Jump to content


Photo

Modification of DI injection engines


  • Please log in to reply
44 replies to this topic

#1 NeilR

NeilR
  • Member

  • 458 posts
  • Joined: October 09

Posted 16 July 2012 - 05:14

ToyBaru have released their FT86/BRZ locally and it has effectively sold out for the next 12months. As you would expect with any bandwagon many have jumped at the chance at the next Big Thing (BT) and BT tuning is all the rage.
forced induction of the engine is the first step of many a tuner, yet I would assume that you cannot simply increase the fuel pressure to get more flow? What would be the differences in tuning between the DI and port injected engine?

Advertisement

#2 kikiturbo2

kikiturbo2
  • Member

  • 627 posts
  • Joined: December 04

Posted 16 July 2012 - 08:47

ToyBaru have released their FT86/BRZ locally and it has effectively sold out for the next 12months. As you would expect with any bandwagon many have jumped at the chance at the next Big Thing (BT) and BT tuning is all the rage.
forced induction of the engine is the first step of many a tuner, yet I would assume that you cannot simply increase the fuel pressure to get more flow? What would be the differences in tuning between the DI and port injected engine?



people have been tuning DI VW TFSI engines for some while.. and it is basically the same as with normal enignes.... larger injectors, different fuel pumps.. larger turbos.. etc.. :)

I suspect, that forced induction for the brz engine will be on the cards but it is a question how much FI tuning can this engine take.. A swap to turbocharged subaru STI engine might be more likely.. :)

#3 Magoo

Magoo
  • Member

  • 2,478 posts
  • Joined: October 10

Posted 16 July 2012 - 11:45

The biggest hurdle is managing the injector pulse width properly. It's very short (sometimes multiple pulses) and must be very precise or the benefit of GDI is lost. Conventional non-GDI boxes can't do it.

#4 NeilR

NeilR
  • Member

  • 458 posts
  • Joined: October 09

Posted 16 July 2012 - 11:50

'tis a single firing on the Toyota system (budget car) but they use an interesting fan shaped spray and unusual piston design. The lack of different injectors will be a hurdle that I'm sure will be overcome soon enough.

#5 kikiturbo2

kikiturbo2
  • Member

  • 627 posts
  • Joined: December 04

Posted 16 July 2012 - 13:58

it will be interesting to see what will the aftermarket do with respect to E85 use in these engines... It is very popular fuel in tuning circles but requires 2x injector capacity..

#6 Magoo

Magoo
  • Member

  • 2,478 posts
  • Joined: October 10

Posted 16 July 2012 - 14:32

it will be interesting to see what will the aftermarket do with respect to E85 use in these engines... It is very popular fuel in tuning circles but requires 2x injector capacity..


I think the most viable approach in that case might be simply to piggyback a set of port injectors.

As a GDI boffin at one of the OEs suggested to me recently, it's not the end of the world that the technology is out of reach to much of the aftermarket. At the end of the day, the net benefit of GDI for performance purposes is a half-point of CR and there are lots of ways to get that.

#7 kikiturbo2

kikiturbo2
  • Member

  • 627 posts
  • Joined: December 04

Posted 16 July 2012 - 14:53

I haven't followed the current DI tech much, apart from playing with parts of the latest mazda skyactive petrol, but don't they revert to normal fuel injection timing at mid to high loads anyway ?

#8 munks

munks
  • Member

  • 344 posts
  • Joined: January 03

Posted 16 July 2012 - 15:52

This presentation from 2008 shows a graph with the potential of GDI in various formats. It also suggests that GDI technology in the pipeline could offer 2 points of CR, not sure if it was just over-estimated at the time or what.

#9 NeilR

NeilR
  • Member

  • 458 posts
  • Joined: October 09

Posted 17 July 2012 - 01:28

magoo I believe the subaru engine already has a set of port injectors...so I suspect tuning will involve these.

#10 Magoo

Magoo
  • Member

  • 2,478 posts
  • Joined: October 10

Posted 17 July 2012 - 04:44

magoo I believe the subaru engine already has a set of port injectors...so I suspect tuning will involve these.


So it's D4-S?

Strictly speaking, I'm not sure it qualifies as either GDI or port injection, or both, though with two injectors per cylinder it should have the cost of both.


#11 NeilR

NeilR
  • Member

  • 458 posts
  • Joined: October 09

Posted 17 July 2012 - 10:34

yes it is. Perhaps the port injection is a means of reducing the carbon build up on valves that has plagued early versions of GDI engines such as the VW products?

#12 Magoo

Magoo
  • Member

  • 2,478 posts
  • Joined: October 10

Posted 20 July 2012 - 08:36

yes it is. Perhaps the port injection is a means of reducing the carbon build up on valves that has plagued early versions of GDI engines such as the VW products?


That it will do, thought I expect this setup is more about low-speed drivability.

But to be honest, I don't understand a system with two complete sets of injectors. That makes no sense at all. Unless I am the injector vendor, in which case this is the most brilliant system ever conceived.

#13 J. Edlund

J. Edlund
  • Member

  • 1,295 posts
  • Joined: September 03

Posted 21 July 2012 - 00:55

I haven't followed the current DI tech much, apart from playing with parts of the latest mazda skyactive petrol, but don't they revert to normal fuel injection timing at mid to high loads anyway ?


Direct injection can be used in either homogeneous or stratified charge mode. With a homogeneous charge (lean, stoichiometric or rich) the fuel is injected during the intake stroke. This mode is always used at high load and speed, and some direct injected engines are restricted to this operating mode. Direct injected racing engines for instance only use homogeneous charge mode. Stratified charge is used at low speed and low loads with the intention to improve efficiency at those operating points. With a stratified charge, the fuel is injected late during the compression stroke. The fuel can be wall guided, air guided or spray guided to achive a burnable mixture around the spark plug.

Fuel injectors for direct injection systems require high voltage injector drivers. While these are more expensive than most after market units, they're not out of range for someone with a decent budget. Currently the units from Bosch Motorsport cost from 1500 euros. Fuel pressure is normally under electronic control.

#14 TDIMeister

TDIMeister
  • Member

  • 235 posts
  • Joined: January 06

Posted 22 July 2012 - 12:44

All contemporary GDI engines with 3-way catalytic converters (TWC) are mapped to operate at stoichiometric homogeneous operation throughout the operating range (except for enrichment at full-load conditions). That's because the TWCs are ineffective at reducing NOx on anything but stoichiometric conditions and emissions regulations limiting NOx are the biggest challenge in the largest jurisdictions (USA, Europe) to meet.

#15 NeilR

NeilR
  • Member

  • 458 posts
  • Joined: October 09

Posted 23 July 2012 - 00:19

so enriching for more power or turbo charging (given most private performance tuners will be looking at a dyno and not considering emissions map) will mean that different catalytic converter is required as std cat will not suffice?

Edited by NeilR, 23 July 2012 - 00:20.


#16 Greg Locock

Greg Locock
  • Member

  • 4,507 posts
  • Joined: March 03

Posted 23 July 2012 - 00:36

so enriching for more power or turbo charging (given most private performance tuners will be looking at a dyno and not considering emissions map) will mean that different catalytic converter is required as std cat will not suffice?

I don't think so, since you will be looking at WOT performance, whereas the standard cal is closed loop only at idle and part throttle, and that's the only bit where the cats will work anyway. Admittedly there is a transition from closed to open loop, probably outside of the part of the engine's performance ithat is used in an emissions test.



#17 NeilR

NeilR
  • Member

  • 458 posts
  • Joined: October 09

Posted 23 July 2012 - 01:11

ahh, thanks.
I'm really curious as to what the 'tuner' market will do in relation to this engine. Toyota reputedly released it to the Japanese tuners some 12 months before official release to allow the various 'exciting' add-ons. curiously none of them offer different GDI injectors or tuning tools on their list of products, but all have some 'in development'. Normally there would be a big rush to be the first.

#18 Magoo

Magoo
  • Member

  • 2,478 posts
  • Joined: October 10

Posted 23 July 2012 - 01:14

I would add that three-way cats don't function most effectively fixed at exact stoichiometry. Rather, FAR perturbation is used to swing the mixture through a narrow-ish range on both sides of stoich to manipulate the O2 reserve in the cat. Max CO reduction is at one end of the range while max NOx reduction is at the other.

#19 TDIMeister

TDIMeister
  • Member

  • 235 posts
  • Joined: January 06

Posted 23 July 2012 - 03:18

Lambda=1 +/- 1-2% is about control range of conventional O2 sensors - close enough.

Advertisement

#20 kikiturbo2

kikiturbo2
  • Member

  • 627 posts
  • Joined: December 04

Posted 23 July 2012 - 05:33

I don't think so, since you will be looking at WOT performance, whereas the standard cal is closed loop only at idle and part throttle, and that's the only bit where the cats will work anyway. Admittedly there is a transition from closed to open loop, probably outside of the part of the engine's performance ithat is used in an emissions test.


but turbo aplications, especially in the aftermarket, usually use some sort of free flowing metallic cat.... also, I dont' think that the cat and exhaust made for 200 NA horses will work for 300 or so turbocharged ones..

#21 Magoo

Magoo
  • Member

  • 2,478 posts
  • Joined: October 10

Posted 23 July 2012 - 13:58

Lambda=1 +/- 1-2% is about control range of conventional O2 sensors - close enough.


It's not really true that a TWC is most effective at reducing CO and NOx at constant or precise stoichiometry. That is a generalization/myth/rule of thumb/lie. In truth, below stoich the cat is performing reduction; above stoich, the cat is performing oxidation. And obviously, it can't be doing both at the same time. So we use FAR perturbation aka fuel dithering to swing the fuel mixture through a range above and below stoichiometry to manage the O2 reserve in the cat and optimize cat efficiency in both reduction and oxidation -- so it is effective with both CO and NOx. If we do this right, among other things we no longer need an AIR pump blowing fresh air into the cat.

#22 Magoo

Magoo
  • Member

  • 2,478 posts
  • Joined: October 10

Posted 23 July 2012 - 14:08

ahh, thanks.
I'm really curious as to what the 'tuner' market will do in relation to this engine. Toyota reputedly released it to the Japanese tuners some 12 months before official release to allow the various 'exciting' add-ons. curiously none of them offer different GDI injectors or tuning tools on their list of products, but all have some 'in development'. Normally there would be a big rush to be the first.


A year or two ago I had my nose stuck in an outside-the-OEs engine program involving GDI. At the start of the project, several major players in the aftermarket performance EFI world said they could handle the GDI system. But as deadlines came and went, one by one their claims grew more modest and they stopped answering their phones. The lead contractor was forced to tackle the job using a shelf OE diesel ECU, unsupported app.


#23 TDIMeister

TDIMeister
  • Member

  • 235 posts
  • Joined: January 06

Posted 23 July 2012 - 14:08

I'm not arguing with you. I'm only saying that the actual magnitudes of the deviations below and above stoichiometric are very small in percentage terms that, averaged out, can be considered stoic. The integral result of those fluctuations is also very close to stoic. By definition, if you're operating outside of global chemical stoichiometry, you are generating more of something that's probably undesirable, or more germane to what's going in a catalyst, hindering reactions to reduce those undesirable species. When we are talking about full-load enrichment magnitudes in the order of lambda of 0.8 and enleanment during overrun briefly to the order of 1.2 , a few percent around 1 during closed-loop control is negligible.

Besides, anyone with any knowledge of control systems knows it's impossible to regulate to exactitude such a complex, dynamic and interacting system as an engine.

Edited by TDIMeister, 23 July 2012 - 15:57.


#24 Powersteer

Powersteer
  • Member

  • 2,460 posts
  • Joined: September 00

Posted 23 July 2012 - 16:45

people have been tuning DI VW TFSI engines for some while.. and it is basically the same as with normal enignes.... larger injectors, different fuel pumps.. larger turbos.. etc.. :)

Ever tested a hybrid sequential DI + port injection? If so, what's the outcome? Is there such a thing?

:cool:

#25 TDIMeister

TDIMeister
  • Member

  • 235 posts
  • Joined: January 06

Posted 23 July 2012 - 16:58

Toyota has done MPFI and DI together in series production.

#26 NeilR

NeilR
  • Member

  • 458 posts
  • Joined: October 09

Posted 24 July 2012 - 00:54

interesting report to me of a toyota v8 turbo diesel landcruiser that had diesel injector fail and lose compression on one cyl. Apparently they come as a matched set and all 8 had to be replaced at vast expense. Am I correct in thinking that GDI injectors are not usually matched?

#27 Magoo

Magoo
  • Member

  • 2,478 posts
  • Joined: October 10

Posted 24 July 2012 - 12:05

interesting report to me of a toyota v8 turbo diesel landcruiser that had diesel injector fail and lose compression on one cyl. Apparently they come as a matched set and all 8 had to be replaced at vast expense. Am I correct in thinking that GDI injectors are not usually matched?


Hmm, I wasn't aware they were a matched set. Is this an official Toyota policy?


#28 desmo

desmo
  • Tech Forum Host

  • 12,975 posts
  • Joined: January 00

Posted 24 July 2012 - 13:39

Sounds like an official "sucker born every minute" policy!

#29 Powersteer

Powersteer
  • Member

  • 2,460 posts
  • Joined: September 00

Posted 24 July 2012 - 19:21

Toyota has done MPFI and DI together in series production.

Any idea what the concept was behind it? Using fuel to increase intake speed? Is ir possible to have some fuel mixture to add density to effect on sonic wave or pressure reciprocation from valve closing?

:cool:

#30 TDIMeister

TDIMeister
  • Member

  • 235 posts
  • Joined: January 06

Posted 24 July 2012 - 19:43

http://www.ft86club....read.php?t=3172

http://www.ft86club....mp;d=1326052791

#31 Powersteer

Powersteer
  • Member

  • 2,460 posts
  • Joined: September 00

Posted 26 July 2012 - 15:46

Going through those tonight TDIM but had a quick a look, very interesting, thank you.

:cool:

Edited by Powersteer, 26 July 2012 - 15:47.


#32 Greg Locock

Greg Locock
  • Member

  • 4,507 posts
  • Joined: March 03

Posted 26 July 2012 - 23:19

Sounds like an official "sucker born every minute" policy!

my guess is that they are matched for production to give even firing at idle. Quite whether the aftermarket needs the same rigorous matching is rather suspect, although legally there may be an emissions issue.

#33 J. Edlund

J. Edlund
  • Member

  • 1,295 posts
  • Joined: September 03

Posted 31 July 2012 - 22:47

All contemporary GDI engines with 3-way catalytic converters (TWC) are mapped to operate at stoichiometric homogeneous operation throughout the operating range (except for enrichment at full-load conditions). That's because the TWCs are ineffective at reducing NOx on anything but stoichiometric conditions and emissions regulations limiting NOx are the biggest challenge in the largest jurisdictions (USA, Europe) to meet.


NOx can be reduced at all conditions richer than stoichiometric.

Direct injected engines that uses a stratified charge at part load, which many do, usually uses a NOx storing catalyst to handle the NOx emissions. The NOx emissions will be stored when operating with a stratified charge, when the storage is full, the engine uses a stoichiometric/rich mixture to reduce the NOx stored.

interesting report to me of a toyota v8 turbo diesel landcruiser that had diesel injector fail and lose compression on one cyl. Apparently they come as a matched set and all 8 had to be replaced at vast expense. Am I correct in thinking that GDI injectors are not usually matched?


Doesn't sound very clever if true. Usually there are simpler methods to handle that sort of things, like entering an injector specific calibration in the ECU. I know that system is used by some OEM's.

#34 kikiturbo2

kikiturbo2
  • Member

  • 627 posts
  • Joined: December 04

Posted 01 August 2012 - 10:27

Doesn't sound very clever if true. Usually there are simpler methods to handle that sort of things, like entering an injector specific calibration in the ECU. I know that system is used by some OEM's.


yeah, like VW for example.. very expensive business changing injectors on their 2.0 CR tdi, especially as a lot of them did fail..

#35 Magoo

Magoo
  • Member

  • 2,478 posts
  • Joined: October 10

Posted 01 August 2012 - 11:02

This is just me talking, but in my opinion a "matched set" of electronic injectors should be across a production run. If selective sorting is required to get four or eight pieces nominal, that doesn't say a whole lot for the production tolerances. We are not building 1947 Studebakers here.

#36 TDIMeister

TDIMeister
  • Member

  • 235 posts
  • Joined: January 06

Posted 01 August 2012 - 19:26

NOx can be reduced at all conditions richer than stoichiometric.

But conversion efficiency falls off precipitously outside of a tight range of Lambda, which was my original point.

Direct injected engines that uses a stratified charge at part load, which many do, usually uses a NOx storing catalyst to handle the NOx emissions. The NOx emissions will be stored when operating with a stratified charge, when the storage is full, the engine uses a stoichiometric/rich mixture to reduce the NOx stored.

While it is indeed a capability of GDI systems, I know of no contemporary GDI passenger car in Europe and certainly North America, that employ lean stratified calibration and an NSC. In Japan and ROW, maybe. I'd be happy to be proven wrong though.

Edited by TDIMeister, 01 August 2012 - 19:37.


#37 TDIMeister

TDIMeister
  • Member

  • 235 posts
  • Joined: January 06

Posted 01 August 2012 - 19:29

yeah, like VW for example.. very expensive business changing injectors on their 2.0 CR tdi, especially as a lot of them did fail..

I have heard of high-pressure fuel pumps failing and taking out entire fuel systems by contamination with metal shrapnel, but never injectors per se. There was also a recall to address an injector line resonance vibration issue that was resolved simply by added damping elements to them.


#38 TDIMeister

TDIMeister
  • Member

  • 235 posts
  • Joined: January 06

Posted 02 August 2012 - 02:29

VW/Audi are evidently also going with both DI and PFI in the new engine, though not at the same time:
http://blog.caranddr...urbo-diesel-v6/


#39 Lee Nicolle

Lee Nicolle
  • Member

  • 5,825 posts
  • Joined: July 08

Posted 02 August 2012 - 07:49

my guess is that they are matched for production to give even firing at idle. Quite whether the aftermarket needs the same rigorous matching is rather suspect, although legally there may be an emissions issue.

Or in some cases the dealer is just stitching up the customer. Toymota dealers have done that for decades on parts, especially late model cars. Toyota policy, maybe or maybe not. But when it is a bit older they will sure sell you one or the customer will buy aftermarket, when they become advailable.

Advertisement

#40 kikiturbo2

kikiturbo2
  • Member

  • 627 posts
  • Joined: December 04

Posted 02 August 2012 - 08:28

I have heard of high-pressure fuel pumps failing and taking out entire fuel systems by contamination with metal shrapnel, but never injectors per se. There was also a recall to address an injector line resonance vibration issue that was resolved simply by added damping elements to them.



there was a major problem with injectors on last generation passat / seat leon CR tdi, and was related to a specific injector manufacturer... I don't remember if it was siemens or bosch.....

#41 desmo

desmo
  • Tech Forum Host

  • 12,975 posts
  • Joined: January 00

Posted 02 August 2012 - 17:38

I think so too. In fact Quattroruote did an entire article about the problem.

#42 Greg Locock

Greg Locock
  • Member

  • 4,507 posts
  • Joined: March 03

Posted 02 August 2012 - 23:36

This is just me talking, but in my opinion a "matched set" of electronic injectors should be across a production run. If selective sorting is required to get four or eight pieces nominal, that doesn't say a whole lot for the production tolerances. We are not building 1947 Studebakers here.

Getting good idle characteristics from high torque fuel injectors has been a problem from 1985, and the requirements and the precision required have more or less kept pace since then. It may be that they had an idle shake problem they solved by using matched injectors, which is not a production issue since every injector is tested. Off the top of my head conrods, pistons and bearing shells are still matched on some engines. I tried to get them to go to five grades of big end shell at one point.



#43 kikiturbo2

kikiturbo2
  • Member

  • 627 posts
  • Joined: December 04

Posted 04 August 2012 - 15:04

Getting good idle characteristics from high torque fuel injectors has been a problem from 1985, and the requirements and the precision required have more or less kept pace since then. It may be that they had an idle shake problem they solved by using matched injectors, which is not a production issue since every injector is tested. Off the top of my head conrods, pistons and bearing shells are still matched on some engines. I tried to get them to go to five grades of big end shell at one point.



hell, even skis are individually sorted to produce matched pairs, I hope rods and pistons are too... :)

#44 Magoo

Magoo
  • Member

  • 2,478 posts
  • Joined: October 10

Posted 04 August 2012 - 16:36

Connecting rods do not require selective sorting in 2012. There shouldn't be enough variation in 10,000 pieces to identify 8 pieces with any significant commonality over the other 9,992. They should all be functionally nominal and the ones that aren't are scrapped.

In the case of electronic injectors: If the pieces are individually tested and require sorting to obtain a positive customer outcome, that means that over the production run, the normal distribution of the manufacturing variation is greater than the true nominal tolerance. In that case, you are aren't fixing rough idles, you're making them.


Edited by Magoo, 04 August 2012 - 19:12.


#45 Lee Nicolle

Lee Nicolle
  • Member

  • 5,825 posts
  • Joined: July 08

Posted 04 August 2012 - 23:35

Getting good idle characteristics from high torque fuel injectors has been a problem from 1985, and the requirements and the precision required have more or less kept pace since then. It may be that they had an idle shake problem they solved by using matched injectors, which is not a production issue since every injector is tested. Off the top of my head conrods, pistons and bearing shells are still matched on some engines. I tried to get them to go to five grades of big end shell at one point.

Yeah, production tolerances on engines. With bearings unadvailable aftermarket with thick backs to compensate for woppy main tunnels, poorly sized rods. And pistons with about 4 sizes of standard. Still happens even now on most popular makes.
If the injectors have that much tolerance it is surprising the engines run very well at all.