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?

 

[jcd06]

Pokeytemplar,

This project has been on hold for a while because of other priorities and I don't expect it to restart in next couple of weeks...

During summer I had the car for a series of laps on the track and I'm glad that it all held together

For increasing the boost I am hesitating to make a larger crank pulley. I might as well decide to go for a second belt and an intermediate shaft.

About your plan.
There was an article somewhere where was explained the differences between these vacuum based bypass valves and turbo bypass valves. I'm talking here about bypass valves for the compressor side where as you're talking about wastegate which is for the turbine side. Perhaps you just confused names.
Anyway, the main difference (that I can remember of) was the spring. So this might need some thinking and tweaking.

The advantage you could have with a vacuum operated valve over an electric actuator like in my setup will most probably be the response time, which I must admit is at the limit for track use.

The inconvenient will be the need for a vacuum tank but these can be cannibalised from stock cars where they are often used for various actuators, for example variable inlet runner length.
On my BMW it's called DISA and I've seen a similar thing in an Audi.

What solenoid valve will you be using for this PWM?
Those used in turbo boost controllers usually work with a frequency in the 10-30Hz range but all specs are top secret. I've always thought that a beefy fuel injector (1000cc/min or so) could do the trick. Response time of such devices is in the ms range.

Good luck and keep us informed about your project.

 

[markburlington]

jcd06,
very interesting project, giving me ideas for mine
I am working on fitting a Mercedes/Eaton M62 supercharger to an old Triumph TR6 engine
This supercharger, from a 2008 car, has a PWM controlled bypass valve and no clutch.
The bypass valve has 2 potentiometers, one to feedback the position of the actuator and the othr to feedback the actual position of the valve.
I am going to control the car ignition/fuel with a 'Megasquirt III' ECU. This kit ECU has progammable PWM ouputs.
I just have to decide which parameters to use to control the bypass valve position. A combination of TPS and MAP, but not sure what algorithm to use.
I look forward to seeing your progress

 

[jcd06]

Mark,
Making this work with a Megasquirt is a nice project.
I've been looking at the Megasquirt ECU at its very beginning but I didn't follow it up. It seems to have very useful features like PWM outputs as you mentioned. I don't know what is the power capability of these outputs. In my setup the valve actuator motor has a resistance of 2ohm and it needs about 2A to open completely.

We might just have the same bypass valve but what you're telling about the potmeters confuses me.
Before mounting it, I have opened the actuator housing and as far as I can remember both potmeters were on the same shaft, one having a longer stroke than the other, probably for some kind of a fail-safe system.
Unfortunately I did not take a picture of the internals, I'll make sure to do so whenever the charger has to come off the engine.
I added a picture of the assembly below

In terms of drivability I have found the pressure/TPS control parameters giving the best results for the valve control.
You can see a couple of graphs in an earlier post. As my approach is completely analog it would drive me too far to define an algorithm but I think the graphs speak for themselves.

I still have some reserve about this approach. When accelerating, for the short time that the bypass is partially closed, the valve is regulating pressure and I am not sure about the practical inconvenience this might bring. For that very short period, the engine charge is air that has been bypassed around the compressor and so far I did not pay too much attention at its absolute temperature. A temperature peak high enough might as well start a pinging in the engine.

I want to use a faster logger and check for unwanted temperature rise during this transition.
With a faster logger it will also be possible to monitor the valve reaction time and perhaps try to find a way to speed it up. For normal driving its response is fast enough but for track work it is a noticeable -although not insurmountable- lapse but for a moment I thought I had reinvented turbo lag
It's not at all sure that this is caused by the valve response, it might as well be the volume of the intercooler and ducts being taking time to get pressurized or just a combination of both.

Is there a place where I can follow the progress of your project?

 

[markburlington]

We have indeed got very similar superchargers and bypass valves = hopefully you can see by my photos. Mine has the same electrical characteristics that you quoted. I was mis-informed about the potentiometers - I took mine a part, as you can see, there are 4 brushes for the pots connected in pairs, shorting the brush for track 1 to the brush for track 2 and likewise for tracks 3 and 4. They are wired so that when one pot gives a max voltage, the ather gives a min, I dont know why. I plotted a graph of the 2 pot values for the complete range of movement of the valve. I am having difficulty accurately controlling the valve. I have built a 2A PWM test circuit, but the positioning is very vague. I will try a different frequency as the 12KHz I have at the moment is probably too high, although DC seemed vague too.
Do you drive the valve with PWM signal or DC and do you find it accurate?
The Megasquirt has a built in fucntion for control of a turbo charger valve which would achieve the same result. The position of the valve can be defined by RPM and Throttle position and you can set a max boost pressure.
I have contacted Bosch in England for some info on the valve and the guy there has in turn applied for info from Germany as he didnt have the data due to it being a 2009 part number. I will pass on the info when I get it.

We haven't progressed much recently due to ill health, but hopefully we will get stuck into it again soon

Mark

 

[markburlington]

Some more photos and data

 

[jcd06]

Hi Mark,

Thanks for the pictures, indeed the chargers are similar.
As for the potmeters, R1 and R2 are 270? and the pots themselves are 1k?.
The reason why there are two pots is most probably for redundancy.
I'll draw a sketch of the circuits used. I'll do this when I'm back home end of April. The first one I built was TPS based and in fact nothing more than just a servo drive. Using an external pot to simulate TPS gives a circuit that allows to check the working of the valve on the bench. Throttle flap follows the pot angle. One wonders why Japanese car builders have problems achieving this.
The PWM driver is the DRV101 with a 24kHz clock. I can't see why 12kHz would prevent accurate positioning.
The valve control was very good in my case but this was probably because I was using other parameters than you.

As you can see in my previous posts, positioning the valve as function of the throttle position did not result in smooth engine operation.
What works well is boost as function of the throttle position. Here absolute valve position is not important but on the other hand in this case the amplification of the differentiator is more important. If it has a wrong value, the reaction to the throttle will be vague (under reacting) and not completely closing the bypass, or oscillating (overreacting).
Unfortunately I can not give you a value for this amplification factor.

I seem to remember aftermarket stand-alone boost controllers drive their solenoid with a rather low frequency. I don't think it would be a problem to drive this bypass valve with a low frequency.
I hope for you that the Megasquirt is flexible and transparent enough.


Train dogs,
Teach people.

 

[jcd06]

Hi Mark,

I attached the schematics in pdf format.
TPS.pdf is the one I first used for controlling the valve. Although the positioning was very clear and reproducable, this was not the good way to go in terms of drivability. On the drawing I replaced the TPS by a 1Kohm pot on the bench and this makes an easy testcircuit. I must admit that I have been a bit sloppy with taking notes so it might be possible that you'll have to swap the connections #2 and 3 of the valve pot.
MAP.pdf is the circuit that is currently in use.
P1 is set at 13.5Kohm and P2 at 3.7Kohm to give the course from the lower graph attached to my post of 5 May 2009.
MAP sensor gives an output voltage in volt for absolute pressure in bar: V = 0.5 + (Pabs / 0.6)
In this circuit there is no feedback from the valve position. It controls pressure as function of TPS. I connected the voltage from one of the valve pots to a LM3914 LED dot display driver so at least I have a visual indication of the valve movement.
Let me know if you have questions.

I just realised how long this project has been going on and I hope to be able to do some further testing before summer because after August there won't be much time for this kind of fun.
From the calculations I have done sometime ago I could conclude to have reached the limits of this charger. I'll see if discharge temperature will confirm this and if necessary I might go for a LYS1600AX.

Keep us updated.
Regards,

Jean


Train dogs,
Teach people.

 

[jcd06]

and the other attachment

Train dogs,
Teach people.

 

[type49]

HI just came across this excellent site and this very useful post
I have a audi tvs charger that also uses an electronic bypass that i was going to try to convert to vac control
the tps feedback looks to be magnetic not a swiper , so is a bit different
Now i have found this i just require some one to translate as i have not a clue when it comes to electronic's
I have a DTA S60 ecu but i dont think the low current ‘simple’ PWM output will power this ? , i was also informed I would require a H bridge but looking here maybe not ?
can an add on module be built that could power this and reference the dta tps / map control ?
regards bob

 

[jcd06]

Hi Bob,
I'm not sure I understand your question and the reason why you want to modify.
Do you want to convert the valve actuator to a diaphragm one or do you want to modify the electronic drive so it takes into account the MAP signal and not only the TPS?
Regards,

Jean


Train dogs,
Teach people.

 

[type49]

Hi Jean
I would want to use the electronic bypass if I can and after showing your diagrams with the DRV101 to an electronic friend ,
we have decided to use the dta ecu via turbo control but with an external drv101 to boost the pwm and with flyback protection
hopefully this will work
bob

 

[jcd06]

Hi Bob,


If the DTA ECU has a configurable analog output then it sure can drive the DRV101 input. In that case it might even be possible to incorporate a RPM component in the process.

The point along the TPS axis where the valve starts reacting could be different for low or high RPM.

Personally I don't plan to go that far. I am satisfied with the fact that I found a way to actuate the bypass valve in a more delicate way than the vacuum actuators where only the spring pretention can be adjusted.

I also realise that the latter have the advantage of simplicity and reliability.

Good luck and keep us updated.

Jean


Train dogs,
Teach people.

 

[jcd06]

Here's a brief report of the tests carried out lately.

There is no noticeable rise in charge temperature when the bypass valve is on its way between open and closed.
I feared that this would have been the case for the short time the system is operating on the slope part of the boost graphs posted earlier.
In real world all-day traffic situations this might be happening more than for a short time.
Also I just found out that on the twincharged VW TSI engine there is a "regulating flap" that is bypassing the charger and regulating the boost delivered. So it's probably not such a bad idea.

Air coming out of the charger does not exceed 90°C.
That is at WOT with the M45 spinning at 16200RPM and a boost of 10-11PSI.
I would have expected the air getting hotter when stepping into the lower-efficiency part of the charger performance map.

As for the timing related issue that's where I found it getting particularly interesting.

(average numbers, all counting from start=0)
Time needed to floor the throttle 130ms
Time needed for the bypass valve to close 215ms
Time needed for the boost pressure to rise till 95% of its nominal value in the plenum 510ms
That last one explains the lag that I noticed when testing on the track and I believe this can be improved only by decreasing the volume of the inlet path, in particular intercooler and tubes.


For reference, Kenne Bell states that supercharger reached maximum full boost in approximately the same "split second" (400ms) as the throttle being floored.

http://www.kennebell.net/techinfo/general-info/boost-vs-rpm-comparison.pdf

I could conclude that reaction time of my electric bypass setup is not bad at all and my right foot is almost supersonic because it reaches WOT 3 times faster than his
Under these circumstances I don't find it necessary to look for a way to increase the speed of the valve actuator, especially knowing that it's a straight DC motor and not a stepper (these are easier to speed up) and the drive consists of plastic gears.

An air-to-water intercooler might help in decreasing this lag because it has a lower volume.

So what's next? When I have some time I will try to fine tune the response curve of the valve (the graphs posted earlier).
Some fellow members have taken an effort in explaining that placing the throttle downstream of the charger is dangerous.

They are right.
If men didn't do dangerous things, we were probably all driving bicycles (which should certainly not be so bad after all).
I will think about building in some security systems but I'm wondering if this shouldn't be the object of a separate thread.
And last but not least, I'll look out for a larger supercharger because this one is far not dangerous enough

Jean


Train dogs,
Teach people.

 

[type49]

Hi jean
I have the Audi V6 TFSI , TVS charger as shown here with in built intercooler

http://www.fourtitude.com/news/publish/Audi_News/article_4110.shtml

this uses a half open boost valve to get its required 10 psi
after market tuners have cracked the audi ecu to close the valve further hence more boost / more power
sadly I may not be fitting this due to frabraction required and where I live ,easy option would be to fit a rotrex
but I will see if I can get this cut and welded and mounted before i give up
the dta should be able to regulate with rpm via turbo control but will be protected with a flyback diode and use an external pwm
I will let you know what happens
bob

 

[vertolot]

Hi Jean,

I'm in the process of fitting my old Porsche 944 -86 (2.5 litre engine) with an Mercedes Eaton M65 (looks the same as in your pictures) with the same kind of electrically controlled (Bosch) by-pass valve. I want to buid the same kind of circuit that you posted with the signals from the TPS and MAF. I'm not to good on electrics so could you give me a little more information how to build these application and what I should take into account?

Thank You,

Frank

 

[jcd06]

Hi Frank,

To be honest if you don't have enough information with a schematic, I don't think you should try to build the circuit. Another engine will have a different behaviour and might need tweaking of some components values. Also you will be using a different MAP sensor and who knows is the TPS fed with another voltage as the 5V on mine.
What I want to say is that there is much chance that even if you blindly copy my circuit you will need to have the insight on how it works to be able to adapt it to your setup.
If you see when this thread has been started and the frequency of my updates you will understand that I am not able to give you much support - just because I don't have enough time, sorry.

I must admit that this was my simple approach to the challenge of making this thing work on my engine. From the electronics point of view you could have a far more performant drive system in terms of stability and ease of adjusting with a more modern design, PIC controller or the like. That's not my branch of electronics, otherwise I would have gone that way myself.

For my new project I will be using this same bypass valve again on another charger but I will make sure that a classic vacuum operated one can be fitted in as well. I have to build the intake and outlet piping from scratch which means there is the opportunity to do so.

Good luck,

 Jean

Train dogs,
Teach people.

 

[vertolot]

Jean,

Thanks for your reply. My car have a 0-5 Volt linear signal on the MAP and a 0-5 V signal on the TPS as well. As you have 2 adjustable resistors P1 and P2, are you then able to tune the caracteristics of by-pass valve? How have you physically installed the circuit?

I understand that this circuit will not work on my car necesserly in the same way than your car. But I really like the idea of having this bypass function and that is electrically controlled. So this is the only option I have at the moment because I want to use the original bypass function.

 

[jcd06]

Hi Frank,
The circuit is built in a plastic box laying on the floor at the passengers side. It's connected to the valve with a shielded cable and powered from a plug I installed for that purpose.
The resistors P1 and P2 are 10-turn Bourns potmeters with a dial knob on the front of the box.
Increasing the value of P1 will lower the point on which the bypass starts to close when opening the throttle.
Increasing the value of P2 will lower the angle of the slope but will also increase the point on which the bypass starts to close when opening the throttle. (referring to the graphs in my post of 5 May 09)

Depending on what the car is intended for, you might as well control the bypass valve motor with a switch or a TPS threshold. That will be easier but less comfortable for a street car.

Meanwhile I came to the conclusion that this valve has a relative large diameter compared to other butterfly bypass valves used for this purpose. I have no idea why it has been made like this.

The exact location of the MAP sensor on the plenum or the length of the hose connecting him to the plenum has an influence on how its output signal is a smoothed image of the absolute pressure which is pulsating. Rough idle with aftermarket ECU's can sometimes be cured by intervening on this point.
I expect it to have an influence on the working of the bypass valve driver as well. I haven't tested this yet but I'll post it here when I have results. That will not before spring anyway.



Train dogs,
Teach people.

 

[1gibson]

Hope I'm not just stating the obvious, but Eaton style bypass valves can be physically adjusted up/down, which affects their function. Push down, boost comes in sooner, push up it comes in later.

Loosen the mounting screws to the blower, there will be a little bit of play. Push all the way up/down, tighten screws, test drive with your calibrated seat-of-the-pants meter (and hopefully a datalogger or scantool.)All the way down can lead to boost before open loop / power enrichment, and cause pinging.

"Egging out" the mounting bracket for more adjustability might not be out of the question.

This is based on experience with L67 series 2 / L32 (GM's 3.8, with M90)

 

[1gibson]

This might be useless (very vehicle specific) but I'll mention it anyway: PCM for those vehicles has a parameter to command bypass open below 1% throttle (comes into play in decel situations.)

If this parameter is changed to say 25% throttle, it will command the valve open until that throttle % is reached. If you physically move the bypass valve down as I described above, and set aggressive boost ramp rates (if you're able to alter the original PCM programming) then the overall effect could be what you are looking for.

This obviously relies on access to the PCM software, which if you had, an external control box wouldn't be needed in the first place. So probably not relevant, but still interesting stuff.