how do i stop it hunting for its idle point?
it oscillates either side of idle point around 200rpm then over 6 or 7 oscillatons, it reduces 180,150,120,80 etc and around 8 secs it settles at idle. very annoying
idle control
-
pat
- Syvecs Staff - Cleaner
- Posts: 356
- Joined: Fri May 23, 2008 10:23 am
- Location: Out there... somewhere
- Contact:
Re: idle control
Oscillations are a closed loop control phenomenon and are nornally an indication of too high loop gain. You will need to adjust the proportional and integral gain to suit your setup. Also you will need to check that your fuelling is OK in that area of the map as a rich to lean swing can cause a misfire which will cause oscillations as the operating point swings between missing and not missing, all the while adding and removing idle duty and thus air.... also if closed loop lambda control is on then that can further amplify the problem as it over corrects the fuelling.
It is best to start with closed loop idle and lambda turned off and try to tune the fuelling in the idle area so as to get a uniform AFR as idle load is varied. Once the fuel table is good in that area you can then move on to closed loop control of the idle speed but with the closed loop lambda still off. Start with low gain and slowly increase the gain while small perturbances to the idle do not cause oscillations. I find that proportional gain can cause issues because of the delay between the change and the effect - by that I mean that more air is added to the plenum but it takes time for the torque to rise and bring the idle back up and then it overshoots since now there is too much air in the plenum. It is thus easy to break into oscillation if you have too much gain in the proportional table.
The integral table can work quite well as long as the gain is kept sensible. I would start with the base duty above what is needed for natural idle and then bring the idle back down using a small amount of integral gain once closed loop idle becomes active. This works quite well but has the small drawback that the natural idle sits a few hundred RPM higher than it otherwise would, until closed loop is entered.
One issue you can find is that return to idle undershoots as a lot of vacuum is drawn and that can cause a misfire due to poor charge quaiity and bouncing off flammability limits. There is no torque being generated by the engine as it returns to idle which then fails to slow the descent to idle and so undershoots. The only real solution to this phenomenon is to dump a load of air into the plenum just before reaching target RPM - there is no specific strategy to do this so you would need to emulate this behaviour using the tables that you do have - eg if you were to put a lot of duty in the proportional table at the highest RPM breakpoint, still have some duty at 0 error and the base table lower than the requisite duty for natural idle - before entering closed loop the valve would be quite tightly shut and so vacuum would also be high. But as closed loop turns on there would be a sudden jump in duty causing the valve to dump air into the plenum - however that will not have an immediate effect thus allowing the RPM too fall past the high duty part of the map and get to the area of natural idle where the additional duty is now correct for natural idle. The danger is that it falls too slowly and thus doesn't fall fully.
Hope this makes sense
Pat.
It is best to start with closed loop idle and lambda turned off and try to tune the fuelling in the idle area so as to get a uniform AFR as idle load is varied. Once the fuel table is good in that area you can then move on to closed loop control of the idle speed but with the closed loop lambda still off. Start with low gain and slowly increase the gain while small perturbances to the idle do not cause oscillations. I find that proportional gain can cause issues because of the delay between the change and the effect - by that I mean that more air is added to the plenum but it takes time for the torque to rise and bring the idle back up and then it overshoots since now there is too much air in the plenum. It is thus easy to break into oscillation if you have too much gain in the proportional table.
The integral table can work quite well as long as the gain is kept sensible. I would start with the base duty above what is needed for natural idle and then bring the idle back down using a small amount of integral gain once closed loop idle becomes active. This works quite well but has the small drawback that the natural idle sits a few hundred RPM higher than it otherwise would, until closed loop is entered.
One issue you can find is that return to idle undershoots as a lot of vacuum is drawn and that can cause a misfire due to poor charge quaiity and bouncing off flammability limits. There is no torque being generated by the engine as it returns to idle which then fails to slow the descent to idle and so undershoots. The only real solution to this phenomenon is to dump a load of air into the plenum just before reaching target RPM - there is no specific strategy to do this so you would need to emulate this behaviour using the tables that you do have - eg if you were to put a lot of duty in the proportional table at the highest RPM breakpoint, still have some duty at 0 error and the base table lower than the requisite duty for natural idle - before entering closed loop the valve would be quite tightly shut and so vacuum would also be high. But as closed loop turns on there would be a sudden jump in duty causing the valve to dump air into the plenum - however that will not have an immediate effect thus allowing the RPM too fall past the high duty part of the map and get to the area of natural idle where the additional duty is now correct for natural idle. The danger is that it falls too slowly and thus doesn't fall fully.
Hope this makes sense
Pat.
-
mapping dabbler
- Posts: 20
- Joined: Sat Dec 07, 2013 9:28 pm
Re: idle control
thx pat. can spend lots of time in these areas eh!
-
pat
- Syvecs Staff - Cleaner
- Posts: 356
- Joined: Fri May 23, 2008 10:23 am
- Location: Out there... somewhere
- Contact:
Re: idle control
The idle area is one of the most difficult to get right becuse there is very little air there and so even a small error tends to be quite noticeable. If that wasn't enough then the fact that the battery voltage is not very predictable makes the situation worse again.
The battery added for the injection opening time comes into play and if that table is not right then you get even more fluctuations in the AFR and also more misfires. The battery voltage also has some unexpected effects... say for argument sake your fuel pressure regulator was marginal for the flow that your fuel pump can deliver at idle - then with the RPM high enough for the alternator to raise the battery voltage the fuel pump will raise the fuel pressure even when there is no change in the reference port pressure since the return port is too small to return all that fuel whilst maintaining the correct rail pressure. As the RPM drops the alternator stops charging and so the pump slows down and the rail pressure drops. Given that the rail pressure is now variable but you are still on the same row on the fuel table, the numbers in the table start to look a little odd.
Furthermore, if that wasn't already enough, the change in the pressure across the injectors will alter the dead time : now even your opening time table is not forced to be right any more! Such are the problems people often create inadvertently when trying to upgrrade their fuel systems..... If I had a pound for every time that I had seen a badly designed fuel system then I think that I might be able to retire, LOL
Cheers,
Pat.
The battery added for the injection opening time comes into play and if that table is not right then you get even more fluctuations in the AFR and also more misfires. The battery voltage also has some unexpected effects... say for argument sake your fuel pressure regulator was marginal for the flow that your fuel pump can deliver at idle - then with the RPM high enough for the alternator to raise the battery voltage the fuel pump will raise the fuel pressure even when there is no change in the reference port pressure since the return port is too small to return all that fuel whilst maintaining the correct rail pressure. As the RPM drops the alternator stops charging and so the pump slows down and the rail pressure drops. Given that the rail pressure is now variable but you are still on the same row on the fuel table, the numbers in the table start to look a little odd.
Furthermore, if that wasn't already enough, the change in the pressure across the injectors will alter the dead time : now even your opening time table is not forced to be right any more! Such are the problems people often create inadvertently when trying to upgrrade their fuel systems..... If I had a pound for every time that I had seen a badly designed fuel system then I think that I might be able to retire, LOL
Cheers,
Pat.