Interesting differences between EJ251 and EJ253 engines

Ratbag

Ratbag

Administrator
Joined
Jan 1, 2012
Messages
7,473
Location
Bayside, Melbourne, Vic
Car Year
MY06, MY10
Car Model
Forester SG & SH
Transmission
5MT/DR & 4EAT Sports
Gidday All

One way and another, there has been a lot of discussion on the forum about how various models (i.e. engines) perform.

After talking to NachaLuva yesterday and comparing the EJ20 in Oswina with the EJ253 in Roo2 (SG series II), I decided to do some basic research ... :iconwink:

So I started here:
https://en.wikipedia.org/wiki/Subaru_EJ_engine#EJ253

That article states that the later EJ253 donk produces about 5% more KWs (121 -> 129) and about 1 Nm less torque than the earlier EJ251 and the EJ252. The maximum torque is produced at 4000 rpm instead of at 4400 rpm, however.

So what causes this?

I have mentioned before that I thought the introduction of the VVLT in the 2006 Forester had a significant impact, but it also appears that Subaru changed from using a MAP sensor to using a MAF sensor with this engine upgrade.

NL also noticed that Roo2 has what appear to be tumbler valves at the bottom of the intake manifold.

Anybody know of/suspect any other changes to the engine at this time?
 
Last edited:
Ratbag said:
NL also noticed that Roo2 has what appear to be tumbler valves at the bottom of the intake manifold.
Click to expand...

Tumbler valves are for cold start emissions, more plumbing! :rolleyessarcastic:
They change the way the air flows into the cylinder and somehow that reduces emissions. Getting rid of them supposedly gains some low end HP, since they only effect anything at low rpm what's the point.

More readings:

https://forums.nasioc.com/forums/show...6&page=2&pp=25

https://www.awdpirates.net/phpBB2/vie...953&highlight=

https://www.wrxfanatics.com/index.php...c=36257&hl=tgv

...more pain for no gain.
 
the EJ251/252/253 engines have a SERIOUS problem with the 11044AA632 head-gaskets. They will leak coolant and/or oil externally.

If you install the 11044AA642 head-gaskets (stock for 2004~2006 EJ257 and EJ255) in the EJ251/252/253 engine...you will NEVER experience the external leak EVER again.

Also the ej253 is similar to the 251 but has the 'active valve control system' (AVCS).
 
^ Forget the low-end power ...

I suspect that they also contribute to the better low-end torque ...

If they are giving even slightly more low-end power, then logically they must be giving more low-end torque ...

Fact is, the series II SG does have more low-end torque than the series I. This must be coming from somewhere/something! While I have put it down entirely to the VVLT in the past, it appears that there are also other modifications contributing to this.

NL also noted that the extractors have changed from the earliest SF to my SG. Now that's hardly surprising. The '97 SF donk is very similar in lots of ways to the EJ18 in Roo1. I'm wondering if the extractor design changed from the SG series I to the SG series II?

When I modified my Morris 1100 (way back ... ), one of the changes I made was to fit the BMC Works full-bore extractors. They made a significant difference to the power and torque over and above non full-bore extractors on the "stock" Mini Cooper S ...

It also appears that the SG series II can get somewhat better fuel economy around town than the series I. e.g. If I am careful with my rather leaden right foot, I can get around 10L/100 kms around town - just so long as the entire tank isn't used going to the local shops and back! If I am heavy-footed, that drops to around 12.5L/100 kms (or if all the trips are to the local shops and back ... ).

So I am interested in any other differences that more experienced people than I may have noticed that could contribute to the changes in torque and fuel economy.
 
Gidday Pigs

PigSti said:
the EJ251/252/253 engines have a SERIOUS problem with the 11044AA632 head-gaskets. They will leak coolant and/or oil externally.

If you install the 11044AA642 head-gaskets (stock for 2004~2006 EJ257 and EJ255) in the EJ251/252/253 engine...you will NEVER experience the external leak EVER again.
Click to expand...

According to the sources I have read, that problem only occurred with the 2003 version of the EJ25. It was the subject of a recall by Subaru. Didn't occur in the later engines.

Also the ej253 is similar to the 251 but has the 'active valve control system' (AVCS).
Click to expand...

AFAICS, only the turbo motors use variable valve timing technology (VVTT = AVCS). The N/A series II donk uses variable valve lift technology (VVLT) instead. Don't know if the FB engine uses either of these, or not.

VVLT uses one rocker arm and one cam lobe for each of the two inlet valves in each cylinder. VVLT, OTOH, uses three cam lobes per inlet valve pair.
One valve rocker arm runs on a single lobe at all revs, the other valve has two rocker arms, and which one is in use (and which cam lobe it is following) at any given time is controlled by an oil pressure switch that locks one or other rocker arm to the rocker shaft dependent on the engine revs. This valve is controlled by the ECU. This allows the VVLT engine to use three different cam lobe designs instead of two. THIS DESCRIPTION IS WRONG - see post below ... RB

I am also rather interested in why one cannot use both VVTT and VVLT in the same engine, but maybe that really does make things too complicated!
 
Last edited:
Ratbag said:
I am also rather interested in why one cannot use both VVTT and VVLT in the same engine, but maybe that really does make things too complicated!
Click to expand...
I may be way out & on the wrong track, but my old Celica (2001 yr model) had VVTLI (Variable Valve Timing with Lift Intelligence)
So I can't see as to why it can't be done :shrug:

Regards
Mr Turbo
 
vvl and vvt can be used on the same engine, I think its just a matter of cost. I can't remember what engines exactly so take this with a grain of salt, but iirc there's some aircraft engines and a honda and bmw engine, and more than likely f1. The benifits of having both i was reading was highlighting the ability to almost completely control the engine timing on the fly.

I also found an interesting animation explaining some of this vvt/vvl https://www.youtube.com/watch?v=K0FUWFLTWCA
 
Last edited:
hahah! my ej20 only manges to get to 10/100 when im driving like a hoon
 
I reach my max torque of 225nm at 3600rpm. 80% Peake of thqts is reached at 2200rpm.
I never use more than 11.5l per 100km unless towing.
 
Gidday Taza

That's pretty impressive torque mate (225 Nm)! And at 3600 rpm. That's 400 rpm lower than the same torque figure from the EJ253 series II at 4000 rpm.

Which block do you have?

Regardless of that, you are still using your EJ20 heads, so the VVLT etc doesn't come into it ...

It indicates to me that one should be able to get considerably more from the series II EJ253 donk with a tune.
 
I have the Ej251 first of the SOHC 2.5s. Yeah I reckon with my tune I've squeezed abit more torque and Kw out of it too. I am running ej25 heads, not ej20.

Yes you are right. You can squeeze a decent amount of torque out of your engine with a tune.

I am not truly happy with my piggy back computer. I reckon with a full art

Aftermarket. One I could achieve much better power, torque and economy gains. ..
 
Interesting, Taza.

Which EJ25 heads are you using?

That torque at such low revs is consistent with having both inlet valves on relatively low lift cam lobes.

From my Batphone
 
Ej251 heads. SOHC 16valve. Heads and block are from a 2003 Outback. Supposedly has 115kw. In comparison to my stock 2.0l which had 92kw and 180nm of torque it's a very good improvemwnt. These specs are on paper.
Yet although not that impressive or powerful in the real world it's a fabulous motor. The low peaking torque makes general driving and hills effortless. So much you don't even think when it comes to them. Jump in a stock SF foz and then I realise just how many gears you have to change through the same roads. .
The only thing I find is that it still lacks a little torque but that is due to the oversize tyres. Put stock size on and it moves along really swiftly.
 
I would think that most of the power/torque differences would be in the tune.

Yes there are other factors involved in the physical construction of the engine - cam profiles, intake design, head flows - and the different engine speeds with different flow rates etc; combustion chamber design, piston dishing (or not); valve position, design and opening depth and time...

Also add in injector size and flow rates with the ECU tune and you'll be able to talk about it until the cows come home and beyond - and you'll still not get to a definite answer IMO.

Gearbox doesn't have anything to do with engine output - engine out put is just that - what the engine can produce at the crank. If you were talking about power/torque at the wheels then that's another kettle of fish!

Just putting it out there.

Cheers

Bennie
 
If the extra power and torque comes from a tune, why don't they do it at the factory?
Is the tune you guys mention purely in the ECU?
 
Gazoola said:
If the extra power and torque comes from a tune, why don't they do it at the factory?
Is the tune you guys mention purely in the ECU?
Click to expand...

Yep, the way they programme the ECU.

Most engines are detuned for the domestic market to give better fuel economy & reliability. Best example are some of the 80's Japanese 4 cyl turbo engines designed from F1 engines, then detuned for sale in "pocket rockets"
 
Gidday Bennie

El_Freddo said:
I would think that most of the power/torque differences would be in the tune.
Click to expand...

It probably accounts for a lot of the improvement over and above the characteristics of the stock N/A 2.5L donk, in Taza's case. However, you can tune a N/A 2.0L donk till the cows come home, and it will not give you this sort of torque, or torque curve, more importantly.

A classic example of how not to do the electronic tune was my '93 Impreza (EJ-18). It was relatively gutless until one got to around 4,000 rpm. It was like driving a car with a full race camshaft in many respects.

Yes there are other factors involved in the physical construction of the engine - cam profiles, intake design, head flows - and the different engine speeds with different flow rates etc; combustion chamber design, piston dishing (or not); valve position, design and opening depth and time...

Also add in injector size and flow rates with the ECU tune and you'll be able to talk about it until the cows come home and beyond - and you'll still not get to a definite answer IMO.
Click to expand...

Without these things - and the extra engine capacity, bore/stroke ratio, compression ratio, etc - you can re-program the electronics, etc ... ^ :poke:

Gearbox doesn't have anything to do with engine output - engine out put is just that - what the engine can produce at the crank. If you were talking about power/torque at the wheels then that's another kettle of fish!
Click to expand...

Agreed. But the gear ratios that the manufacturer chooses have an enormous impact on the way a car drives. My Austin Kimberley had a very nice SOHC, cross-flow head, 2.2L straight 6 engine, with plenty of carburettors (twin 1 3/4" SU), extractors, etc.

Then they put a 4 speed box under it, with FD gearing of 16.9 mph/1000 revs in top ... !! UGLY. Real fugly. That's only 27.2 kmh/1000 revs for the Imperially challenged ... Revving its guts out at 100 kmh! Even my '93 Impreza could manage about 32.5 kmh/1,000 revs in 5th!

So I had FW Hercus in Adelaide make me a new pair of final drive gears (sidewinders) that gave me 22.5 mph/1000 revs in top (or could have been 23.5 ... It was a long time ago - cost me $284, IIRC. A King's ransom in the early 1970s!).
22.5 mph/1000 revs is around 36 kmh/1000 revs, a whopping 33% increase! After that, I was doing around 3,000 rpm at 110 kmh, vs. just over 4,000 rpm at that speed before.
Given what I know now about torque curves and gearing, the engine would probably have stood up for ~25 mph/1000 revs - just over 40 kmh/1000 revs.

Keep in mind that the N/A 2.5L Forester is a heavier car with a bigger engine, and is geared at 38 kmh/1,000 revs (MT) and 40 kmh/1,000 revs (AT).

Also, my '93 Impreza had gear ratios that were all wrong for its weight and engine capacity and torque curve (specially with the rotten factory tune ... ).

So gearing does have an effect ... Particularly when it's completely wrong.

Just putting it out there.
Click to expand...

Ditto ... :iconwink: :biggrin:
 
You must log in or register to reply here.
Back
Top