Kevin's Compound Sequential Twin Turbo Design

[V8KILR]

...

Edited November 17, 2013 by V8KILR
Double post (see edit history)

[David P]

Thanks. I'm planning on doing closed loop for each wastegate based on the boost pressure feed directly from each turbo's compressor housing. I guess if that fails, I could try setting it up to use turbine speed.

 

I'm thinking the IACV control (which probably needs to use 3D tables) will be even more difficult to get right for all driving conditions.

 

IACV only functions for cold-start cycle and tickover.

[Digsy]

When you open the first wastegate to slow the first turbo down you will gain alot of energy to the second turbo which will spin faster and raise boost. Then when the second wastegate opens you get lower backpressure in in the pipe between the turbines and that will make the first turbo spin faster. So at the same boost pressure you will need different pulsewidth controlling the first wastegate depending on backpressure in the pipe between the turbos. The only way I see it possible to get any control over boost is to base it on turbo speed. And the cost of that setup for 2 turbos would easily be 1000£

 

How about adding compressor bypass to turbo #2? That way you could spin it up propery without building boost and without deadheading it against the flow control valve.

 

Here's how I see it working:

 

Off boost

WG1 open

WG2 open

Control valve closed

Comp2 bypass open

 

Turbo 1 boosting

WG1 controlled

WG2 open

Control valve closed

Comp2 bypass open

 

Turbo 2 pre-spool

WG1 controlled (ramping from closed to 50% open)

WG2 controlled (ramping from open to closed)

Control valve closed

Comp2 bypass open

 

Parallel twin

WG1 50% open

WG2 closed

Control valve open

Comp2 bypass closed

[V8KILR]

Good thinking. It's also a thought that has crossed my mind as well. Toyota didn't use one on the 2JZ-GTE sequential system (because of the reed valve) so I'm not sure if it would be necessary or not which is why it has been left off the diagrams. Its very easy to add a very simple boost controlled compressor bypass afterwards if needed. Here's a pic of what I could use (by rotating the lever on the shaft 90 degrees) to do the job.

 

https://jonbondperformance.com/images/TJ%20bypass%20assembly.JPG

Edited November 17, 2013 by V8KILR (see edit history)

[Pipedreams]

Do you mean bypass it back to the compressor inlet? Without any restriction it will overrev and break. And once again it comes back to turbospeed to control it

[V8KILR]

Do you mean bypass it back to the compressor inlet? Without any restriction it will overrev and break. And once again it comes back to turbospeed to control it

 

No, if I did do this it would just be vented to atmosphere and with the #2 wastegate fully open there should not be much drive pressure anyway. Also, I would close it either before or at the same time that the #2 wastegate is closed, so that there was no risk of overspeeding. As Toyota did not see the need to do this with the 2JZ-GTE, I'm not convinced yet that it is necessary to do this.

[Digsy]

Do you mean bypass it back to the compressor inlet? Without any restriction it will overrev and break. And once again it comes back to turbospeed to control it

 

Many modern turbos have an integrated compressor bypass that recirculates back to the intake. They work fine. It's not a perpetual motion machine :-)

[V8KILR]

Many modern turbos have an integrated compressor bypass that recirculates back to the intake. They work fine. It's not a perpetual motion machine :-)

 

Yes, I guess that makes very little difference if it recirculates as there is not really any vacuum in the compressor inlet. I think it would be heating the intake air a bit though, but recirculating may be a legal requirement in some countries.

[Digsy]

Yes, I guess that makes very little difference if it recirculates as there is not really any vacuum in the compressor inlet. I think it would be heating the intake air a bit though, but recirculating may be a legal requirement in some countries.

 

 

I guess the issue is that you would be running a load of airflow with little or no pressure differential across the compressor, which would run you straight into the choke line on the compressor map. IIRC compressor bypasses are usually used when you are not generating boost, so maybe dumping it oiverboard would be the better option. Might be noisy though :-)

[V8KILR]

I guess the issue is that you would be running a load of airflow with little or no pressure differential across the compressor, which would run you straight into the choke line on the compressor map. IIRC compressor bypasses are usually used when you are not generating boost, so maybe dumping it oiverboard would be the better option. Might be noisy though :-)

 

Checking with Google, a CBV is exactly the same as a BOV so they both operate when the throttle is closed and vacuum occurs in the intake manifold. I think that the reed valve that comes with the factory IACV is all thats needed for a sequential setup.

[Pipedreams]

Many modern turbos have an integrated compressor bypass that recirculates back to the intake. They work fine. It's not a perpetual motion machine :-)

 

It wasn´t that it was recirculated that was the problem, it was having the turbo spinning without any restriction on the compressor outlet. But with the turbine bypassed also that wouldn´t be an issue. But then the turbo won´t spool and then there is no point in venting the compressor side either. The reed is a much better option.

[Digsy]

The reed is a much better option.

 

Agreed. TBH I missed how that fitted into the system schematic

[V8KILR]

The only part of my design I'm not happy with is the complexity of controlling the IACV. Toyota used a 3D map with engine speed (rpm), vehicle speed and throttle position as the three axis (axes) for controlling their IACV. The 3D map method seems to be a very complicated way of calculating the load on the engine and this makes for a very complex tuning requirement for a one-off setup like mine. So, in line with the general theme of keeping my compound sequential design as simple as possible, I've been thinking of a way to reduce this complexity.

 

The factory setup seems to use the reed valve as a simple way to let the flow from the #2 turbo through to the intercooler until there was sufficient flow from the #2 turbo, that it could then completely open the IACV in one single step (is this correct?), or in other words the IACV is either open or closed. Perhaps a simpler way to do this would be to control the IACV at all times as if it was just like a wastegate, which would only require a closed loop 2D map.

 

The steps I see are:

 

1. The #1 turbo is making full boost of 35psi (taken from the #1 turbo compressor outlet) and the #1 wastegate is opening (under 2D closed loop control) so as to maintain this boost level.

 

2. The #2 turbo should only be making around 20psi (taken from the #2 turbo compressor outlet) at this point (assuming an initial 15psi lag) but because the IACV is closed, pressure quickly builds up and reaches 35+ psi.

 

3. The IACV now starts to open (under 2D closed loop control) just as if it was a wastegate set to 35psi (taken from just before the IACV). However it only needs to open a little bit at this stage as flow from the #2 turbo is still quite low.

 

4. As the #2 turbo continues to spool up, the IACV will be opened more and more until it is 100% open, so as to keep the boost from the #2 turbo at 35psi.

 

5. Once the IACV is fully open, the #2 wastegate control is now activated (using a physical or virtual switch activated by the IACV reaching 100%) as it now needs to take over boost control of the #2 turbo.

 

6. The #2 wastegate now controls the #2 turbo at 35psi boost (while the IACV is 100% open), until easing off the throttle occurs. Then everything resets waiting for step 1.

 

 

You have probably realized that I have not mentioned the reed valve at all in these 6 steps. This is because it is now redundant and can be completely replaced by the wastegate type control of the IACV, which is acting exactly like it is a huge reed valve. This means I could use the Turbonetics Newgen (or other type) wastegate in place of the factory IACV as it would probably allow finer control.

 

Edit: Thinking about the reed valve a bit more, there may still be a small need for it under certain conditions, so probably a good idea to leave it in the setup as it works automatically anyway.

 

Can anyone think of a reason why this simple 2D closed loop control of the IACV would not work?

[jagman]

It's not really a compound system , however terminology apart , Ian c has highlighted the primary issue - using exhaust gas after its passed through the first turbine and lost energy .

Second issue is the compressors must always flow if spinning and have similar output if joined .

The system should also cope with instant changes ie wot , followed by zero throttle and then wot once again .

Lastly component fail safe - what happens if one component fails ? Will it cause a major problem elsewhere or fail safe ?

[V8KILR]

All very good points.

 

It's not really a compound system , however terminology apart , Ian c has highlighted the primary issue - using exhaust gas after its passed through the first turbine and lost energy .

 

Agreed, it's only a compound setup on the exhaust side. Using secondhand exhaust energy is how any compound diesel system works so that should be okay. This setup also gets firsthand exhaust energy that bypasses the #1 turbo once the #1 wastegate opens so it should spool the #2 turbo better then any diesel compound setup would. Also the #2 turbo is the same size as the #1 turbo, so easier to spool then a larger #2 turbo in a normal compound setup.

 

Second issue is the compressors must always flow if spinning and have similar output if joined .

 

I might not have mentioned it on this thread, but I have added an exhaust bypass to my design (using an altered factory EGCV) to bypass exhaust from the #2 turbo (rather then using a more complicated way of holding the #2 wastegate open) so as to improve the #1 turbo spool. This will mean the #2 turbo will only be idling (and not making any real flow) until the EGCV closes when it gets around 7psi boost from the #1 turbo. At that time there should be enough exhaust energy to start spooling the #2 turbo, so that it can start building up enough pressure to open the reed valve.

 

I don't think you need similar flow while the #2 turbo is spooling as that's how a reed valve works. All you need is the same or higher pressure from the #2 turbo to be able to join the flows.

 

The system should also cope with instant changes ie wot , followed by zero throttle and then wot once again .

 

I guess that will come down to tuning and how quickly the two wastegates and IACV can react to the TPS changes.

 

Lastly component fail safe - what happens if one component fails ? Will it cause a major problem elsewhere or fail safe ?

 

If either of the two wastegates fails then it would over boost the associated turbo but this would happen on most turbo setups. If the IACV fails then #2 turbo will stall once it reaches the max psi it can create for the exhaust pressure it is getting.

[V8KILR]

No replies on my IACV thoughts! Does that mean everyone thinks the IACV idea will work?

[Pipedreams]

No, I means my phone wouldn´t let me post I don´t think it will work in closed loop against pressure since the nr1 turbo will equalise pressure in the pipes if nr2 is slowing down so the valve won´t know that it should close. It will only close when the overall pressure is lowered.

[V8KILR]

No, I means my phone wouldn´t let me post I don´t think it will work in closed loop against pressure since the nr1 turbo will equalise pressure in the pipes if nr2 is slowing down so the valve won´t know that it should close. It will only close when the overall pressure is lowered.

 

My first thoughts were the same as yours, but then I realized that the pressure on the other side of the IACV will not be 35psi despite the #1 turbo producing 35psi. If you type in "use air compressor as vacuum pump" in to Google, you will see my reasoning on this. Basically it comes down to the design of the merging pipes and using Bernoulli's principle to ensure the pressures are lower (venturi effect) in the #2 intercooler pipes between the IACV and the intercooler.

[Pipedreams]

Sure it will help to some extent but when using a compressor as vacuum pump you don´t have any restriction at the end of the air pipe. If you put a restriction at the end to raise the pressure to 35psi I don´t thing you will get much suction in the second pipe.

[V8KILR]

Probably, but it only needs to be 1-2 psi lower for it to not effect the IACV control.

 

Edit: as long as there is a good high speed flow of air through the #1 intercooler pipe then it shouldn't matter what the psi is. E.g. If flowing 600cfm through the 2" intercooler pipe from the #1 turbo, then air speed will be 458ft/sec at 14.7 psi, which I'm guessing is around 190ft/sec at 35psi.

 

I was thinking of joining the flows like this with the #1 turbo flow being the straight down pipe:

 

http://www.industrialplasticpipe.com/pages/images/35129p.jpg

Edited November 23, 2013 by V8KILR (see edit history)

[V8KILR]

Does this make sense as to why I hope to be able to get a lower pressure in the #2 intercooler pipe when the #1 intercooler pipe is at 35psi?

 

A = #1 intercooler pipe

B = combined intercooler pipe just before intercooler

C = #2 intercooler pipe

 

"As air is ejected from nozzle A, it mixes with the air in the tube at B. Imagine that the air from the nozzle mixes with two times as much air. Momentum is conserved, so this cloud of air, now two times as much as came out of the nozzle, is now moving at one half the speed of the jet from the nozzle. All that air moving to the right requires replacement air to be pulled in from the left. So we now have suction at C."

 

 

The page that explains this is here: http://woodgears.ca/physics/venturi.html

 

I would do this just as the two pipes join going in to the intercooler. This means there will be two, 2" pipes from the turbos combining into one 3" intercooler pipe. This will give twice the air volume for the 35psi air from #1 turbo to push along and create a pressure reduction in the #2 intercooler pipe. If this works, then air from the #2 turbo may be pulled through the reed valve at quite low pressures resulting in the #2 turbo creating air flow for the engine, even when it is at low boost compared to the #1 turbo.

 

Just had a crazy idea. If it did work like this at all times (like a one way valve), that means I could eliminate the reed valve and the IACV altogether!

Edited November 24, 2013 by V8KILR (see edit history)

[jagman]

You can't equate diesel turbos to petrol turbos

The turbines differ , derv produces lower heat more volume , the rev range is limited .

Compounding in the diesel works because the second turbo is smaller , it's using the residual energy from after the first turbo so has a smaller turbine - but because the output from the compressors are fed through each other - they compound , you get a multiplying effect on boost .

You still only have one heat source to feed 2 turbines - in your case 2 large turbines? Ie you have the mass of 2 turbines to accelerate using the same heat source .

This is why twin charging works , you introduce a second energy source - the crankshaft energy to spin a charger .

It's only the inefficiency of turbo turbines that allow any post first turbine heat to remain at all - a perfect turbine would convert all heat energy to rotation .

The jet pump idea for the outputs also isn't how it would work .

Keep drawing out layouts and you will get a solution , but it will involve a super charger ( second energy source) and a check valve/ pressure regulation system

[V8KILR]

You can't equate diesel turbos to petrol turbos

The turbines differ , derv produces lower heat more volume , the rev range is limited .

Compounding in the diesel works because the second turbo is smaller , it's using the residual energy from after the first turbo so has a smaller turbine - but because the output from the compressors are fed through each other - they compound , you get a multiplying effect on boost .

You still only have one heat source to feed 2 turbines - in your case 2 large turbines? Ie you have the mass of 2 turbines to accelerate using the same heat source .

This is why twin charging works , you introduce a second energy source - the crankshaft energy to spin a charger .

It's only the inefficiency of turbo turbines that allow any post first turbine heat to remain at all - a perfect turbine would convert all heat energy to rotation .

The jet pump idea for the outputs also isn't how it would work .

Keep drawing out layouts and you will get a solution , but it will involve a super charger ( second energy source) and a check valve/ pressure regulation system

 

FYI, the second turbo in a diesel compound setup to receive the exhaust gas is often the large one. That's how they are normally setup.

 

Compounding the exhaust flow works perfectly fine on petrol turbo engines. Boost Logic have done it and so have others and they all worked great! There is no question that the exhaust compounding part of the design will not work.

 

What's wrong with the "jet pump" idea? It was invented in NZ after all.

Edited November 25, 2013 by V8KILR (see edit history)

[V8KILR]

In case the idea above does not work, here is the backup plan.

 

The IACV and the reed valve are really just one way valves but the IACV requires complicated ECU setup to control it. A far better way that requires no ECU control is to use a "swing check valve". e.g.

 

 

Edit: Here is a better non-surge design:

 

http://www.asiawaterbusiness.com/images/img_library/Untitled_1-6807.gif

 

Using a valve like this will result in very low flow and pressure losses from the #2 turbo, probably not much more then the butterfly from the IACV does anyway.

 

Without the need to control the IACV, the ECU setup has just been simplified by 90% and all that's now needed is two wastegates being controlled by the ECU. This has got to be the simplest sequential setup ever!

Edited November 25, 2013 by V8KILR (see edit history)

[V8KILR]

Here's an updated design drawing based on the feedback from this and other Supra forums.

 

 

The modified factory 54mm EGCV is always open until the #1 turbo boost reaches 7psi. This is to help spool the #1 turbo.

 

The 65mm swing check valve is to stop reversion from the #1 turbo flow back to the #2 turbo. This replaces the reed valve and the IACV as it operates automatically and doesn't require any ECU control.