Ibreakstuff

Thanks for confirmation Karhu. So it would appear there are 2 different TK8 EDM heads, one underlined (gen2) and one boxed (gen3). I ended up coughing up the dough for Drischnie's heads, when they arrive I'll put some valves in and CC the combustion chamber. Then we can compare it to the USDM spec heads (which I think are around 39cc in the heads). I have a piece of plexi with a hole drilled for that.

Odd, so it seems that TK8 head can come both in underline and boxed EA82 then? Are you sure it didn't say "6 JAPAN"? Any idea what the combustion chamber looks like?

This was an enigma to me, surprised no one else has pointed it out over the years. There is only one thread that points out the casting numbers and its on ausubaru.

I came to similar conclusions, it would be interesting to compare the actual changes in chamber volume and how that would effect the compression ratio (lower in the TK8 head). The one flaw to the theory is that the even older molds have the same flat shape as the TK9 head. Over the years on the USMB, it has been quoted that Subaru made changes to the bridge area in the heads. But I see nothing to indicate that they did, at least between gen2 and gen3, other than the casting quality. I know the pressure and temp of the casting will have an effect on ductility and some of these are definitely better castings than others. This is hard to show in a picture too.

My wife only allows one room of the house to have car parts, nothing with smelly transmission fluid either. I traded some closet space for that clause in the contract LOL. I had to have a place to measure/blueprint parts at a normalized temperature, the garage temps swing wildly. With the heads, I'll make an effort to lay them out later and take pictures later today. Most are super dirty, I still have a lot of cleaning to do. Regarding the Euro spec cars, I thought that it was just the computer (EGT vs O2 controlled), and cat free exhaust that opened up the extra power. That TK8 head is the only one I have ever seen with that chamber shape, I sent her a offer on them.. if nothing else to document and cc the chamber volume for comparison. I appreciate the kind words Jeszek

100% positive. All of the heads I own came from turbo cars, RX's mostly. I have 9 heads sitting in the dining room (my wife hates me), all are dual intake heads. I can take pictures of each if you desire.

TK9: TK8: *note the combustion chamber shape

Just looking for some confirmation based on a recent epiphany. We know there are 3 "generations" of MPFI heads, but I am looking for how many molds there were and if each/any have subtle changes. I have found at least 8 different casting numbers and the molds seem to vary slightly. In my possession are MPFI heads with castings numbers TK4 TK6 TK7 TK9 TK10. And recently some EDM heads were posted with pictures showing TK8 (Thanks Drischnie). I know that TK3 and TK5 exist also, just from my stash of pictures. I should note also that most of my heads are from turbo cars, I would be interested to see some heads from NA mpfi EA's (XT4's in the states). The only actual differences between the head revisions seems to be the casting material and the quality of the cast itself. I think it is a myth that Subaru beefed up the bridge between valves and my calipers can't measure any differences. The TK9 heads seem to be the best quality castings imho and I have found a few of them without cracks over the years. I will avoid TK10 heads in the future. What my research has found: TK3 EA82 No lines (gen1) ? TK4 EA82 Underlined (gen2) USDM dot high above stamp TK5 EA82 Underlined (gen2) ? TK6 EA82 Underlined (gen2) USDM line above stamp TK7 EA82 Underlined (gen2) USDM dot just above stamp between E and A TK8* EA82 Underlined (gen2) EDM (thanks Karhu for posting) TK8* EA82 Box (gen3) EDM (thanks Drischnie for posting) TK9 EA82 Box (gen3) USDM smooth/worn casting lines but tight aluminum grains TK10 EA82 Underlined (gen?) USDM both examples I have sharp casting lines but are porous and brittle Notes/etc: -I ordered them by casting numbers, but I am not sure this would be the order they were molded chronologically. TK10 being underlined (often referred to as gen2) is what makes me question this. -The casting number can actually be viewed on the top of the head externally, it will say JAPAN #. The number will match the TK# stamp inside. -The combustion chamber shape appears to be the same on every head except the TK8 head posted by drischnie. This mold intrigues me. -Above the EA82 stamp are 2 numbers, these appear to be the date of the casting. Just month and day, not year unfortunately. I could be wrong and this could just be batch numbers, but the top number is lower than 12 and the second is lower than 31 (that I have found) Any input? Comments? edit: -Added second TK8 EDM head

From what I am finding, there are actually 6 revisions of the MPFI heads. Casting numbers are TK3 (gen1), TK5/TK6/TK7 (gen2?) TK8/TK9 (gen3).

Negative, mine are dual port MPFI gen3 ea82 turbo USDM heads.

From my local stealership and the 20% VIP subaru discount, a bare NOS ea82 turbo head is only $323.96.

This would be news to me, I never knew there were 2 different versions of the gen3 heads. Seems like the USDM ones have more meat between the valves just from visual clues. I wonder now if the EDM pistons are shaped different to offset the head volume. My eyes may be tricking me also, but the actual valve angle looks different too. Edit: My gen3 heads are stamped tk 9 and have slightly different casting lines on the cam side.

Could be as simple as loose or dirty pins in a connector. But the advice above is sound, there are many things that could cause it. I had a random start issue that was resolved by slightly bending a couple pins to make a better connection. I found the culprit by jiggling wires on the side of the road in the dark.

I have a set of USDM gen 3 heads and the combustion chamber is shaped a bit different than these. That V shape between the valves is different, if it increases the volume it would yield a lower compression ratio.

I have a few thousand miles on one of set of the MLS gaskets, but only on stock pressures and stock pistons. I still have not put one together with the head studs and SLA's yet. My wrx (motor #3) and ralliart (crash) have been drawing my attention away. The wife's 02 outback needed head gaskets too (ez30 not fun)... Trying to get back to tinkering with the RX's, the 88 is back in the garage at home. Uncle Sam (tax return) bought me a tig welder and a tube bender this year. I'll be making a couple of EA turbo headers this month.

Most of the squareback swaps go the other way, but would be a cool idea. Or a corvair drivetrain rear mounted?

450hp on stock internals is asking a whole lot from an ej20g. You would need a HUGE turbo to hit those numbers too... like a gt30r (~400hp) or gt35r (~450+ re: LAG).

And Gloyale was right, if they are not dished.. they are not turbo pistons. Would lead to a much higher CR. Even the ej205 pistons I'm playing with, they have a ~14cc dish.

I usually take your word as gold around these parts, but I have to disagree with the performance potential. The "factory" rally spec ea82t made ~181hp and just had a few basic modifications.. intake, lightened rods, and solid lash adjusters. Pretty much everything else was stock. I have the tech docs for modifying for rally spec, but none of the parts are avail. When I get out of the poor house I plan to prove you wrong with head studs, mls hg's, and cooling upgrades. Edit: I've posted these, but I'll just steal the link from ausubaru (Thanks Gannon): http://www.ausubaru.com/coppermine/albums/userpics/rxModManual.pdf

Clean car. Like you, I love the lines and the look of the 3 door. I still can't say exactly what draws me to this car shape. Looks like FWD from the way it's sitting. So 1987-89 GL/GL10 FWD?

Yes, you could easily disconnect it to test also. Edit: The HH is the first thing I remove, I wouldn't trust any HH on a steep San Francisco hill anyways.

Definitely don't run lower octane gas, you are asking for detonation there. If you connect a knock sensor back up, it should pull timing if a actual knock is detected. Can you take a video or sound recording of it? Rod knock is pretty rare on these motors. My guess is HLA tick tho, 5w30 is probably too thin. I would try 5w40 or 10w40 if I lived in your region (maybe even 0w40 for winter). Amsoil 15w50 dominator is my new favorite ea82t oil, but I live in a much warmer climate.

Sounds like the clutch cable is too loose and it is not fully disengaging when you press the clutch in.

For hsoj and 175eya: (thanks to numbchux) http://www.ultimatesubaru.org/forum/topic/63821-pics-of-lsd-conversion/ The carrier is the guts of a diff, the part that does the work. The idea is to replace the open carrier with the lsd carrier but keep the ring, pinion, and case matched. Edit: More good reads.. http://forums.nasioc.com/forums/showthread.php?t=2597453 (the clsd shown here is ours, 2way plated goodness) http://www.scoobymods.com/showthread.php/diy-upgrading-replacing-rear-differential-14270.html

You could do the old suicide trick, pump the exhaust gas into the garage for a few hours. Try not to kill yourself in the process. Nothing survives the summer in my storage unit.. mice, spiders, and moths all succumb to the heat. I prefer to have a car survive the summer in the box before working on them lol.