Robbobnob wrote: I would imagine that this is quite achievable, giving them the benefits of stabilising the turbine speed to fluctuations in rotational velocity by dampening out the peaks and regenerating power from the smoothed response.
What type of fluctuations are you talking about? For example: the slight drop in pressure between individual cylinders firing?
Brian
Greg Locock wrote:Re tuning turbo response to pulses, I've seen it, and measured it. The turbo spins up and slows down by about 50% in response to the incoming pressure pulses. I only found this out after many months working on a related issue, it killed that project stone dead.
Greg mentions that the turbine speed will change speed by 50% during a pulse's duration.
"I continuously go further and further learning about my own limitations, my body limitations, psychological limitations. It's a way of life for me." - Ayrton Senna
I looked a bit around for pulse-to-pressure converters, which are essentially converting the pulse velocity of the single cylinder outputs into pressure by joining them into a diffuser.
According to a paper (from Basshuysen IIRC) in best case a pressure increase of 1.4x cam be achieved in the diffuser without negative effects on backpressure to the single cylinders.
I do not understand how a std turbo can have a 'significant' response to individual cylinder pressure pulses. This does not seem to correlate to the what is known as turbo lag.
This all seems less relevant when you put a motor/generator between the turbine and compressor.
brilliant question. I don't know the answer. Maybe the fluctuations in speed of the compressor wheel are ineffective in increasing the bulk flow of air?
So the puzzle remains. If the boost pressure responds to the individual exhaust pulses, that is with a characteristic time of the order of 20 ms, why does turbo lag have a time constant of the order of seconds?
Greg Locock wrote:here's a picture showing the exhaust pulses and the distinct rise and fall in boost in response to them, over the course of one cycle of the engine.
So the puzzle remains. If the boost pressure responds to the individual exhaust pulses, that is with a characteristic time of the order of 20 ms, why does turbo lag have a time constant of the order of seconds?
If you look at Fig 14 from ASME99.pdf, you can see that the mean pressure for the two configurations is quite different, with an average of about 500mbar "without capacity" and 400mbar "with capacity". The area under the lines of the graphs represents the possible work which can be done using this pressure (^= energy). This difference is the macro effect (one second lag).
The micro reaction in the 20ms cycles is then realized as a drop in the pressure difference across the turbo blades which results in a drop in the actual rpm of the turbo. This constant drop causes the lag.
I have a diagramm of a six cylinder diesel exhaust manifold (from a truck) in log form which was layed out using the spacing between the header's ports to create the puls timing. I am now wondering if it was a Mercedes engine...
Or am I talking about something completely wrong? Wouldn't be the first time...
“Strange women lying in ponds distributing swords is no basis for a system of government. Supreme executive power derives from a mandate from the masses, not from some farcical aquatic ceremony!” Monty Python and the Holy Grail
Greg Locock wrote:So the puzzle remains. If the boost pressure responds to the individual exhaust pulses, that is with a characteristic time of the order of 20 ms, why does turbo lag have a time constant of the order of seconds?
What do you mean by turbo lag?
Are you talking about boost threshold or a delay between the driver using the throttle and receiving full power, when the turbo is well within its operating range?
surely the work required to accelerate the turbo rpm eg by 5% depends hugely on the rpm ?
(at 100000 rpm a turbo will be '% accelerated' far, far less by a pulse than it would at eg 1000 rpm)
in F1 we would be worried by such rpm ripples ? (if they occurred)
they would tend to disrupt the mgu-h efficiency (with the likely mgu types)
nearer to the OP ......
when we say a 'Log' exhaust system, what do we mean ? ......
1 non pulse-conserving and non tuned length (ie pressure working the turbine) ..... or .....
2 pulse-conserving and non tuned-length (ie blowdown working the turbine)
there's a big difference
also .....
NA tuned length works by using 'negative' (sub-ambient) 'reflections' of exhaust pressure pulses
true their propagation would slightly affect the turbine
this is unimportant in a turbocharged engine as there is more exhaust energy than the turbine needs
and all SI turbo race engines have used this tuned length
though it's possible to think that 2014 F1 is/should be different ?
very short pipes will still conserve the pulses
turbo lag has little to do with the compressor/turbine response to individual pulses, and the reason is - it requires mass of gas to operate, and gas generator (IC engine) will not "instantly" increase the gas it will generate simply because someone opened throttle valve, it needs time, and time is your "turbo lag"
turbine compressor and turbine act as a restrictor in the intake/exhaust of the IC engine and this contributes to the "turbo lag" effect by making it worse (right up until the engine reaches its effective rpm range and the turbine is in its effective range, at that point the engine/turbine are very responsive to throttle even without ALS systems), to combat this, daily street cars use narrow/long intake channels and low/short cam duration to increase volumetric efficiency of the engine at low rpm - in an effort to make the engine respond to gas pedal a bit quicker - at the same time reducing the turbo lag - but this comes at a price of the top end power the engine can produce, also the turbine itself has an effective flow range where it works most efficiently, and this must be matched to the desired power curve of the IC engine
all these factors affect the final engine characteristics - the total power it can produce, the rpm range where it will be at, and the response to throttle body opening/closing - turbo lag
edit:
p.s. there was a really good documentary from the first F1 turbo era, where Keith Duckworth went over this basic stuff
jz11 wrote:
turbine compressor and turbine act as a restrictor in the intake/exhaust of the IC engine and this contributes to the "turbo lag" effect by making it worse (right up until the engine reaches its effective rpm range and the turbine is in its effective range, at that point the engine/turbine are very responsive to throttle even without ALS systems)
When you are talking about the engine being at too low an rpm for the turbo to be responsive you are referring to boost threshold, not to turbo lag.
The engine can be at maximum rpm on partial throttle and have no boost generated by the turbo e.g. revving the engine in neutral. It is only when the engine is under load that you generate boost. (ignoring ALS systems and in this case mgu-h)
in my mind the threshold and lag are essentially the same thing (and something that competition turbocharged engine doesn't care about, since it it not there, or doesn't matter), just that at low rpm the engine cannot reach the intake pressure (or more precisely, it can, but takes way too much time, and the engine rpm will rise faster) needed for cylinders to fill up properly, turbocharger itself is not the cause (it only contributes to the "problem"), its the combo of intake/head/cams/bore/stroke/exhaust that is simply is not "working" yet, and this is where you can start to talk of exhaust header length and volume + intake i/c type (air/air usually has quite significantly larger volume) designs, just maybe MB opted for not the most efficient type of exhaust and a bit heavier i/c setup in favor of smaller volumes and more responsive engine, either way, without enough facts (which will maybe revealed in couple years or so, when others catch up), this is all pure speculation, and the "lag", you're referring to, is a non-issue, it is something that a street car t/c engine might have to be dealing with
Greg Locock wrote:here's a picture showing the exhaust pulses and the distinct rise and fall in boost in response to them, over the course of one cycle of the engine.
So the puzzle remains. If the boost pressure responds to the individual exhaust pulses, that is with a characteristic time of the order of 20 ms, why does turbo lag have a time constant of the order of seconds?
The boost could vary in response to exhaust pressure, i.e. pulses. But it would seem that opening the intake cycle could also bleed off boost pressure.
