Compound boost not as expected

[TimIacobucci]

I doubt this setup would be happy with backpressure 2:1

Jon does beautiful work.

Yes that car is just amazing. Perhaps only topped by this?



This is Kevin Jewer's compound turbo talon I was referencing.

I am only dragging it into this discussion because I believe turbo compounding is more related to twincharge compounding than any single stage turbo or supercharger system. I know that there certainly are differences but there looks to be more similarities in compounding anything relative to any other type of forced induction.

[Big Al]

Draw a quick drawing of the tripple turbo setup and a picture

itsfun2.jpg

35077340_10157574118053378_5497759413932392448_o.jpg

[ptuomov]

So the compound turbo works because the wastegate of the first (small) turbine goes to the second (big) turbine's inlet. With a gasoline powered car, you'll need a second wastegate there as well for the big turbine. How are the wastegates controlled? I would think that you could run the first, small turbo wastegate based on the small turbine inlet pressure and the second turbine wastegate based on final intake manifold boost. This way, the small turbo would after spooling see about constant turbine speed, just the mass fraction bypassed (to the second turbine) would increase. Or is there a better control strategy?

[ptuomov]

I am getting this a little bit better now. There are many ways to do this, but the best way (after thinking about it for a couple of hours) let's the upstream, small turbine exhaust to go out the tail pipe and only feeds the small turbo wastegate circuit into the downstream, big turbo. This way, both turbines see near-ambient pressure at the outlet and can therefore do work with minimal inlet pressure. This will make the engine run like with just a big turbo from the exhaust back pressure perspective.

On the inlet side, I think one can just feed it thru both turbos without much ill effect, with a bypass valve going from near the throttle to the big, upstream compressor's inlet.

The control is going to be critical. The best control strategy that I can think of is the following. Use a turbocharger speed sensor to control the small turbo, basically having a set point at a given rpm and always targeting that. The actuator for this control strategy is the small turbine wastegate and a solenoid valve. If there's room to cool exhaust gas, one could drive this off the exhaust manifold pressure. If not, use the intake manifold pressure. But the point being target the turbo speed for the small turbine. Then, run the big turbo wastegate the same way as in a single turbo setup based on the intake manifold pressure.

[ptuomov]

With all the foolish confidence of a person who's thought about compound turbocharging for a couple of hours, I wouldn't do it this way:



Instead, I would run the small turbine outlets directly to tailpipe. I'd only send the small turbine wastegate circuits to the big turbo. This way, both the small turbo and the big turbo operate at the same pressure ratio. You're only just putting more turbine capacity in line with the wastegate. You'll need big wastegate, though.

Matching the turbines and compressors for this setup with be difficult. The best combo may be a regular turbos for the small turbos and weird hybrid turbo for the big turbo with an oversized compressor that usually wouldn't be matched to that kind of turbine.

Also, doesn't the bypass valve need to go to the big compressor inlet? It'll still stall the big compressor and turbine with the arrangement of the diagram.

[user-23911]

The 2 small turbos, it would be better if one was fed from 1 and 4, the other fed from 2 and 3.

As to WHY, you can work it out yourself.

Making something on paper and really making something are 2 different things.

[Big Al]

Joe 90: It's actually is made that way, I made some errors in my drawing. The next error is that the big turbo have a waste gate.
But I don't really know if it is such a big deal with the 180 degree pulse split when you have full pulse split?
As for the wastegateing to the other half of the big turbo (the one closer to the core) it has already been tested, they made earlier a 2 stage turbo setup with the same principle and that worked just fine. 300rpm better spool that way on that particular setup.

ptuomov: I think the WG's is controlled the way you are saying in first post. Have to check to confirm.
I don't know if waste gating from small turbos to tailpipe is such a great idea, too much energy is going to waste so to drive the big turbo it becomes hard.
It could stall the big one with the BOV reconnected from stage 2, but that's nothing that they have noticed (on another setup). Only that it is easier to build boost after you let go of the throttle and then floor it again.

On a turbo compound setup the most important thing is to have enough WG flow from the small turbo. The WG should be placed so it gets most flow otherwise you are driving the small turbo too hard and gets too high back pressure. It's the most common problem when people build 2 stage turbo.

[Wolf_Tm250]

Sorry to the OP, as I will not help his issue, but go Off Topic.
Just wanted to say that it's very difficult for me to understand why you should need a compound system nowadays in a well geared car and with the new Garrett GTX2 twin scrolled.
Off Topic End.

Result was just over 700ft/lbs torque at 5500rpm, no lag at all and a very healthy spread of torque from 2500rpm up to 7800rpm. Power at the flywheel estimated at 700hp. Estimated because the dyno only gave wheel figures and wasn't setup for coast-down losses. Going by the previous visits to a different dyno, we have estimated 100+ hp and torque losses through the 4wd system.

It's incredible to drive now. From 2000rpm it pulls hard even in 5th and picks up a fair pace very quickly! Overall i think a very worthwhile modification as the car is very very quick on the hillclimb tracks. No waiting for the turbo to spool on the exit of the hairpins, it's instant torque at any revs and speed!

The last few events i've suffered clitch slip so a triple plate Superclutch has been fitted. There's also been the odd transmission breakage due to fatigue. I stripped first gear on the line and then twisted the centre diff output shaft at Doune. Both breakeages have evidence of class

From this thread, https://www.eng-tips.com/viewthread.cfm?qid=267889

That's on a 2.0l evo for reference.

Let me know which 700hp gtx does that.

Ah, yes, my bad... never thought about the case someone would want to stay in the longer gear forever and ever!

I wrote "well geared car" because I was obviously talking about race cars/race driving.
Being that said, GTX gen2 are smaller turbos with the same power output of the older next step up, with almost instant boost when you're over the spool up threshold.
Talking about the example you took, I'm pretty sure he had HUGE lag before twincharging... but that was for the very old schhol T04Z turbo
A GTX3582 gen2 or even an EFR8374 would be a completely different power band and response.
There are not so many GTX2 around as these are quite new, but few EVO already made 650/700whp with a GTX2 3582 and top boost and torque already at 4k rpm.
So i like the effort and the cheapness of this project, but my opinion is that the complexity is no more worth it

[user-23911]

Unfortunately the more complex you make something, the easier it is for something to go wrong and the end result is usually the same. It might take several attempts to get it to work properly and reliably.

The evo above....it blew up like they all do.
But then that's part of the learning curve.

[Wolf_Tm250]

Sorry to the OP, as I will not help his issue, but go Off Topic.
Just wanted to say that it's very difficult for me to understand why you should need a compound system nowadays in a well geared car and with the new Garrett GTX2 twin scrolled.
Off Topic End.

https://youtu.be/5HlqEkJ3VGM?t=3m37s
it makes a 2 liter 4 cylinder run like this

and smoke 6 cylinder turbo motors...
https://youtu.be/u9FpUeoiJXg?t=7m24s
watch the porsche and then the lancia follows, Ill take the lancia.....

Yes, I pretty know the S4... and even the Predappio's road of the video...
but, again, the twincharging was requested for those ancient school turbo... IMHO

END OF OFF TOPIC.

[ptuomov]

This is diesel oriented and heavily focused on fuel consumption, but nevertheless interesting on compound turbocharging. The three-valve system is complex, though:

https://www.gtisoft.com/wp-content/uplo ... ystems.pdf

For gasoline powered car, I still think that running the turbines in a more of sequential fashion while running the compressors in series would be the way to go, as the gasoline powered cars can't use very high boost and greatly benefit from exhaust manifold pressure below boost pressure.

[TimIacobucci]

That last article is interesting, I will read it when I get time.

Compound turbo needs the secondary turbine wastegate fed into the primary as this is how most of the work gets done for compression at higher power. It is much more efficient and effective to let the turbine speed of the secondary drop and primary take the lead after its spooled. Kevin has gone over all this with turbo shaft speed and backpressure and temperature/pressure data and different control strategies if you read through his big thread on Yellowbullet I linked to.

The 3 turbo setup, are the twins both twinscroll, so each cylinder is completely separated?

Being that said, GTX gen2 are smaller turbos with the same power output of the older next step up, with almost instant boost when you're over the spool up threshold.
Talking about the example you took, I'm pretty sure he had HUGE lag before twincharging... but that was for the very old schhol T04Z turbo
A GTX3582 gen2 or even an EFR8374 would be a completely different power band and response.
There are not so many GTX2 around as these are quite new, but few EVO already made 650/700whp with a GTX2 3582 and top boost and torque already at 4k rpm.
So i like the effort and the cheapness of this project, but my opinion is that the complexity is no more worth it

I've daily driven turbo cars with a power band like that. They are nice, I actually like that it's just a normal car under 4k, kinda of jekyll hyde ish. Have you driven a blower car? The response is just not comparable, even in boost modulating the throttle is completely different. Not to mention any low speed turns you are going to be out of boost and a 4k turbo would get absolutely destroyed in those hill climb events if you watch some of the excruciatingly slow speed corner exits.

I don't see any evidence a 35r range turbo would even be in the same league as that twincharged evo for the type of racing he does.

If you look at the older discussions about this subject and this very discussion, the control strategy is the most difficult thing and even that's been sorted out. It's just a matter of packaging everything. If you can package it, what's so horribly complicated about a mechanical blower? It's much less work and complexity than compound turbos and with better response.

[Big Al]

TimIacobucci: Yes the 3 turbo setup have twinscroll on the small turbos so each cylinder is separated

One type of compound that never been tested (of what I know?) are two turbos where the turbines are parallel and the compressors in series.
My turbobuilder friend have sorted the theoretical problems and built the turbos but newer made it yet. (same that built the turbos for the 3 turbo engine)
One turbo has a big compressor and have Trim 100 on it. And the next is a smaller one that have really low compressor trim. Both have the same size turbine (a really big one) and exhaust housing.
In theory the backpressure should be extremely low because of the huge turbines. And you should have really fast spool when one turbo "makes the airflow" and the other "makes the boost"

[ptuomov]

TimIacobucci: Yes the 3 turbo setup have twinscroll on the small turbos so each cylinder is separated

One type of compound that never been tested (of what I know?) are two turbos where the turbines are parallel and the compressors in series.
My turbobuilder friend have sorted the theoretical problems and built the turbos but newer made it yet. (same that built the turbos for the 3 turbo engine)
One turbo has a big compressor and have Trim 100 on it. And the next is a smaller one that have really low compressor trim. Both have the same size turbine (a really big one) and exhaust housing.
In theory the backpressure should be extremely low because of the huge turbines. And you should have really fast spool when one turbo "makes the airflow" and the other "makes the boost"

I think that's the most interesting approach for a gasoline engine (after thinking about this approach for a day). With a gasoline engine, you never want the exhaust gas to go thru two turbines and never want it to not go thru any turbine. This is because gasoline engines hate back pressure, especially when run on pump gas. Furthermore, the turbine efficiency is relatively independent of the shaft rpm, so what you want to do is instead of compounding the turbines or bypassing a lot of mass from the wastegate, you want to bring more turbines on line as more exhaust gas is being produced by the engine.

On the intake side, most factory setups seem to have a bypass for the high-pressure small compressor. I'm not sure how required that is. I think that the intake side bypass can be just a passive check valve, bypassing the small compressor when it's not needed. Whether the intake side needs compounding in the first place depends on the efficiency of the whole turbos, it may be possible that a pure sequential approach on the compressor side would get one to back pressure below boost without any compounding.

[englertracing]

Sorry to the OP, as I will not help his issue, but go Off Topic.
Just wanted to say that it's very difficult for me to understand why you should need a compound system nowadays in a well geared car and with the new Garrett GTX2 twin scrolled.
Off Topic End.

https://youtu.be/5HlqEkJ3VGM?t=3m37s
it makes a 2 liter 4 cylinder run like this

and smoke 6 cylinder turbo motors...
https://youtu.be/u9FpUeoiJXg?t=7m24s
watch the porsche and then the lancia follows, Ill take the lancia.....

Yes, I pretty know the S4... and even the Predappio's road of the video...
but, again, the twincharging was requested for those ancient school turbo... IMHO

END OF OFF TOPIC.

https://youtu.be/9x-pkymzDOA
i dont know turbochargers is a gt3076r ancient school, old school, current, or new tech?
look at the dyno chart there is a difference of 150HP at 2300rpm.............
why wouldnt you want that? why are you so opposed to superchargers?
do you not enjoy immediate throttle response?
do you like to use throttle to correct under steer?

I simply don't understand your opposition?
the cons are what? 35 extra lbs of weight?
I guarantee you that is more fun to drive with the supercharger except to maybe a novice that finds the boost coming on late and suddenly amusingly scary and would be afraid to floor it with the supercharger.

I dont care how advanced the turbocharger is, but i suppose you can downsize the supercharger the better the turbo is. To me its an efficient, quiet reliable, FAST, anti lag system with a small weight penalty.

perhaps I am biased towards superchargers in the way that you are biased towards turbochargers, If a twin charge system is too complex, then just leave off the turbocharger . I am going to build a 2500cc-2800cc 4 cylinder with a whipple 510R twin screw supercharger, targeting 700hp on e85 for something like a street legal hill climb car. That way you really get to hear the exhaust , No lag, and pretty simple, consistent 2d throttle response, off idle power, ability to change boost with pulleys