MR2 Owners Club Forum banner

21 - 40 of 42 Posts

·
Registered
Joined
·
617 Posts
Yeah, my buddy on the East coast is running 17.9psi w/ electronic controller on his stock ECU. He drag races it. He runs 13 flat at the moment.
Eventually, years from now when the car is not my daily I plan on drag racing it, I'd like to see it do 10s which is what I was talking about. I'm not here to talk about what mods will support that.

Anyway, his car will hit peak boost at 2750rpm and pull pretty much to redline. He dyno's in at over 300, looking at close to 320. We will have the results in a few days when he finishes getting it tuned on his EMU Black. He's gonna post up the results for his stock ECU run, and his post-tune run on the EMU Black. The turbo can do 20psi no problem. There are people making ~330hp on the CT15B with the billet compressor wheel. Retaining that spool as well.

My thoughts are that you're biased towards your gen2 build, you fail to see the benefit of running that cast-in-manifold turbo. I can always do aftermarket ECU, internals, whatever I want later on but for now I want ~300 horsepower and a drop-in gen5 will do that for me.


This is a stock boost ST215 3S-GTE dyno, looks like 13 psi comes in at 3200 RPM or so. 18+ psi should be in the 3400-3600 RPM range. I've never really seen a 2L make that much boost that early on a turbo that can make 300+ rwhp, so maybe he gets that much boost that early, but it'd be highly unusual based on my experience with lots of 2L turbos.


I think you misunderstand me, I don't think a modded gen2 3S-GTE is a great engine. In fact, I'd put it behind many other Japanese 4 cylinders of the era. The head doesn't flow well, the timing belt + distributor is a pain (biggest limit to my setup now, and yes, I have coil on plug), it weighs a lot, and the early 3S block isn't that reliable beyond ~350-400 rwhp.

I only mention that if you're wanting to really mod your car, I'd say mod towards those goals. I went through 5 or 6 turbos on my old SR20DET in an S13, and looking back on it, I should have been a little more deliberate in what my end goals were. If you want to run 10s, I'd look at a turbo you can use to get you there going forward (a CT15B, or a CT15B with a billet compressor wheel won't do it). Same with a pushed to the hilt stock ECU vs. a standalone - the standalone will get you there, so why not jump in with both feet. It takes too much time to do this stuff and get it all working well to constantly be trying to switch big parts out IMO. Just my 2 cents from someone who's been there, done that.
 

·
Registered
Joined
·
617 Posts
Oh, and he drives it quite hard and pretty much daily. I do the same, I am persistently driving my car hard. Every night or every other night I try to take it out and I'm racing Wednesdays and weekends. I am pretty sure the real world results I've seen from these cars is a testament to how reliable they are. If you're scared of the A/F ratio just get a gauge that lights up red above or below a certain number so it alerts you when something's wrong. Check the car over before driving it hard, replace wearable parts often, regularly maintain things and replace before they're likely to break. You shouldn't have a problem with a properly maintained car especially if you leave it mostly stock. I consider running 17.9psi on a gen4/5 practically stock, because it is. The ECU knows how to adjust and does so. The gen4/5 has very good knock detection with ion sensing coils. It can actively advance or pull timing to account for it. It can also cut ignition / fuel should the "knock" be really rough. Especially if the stock knock sensor detects it as well.
Are you sure it has ion sensing coils? That was cutting edge stuff on the S85 V10 BMW made in the mid 2000's (on a car with a $100k+ base price). I googled and came up with nothing. If so, that sounds really really advanced for the time.

A wideband AFR gauge won't save your 3S-GTE at 450+ rwhp if something goes wrong and it goes lean. By the time you would register something was really wrong, your engine is toast. Even at 300 rwhp there's not a ton of time the engine can tolerate something really being off AFR wise if it goes lean. In general humans are poor engine protection schemes - everytime my Infinity's engine protection kicks in, I had no idea something was going on since it's instant from the driver's perspective (i.e. problem encountered, engine/boost cut). Only looking over logs does it become obvious what it was doing to save the engine. A major push in aftermarket ECUs in the last 10 years has been increasing engine protection schemes, so obviously the whole industry feels it's an important feature in standalones and really modified engines in general.

Likely going to build a 5sgte hybrid down the road... that's what DefSport should REALLY be recommending. I'll look up all that info later on. For the next few years I'll just run a stock gen4/5 setup and focus on getting it to run on aftermarket EMS and getting the chassis / suspension figured out in preparation for the 5sgte. It's a long road I'm on but just trying to filter through misinformation and whether I'm disregarding safety precautions or not is something I'll learn the hard way I suppose. Some people will say 17-18 psi on a gen4 is dangerous, they'll cite sources where something failed and caused it to blow. They'll do the same for gen2. The thing that I recognize is common with all these failures and blown motors is that something FAILS before it blows. They don't tend to just blow up because of the boost... so if you take your time and ensure everything is running right before you start turning the boost up... not an issue. Besides, a gen2 long block is like $600-800... minus turbo and accessories. Once covid import restrictions are eased up you'll see gen4/5 at $1000-1200 or so again.
A little more displacement is good on a 5S-GTE, but you really need the later model head, and even better, the 4th/5th gen BEAMS head (which is why I recommended that engine if you want to stay 3S). The 3S heads, even the later gen3/gen4/gen5, don't really flow that well so at higher power they're a fairly large restriction. 10% more displacement is better, but adding a solid 30-40% more head flow is even better at 400+ rwhp IMO.

TBH, if you want 400+ rwhp in an SW20, I'd say a single turbo 2GR-FE is probably a better starting point. Lots of displacement to have good low end torque, you can go to a free flowing turbo and still have a solid powerband, and a 3.5L at 400-500 rwhp is going to be fairly low stressed as long as the setup can support it (good cooling, low exhaust back pressure etc.). I'd also choose an NA 2GR-FE over a stock CT15B gen 4 3S-GTE as well.

I'm kinda meh on my gen 2 3S-GTE, so if I had a good reason to put a new engine in, I'd put a 2GR-FE in and try to get a turbo on it.
 

·
Registered
Joined
·
7,628 Posts
A major push in aftermarket ECUs in the last 10 years has been increasing engine protection schemes, so obviously the whole industry feels it's an important feature in standalones and really modified engines in general.
Very few aftermarket ECU’s have decent detonation detection and knock control. This is a very complicated system that needs to be integrated and tuned for a particular engine. I’m not saying this is the sole protection but is generally the last line of defence. Objective is to get your tune dialed in to reduce likelihood of detonation. This is something aftermarket ECUs can do rather than just adding parts and expecting the OEM system to compensate.

jim
 

·
Registered
Joined
·
53 Posts
Discussion Starter #24
This is a stock boost ST215 3S-GTE dyno, looks like 13 psi comes in at 3200 RPM or so. 18+ psi should be in the 3400-3600 RPM range. I've never really seen a 2L make that much boost that early on a turbo that can make 300+ rwhp, so maybe he gets that much boost that early, but it'd be highly unusual based on my experience with lots of 2L turbos.


I think you misunderstand me, I don't think a modded gen2 3S-GTE is a great engine. In fact, I'd put it behind many other Japanese 4 cylinders of the era. The head doesn't flow well, the timing belt + distributor is a pain (biggest limit to my setup now, and yes, I have coil on plug), it weighs a lot, and the early 3S block isn't that reliable beyond ~350-400 rwhp.

I only mention that if you're wanting to really mod your car, I'd say mod towards those goals. I went through 5 or 6 turbos on my old SR20DET in an S13, and looking back on it, I should have been a little more deliberate in what my end goals were. If you want to run 10s, I'd look at a turbo you can use to get you there going forward (a CT15B, or a CT15B with a billet compressor wheel won't do it). Same with a pushed to the hilt stock ECU vs. a standalone - the standalone will get you there, so why not jump in with both feet. It takes too much time to do this stuff and get it all working well to constantly be trying to switch big parts out IMO. Just my 2 cents from someone who's been there, done that.
When you turn the boost up with an EBC it can hold the wastegate shut all the way until it needs to let off. The stock actuator will let the wastegate flapper door open up slightly even at 9-10psi. Maybe as late as 12psi or 13 when using a manual boost controller or just a restrictor in the actuator vacuum hose. I'll ask for his dyno sheets to verify since you've presented one as well.

First car with ion sensing ignition was a Saab in 1988, I believe the "9000" series? Saab's "TRIONIC" EMS was later using a proper ion sensing coil starting in 1993. This technology has been around for a minute.

Anyway, the stock ECU on the gen4/5 uses the coils to keep the timing as advanced as possible to the point that his tuner finds it will knock if he adds any timing whatsoever over what the stock ECU was doing. They literally can't make more power on the stock turbo because the stock ECU leaves no headroom if you have the ancillary mods. Such as a bigger intercooler, airbox, straight pipe exhaust, and whatnot. With the stock turbo actuator he hit a peak of 275whp due to the wastegate door being pushed open by pressure past 13psi, will peak at 18 regardless of his settings. Peak 275whp on high boost, low boost, whatever. With his EBC and aftermarket billet actuator he's looking at closer to 320whp but they haven't done a full run as they're tuning on the EMU Black now and had to create their own basemap. Unfortunately for them I don't think any aftermarket ECU on the market has tuning or can make use of ion sensing coils at the moment.
 

·
Registered
Joined
·
53 Posts
Discussion Starter #26
The ion sensing coil thing has me interested, are we talking about the IGF signal on the individual coils?
No one knows how Toyota does it as compared to Saab because Toyotas ECU has yet to be unlocked for the most part and it was only used on the Caldina and who knows what other cars / motors. From what I understand the gen5 improves itself over the gen4. Gen5 w/ immobilizer disabled will put down slightly more horsepower stock than a gen4 despite being labeled the same power rating. No clue if that has to do with the longer injectors or the coils or some of the other changes.
 

·
Registered
Joined
·
617 Posts
Ah, the 90s ion sensing was extremely crude due to the lack of processing power available. It essentially just looked at the ion sense current and triggered a “knock” event. This works for a pretty stock engine, but significantly higher cylinder pressure looks the same.

The S85 was the first engine I read about where they were doing basic FFT to try to more accurately distinguish knock based on the ripple in the cylinder pressure profile. This takes way more processing power, and is what I’d actually call a true implementation of ion sensing for actually measuring knock.

I didn’t think the gen 4 and gen 5 coils were all that different, and all the knock sensing stuff would have to be on them. Does it feed back through another line? Honestly seems hard to believe, but I guess it might be true as there’s nothing technically stopping them from doing a basic knock sense circuit and triggering on some simple values. I highly doubt there was advanced FFT done on the signal, as Toyota really didn’t have to do it to hit the power level, and the compute cost would have been insane, especially in the gen4 era.
 

·
Registered
Joined
·
53 Posts
Discussion Starter #28
Ah, the 90s ion sensing was extremely crude due to the lack of processing power available. It essentially just looked at the ion sense current and triggered a “knock” event. This works for a pretty stock engine, but significantly higher cylinder pressure looks the same.

The S85 was the first engine I read about where they were doing basic FFT to try to more accurately distinguish knock based on the ripple in the cylinder pressure profile. This takes way more processing power, and is what I’d actually call a true implementation of ion sensing for actually measuring knock.

I didn’t think the gen 4 and gen 5 coils were all that different, and all the knock sensing stuff would have to be on them. Does it feed back through another line? Honestly seems hard to believe, but I guess it might be true as there’s nothing technically stopping them from doing a basic knock sense circuit and triggering on some simple values. I highly doubt there was advanced FFT done on the signal, as Toyota really didn’t have to do it to hit the power level, and the compute cost would have been insane, especially in the gen4 era.
I think cylinder pressure would be better to go off of anyway, but I believe you're correct. At this power level though, the stock ECU advances the timing so far that any degree further would cause knock. Like I said, below 21psi or so it's gonna make as much possible power as it can on the stock ECU. Aftermarket won't improve power any. Might make it easier to control boost and whatnot. Like you said, would make a lot of improvements in safety though.
 

·
Registered
Joined
·
617 Posts
More power = more cylinder pressure.
That’s why a bit of light knock at say 250 rwhp isn’t a big deal on a 3SGTE, but the same knock at 500 rwhp will probably blow stuff up. The 250 rwhp knock probably has peak cyl pressure close to the normal running 500 rwhp case. The 500 rwhp knock case... well, a stock block 3S probably won’t hold that reliably for very long.
 

·
Registered
Joined
·
152 Posts

The 2005-2007 ST246 has a normal knock sensor, wideband piezoelectric device.


The coilpack is a match for the MR2 Spyder. I don't think the MR2 Spyder has ion sensing knock control. What is most likely is that you are just running at the ragged edge of AFR/timing because Japan has better fuel.
 

·
Registered
Joined
·
53 Posts
Discussion Starter #31

The 2005-2007 ST246 has a normal knock sensor, wideband piezoelectric device.


The coilpack is a match for the MR2 Spyder. I don't think the MR2 Spyder has ion sensing knock control. What is most likely is that you are just running at the ragged edge of AFR/timing because Japan has better fuel.

The coils share parts numbers.. but don't even look alike. One of those things, I guess. The tuner says ion sensing coils, the polish documents they were given for the EMU Black and the gen4 PNP adapter mentions that you lose that when switching to aftermarket EMS among some other things. Maybe it's a bad translation? I'm inclined to believe that they have them due to the lack of data anyone has for the gen4/5 especially on aftermarket EMS.
 

·
Moderator
Joined
·
6,318 Posts
The coils share parts numbers.. but don't even look alike. One of those things, I guess. The tuner says ion sensing coils, the polish documents they were given for the EMU Black and the gen4 PNP adapter mentions that you lose that when switching to aftermarket EMS among some other things. Maybe it's a bad translation? I'm inclined to believe that they have them due to the lack of data anyone has for the gen4/5 especially on aftermarket EMS.
Watching CartuneNZ on youtube he does 1uz and 3uz work with LINK ecus and when he makes a harness for them he doesn't run the IGF wire I guess because like you said aftermarket ecus don't accept that signal.
 

·
Registered
Joined
·
53 Posts
Discussion Starter #33
Watching CartuneNZ on youtube he does 1uz and 3uz work with LINK ecus and when he makes a harness for them he doesn't run the IGF wire I guess because like you said aftermarket ecus don't accept that signal.
I'll have a buddy with the gen4/5 in their car actually look through their harness and see if there is an extra wire or lead going to the coils, but that's also no guarantee that they have ion sensing coils. I'm going off a tuner and a bad translation from an aftermarket ECU. So little data is available for these. I would also like to point out that the gen4/5 doesn't make more power on aftermarket EMS over the stock ECU even with high grade fuel and more boost than stock. Unfortunately the 18psi boost cut on stock ECU doesn't let us explore how much more boost it can handle before aftermarket EMS will start to extract more power than the original ECU can.
 

·
Moderator
Joined
·
6,318 Posts
I have a gen5 there are only 4 wires going to each coil, 12v, gnd, IGT (turn on), and the IGF ignition confirmation signal shared with all of the coils. Compare that with my ls430 with the 3uz, it has 4 separate ignition confirmation signals with 2 coils on each signal wire.
 

Attachments

·
Registered
Joined
·
53 Posts
Discussion Starter #35
I have a gen5 there are only 4 wires going to each coil, 12v, gnd, IGT (turn on), and the IGF ignition confirmation signal shared with all of the coils. Compare that with my ls430 with the 3uz, it has 4 separate ignition confirmation signals with 2 coils on each signal wire.
Sooo.. no ion sensing coils on 4/5th gen?
Going to have to ask the tuner to clarify what he meant by ion sensing coils, if they even have it. Strange that the ATS plug and play kit for EMU Black comes with a pinout and instructions in polish that says something about not being able to use ion sensing coils on aftermarket EMS.
 

·
Registered
Joined
·
152 Posts
Sooo.. no ion sensing coils on 4/5th gen?
Going to have to ask the tuner to clarify what he meant by ion sensing coils, if they even have it. Strange that the ATS plug and play kit for EMU Black comes with a pinout and instructions in polish that says something about not being able to use ion sensing coils on aftermarket EMS.
By far the most likely thing going on here is that IGF (which exists even on the 3S-GTE gen 2 btw) is actually used for misfire detection, not knock detection. To do knock detection with the spark plug requires a model of expected current flow across the spark plug gap and to measure that profile vs the expected profile. If it deviates in a way that resembles knock then the engine has to react and adjust fueling/timing.

For misfire detection it is much, much easier to do. You just have to see if the circuit closed across the spark plug at all. If it did then it probably didn't misfire. You just need a rising edge detector which is simple and cheap.

Modern ECUs still don't use ion sensing knock detection by and large, it's just cheaper to use a microphone knock sensor instead. You do have stuff like the E90 M3 using these kinds of knock sensors though: The new V8 Power Unit for the BMW M3 - Ion Flow Technology and Spark Plugs

Also keep in mind that this kind of knock detection is very sensitive to changes in the engine. You probably have to recalibrate the knock detection for any change that would warrant recalibration in a speed density system.
 

·
Moderator
Joined
·
2,417 Posts
Like @joshuaho96 says. IGF is Toyotas IGnition Fire sensing. It tells the ECU that the coil has fired. This is why the GEN4 and 5 coils have 4 pins instead of 3 like other coils you can get. ION is actually the signal return like IGF on the single coil engines.

Pin 1 IG1/2/3/4 (Power)
Pin 2 ION1/2/3/4 (Return signal)
Pin 3 IGT1/2/3/4 (Igniter signal)
Pin 4 C1- (C2-/C3-/C4-) (Earth)
 

·
Moderator
Joined
·
2,417 Posts
Like @joshuaho96 says. IGF is Toyotas IGnition Fire sensing. It tells the ECU that the coil has fired. This is why the GEN4 and 5 coils have 4 pins instead of 3 like other coils you can get. ION is actually the signal return like IGF on the single coil engines.

Pin 1 IG1/2/3/4 (Power)
Pin 2 ION1/2/3/4 (Return signal)
Pin 3 IGT1/2/3/4 (Igniter signal)
Pin 4 C1- (C2-/C3-/C4-) (Earth)
Having a look at what an ION sensing coil is, they are used for sensing misfire OR knock. I'm not sure what the exact function of the GEN4 coils ION output is but the GEN5 coils use the IGF pin name so it's possible they just changed the name of the output to be ION misfire detection instead of misfire detection. On the GEN5 there is only 1 IGF pin to the ECU from all 4 coils as they are joint. The MR2 single coil use IGF as the pin name. The knock sensing for the ION coil is possible but less likely although I have no proof of this.
 

·
Registered
Joined
·
53 Posts
Discussion Starter #39
What kind of power could someone realistically make on a gen2 3sgte with stock internals?
(If I had set it up on aftermarket EMS and had all the supporting mods / upgraded turbo and fuel system)...?

This may be a big reason I'll decide not to swap, if I have to build the internals at all I might as well grab a newer motor to put nice parts in.
 

·
Moderator
Joined
·
6,318 Posts
If I remember correctly there were some rare cases of unopened gen2s with 400whp or more. It really comes down to how good your tune is, how safe it is, and keeping IAT's down. How long a stock unopened gen2 will make that kind of power? Who knows. Get yourself an efficient turbo, I don't know where to point you for a good one but don't try to do this with an ancient turbo.
 
21 - 40 of 42 Posts
Top