When you click on links to various merchants on this site and make a purchase, this can result in this site earning a commission. Affiliate programs and affiliations include, but are not limited to, the eBay Partner Network.
You should ask:
- what is the best setup for ~330 hp for a 1.8L gasoline engine that makes 140 hp when stock / naturally aspirated for a (street/track) application, at a constant full boost of 2.2 bar of MAP?
Last edited by JasonC SBB; 01-08-2018 at 09:49 PM.
You should ask:
- what is the best setup for (insert hp target here) for a 1.8L gasoline engine that makes 140 hp when stock / naturally aspirated for a (street/track) application ?
I understand you completely, because you make sense.
Me, **** sense.
I want the challenge. I want to play. I want to think about this till steam comes out of my ears.
I do not want to be sensible. Uncharted waters are excite.
When all is said and done, I will be making something between 200 and 250 Hp at the wheels. That is it.
But, it will be fun.
(If I had access to 6 speeds and spare diffs like you guys over there, I'd maybe chase more power, but you just can't get that stuff over here.)
These turbos are not unusual to anyone with diesel experience. Caterpillar plumbs some of theirs very similarly on larger engines than this. Regular over the road dump trucks have them. Some of the valving is simpler on the ones I know. Two instead of three.
"If this system is for a diesel, will it be happy with the high EGTs of a boosted miata motor? (upwards of 900°C) Diesels have much lower EGTs." Not a problem. I have access to literally thousands of impellers - they can put together whatever is necessary, at no cost.
--- It's not just the turbine wheels that need to withstand the high EGTs, it's also the center section / bearing / seals / oil cooling, and the turbine housing itself (hotside snail)
You will most likely need a pressure sensor between the 2 compressors, and probably one at the throttle body inlet.
The HP turbo is the smaller, low power turbo, which sits at higher pressure, thus the name. The LP turbo is the bigger, higher power turbo that sucks in air from, and exhaust to, ambient.
The strategy is to run both turbos at low RPMs to get full boost at as low an RPM as possible. At low flows turbos have difficulty generating much boost; when the turbos are in series their pressure ratios multiply.
Looking at the maps in the PDF (pages 10,11,12), at lower RPMs the control is exclusively done by the HP wastegate. The HP compressor bypass valve is closed of course, and the LP turbo wastegate (which is also a bypass valve) is closed, so both turbos are running in series.
Above some amount of airflow (power), the HP turbo is bypassed quite abruptly (both the compressor bypass and its wastegate are opened), and then boost control is performed by the LP wastegate.
The switchover point is as a first approximation some amount of airflow (approximately MAP x RPM); it appears to roughly coincide with the point where the HP wastegate is close to fully open. I guess you blend it over, and when the bypass valves of the HP turbo are almost fully open you then blend in LP wastegate operation. Maybe you don't need the control strategy to use the pressure information between the 2 compressors, if there's no chance of either one overspeeding using the above strategy.
No, don't use electrical actuators for the wastegates. It might be easier to find an electric vacuum pump and regulator.
Last edited by JasonC SBB; 01-09-2018 at 02:32 AM.
Jason I think you have all your answers in this document:
OK I went through that.
The diagram below says that you do a mix of control of both wastegates at lower power; at higher power the HP turbo is abruptly completely bypassed.
I suppose that the blending of the 2 wastegate controls tries to target a certain pressure between the 2 compressors. This was why I originally suggested a pressure sensor there.
I read a bunch of literature and tried to grasp (and visualize) the handover between the two turbos under varying conditions.
It seems controlling can be achieved several ways:
A- Vacuum pump, reservoir, pwm driven vacuum solenoids, driver, control unit.
B- Vacuum pump, reservoir, simple vacuum valves, on-off signal from MS
C- Boost actuated controls (wastegate actuators) of different calibration, manifold pressure dependent control
D- Boost actuated linear motors or servos, either pwm or simple voltage driven, driver, controller.
Option A is what Mercedes is using. I would need a dedicated controller with custom code. Forget this one.
B would be inefficient and a nightmare to get it right under varying driving conditions.
D would require an Ardunio setup at the very least, and that means learning a new programming language, developing a custom GUI AND calibrating everything on top of that.
So, I am thinking I will go with option C.
Basically, I will use manifold pressure and feed that pressure into 3 separate WG actuators:
The "handshake" actuator will be adjusted with considerable preload so it will crack open early and be fully open at rated pressure level.
This is critical, because the first few degrees of flap opening between the HP and LP compressors actually determines the smoothness (and efficiency, and low EGTs, and avoiding excessive HP shaft speed, etc) of the handover process.
This could be achieved at around 5 or 6 PSI boost pressure. The valve will start opening at maybe 3 PSI, and be fully open at rated PSI value.
At this point, the LP compressor will be blowing into the HP. They may not be perfectly efficient on their own in this configuration, but since they are multiplied and one is feeding dense air into the other, there is no loss in terms of overall efficiency and output.
Once the LP gains enough "momentum", I will open the LP bypass at around 8 PSI so it will bypass the HP and feed directly into the engine via the IC.
A really cool thing happens after the final LP bypass opens:
HP is left freewheeling with little to no load on it, and its vanes create a scavenging effect, increasing efficiency big time. ("Big time" is a technical unit, like "oodles of power")
Finally, the main wastegate opens at 14 or 16 PSI. Final system boost will be largely determined by EGT, whether or not my 5 speed can hold itself together, and whether or not I am able to keep the whole car between the white lines under heavy acceleration.
I am not pulling these control scenarios out of thin air, BTW.
I was able to get a hold of the chief engineer in charge of "Development of R2S Turbo Systems for Gasoline Engines" at Borg Warner in Germany, and spent about an hour on the phone with him.
He was genuinely interested in my project and control ideas, and gave me some pointers along the way.
His initial concern was EGT management, and he was pretty relieved when I told him I would be running about 14 to 16 PSI max.
Apparently, much higher boost levels are possible.
So anyway, I feel like I am on the right track, especially after talking to the chief engineer.
There will be much testing, many joy and a whole new stable of new ponies to play with.
Please feel free to comment and yank me in the right direction if you think I am going in a stupid direction.
What you are trying to control is the position of the actuators. Simply putting a solenoid using the HP turbo compressor outlet as a pressure source, with the computer sending a duty cycle to the solenoid(s), (standard wastegate plumbing) gives poor control over said actuator position. I suspect your system would benefit more from an improved approach.
The motion of the actuator after applying a given duty cycle is slow to reach its final position, because the air coming through the solenoid needs to fill the wastegate canister. You can think of the solenoid with some duty cycle (say 30%), as a restrictor. Let's say you suddenly go from 0 to 30% duty cycle. The restriction slows down the filling of the can. If you look at the pressure that reaches the wastegate can after suddenly applying a duty cycle, it rises up quite slowly and takes almost a second to reach the final pressure (and the position tracks that slow rise). Additionally the can pressure (and thus position), at a given solenoid duty cycle, will vary as the applied manifold pressure rises. So it adds an extra layer of uncertainty.
I built a circuit that sits between the ECU duty cycle output and the solenoid. This circuit uses a pressure sensor that looks at the pressure in the canister (i.e., between the solenoid and the canister). It takes the ECU duty cycle, and interprets it as a canister pressure command. It will then control the solenoid pulses in order to maximize the speed at which the canister pressure (and thus the actuator position) hits the target pressure command from the ECU. So if the ECU suddenly goes from 0 to 30% which may mean 0 to 5 psi in the can, it will initially keep the solenoid wide open to maximize filling speed then very quickly modulate the pulses to hit 5 psi as quickly as possible.
The result is if you change the duty cycle command from the ECU suddenly you will see the can position change much more crisply and quickly than with the ECU simply sending duty cycle directly to the solenoid. I did the tests in my garage using a garage compressor and pressure regulator as a pressure source. You could also see that the circuit solves the problem of the can position changing (with a given duty cycle form th ECU) when you change the pressure source. In control system parlance this circuit adds a "fast, nested inner local feedback loop" to the overall system, and it improves overall system performance. In my datalogs, it was easier to tune the boost control, and boost much more quickly followed my right foot, with little overshoot. I showed the results somewhere on here on mt.net. I suspect this circuit will make your complex system easier to tune and more responsive.
Lastly you'll want to use larger rather than smaller diameter cans. Larger cans are more resistant to changing their position as the exhaust backpressure changes, which applies a load on the flapper valve they're controlling. This will also make tuning the system easier.
Re: the control strategy your wrote above. It would be very useful if the engineer you spoke to would give you a typical breakdown of the map in the red region in my post #30. i.e. show the individual maps of the 2 wastegates.
Forewarning!
I have been battling a weird flu for about a week now, so my thinking has not been the clearest.
I have somehow managed to cook up a scheme to get the whole thing right.
Funny thing is, I have been dreaming about this, running different scenarios.
I am pretty dead set on choreographing this thing with a purely mechanical approach.
It is 2:27 am, I have a pot of tea in the kitchen and a clean ashtray on my desk. Best time of the day to think and solve stuff for me.
I should have things mapped out by the morning. (fine tuning of all parameters will be done on the running engine.)
One major disclaimer:
My projects usually take anywhere from a few weeks to a few months, and I post them after they are finished.
I am taking a different approach for the first time, and posting everything from the word go.
This will not go as fast as my other threads; it will rather be like a live journal of this effort.
I am being very patient with this one.
Re: the control strategy your wrote above. It would be very useful if the engineer you spoke to would give you a typical breakdown of the map in the red region in my post #30. i.e. show the individual maps of the 2 wastegates.
Cheers.
It's actually pretty straightforward. Or, maybe I am getting the whole thing wrong..
Speaking of the red zone, you are over 100 kpa, producing boost with 2 compressors.
Even though the RPMs are not that high, there is a chance the combination of the 2 turbos may produce pressure levels in excess of the designed final boost level.
That is because the LP turbo bypass is closed, and you get a true compounding effect, with a multiplied output from those 2 turbos.
Just remember the fact that the boost graph will look pretty alien, rising very early on and remaining flat all the way to redline.
It's only normal to expect the wastegate to come into play every once in a while during the red zone conditions.
As for the HP - LP bypass, yes, that is a wastegate by definition, as well, because it relieves the HP compressor.
The most important function of that bypass is protecting HP. It also serves to help spool the LP without creating appreciable backpressure.
This is what I get from all the reading I have been doing.
Figure 2.3.1: Diagram of two-stage regulated charging
The control strategy of the actuators aims to keep the charge-cycle losses of the engine to a minimum; the results shown in chapter 3 were measured according to this strategy. From the illustration it is evident that the high-pressure stage is not required from medium speeds on and only the low-pressure stage provides the required boost pressure. In order to reduce losses, the waste-gate is opened below the naturally aspirated full load line, so that boost pressure is generated above this line only.
Gents, please help me interpret this correctly...
It says "open the wastegate below 100 kPa, not unlike a supercharger bypass valve which lets the pressure generating component to freewheel and allow the engine to run in a normally aspirated manner.
Then, as load increases and the manifold pressure approaches the magical 100 kPa threshold, the spring in the vacuum actuator overcomes the substantially decreased vacuum at that point and closes the wastegate. This allows the system to generate and hold boost.
When desired boost level is achieved, the wastegate comes into play again, this time in the traditional sense, to bleed off excess pressure.
I can build the system this way.
Before I delve into the methodology, I would like to ask the collective mt.net brainpower if such an approach would be absolutely necessary.
I mean, none of the turbocharged gasoline engines I know of cruise below 100 kPa with their wastegates open, right? The thing only opens to bleed off excess pressure, right?
Will I be overcomplicating things by following Mercedes Benz's scenario to the letter and try to keep the WG open below 100 kPa? It says "a possible scenario", after all.
Now, the methodology (if I were to employ a fully open WG under 100 kPa strategy)
1. Use an EWG in addition to the IWG. This would be a preferred method IF I have to keep the WG open in cruise.
EWG would be handled in the traditional manner, with an EBC, controlled by MS.
IWG would be vacuum actuated, and pulled open under vacuum. It would close as manifold approaches 100 kPa (which is easy for me to accomplish) and be held shut real tight as boost pressure loads on the actuator's membrane.
2. Use just the existing IWG (which has a pusher actuator on it), weld another "arm" on it so a vacuum actuator can be added to work the WG arm in a seesaw manner.
Both actuators will be shaped like this to allow movement of the other:
Pin in slot actuator linkage - a very crude representation:
The idea is to is allowing both actuators do their jobs on the same WG without interfering with each other, but I realize I have to think about it further to refine proper control.
The actual wastegate valve is a flapper valve. Looks like a coin on the end of a swing arm. When open, it allows exhaust gas to go around the (exhaust gas) turbine.
The turbine generates mechanical power to spin the compressor. The available power drawn from the exhaust gas flow is approximately mass flow rate through the turbine times the pressure drop across the turbine. Opening the wastegate allows exhaust flow to bypass the turbine, reducing mass flow rate through the turbine, and also reduces the pressure drop across the turbine.
The exhaust backpressure seen by the motor, is the exhaust pressure in the exhaust manifold. It is the sum of atmospheric pressure, the exhaust system backpressure, and the pressure drops across each turbine.
More backpressure seen by the motor reduces volumetric efficiency (and thus torque), for a given manifold pressure. However an increase in intake manifold pressure will increase torque even if there's a greater increase in exhaust backpressure. But for a given MAP, minimizing exhaust backpressure maximizes torque and also minimizes fuel consumption.
The vast majority of mt.net turbo setups have a wastegate actuator that is closed at cruise (i.e. manifold pressure is below atmospheric). This is because the wastegate canister/actuator can't open the flapper when there is no boost (they don't use a boost reservoir). Under this condition there will be *some* backpressure and *some* boost in the compressor outlet (i.e. before the throttle butterfly) even if there's vacuum in the intake manifold (because the throttle butterfly is partially closed). Most OEs *do* open the wastegate when not in boost, because this reduces said backpressure and improves fuel economy. The shaft will freewheel, and the turbine will generate very little backpressure. The disadvantage is when you suddenly floor it, the compressor isn't spinning as fast and will take slightly longer to hit full boost. OE's usually use vacuum operated actuators and a vacuum reservoir.
You can perhaps use the less-common 2-port canisters - applying vacuum on the "other" side of the diaphragm will open the wastegate. Boost on the "normal" side will open the wastegate, which is how a typical 1-port wastegate canister works.
Re: IWG vs EWG. I presume you are thinking internal vs external wastegates. The above text discusses the actuators. It doesn't matter if you're talking IWG vs EWG. The text applies to both. Adding an EWG does not solve any problems for your setup. EWG's are messy to plumb in and expensive. And I'll bet B-W's (internal) wastegates in your setup are awesomely engineered. Adding EWGs would be step down.
The actual wastegate valve is a flapper valve. Looks like a coin on the end of a swing arm. When open, it allows exhaust gas to go around the (exhaust gas) turbine.
The turbine generates mechanical power to spin the compressor. The available power drawn from the exhaust gas flow is approximately mass flow rate through the turbine times the pressure drop across the turbine. Opening the wastegate allows exhaust flow to bypass the turbine, reducing mass flow rate through the turbine, and also reduces the pressure drop across the turbine.
The exhaust backpressure seen by the motor, is the exhaust pressure in the exhaust manifold. It is the sum of atmospheric pressure, the exhaust system backpressure, and the pressure drops across each turbine.
More backpressure seen by the motor reduces volumetric efficiency (and thus torque), for a given manifold pressure. However an increase in intake manifold pressure will increase torque even if there's a greater increase in exhaust backpressure. But for a given MAP, minimizing exhaust backpressure maximizes torque and also minimizes fuel consumption.
This, I know.
The vast majority of mt.net turbo setups have a wastegate actuator that is closed at cruise (i.e. manifold pressure is below atmospheric). This is because the wastegate canister/actuator can't open the flapper when there is no boost (they don't use a boost reservoir). Under this condition there will be *some* backpressure and *some* boost in the compressor outlet (i.e. before the throttle butterfly) even if there's vacuum in the intake manifold (because the throttle butterfly is partially closed). Most OEs *do* open the wastegate when not in boost, because this reduces said backpressure and improves fuel economy. The shaft will freewheel, and the turbine will generate very little backpressure. The disadvantage is when you suddenly floor it, the compressor isn't spinning as fast and will take slightly longer to hit full boost. OE's usually use vacuum operated actuators and a vacuum reservoir.
I really did think about this and wondered why Mercedes would sacrifice performance by opening the WG below 100 kPa. Then it hit me.. This system delivers full boost at 1000 RPM on their engines. They are not worried about a tiny bit of throttle response delay. Emissions gains and fuel economy are more valuable for them.
You can perhaps use the less-common 2-port canisters - applying vacuum on the "other" side of the diaphragm will open the wastegate. Boost on the "normal" side will open the wastegate, which is how a typical 1-port wastegate canister works.
I will have to look around for one of these. Haven't seen one so far, or I am not looking in the right places.
Re: IWG vs EWG. I presume you are thinking internal vs external wastegates. The above text discusses the actuators. It doesn't matter if you're talking IWG vs EWG. The text applies to both. Adding an EWG does not solve any problems for your setup. EWG's are messy to plumb in and expensive. And I'll bet B-W's (internal) wastegates in your setup are awesomely engineered. Adding EWGs would be step down.
Yes, External WasteGate... If I have to use a second WG to duplicate the MB graph in my previous post, I pretty much have to use an EWG. I have no other place to install one inside this unit.
Are they really that expensive? I have seen some anywhere between 50 and 100 bucks. Also, some boost line, a solenoid on it and a bit of wiring to the MS is really not messy.
Or, am I missing something here?
The above talks about the black region in your last post. It's about fuel economy at cruise.
The red area is crucial for response and performance. I would like to see a more realistic breakdown of the red area, for each of the 2 wastegates, for a gas engine.
The first PDF file you linked (post #16) does have them, but it contradicts 2 statements. Albeit this is for a diesel, not a gasoline engine.
Remember 10 bar of BMEP is approximately 100 kPa of MAP.
The first statement is that the wastegates are open in vacuum (below 10 bar BMEP). This is a minor point as you can implement this if you want to.
The second contradiction is that the below map shows that the LP wastegate is completely closed until 3000-3500 RPM. There is no "blending" where both wastegates are partially open below 3000 RPM.
Ergo, the HP wastegate appears to be doing all the boost controlling below 3000-3500 RPM. This is different than the strategy you typed wherein the HP wastegate opens at 3 psi of boost.
What would be very useful is a typical map of the pressure between the 2 compressors as a function of BMEP and RPM for a gas engine. If resembles a simple function of MAP and/or RPM, then perhaps you can build a simple, separate controller for the HP wastegate. It can be done with a pressure sensor and a microcontroller like an Arduino or perhaps a small analog circuit.
The second file (https://www.miataturbo.net/attachmen...ry_532_626.pdf) seems to offer clues. See below graphs.
The issues are it only shows what happens at full boost, and it doesn't show what the wastegates are doing.
Based on the curves:
- full boost is reached at 1400 RPM
- the engine "comes on cam" from 2500-3000 RPM, which is why the target MAP drops; looks like they're trying to maintain a flat torque curve from 1400-4000 RPM.
- the 2 curves suggest to me that from 1400-3000 RPM the LP wastegate is completely shut and boost regulation is done by the HP wastegate. This is consistent with the wastegate maps above, and inconsistent with the statement "HP wastegate opens at 3 psi, regardless of RPM". It looks like it operates at all boost targets at <2500 RPM
- there is no blending of the 2 wastegates.
- above 3000 RPM the HP turbo is completely bypassed, and the compressor bypass is also opened
I would think at part-boost operation the strategy would be the same, but the HP wastegate would fully open (and LP wastegate start opening) at a higher RPM (based on the diesel wastegate maps - first 2 maps of this post).
The entire discussion above is about steady-state operation. No talk of transients, i.e. spoolup (WOT but not yet at full boost).
When spooling generally you keep both wastegates shut for maximum shaft acceleration. But, for high RPM spoolup I suspect you would need to begin opening the HP wastegate at some point to prevent the HP turbo from operating outside its most efficient region (else spool might be slowed) and to prevent shaft overspeeding. This is why I'd like to see the wastegate maps and a map of the pressure between the compressors, with BMEP or MAP on the Y axis, and RPM on the X axis.
-----------
Compressor pressure ratios at full boost:
EWG's have the exhaust valve *and* the canister/actuator in them. You have to do exhaust plumbing.
Your system has the exhaust flapper valve(s), and you just need the right canister to actuate said flapper.
A 2 port wastegate canister/actuator looks like this.
One port (left one) pushes (opens flapper, moves rod to the right) with boost. The other port (right) uses boost to close (pull, moves rod to the left), or vacuum to open. Inside is a spring that preloads (pulls, moves rod to left) flapper shut.
Jason, regarding post# 38, there is a big misunderstanding.
I will write in detail as soon as I am able to sit down by the computer in a quiet / uninterrupted setting.
01-08-2018, 07:43 PM
0
