[sapl]

Originally Posted by tsk1979

Let me get this straight.
From what I understand about ECUs in general, even with a piggyback ECU, everything related to fuel delivery and engine management is done by ECU.
All the piggyback ECU does is "lie to the ECU" to trick it to do something else as opposed to what ECU would naturally do.
A remap would mean rewriting the software inside ECU so the ECU itself does different things than what the mfr intended it to do?

Am I correct?
Or do Piggyback ECUs directly control the fuel pump/injectors etc.,

You are correct!.

However, some piggyback systems reach out a little more. For example in the Unichip Q/Q+, you can control extra injectors, or overdrive injectors. You can also control a pump if you wanted to.

[sapl]

Originally Posted by Sutripta

I take it that this 1 sec period is essentially independent of engine load/ speed? (Assuming that it remains in closed loop)
It would be interesting if one were to also fit a wideband, and output its value on the second beam. And other parameters, one by one.


Why don't we have an architecture where the piggyback sits between the engine/ environment and the stock ECU, monitoring all inputs in parallel with the stock ECU (except, I suppose lambda), and intercepting all outputs. Those which are not its concern are passed through as is. Those which are (esp. injector PW) are modified. (The delay introduced might be a problem, but given the speed difference between electronics and mechanics, surmountable). Will we not get more flexibility this way.

Regards
Sutripta

It may not be independent of speed/load, because the speed with which the NBO2 sensor receives the exhaust gases with the modified AFR could depend on things like engine speed, location of the sensor, geometery of the engine, and of course the ecu's response characteristics.

Different cars havbe different waveforms for the NBO2 signal. See another car's signal below - notice its shape.



And about that idea of yours about the WBO2 (or other signals) and NBO2 together - will do so and post you the images when I get to doing it !

On your second suggestion, whether we need to monitor ALL inputs and outputs is debatable, because it makes the installation complex and unwieldy. However, it is possible for the Unichip system to override things like injector pulsewidth. We do use this in some applications.

[Born2Slow]

Originally Posted by sapl

The first part:
In open loop conditions most cars switch to very rich AFRs, since it is always safer for the engine to run rich rather than lean. We have seen AFRs as rich as 9.5 in some Honda Accords at full throttle, 5000+ rpm.

Sorry .

I am in no position to question a professional tuner but with whatever limited knowledge I have 9.5 AFR looks very rich to me. I know there are many other factors involved which I may not be aware of but still why should Honda run the engine so rich when 12-13 is mostly quoted as the best AFR for making power. Please dont tell me its a safeguard against leaning out. Does lean AFR damage an engine running mild tune, like blowing off the head gaskets and spoiling the pistons etc I know you guys are used to engines producing 800 bhp from 2L but here we are talking about 3.5 L engine just producing 270 odd horses or less.

[sapl]

Originally Posted by Born2Slow

...... 9.5 AFR looks very rich to me. I know there are many other factors involved which I may not be aware of but still why should Honda run the engine so rich when 12-13 is mostly quoted as the best AFR for making power.

I have no idea why it is running so rich - it is only a report of what we observed on a Honda before the Unichip was fitted.

[tsk1979]

Originally Posted by sapl

You are correct!.

However, some piggyback systems reach out a little more. For example in the Unichip Q/Q+, you can control extra injectors, or overdrive injectors. You can also control a pump if you wanted to.

Now this has me confused. You mean the Unichip directly connects not only to ECU input ports, but also to the input ports on the control lines of injectors? If that is the case, how does it interface with different injectors of different companies.
Interfacing with ECU is easier due to standardizing, but interfacing with injectors and fuel pump directly?

[sapl]

Originally Posted by tsk1979

Now this has me confused. You mean the Unichip directly connects not only to ECU input ports, but also to the input ports on the control lines of injectors? If that is the case, how does it interface with different injectors of different companies.
Interfacing with ECU is easier due to standardizing, but interfacing with injectors and fuel pump directly?

The Unichip CAN connect to the injectors/pump directly if need be for a particular application. We generally decide the installation strategy after studying the car. There is no "one size fits all" type of installation.

In some cases, we may need to drive one or more extra injectors (e.g. turbo installations); in others, we may need to overdrive the injectors over and above what the stock ecu does. The injector drivers that come in the Unichip system are capable of driving a wide variety of injector types.

Hope this clarifies.

[pranavt]

Originally Posted by tsk1979

Now this has me confused. You mean the Unichip directly connects not only to ECU input ports, but also to the input ports on the control lines of injectors? If that is the case, how does it interface with different injectors of different companies.
Interfacing with ECU is easier due to standardizing, but interfacing with injectors and fuel pump directly?

They are electrical signals, and all injectors have +ve and Gnd terminals. Same with the fuel pump.

[tsk1979]

Originally Posted by sapl

The Unichip CAN connect to the injectors/pump directly if need be for a particular application. We generally decide the installation strategy after studying the car. There is no "one size fits all" type of installation.

In some cases, we may need to drive one or more extra injectors (e.g. turbo installations); in others, we may need to overdrive the injectors over and above what the stock ecu does. The injector drivers that come in the Unichip system are capable of driving a wide variety of injector types.

Hope this clarifies.

Ok, when you connect the unichip directly to injectors, what is the role of the signal wire from ECU? Where is it connected. So unichip, or do you use some kid of router to either route ECU signals or Unichip signals depending upon need?

Originally Posted by pranavt

They are electrical signals, and all injectors have +ve and Gnd terminals. Same with the fuel pump.

Yup I understand that. I meant electrical ports. this is how we refer to them in VLSI speak.
any chip as three kinds of ports
1. Inputs
2. Outputs
3. Inouts

So when you want to get a signal out of a chip you listen on the output/inout. When you want to give a signal to the chip, you give the signal to Input/inout

That is where protocols come in. For example OBD II read protocol is pretty standardized. you just connect a OBD bus, and run of the mill software can read your chip.
Writing is usually a bit difficult. since mfrs play around with the configs. So while the interface may be standardized, the addresses of the "error codes" may be not.
Thats why world over, forums are filled with queries like I can read data from eCU, but cannot clear error codes.

[CPH]

Originally Posted by tsk1979

Ok, when you connect the unichip directly to injectors, what is the role of the signal wire from ECU? Where is it connected. So unichip, or do you use some kid of router to either route ECU signals or Unichip signals depending upon need?

Yup I understand that. I meant electrical ports. this is how we refer to them in VLSI speak.
any chip as three kinds of ports
1. Inputs
2. Outputs
3. Inouts

So when you want to get a signal out of a chip you listen on the output/inout. When you want to give a signal to the chip, you give the signal to Input/inout

That is where protocols come in. For example OBD II read protocol is pretty standardized. you just connect a OBD bus, and run of the mill software can read your chip.
Writing is usually a bit difficult. since mfrs play around with the configs. So while the interface may be standardized, the addresses of the "error codes" may be not.
Thats why world over, forums are filled with queries like I can read data from eCU, but cannot clear error codes.

I think you have opened a can of worms.

By the time you do all of this you better go with a stand alone like the Emerald, Omex or DTA, which will give you plenty of inputs and outputs as well as quite a few more mapping points and lets you set the peak rpm as high as desired, run it open loop or closed loop switching it as you wish and you can make further adjustments on the road with the free supplied software. The cost factor is not much more than the above mentioned system.

[sapl]

Originally Posted by tsk1979

Ok, when you connect the unichip directly to injectors, what is the role of the signal wire from ECU? Where is it connected. So unichip, or do you use some kid of router to either route ECU signals or Unichip signals depending upon need?

When driving the injectors, we can adopt 2 methods - either separately driving injectors (when additional injectors are used) or overdriving the injectors along with the main ecu (when we just need to extend the existing injector duration under some conditions).

The injectors are driven by small addon modules called the injector driver/overdriver modules which we use only when necessary.

As far as signals are concerned, they can all be routed through the Unichip directly - be it timing, fuelling, boost, tps etc., and each of these may be mapped independently. Even some output signals like boost control etc can be routed through the Unichip.

[Sutripta]

Originally Posted by tsk1979

Ok, when you connect the unichip directly to injectors, what is the role of the signal wire from ECU? Where is it connected. So unichip, or do you use some kid of router to either route ECU signals or Unichip signals depending upon need?

Yup I understand that. I meant electrical ports. this is how we refer to them in VLSI speak.
any chip as three kinds of ports
1. Inputs
2. Outputs
3. Inouts

So when you want to get a signal out of a chip you listen on the output/inout. When you want to give a signal to the chip, you give the signal to Input/inout

That is where protocols come in. For example OBD II read protocol is pretty standardized. you just connect a OBD bus, and run of the mill software can read your chip.
Writing is usually a bit difficult. since mfrs play around with the configs. So while the interface may be standardized, the addresses of the "error codes" may be not.
Thats why world over, forums are filled with queries like I can read data from eCU, but cannot clear error codes.

The actuators receiving 'commands' from the ECU are dumb (we are not talking of CAN connected architectures). These commands are just electrical signals (eg. the injector PWM signal), not encoded datastreams. I don't see any problem in handling them with a mux/ demux, and a linedriver (or amplifier if the signal is analog, doubtful).

Different case when an 'intelligence' (ECU) communicates with another 'intelligence' (scanner/ some other controller)

My understanding only.

Regards
Sutripta

[Sutripta]

Originally Posted by sapl

I have no idea why it is running so rich

How was it measured?

Originally Posted by sapl

It may not be independent of speed/load, because the speed with which the NBO2 sensor receives the exhaust gases with the modified AFR could depend on things like engine speed, location of the sensor, geometery of the engine, and of course the ecu's response characteristics.

......

The complete loop is formed by the sensors, ECU, actuators, and the engine itself. Each having its own contribution to the loop response. However, I think the 1Hz oscillation will be almost entirely due to a slow to respond sensor, and software (the mathematical basis of it) which does not take that into account. At say 3000 rpm on a 4 cylinder, there are 100 ignitions a sec. I don't think combustion characteristics will lead to an 1 Hz instability. Similarly, if such a delay was caused by the ECU/ actuators, the car would be undrivable.

Excuse my being pedantic, but I'm really surprised at the graph. May I know which car/ engine?

Also the TBhp watermark effectively obliterates the X-axis. I can see triggering was set to DC, rising edge. Was the beam set for normal, or auto?

For the Unichip, what processor does it use. Does it come with comprehensive documentation, and a high quality toolchain?

Regards
Sutripta

[drpullockaran]

Dear CPH
whats your opinion regarding the zener diode mod for the MAP sensor for turbo charged diesel cars. The car should behave like a normal car till we floor the accelerator. Is it true or are there other gremlins in this mod. Please note that my car is severely over fueled at WOT. Its a Cruze with the 2 liter diesel with a VNT turbo which is actuated electrically and it has a lot of reserve potential. I have procured all parts for the mod but have not had the courage to carry it out as yet. I intend to do it like the TDI club has done it.

Chevrolet Cruze Irmscher Edition Packs 186 Horsepower - autoevolution

Diodes

AN008 - How to Use Zener Diodes

AnthonyDAnna.com - Diode Mod

MAP sensor MOD... - TDIClub Forums

[CPH]

Originally Posted by drpullockaran

Dear CPH
whats your opinion regarding the zener diode mod for the MAP sensor for turbo charged diesel cars. The car should behave like a normal car till we floor the accelerator. Is it true or are there other gremlins in this mod. Please note that my car is severely over fueled at WOT. Its a Cruze with the 2 liter diesel with a VNT turbo which is actuated electrically and it has a lot of reserve potential. I have procured all parts for the mod but have not had the courage to carry it out as yet. I intend to do it like the TDI club has done it.

Chevrolet Cruze Irmscher Edition Packs 186 Horsepower - autoevolution

Diodes

AN008 - How to Use Zener Diodes

AnthonyDAnna.com - Diode Mod

MAP sensor MOD... - TDIClub Forums

It might work in some way in some ECU and it won't in a lot. Some ECU's will even shut down. However it is a time bomb. The fueling will not go correctly with it, which will in the long run lead to turbo failure.

What Irmscher has done is totally different.

[sapl]

Originally Posted by Sutripta

How was it measured?

Excuse my being pedantic, but I'm really surprised at the graph. May I know which car/ engine?

Also the TBhp watermark effectively obliterates the X-axis. I can see triggering was set to DC, rising edge. Was the beam set for normal, or auto?

For the Unichip, what processor does it use. Does it come with comprehensive documentation, and a high quality toolchain?

Hi Sutripa - brief answers below:

1. The AFR was measured with the car loaded on the dyno, 5000+ rpm, WOT with a wideband AFR meter.

2. The first graph was from a Swift with a G13B engine, the second from a Cedia with a 4G94 engine.

3. I will check on the settings - have to fish out the original waveform stored - it wil have all the associated data stored.

4. I do not know what processor it uses. It is not a DIY system. Yes, the tuning tools are powerful, but they give it only to authorised dealers.