No wonder EA82T's had problems making power :???:

[Gannon]

Images will reveal all,.. of dodgy factory crossover pipe construction

Dad and I borrowed the tractor off the guy i sold my car to, and while we were there, we got talking about my car and the guy showed me the new crossover pipe he was constructing.

He dismantled the factory crossover pipe to see what was inside and was disgusted, the crossover pipe is made up of one pipe inside the other, but to make it, the simply slid the smaller one inside the larger one and then crush bent them together.

The photos show the awful bend of the inner pipe inside the crossover pipe





The pipe on the right has a less than 20mm x 30mm oval gap (its smaller than the siamese port in the head) for the exhaust gasses to flow through. The T Join under head 1/3 is the best flowing part of the whole manifold.

This is the new one he is making, out of all 3/4" pipe, but is keeping the stock 2/4 side bend, and the 1/3 side T join. With a bit of porting with a dremel, they dont look to bad.



He has put a flange in the middle to make it easier to remove the pipe once in the car. There will also be a flange between the 1/3 head and the turbo, so the manifold can be broken down to 3 separate parts

Feel free to ask questions

[twilightprotege]

just cant help but laugh

[discopotato03]

It doesn't look real flash inside the crossover pipe but having the pipe lengths so different makes things awkward .
I think if he tries to have all the diameters the same meaning the port end bends and crossover it may create problems .

It's possible see what they tried to achieve with the std header and always remember that they don't make allowances for extra performance down the road from the showroom .

I like the idea of the TWE design and I get the feeling that they get away with (to a degree) slightly larger pipe diameters because the engine pipes merge just before the turbo rather than into a single up pipe .
The problem is that its difficult to get the two pipes up through the crossmember (or whatever) and things like power steering hard lines get a bit close and probably hot as well .

Having extra joints in the system is a good idea and the most compact way to do the fasteners would be to thread the flange plates on one side and use high tensile cap screws from the other side .

Have to get mine started soon , cheers A .

[ORX-18]

All the talk of how big is too big ect ect has prompted me to share my experiences. On the RX originally we had an equal length header system, 2 1/4" with all mandrels ect. Wasnt all that flash for the VF10 but when i changed to a VF34 it was a godsend. Sadly we heated it up and cracked the manifold beyond suitable repair so i knocked myself one up out of 2 1/4" and one out of 1 3/4" mild steel tube with simple crush/machine bends. The 2 1/4" was good for the top end of the VF34 but when we killed it i went to a Garret GT2841 hybrid BB Turbo. The larger Diameter Manifold was a serious power robber in the midrange so we are now running the 1 3/4 with the garret and its a VERY big improvement. If anyone is seriously thinking of making power out of an 82-T i suggest making up cheap manifolds and experimenting first, Mandrel bends and ubergood flowpaths are expensive! Its also a good idea to try portmatch the manifolds to the heads as good as u can as there a reasonably crude design and can benifit from this greatly. For the majority of the headwork ive been getting i have been using JHH in southeast QLD, John really knows his subaru's and has extensive knowledge on the head design and what to do with them. If its not an option i suggest power porting(the abrasive media paste type) its a great way to cheaply ahd quickly clean up the inlet and exhaust of the heads.

God, i could go on all day, but i wont!

Cheers, Allan

[discopotato03]

I sort of worked a few calcs based on the std header being ~ 31-33mm at the pass side heads 90 deg bend (ID) which I think worked out to be about 1 1/4" . From memory the drivers side tube into the headers "T" piece was 41 odd mm or ~ 1 5/8" ID .

To me it seems obvious that the problem was the difference in lengths of tube between the heads and the T , they went smaller on the long side and about 3/8 (9.5mm) larger on the short side in an attempt to compensate . For a given flow small tube diameters speed up and larger ones slow down the gas speed .
Then of course having a spindly up pipe would dampen the exhaust pulses even more .

Corky Bell published a good reference years ago called Maximum Boost , he says make the tube size the same as the exhaust port size but I reckon he intended this for engines with individual exhaust ports per cylinder . The EA82 "port" has two tight "S" bend ports feeding into the one outlet which is quite large - too large I believe for most purposes .
I chose 1 3/4" tube which would be approx 1 5/8 ID .
The TWE header allows you to use the same tube size for both ports because the lengths are 100 times closer than the factory design , I think if you just copied the factory layout in 1 3/4 it would flow better but not sure about reversion issues .

I can quote 99% of Garretts GT25 and 28 series BB turbos and cartridges but I don't know about a "GT2841R" . The numbers in Garretts system tell you in that case GT28 turbine (of which there are 2 types) and the second pair of numbers mean the major or OD of the compressor wheel and I've never seen a 41mm compressor in any GT ball bearing turbo . The "R" means rolling element or BB center section .
The latest and smallest addition to their BB range as of late last year is a GT2554R and its compressor is 54mm OD . Smaller GT series turbos are all bush bearing .

My hybrid BB turbo started out as the original GT2554R (not the one mentioned above) and I got GCG to remove its turbine and swap in the higher performance GT28 NS111 turbine , 9 blade vs 11 and 76 trim instead of 62 .
The original low temp spec GT25 turbine housing was changed for the S15 200SX/Silvia high temp material spec one and profile machined to suit the NS111 turbine . The idea of this is to have low specific manifold pressure and lots of flow capacity with good turbine response . The good turbine response comes from having a very lightweight turbine with all the aero smarts .

My old FJ20 twin cam turbo engine used a GT2860RS which is the same turbine in the larger 0.86 A/R (vs 0.64 A/R) turbine housing and with a large tip height 60mm 62T compressor in a considerably larger T04B 0.60 A/R compressor housng .
Now an FJ20 has ports you could pass an EJ20 through yet that engine could get positive inlet manifold pressure at about 15-1600 revs so my hybrid should work quite well provided the header tubes are not too big .

In time ...

[steptoe]

When you guys do or have done your experimentation , have you/will you include the size of the hole in the plate the turbo sits on. Being a turbo novice and only based on theory found in here I chose to use the smaller diameter hole in plate (38mm?) and got what I wanted - nice and torquey.Only when I had no choice than to use the larger diameter hole in plate (+5mm) and another motor did I experience something like turbo lag. But the one with turbo lag is more willing to freely rev out to near redline and shot it goes and snorts up the high end revs! Like some others out there I like to torque not to rev. So to cheaply experiment surely a ring to fit in plate to alter the diameter from large to smaller may be an easy option?

[discopotato03]

If you are talking about different engines there could be other reasons why the result felt "laggy" .
If you used a late 4 plug or hotwire type engine in an early car things changed like cam profiles and the early cars engine electrics (dizzy vs ECU timing control) means the advance curve wouldn't be the same .

The fella that fitted my RX's new engine had an 86L and two Vortex (85 and 87) headers to chose from . He said they effectively measured the same so picked the one in the best condition .

From looking at them I think the only difference was the top of the 87 Vortex header was slightly larger because the oxygen sensor screws in to the up pipe not the down pipe and they didn't want any extra restriction around the probes head .

The thing to remember is that while the std header doesn't flow super well the standard miniscule turbine housing doesn't either . Add to that a wastegate swing valve thats pathetically small and the system chokes up quite badly . Mind you it ment you got boost at low revs on a low compression ratio engine and low to mid range torque was what they were after .

Also lots of engine tech water's flowed under the bridge in the last 24 years .

A .