Supercharger bypass valve operation 5

[jcd06]

Roots superchargers usually have a bypass valve that is open at idle.

I want to start playing with the electric bypass valve that came with my Mercedes kompressor
(mentioned here: thread71-205859 )
and I am wondering which strategy to choose. I can make it TPS or MAP dependent.

I was wondering when it should close.
Should it close gradually from above a certain level of inlet pressure on, or does it shuts close over a fairly short stroke of throttle, located somewhere above the cruise position?

 

[patprimmer]

Roots superchargers usually have a bypass valve that is open at idle.

No they don't.

I want to start playing with the electric bypass valve that came with my Mercedes kompressor

WHY.

Did MB get it that wrong.

Alterations to this may have a deadly impact on safety as discussed in previous threads.

Regards
Pat
See FAQ731-376 for tips on use of eng-tips by professional engineers for professional engineers

 

[rmw]

Is this a turbocharged and blown engine? I am familiar with a bypass arrangement on the old 2 cycle Detroits where at cruising speeds the turbo would put up more air than the blower could pass so there was a spring loaded bypass around it to let the extra air get into the airboxes. Only when the pressure between the turbo and the blower started building up would the spring pressure be overcome but what you are describing and wanting to do is diametrically opposite to that situation.

rmw

 

[jcd06]

Quote:
No they don't.

Well, err mine has. It's from a MB M271 engine where the throttle is downstream of the charger. Airfilter - AFM - charger with bypass - intercooler - throttle - inlet plenum.
Now that charger is on a Nissan CA18DET engine but with turbo removed.
Driver fro the bypass had to be built from scratch.
As far as I could find out, superchargers are bypassed at idle mainly for sake of fuel economy.
Fourth and fifth generation Eaton chargers have the bypassvalve incorporated.

http://www.capa.com.au/eaton.htm

http://www.magnusonproducts.com/mp1125thgen.htm

Any idea if such bypass closes gradually or not?

Regards,

Jean

 

[turbomotor]

Lysholm compressors (screw compressors) employ internal compression (the trapped volume becomes smaller). Hence, a bypass will not reduce the work requirement as it does in a roots blower (an air mover with no internal compression).

 

[j79guy]

Back to the OP, mechanically operating the bypass valve, obviously is the most simple method, but can lead to some abrupt boost pressure changes, and thus effect driveability. In a racing application this may be acceptable, but for daily driving, it can get annoying. The Ford Thunderbird 3.8L supercharged cars had pure MAP actuated bypass valves, and yes, the boost rise was fast giving great throttle response, but you could quite perceptably feel the boost come on, and the cars were a bit "lumpy" to drive. The ultimate would be to use a pure microprocessor controlled bypass valve and the necessary programming to control it to give smooth operation, and yet reasonable throttle response, but in your particular application you don't indicate the end-use, so I can only relate what experience I've had with Eaton superchargers that have bypass valves. A pure MAP controlled valve is stone simple, but you may have to slow the valve rate a bit by restricting the sensing line. Yes, the newer cars' bypass valves tend to close "slower", but this is a function of several parameters that the VCM monitors to give the smoothest torque rise.

j79guy

 

[70btdc]

JCD06

The bypass should normally be operated using a pulse width signal to give proportionality. The TPS and MAF sensors can both offer an electronic signal that can be sampled and converted into a duty cycle by a simple chip and circuit, or you could use an already interfaced chip with PWM functionality. It would then be up to you how to control the valve. Using a linear coefficient over a certain rpm range and TPS signal would be the easiest. Closing the valve at very low MAP levels would offer the supercharger to work when it is needed, but the valve must then stay closed as the pressure rises. In reality the control is a mix of these two plus other factors.

The old MR2 SC just operated the valve when MAP determined. It was also fitted with a clutch.

Alternatively, applying full voltage to the actuator will hold it closed. Using this method would result in a very noticeable moment when the gate changes position and would probably break something eventually.

Where is your throttle placed? Do you use the electromagnetic clutch? What is the car used for?

Pat, which charging systems are you referring to? All systems use this bypass open at idle, if the SC is clutched.

 

[patprimmer]

I'm talking about all those systems where the supercharger is not clutched and where the throttle is in the inherently safe position upstream of the supercharger.

A roots blower does not consume much power when it runs in a manifold pressure that is typical at closed throttle.

The bypass valve synchronisation problem does not exist if the blower is not clutched and there is no bypass.

The bent throttle shaft problem does not exist if the throttle is upstream.

I can understand OEM investing huge resources and going to all the trouble to squeeze some extra power out of a restricted space and still retain reasonable economy, but a hot rodder chasing maximum power is really wasting a lot of time and effort and really decreasing safety to save a minimal amount of fuel.

To be frank I really can' be bothered contributing to this.

I also noticed the vast majority of real professional knowledgeable full time working professional engine designers are not bothering with this thread. You might ask why.

Regards
Pat
See FAQ731-376 for tips on use of eng-tips by professional engineers for professional engineers

 

[izzmus]

If one absolutely must put the blower upstream of the throttle, you may do well to copy the throttle body setup as used by VW on the G60 engine. (The G60 used a remotely mounted positive displacement supercharger)

They employed a bypass throttle that was mechanically coupled to the standard throttle plate. When the throttle is closed, the bypass is wide open, and vice-versa. It also appears that the butterfly is biased to blow open with pressure, if the linkage should disconnect. Certainly, they trusted its operation enough to put it into mass production.

 

[patprimmer]

I just looked at the supercharger on a GM 90 deg V6 Ecotech.

It had no clutch nor visible by pass. I think quite a few of them are out there also.

Regards
Pat
See FAQ731-376 for tips on use of eng-tips by professional engineers for professional engineers

 

[70btdc]

All systems use this bypass open at idle, if the SC is clutched.

Pat said:

To be frank I really can' be bothered contributing to this.

I also noticed the vast majority of real professional knowledgeable full time working professional engine designers are not bothering with this thread. You might ask why.

Pat said:

It had no clutch nor visible by pass.

Come on Pat, the original question has some interest and valid discussion. If you only wish to defend your original unhelpful snappy attitude by backtracking then what you post does not assist the persons original question, does it?

The GM blower has neither clutch or bypass.

The VW G-lader system gives good tip-in as the pressure is fed in smoothly, but using electronic controls can give alot more accurate control over the air path. Some of the G-lader positive pressure is 'wasted' in pusuit of smooth transition, more than would be seen using an electronically controlled bypass. Although if the point of the system on the OP's project was complete driveability at the expense of some outright performance then it may be a viable approach. The Mercedes bypass on this particular model is electrically operated though, functionality that might as well be used.

Maybe the OP will chime in with some more details of the project so we can specify what would actually be required.

 

[patprimmer]

My original position was that the OP was acting on a false premise and based on that false premise was wasting his time for little gain at great cost in time, materials and risk.

I stand by that position.

My contribution is why bother.

Remove the throttle blades from the manifold downstream of the blower, add a throttle body before the blower and forget about bypass valves.

You will have a drivable high performance car with no detectable or significant response or fuel economy difference long before you could make the bypass valve work better than an upstream throttle.

I thought that was already strongly implied in this and previous threads.

Regards
Pat
See FAQ731-376 for tips on use of eng-tips by professional engineers for professional engineers

 

[jcd06]

j79guy,
What you're telling about the Ford is very useful to me. People who have an aftermarket supercharger kit on their car will not communicate easily about the imperfections and drawbacks. They tend to accept these as inherent.
In my case it is a project car, it will never be a daily driver anymore.
It should allow me to acquire experience in this domain that is not really mine.
This said, that doesn't mean that it should be lumpy

The motor shaft is connected to the valve shaft over a gear reduction and the motor has to do several turns before the valve is closed. The other way around, the spring needs some time to rotate the valve shaft and make the motor spin in the other direction.
This makes that it already has a certain delay.

70btdc,
There is a pair of potentiometers connected to the valve shaft, basically making it a servo.
The actual circuit is based on a DRV101 PWM-driver and a MOSFET final.
For the time being, I made the circuit react to the TPS voltage. So obviously it reacts differently then the diaphragm actuated valves but I plan to connect it to my MAP-sensor later and compare both strategies in terms of drivability.
But obviously the valve's "start opening" and "complete open" points need to be different for low and high RPM.
The servo approach could be referred to pressure instead of valve position too.
There is no clutch BTW.

izzmus,
I did consider moving the throttle upstream of the charger. What held me back is the intercooler that is probably not up to the task of being under manifold vacuum. It's a core with plastic end caps and it was made for positive pressure. I'll need an all-aluminium for that.
Also the presence of an OEM bypass valve was challenging
Would be interested to see how proportional the VW G60 linkage is.
I'll check out my local VW contact for that.
Installing the throttle before the charger is planned for later.
Corky Bell in "Supercharged" has a bypassvalve even when the throttle is before the charger.
He claims it would avoid the waste of power caused by the charger pulling from the throttle, even partially open.

 

[Surestick]

Is the cost of an intercooler that can handle manifold vacuum that much more than the cost of designing and building a bypass system, especially one that may affect driveability?
Additionally, what is the cost if the bypass system fails?

 

[BrianPetersen]

If you are talking about a traditional air-to-air intercooler installed in front of the radiator or some other place equally remote from the engine, then having the entire intercooler system and all connecting pipes under intake vacuum might have some interesting consequences in the event of a leak, and all that "dead volume" behind the throttle might make throttle response a little interesting.

Depending on application, an air-to-water intercooler can be built directly into the intake manifold with relatively minimal volume. It all depends on the layout of the system in question.

 

[j79guy]

For what it's worth, the Ford Thunderbird bypass type supercharger systems have the throttle upstream of the unclutched supercharger.

"I also noticed the vast majority of real professional knowledgeable full time working professional engine designers are not bothering with this thread. You might ask why."

Hmmmm, guess that's where the rest of us come in.

j79guy

 

[jcd06]

So far, the bypassvalve starts closing at about 60% TPS and is fully closed around 75% TPS, this means that it's position is dependant of TPS.
I had a look at my logfiles and found out the following: the boost builds up over a much smaller TPS range than over which the valve is closing.
This makes engine operation feel jerky and somewhat uncomfortable because boost always build up at the same point, independent of RPM and driving conditions (uphill-downhill).

So, I'll have to extend the TPS range over which the valve closes or go for the other approach.
That will be to drive the valve so as to deliver a certain boost dependant of TPS.
I'm tending to go for that second one, and in the mean time I'll foresee the necessary wiring for all the relevant parameters to be logged by one device and not by two separate like now, because that makes it awkward and unprecise to reconstitute te data.

 

[enginead]

My experience on similar systems (Mini Cooper S) has shown that the bypass (operated on purely manifold vacuum with no active control) is continually variable so even at part load there is still some positive pressure effect from the S/C. On the Siemens system used on Mini this was compensated for by using a MAP sensor before and after the S/C which then fed back into the ETC control and torque model. This prevented any drivability/surge issues as the balance of low pressure/high pressure changed through the engine speed and load range.

We tried implementing electronic bypass control in the interests of pursuing part load fuel economy and improving response but the cost out-weighed the benefit and it never made it into production.

Pat makes a good point in saying that OE's spend a lot of time/effort/money getting these systems to work so unless you're prepared to have the potential of some significant compromise you maybe better off focusing your efforts on something else.

What is it you're actually trying to do?

 

[jcd06]

Enginead,
What is so very useful in the experience that you shared here is that you mention the compensation from the two MAP sensors onto the ETC and torque model.
The ECU takes into account what the charger/bypassvalve is doing by sensing the pressure before and after the charger.
How it compensates engine parameters will be load dependent at first because these absolute pressures are load dependent.

It is to note that in principle the ETC hardware is quite similar to the electromechanical bypass valve as used on the MB engine.

The valve control that I described in my previous post and which is used now is load dependant too, albeit in a more crude way because the cable actuated throttle plays a role in it. Anyway it will be more load dependant than my first approach.

Winter was long and the opportunities to test on dry roads were very rare.
The first -short- tests with this system look promising.

The attached picture shows the path to a certain improvement.
X is MAP in bar (scaled to 0 = atmospheric), Y is TPS in volts.

http://files.engineering.com/getfil...a-8b9f-473d-956c-fb5f2e2cd8b4&file=Image4.jpg

On the second graph, compared to the first one, the gain of the MAF amplifier has been increased what makes boost build up over a wider range of TPS.
Then for the third graph, the offset between the MAF and TPS signals is reduced.
Boost buildup is more "the extension" of the manifold vacuum.
Obviously, that was the setting yielding smoothest throttle response.

What has to be refined? The slope of the positive pressure as function of the throttle position is still steeper than the vacuum as function of throttle. I think drivability might benefit from stretching even more this part of the action.

As to your question what I am trying to do I can answer you the following.
This has nothing to do with my job.
I'm working full time professionally, who knows even knowledgeably, doing something completely different. Don't ask me why. Partially it was destiny...
I have been charmed by the technique that I found when I bought this car (Nissan S13 with CA18DET engine), meanwhile ten years ago.
That was quite a difference compared to my old twin-DCOE commuter.
Since then I have walked different ways and meanwhile it became a project car.
After being tired of playing with turbo's I mounted this charger that I had available.
The challenge now is to optimise its control.
A first ECU remapping has been done last year and once I will have satisfactory results with this bypass, I plan to increase the boost.
What will be next? Who knows, I have other hobbies too, and a lack of space and time.
You will understand that given the situation, the potential of significant compromise is virtually inexhaustible

 

[pokeytemplar]

JCD06,

I am interested in what you are doing as well. I am a hobbyist that likes to think "outside of the box". I am looking into using a Whipple 5.0L supercharger and have thought the same things about the bypass valve. Naturally their solution is vacuum based (cheapest/easiest solution). Like you I feel that there has to be a better way (not necessarily cheap or easy). I was thinking of utilizing a "wastegate" that could be PWM by the PCM to vary the amount being bypassed. More info on your progress would be greatly appreciated.