open-loop function and a PID control to control the deviation from the target rpm. You could make the PID control parameters such that the PID controller has very little impact on the ISV % and that would then just be open-loop idle controller.
Old 10-21-2020, 12:26 AM
  #2526  
Tony
Addict
Lifetime Rennlist
Member
 
Tony's Avatar
 
Join Date: May 2001
Location: Las Vegas
Posts: 14,676
Received 584 Likes on 305 Posts
Default

Originally Posted by ptuomov
It's built into the JDS idle controller. Basically, it's always an open-loop map for the target rpm and then an open-loop function and a PID control to control the deviation from the target rpm. You could make the PID control parameters such that the PID controller has very little impact on the ISV % and that would then just be open-loop idle controller.

i wired a switch in to disable the O2 sensor input when tuning. Just disconects it.
Old 10-21-2020, 04:26 PM
  #2527  
dr bob
Chronic Tool Dropper
Lifetime Rennlist
Member
 
dr bob's Avatar
 
Join Date: Oct 2001
Location: Bend, Oregon
Posts: 20,506
Received 547 Likes on 410 Posts
Default

I'm not sure I've ever heard of or seen a "partial-integral-differential PID controller". Is there some background available on this?
Old 10-21-2020, 05:06 PM
  #2528  
ptuomov
Nordschleife Master
Thread Starter
 
ptuomov's Avatar
 
Join Date: Jan 2009
Location: MA
Posts: 5,610
Received 81 Likes on 64 Posts
Default

Originally Posted by dr bob
I'm not sure I've ever heard of or seen a "partial-integral-differential PID controller". Is there some background available on this?
I mistyped it. I of course meant proportional-integral-differential (or derivative) controller. It's an age old controller that can be implemented in all sorts of digital and analog computers whether mechanical, electronic, hydraulic, etc.

In case someone is interested, here’s where I ended up:





Last edited by ptuomov; 10-22-2020 at 09:13 AM.
Old 10-22-2020, 01:09 PM
  #2529  
Thom
Race Car
 
Thom's Avatar
 
Join Date: Mar 2002
Posts: 4,329
Received 41 Likes on 31 Posts
Default

Originally Posted by ptuomov
make sure your idle ignition map is V shaped or U shaped with a downward slope of ignition advance as a function of the rpm in the idle range of rpm.
V or U shaped?
We want an upward slope of timing when engine speeds falls below the target idle speed, but a downward slope when engine speeds shoots over target idle speed.

AFR of 10.95 at idle?

Last edited by Thom; 10-22-2020 at 01:10 PM.
Old 10-22-2020, 01:09 PM
  #2530  
ptuomov
Nordschleife Master
Thread Starter
 
ptuomov's Avatar
 
Join Date: Jan 2009
Location: MA
Posts: 5,610
Received 81 Likes on 64 Posts
Default

Fiddled with it some more. The prior belief is that if Porsche can get the GT cams to idle at 775rpm, I should be able to get these Elgin 65-6 cams to idle equally well at 800rpm. I think we met that goal:

View this post on Instagram

Last edited by ptuomov; 10-22-2020 at 01:10 PM.
Old 10-22-2020, 01:13 PM
  #2531  
ptuomov
Nordschleife Master
Thread Starter
 
ptuomov's Avatar
 
Join Date: Jan 2009
Location: MA
Posts: 5,610
Received 81 Likes on 64 Posts
Default

Originally Posted by Thom
V or U shaped?
We want an upward slope of timing when engine speeds falls below the target idle speed, but a downward slope when engine speeds shoots over target idle speed.

AFR of 10.95 at idle?
Yes, downward sloping in the idle rpm “hunting” area and then rapidly upward sloping at higher rpms above the idle hunting area.

A picture is worth a thousand words:





The WBO2 controller is not hooked up, so that’s a default value. The narrow band fluctuates so it’s almost exactly lambda 1 with 93 octane MA E10 pump gas at idle.

Last edited by ptuomov; 10-22-2020 at 01:24 PM.
Old 10-22-2020, 01:23 PM
  #2532  
Thom
Race Car
 
Thom's Avatar
 
Join Date: Mar 2002
Posts: 4,329
Received 41 Likes on 31 Posts
Default

Originally Posted by ptuomov
Yes, downward sloping in the idle rpm “hunting” area and then rapidly upward sloping at higher rpms above the idle hunting area
?... I said the opposite - when rpm overshoots the target idle speed we want a downward ignition curve slope above the target idle speed, obviously only active when idle mode is enabled.
Old 10-22-2020, 01:27 PM
  #2533  
ptuomov
Nordschleife Master
Thread Starter
 
ptuomov's Avatar
 
Join Date: Jan 2009
Location: MA
Posts: 5,610
Received 81 Likes on 64 Posts
Default

Originally Posted by Thom
?... I said the opposite - when rpm overshoots the target idle speed we want a downward ignition curve slope above the target idle speed, obviously only active when idle mode is enabled.
Hunting area = 500-1000 rpm.

At high rpms when coasting with the throttle closed you probably want to give it 40-50 degrees of advance? With deceleration fuel cutoff it probably doesn’t matter that much, but for the sake of the principle.
Old 10-22-2020, 01:37 PM
  #2534  
Tony
Addict
Lifetime Rennlist
Member
 
Tony's Avatar
 
Join Date: May 2001
Location: Las Vegas
Posts: 14,676
Received 584 Likes on 305 Posts
Default

Originally Posted by ptuomov
Hunting area = 500-1000 rpm.

At high rpms when coasting with the throttle closed you probably want to give it 40-50 degrees of advance? With deceleration fuel cutoff it probably doesn’t matter that much, but for the sake of the principle.

but them you dont get that nice gurgling pop pop in the exhaust....i left mine on.
Old 10-22-2020, 01:38 PM
  #2535  
ptuomov
Nordschleife Master
Thread Starter
 
ptuomov's Avatar
 
Join Date: Jan 2009
Location: MA
Posts: 5,610
Received 81 Likes on 64 Posts
Default A riddle

Q: How do you merge into 90mph highway traffic when you are initially traveling at 45mph?
A: Slightly sideways.
The following users liked this post:
SwayBar (10-23-2020)


Quick Reply: Twin Turbo 928 fixed and back out there terrorizing the streets!



All times are GMT -3. The time now is 01:32 AM.