SupraTT - Boosting the Beast

Twin-turbo, is like two rockets firing, right?

Well, not quite...
If you're used to fast, tuned turbo cars, then driving a stock supra twin turbo leaves you scratching your head and wondering if the turbos really do work. Especially if it's an auto, that tries to start in 2nd gear (equivalent of 3rd gear in a 6sp), change into 3rd gear immediately after, and go to top gear at the first opportunity - good for 200odd mph. Now if you're doing 50mph and the gearing is for 200mph, it's no wonder that it feels sluggish.

The car is transformed when boost is raised around 1 bar or more. Then when it hits 4000rpm it lurches forward, the second turbo really feels kicking in.

....But soon the novelty wears off, and you're left with a sense of flatness and lethargy at low revs, followed by the rush at 4K rpm. But fear not: this can be fixed with the EBV mod described later on, leaving the Supra with a real feeling of a twin turbo 3-litre engine.

 

Running higher boost on the SupraTT

Everybody wants to run higher boost, because that's where all the power is hidden, right?

That's roughly true, but if we're trying to interfere with the sequential system, at least we want to keep the disruption to a minimum.

Having seen how the sequential system works, we now appreciate the complexity underneath. The orchestration is conducted by the ECU but the result is also determined by the performers themselves. The way the actuators are connected to the pressure signal and the strength of their individual springs dictate some of the timing of the events. Also the airflow through the airpipes (especially the 'bypassing' or 'bridging' ones) is pressure-driven, and that is not always obvious just by looking at a diagram.

It is similar to aerodynamics --- just because we create a duct and want the air to flow through it, doesn't mean that under operating conditions the air does what our common sense feels it should. Very often it does something totally different. Pressure differences also depend on temperature differences, not just airspeed. As air moves down a path it cools down, especially if the path is narrow and the pipe is metal. So pressure differences can exist even on (what appears to be) the same boost signal path.

The Toyota engineers had thoroughly measured, tested and debugged the original setup, so it works 100%. But when we start restricting pipes and blocking paths, the whole balance changes and it's not easy to foretell what repercussions we are creating. Chaos theory for the paranoid.

 

Enough theory dude, where do I fit my boost controller now? In all this maze of hoses, pipes and actuators I don't know where to start!

The typical point that boost controllers intercept is the one below:

The wastegate actuator of turbo#1 is always connected to the boost signal. So when the engine is doing 2 psi, that's what the back of this actuator is filled with.
Or is it?
Looking closely we see another connection at the back of the actuator, on the same side of the diaphragm. It goes back to the unpressurised side of the intake, via the VSV.
Aye, 2 psi may be squirting through the actuator, but if the VSV is wide open then there is a generous 'relief' path for the pressure, so in reality the actuator never sees more than a fraction of a psi.
If the VSV has been shut by the ECU however, then the actuator will be filled with boost gases, and will be at the same boost pressure as the intake manifold (a bit higher actually, since it receives hot gases before the intercooler, which have a higher pressure compared to the manifold, where boost gauges typically get their signal from).

If you use a 'bleed valve' it won't matter a lot if you use either of the two pipes coming off this actuator. In both cases you'd be depressurising (partially) the actuator. How much would depend on the position (pulse duty-rate more accurately) of the VSV.

If you use a 'ball and spring' boost controller then it does matter which pipe you use. It has to be the one shown above - the 'pressurising' one.

Electronic boost controllers usually ask for the VSV to be blocked off so that they can fully control the boost pressure behind the actuator. This is normal, otherwise the ECU would be trying to undo the actions of the boost controller, and the result would be a funny boost curve.

 

I've done this mate, but I only see higher boost at high rpm, say 4Krpm upwards. What's the vibe then?

The vibe is that you didn't bother to read the long page on how the system works. There are two wastegates operational until 4Krpm, the WGV and the EBV.
They are both upstream of turbo#1.
They are also both continuously connected to the boost signal taken from before the intercooler.
They also both have a VSV downstream which is pulsed by the ECU.
Therefore at low revs (below 4K) there are two wastegates in the system

You have only fiddled with the workings of the WGV, so the EBV still does what it's supposed to do: keep max boost levels down to stock. To put it simply, the EBV is bleeding off gases that could be used to spin turbo #1 even harder.

Since both of these valves operate in similar ways, we can replicate what we did with the WGV, i.e. stick a "ball and spring" controller on the pipe pressurising the EGV actuator. However, there are two extra issues here:

1. Accessibility. This actuator is tight between the engine and the firewall, so accessing the two nipples is a fiddly job.

2. The fact that the EBV also controls the prespooling of turbo #2.
This is more serious, because in order to get more boost out of the whole system at low revs, you need this valve to open later. But this means that there will be less gases available to prespool turbo #2. Even worse, turbo#2 will then have to jump suddenly and catch up with the new, higher boost levels we've created. Not very straightforward then.

 

An unsatisfactory approach found on the internet

The setup below is easy to implement, because that intersect point is easily accessible at the top of the engine. A 'ball and spring' controller cannot be used here, because it is downstream of the actuator. So the only option is a bleed valve or an electronic boost controller.

What we'll be achieving is that the EBV VSV will be pulsed shut by the ECU, but some of the pressure will be escaping from the bleed valve. So we're effectively reducing the duty rate of the VSV. This is preferable, as it gives a higher degree of fine-tuning. We do need this, because we need a safe balance between higher boost at low revs and retaining some prespool for the second turbo.

Unfortunately it doesn't really work, because the return path is long and torturous and doesn't bleed off quickly enough. This leads to the need to increase the rate of bleeding leading to an on/off situation, where the actuator is either under full boost pressure, or empty.

There are other ideas floating on the internet, none of which have an advantage to this solution (quite funny to see a bleed valve up the list of mods for such a complicated engine!)

The most dangerous of all is this one:

You are 'advised' to Tee the two hoses below A and B. This gives maximum boost at low revs.

Looking at these locations on the vacuum diagram we see why this is dangerous. The 'mod' totally bypasses the VSV, so any pulsing by the ECU is in vain.
Point A is always in vacuum, from idle upwards.
Point B is usually under positive pressure, roughly at what the boost gauge would show.

What this means is that there will always be a rush of air from B to A and this will be much stronger when the engine is under higher boost. Therefore the EBV actuator will never be pressurised, and it will only open the EBV when the backpressure after turbo #2 manages to overcome the actuator spring pressure. While the EGCV is shut (below 4Krpm) this ain't gonna happen any time soon.
...so the EBV will always be shut, at least until 4K rpm. This means that the prespool for turbo #2 will be zero (save the odd pressure leak from the EGCV or the EBV)

But wait, it gets worse: a couple of hundred revs after the EGCV has opened, the IACV is ordered to open, because the ECU assumes that by that time the boost pressure created by turbo#2 is well up to the levels of turbo #1. But that assumption is with prespooling in mind. Without any prespooling, it will certainly take longer for turbo #2 to catch up with #1. So we risk the ECU opening the IACV wide open while turbo#1 produces more boost than turbo#2:

In the example above, turbo #1 produces 11psi (pressure left of the IACV) while turbo #2 produces 7psi (pressure right of the IACV) This is possible while the IACV is kept shut -- note that the reed valve can only flow the other way, so that is shut as well.
The moment the ECU orders the IACV to open, the full blast of the pressure tank fills the IACV actuator and it tries to crack open, since now it has to overcome the 4 psi (11psi - 7psi). Let's put aside the extra stress of the IACV flap for now. Even if it doesn't manage to open fully, we'll be left with higher absolute pressure on the left of the valve, and lower on the right. No guesses as to which direction the air will try to flow. But that is against the direction the compressor of turbo #2 is trying to flow. So the turbine of #2 is trying to turn in one direction, while the compressor faces a sudden brake, trying to turn the shaft the other way. That can't be good!
This is compressor surge caused deliberately by the ill-advised 'mod' and the stock 'dump' valve can do nothing to relieve it (apart from being downstream of the intake 'meeting point' it would never react anyway since its pressure reference shows 'boost').

How long turbo #2 will survive is anyone's guess.

There are other variations on this theme, where the connecting pipe between A and B is partially restricted, so that (hopefully) some prespool will take place --- at the expense of running higher boost at low revs of course.

Another hack would be to fit a bleed valve on pipe B and adjust carefully, slightly increasing the leak and checking on the road and boost gauge that only a couple of psi have been gained over stock. This might keep some prespool and hopefully find a happy mid-point where turbo #2 won't be kicked to work from a total idle state, then be surged to death.

What all the above methods fail to do is intercept the EBV pressurising signal (seen below)

That's something that has to be done. It's fiddly but not impossible.

 

A better way to increase low-rpm boost

An internet favourite is Stuart Hagen's EBV mod. This is a bit more involved to implement. You need to crawl between the firewall and the back of the engine and disconnect the vacuum hoses from the EBV actuator. You then run a new hose (after blocking off one of them) and connect it to the pressure tank feed --- by using a "Y" splitter at the top of the engine. This creates a new path to the pneumatic activation of the EBV, so a boost controller can be used to open the turbo#2 prespool whenever we like. (The EBV VSV is also blocked off)

This mod allows max boost at low revs, well over 1 bar is possible. If a ball&spring valve is used, then it should be set to open at least 3-4psi below the system's max boost setting on the wastegate, allowing allowing some prespool. (local copy here)

 

An even better way to increase low-rpm boost

This is my favourite, a variation on the theme of the previous 'EBV mod'

Again, access to the EBV actuator is quite restricted by the firewall, you may wish you had fingers like Edward Scissorhands.
Let's put it this way: if Chopin had a Supra, all his friends would go to his place to get the EBV mod done.

First there is a water pipe and an air pipe (idle control circuit) that need to get out of the way. Thankfully the coolant leaks will be minimal as this pipe is at the highest part of the engine. The electric loom for the two VSVs has to move on the side, too. Easy. Then you stare wondering Matrix-style: "is it hose #1 or hose #2 that I need to pull off?" The answer is "neither!". These hoses are irrelevant to the EBV mod, although they'd be convenient if they indeed went to the back of the EBV actuator, as they appear to do.
Under closer inspection and by using a bright torch you see the two hoses coming off the actuator. As we're looking down the firewall, the one to the right is the 'feed', we can see the "T" junction of the pressurised metal feed as it goes to the pressure tank. The one to the left is the 'relief', which the EBV VSV pulses open according to the ECU's commands. The more it's pulsed open, the more boost is leaked back into the pre-compressor intake and the EBV actuator is forced to close again. Note that in this diagram view pressure to the left, relief to the right. Don't mix them up, we leave the relief alone.
Here we can see some long-nosed pliers that will help cut down on bad language as we try to negotiate hoses and clips hidden down the firewall...
In an ideal world we'd keep the original EBV actuator feed and relief performing the same tasks. We'd just want to insert a Manual Boost Controller (MBC) in the feed so that it doesn't flow anything below a preset pressure level (say 1 bar) That's exactly what we're going to do.
Apart from various lengths of vacuum hose, we need two connectors a "T" and a "I". Make sure that their dimensions are proper for the internal diameter of the hoses used. You don't want them too thick (they won't fit, and you may not even be able to push the "I" inside the remote feed) and you certainly don't want them too thin (they'll pop off)
Here is the "T" junction that has been introduced to the original boost feed. It still feeds the metal pipe as it used to (ultimately ending up to the pressure tank) and now it also feeds our rendition of EBV actuator control
1. Hose coming from the EBV actuator's pressure side. No need to take off the old one completely, just undo it from the metal hose and use the "I" connector to extend it. Don't forget to block the metal ending ofcourse. 2. MBC ensuring that boost only pressurises the actuator when it exceeds the limit we set 3. "T" connector diverting boost pressure towards our new EBV boost feed.
	Another view of the same setup. 1. MBC 2. "T" junction

The beauty of this setup is that the EBV relief is still intact, so the ECU still has partial control of the EBV. If it tries to pulse the VSV at boost pressure below our MBC setting, obviously nothing will happen because the actuator is empty.
But if it tries to pulse the actuator above our setting, it will work fine! So if at high boost (over 1bar and over 4Krpm) the ECU wants to shut the EBV actuator, now it can do so. Also if during a downshift it decides to open the EBV it can still do something, it's not castrated as it is with the previous plumbing.

As a bonus, it's very easy to turn the setup back to standard --- just eliminate the MBC, or take out the ball and spring and refit the adjusting screw.

The results obviously depend on the settings of the MBC. A logical setting would be to have it set to 15psi for example, and the wastegate set to over 15psi via another MBC.

Then at low revs (up to around 3800)turbo #1 will hit a boost level of 15psi and stay there (the excess gases will be prespooling turbo #2). At around 3800rpm overall boost pressure will drop a few psi as the second turbo comes fully online, then it will shoot up to whatever level the wastegate MBC is set to.

If the EBV MBC is set to 13psi, then turbo #1 will hit 13psi boost, and the drop at 3800rpm will be less obvious, as more gases were allowed to prespool turbo #2 and the 'gap' from #1 was not so wide this time.

There is some fine-tuning to be done with the newly-installed EBV MBC, don't be greedy, start from a low boost level and gradually go upwards until the 3800rpm boost drop becomes an issue.

 

Stealthy version of this mod

The "T" piece right in front of the engine is a giveaway that the engine is not stock. This can be avoided easily for those who like sleepers, or the 'factory' look.

First of all don't use blue silicon hoses, rather black fuel hoses. They are bound to attract less attention and be at least as durable under the harsh engine bay environment.

Second you can completely avoid the "T" at the front and extend the stock EBV actuator metal feed via another hose (instead of blocking it off and using the "T" described above). The MBC could be secured with cable ties near the firewall and nobody would be any the wiser.

Technically it is exactly the same thing, there is no reason to prefer one approach to the other.

 

Faster spoolup too - as a bonus!

An interesting by-product of the above modifications is that the car builds boost quicker overall.

It feels faster, and that's because if you use MBCs on both actuators (remember at low revs turbo #1 thinks it has TWO wastegates) no boost whatsoever enters these actuators until 1 bar is reached (or whatever the lowest setting is for any of the two). Therefore ALL exhaust gases are available to turbo #1 until the very last moment, when the EBV actuator is filled with boost and opens up.

One would expect the ECU in stock form to do the same thing - i.e. keep these actuators free from boost until the preset level are reached, then pulse them opened. That's what the VSVs are there for. But in real life, standard ECUs are programmed to be smooth and gentle, so they start pulsing the actuators open well before the full settings have been reached, diverting exh gases that could be better used to spool up the turbos quicker.

By using MBCs on both actuators this 'smoothness' is gone. Note that if you use an MBC on just one actuator you won't feel any difference, because the other one still diverts gases before their time. You need MBCs on both wastegate and EBV to get this extra edge in spoolup.

 

Yikes dude, I didn't understand half of the stuff you've said, but now I'm kinda scared.

Makes you think twice before changing anything off-standard, doesn't it?

 

....on to 2JZ GTE Intercooling