fast951

I have been asked a few questions regarding overboost and the functions of the DME & KLR. It appears that there is a bit of unknown or incorrect information floating around. The factory manual is a bit unclear on some of the issues. All of these findings are based on actual testing under a controlled environment, wheree we can generate errors and capture the DME & KLR behavior at will. We tested with various KLRs from different years. Hopefully this helps!

- Overboost Protection - DME: This is a function of the DME. it should be called overload protection as the DME does not read boost. At the same overboost settings, a smaller turbo running high boost may not trigger overboost protection, yet a large turbo running lower boost might trigger it. It's a bit more complicated, but you get the idea.

- Overboost Error - KLR: This is an error condition detected by the KLR when it detects an actual boost that is higher than it expects. This error will be triggered when you use the stock KLR chip and run higher boost by using an external boost control instead of the cycling valve. The KLR throws a code for this error. However the KLR does NOT retard timing based on this error.
* So if you are using the stock KLR and running high boost by using an external boost controller, the ignition timing is not being retarded UNLESS there is knock.

- Ignition Retard - KLR: The KLR retards timing when it detects KNOCK. However the KLR does NOT retard timing when it detects : Low battery/voltage, bad TPS, HI Boost or Low Boost errors. These errors will generate a error code.
* When the KLR detects knock, it retards timing by up to 6deg. When knock is gone, it will restore timing "gradually" not all at once.
* When errors are present (overboost, underboost, TPS,..) the KLR WILL pull timing if there is knock.

Geneqco

Thanks for the clarification John.

Tom M'Guinn

John, do you know if the KLR throws a code if the DME goes into its own "overboost protection" mode (i.e., the first one you list above)?

eman930

How does the DME know to activate "overload" protection, if its not reading boost. Is it talking to the KLR since that's the only thing with a pressure sensor?

divil

The DME's overload protection is based on airflow vs rpm. It has a hard coded map of maximum airflow for a given rpm value, and if the actual airflow value exceeds that, it cuts fuel. I don't think the DME and KLR communicate in any way beyond the exchange of the ignition signals (DME->KLR) and full load (KLR->DME) signals, but I could be wrong.

Tom M'Guinn

And, totally separately, the KLR does have a pressure sensor and triggers its fault mode if the boost exceeds its pre-programmed thresholds. I'd be curious to know which happens first -- KLR excessive boost protection, or DME excessive load protection.... ?

divil

Good question. I'm guessing here but I would say the KLR would detect the error first. The DME overload protection is basically guaranteed to work, so if that has to fail in order for the KLR to register an error, then it would never happen. I suppose that would mean that if the DME overload kicks in, then you would expect to have a KLR code too - not because the 2 units are communicating, but because the KLR would have experienced the error first. Unless the DME's overload is triggered because of a faulty AFM signal - in that case, the KLR might not think there's a problem at all.

Tom M'Guinn

It does seem that the KLR alerts the DME when the KLR sees too much boost though. I'd have to guess that the KLR kills the ignition and the DME in turn kills the fuel so as not to flood the motor. See the Technik page's description (which only talks about injection...?).

I guess if the limits are in the same ball park, you could see either tripping first depending on circumstances. High PSI without much load (inefficient boost), KLR wins? Oxygen-rich boost just under the KLR limit, DME wins? Sounds like both abruptly cut off the engine, so probably hard to know in practice which triggered. Or maybe one is set so far below the other it always wins. I dunno. Although to your earlier point, if you hit overboost protection and have no KLR codes, I guess that's a strong clue it was the DME....?

divil

I didn't think the KLR ever kills the ignition signal. If you trigger a big KLR error by doing something like unplugging the throttle position sensor, then he KLR's fail safe mode is to drive the DME into wide-open-throttle (WOT) mode and open the cycling valve to direct all intake pressure to the wastegate, limiting boost to around 3-4psi. The car still runs in this mode though. So here's how I think it all works:

* The KLR has a target boost curve, mapped against rpm and throttle position. If the boost deviates because of a slightly leaky or slightly stuck wastegate, or just variations in the wg spring, or something similar, the KLR corrects it with closed-loop control

* If the KLR can't keep the boost to the proper level, it triggers an error code and enters the failsafe mode I described above (actually I don't know if all errors cause it to enter failsafe mode, but some certainly do)

* The KLR's failsafe mode is not bullet-proof though; the wastegate could be physically stuck closed something like that. If the engine still overboosts even with the KLR in failsafe mode, then DME detects that something ridiculous is happening and cuts fuel.

The DME's failsafe should be a last resort because it would be very abrupt. I could be wrong here but I've never heard of the KLR cutting the engine abruptly by killing the ignition signal (although that would have to be how it would do it if it was to kill the engine)

Tom M'Guinn

Yeah, I don't know, I was guessing as I said. But if the KLR doesn't kill the ignition when the boost exceeds its limit, how does it shut off the injectors as described in the Technik?

divil

I think they're actually talking about the DME in that part. Confusing I know since it's in a section specifically about the KLR. But I bet the boost curve shown in the picture is just to show the amount of boost that would correspond to the actual airflow safety curve in the DME. I'm guessing too of course...hopefully fast951 will chime in But incidentally I did start taking a look at a spare KLR I have and reading the disassembled code after the discussion we had about the trigger signal in the other thread. I'm finding it pretty interesting, although I have only scratched the surface of how it works so far. I might start another thread for that.

Tom M'Guinn

I guess that's possible, but they use the same graph when discussing boost pressure. And they specifically say "when the charging air pressure" exceeds the C line.... Maybe if the "C" boost curve and the DME load limits are designed to be the same, they just blurred the concepts in the description, but that doesn't sound very German engineer-like. Also, do we know if the DME waits 3 seconds as described before it cuts power? That seems like something to avoid triggering on a boost spike, and weird that they'd let the motor go for 3 seconds if it's running beyond the map's range. Where's John when we need him anyway?

Oddjob

For a stock car, the DME overboost protection hits first. I've only had the KLR overboost protection kick in when the KLR pressure sensor fails, The DME overboost protection is the hard fuel cut off stumble. There is no KLR blink code fault generated with the DME overboost protection. The KLR overboost kicks into limp mode and gives the blink code fault.

divil

Interesting! So are you saying you've had DME overboost kick in on a stock car?

Oddjob

The k26/6 car overboost protection is really conservative. Doesn't take much to trip it. Maybe as little as 1 psi over bone stock. So cat out and with just a standard washer shim in the waste gate I would hit the DME overboost protection in 4th gear.

The k26/8 car (Turbo S) overboost seems to be mapped higher on the factory chips. I've run higher boost on those cars (not on purpose) without hitting the factory DME overboost protection.