On Tuesday I took my car up to KTR Performance in Massachusetts to make use of their Dyno Dynamics AWD dyno. The last time I had the car dyno'd was down in Maryland about six months ago on a Dynojet. (http://forums.nasioc.com/forums/showthread.php?t=1021186&highlight=obx+dyno) Some additional info is here. (http://forums.nasioc.com/forums/showthread.php?t=1031584&highlight=pat+olsen+dyno+obx) As that first linked thread mentions, I've had the car on a Dyno Dynamics before, but that was back in 2002, so I wanted to see what the car would do now. No real changes to the car since I did the Dynojet runs, other than mileage (about to roll 179,000 miles vs. 167k on the Dynojet 6 months ago vs. <100k the last time it was on a Dyno Dynamics), bad head gaskets :rolleyes: , and a throttle body spacer (which, in addition to re-baselining the car on a Dyno Dynamics, was the other reason I went to the dyno).

The last time the car was on a DD it put down 142.5hp and I don't remember how much torque. Unfortunately, the computer at the dyno hiccuped and I never got the data for that run, I only saw the number on the screen. So, here we are, 4-1/2 years and 80 or 90k miles later, and it throws down:
http://www.submariner.org/thepno95/Pictures/Subaru/Dyno%20curves/TBS%20vs%20nonTBS.jpg

I drove up to the dyno and did the first 3 runs with a 1" throttle body spacer installed. I then removed the spacer and did 3 additional runs. The graph shows the average of the 3 runs in each condition, although the data's been manipulated a bit at the low end because of some dyno difficulties. Long story short, the dyno didn't seem to think to highly of the Cusco Tarmac Gear 35/65 center diff, and was getting pretty confused. In the end, two guys had to do the runs (one in the, one manually operating the load controls), and due to the way they were tricking the dyno the runs weren't very consistent below about 3000rpm or so. One of the TBS runs didn't even start until ~2700rpm, and doesn't join up with the other 2 curves until ~3500rpm, so that data isn't included in the average curve displayed. Similarly, the huge dip in the non-TBS curve from 2200-3100rpm isn't "real", that was just a factor of trying to trick the dyno to do the runs properly.

Although the low end wasn't very repeatable, from 3500rpm on up the curves all ran pretty much right on top of each other, so that was nice to see. The only misfit run was the first run after removing the spacer, which only pulled 141.3hp, but the 2nd and 3rd runs were identical at 145.7hp at the same 6400rpm. AFRs were pretty fat at the top end, slowly dropping off from 12.5 @ 5000rpm down to 12.1/12.2 @ redline, so I could probably have picked up a little more power by leaning it out, but I didn't bother to mess with the S-AFC settings.

Due to the poor low end repeatability, I can't definitively say whether the TBS is a worthwhile modification or not. Is the little dip from 3100-3800rpm just an anomaly due to the uneven dyno loading, or is it indication that the TBS was smoothing out the torque curve there a little bit? I really can't say.

Next on the project list is to pull the current engine to be rejuvenated, and in will go a stock '98 DOHC engine for the meantime. I still plan at some point to do a torque box vs. CAI intake test, and I may re-perform the TBS test (assuming I can get some answers on how the dyno should be set up to get proper runs.)

Pat Olsen
'97 Legacy 2.5GT sedan

what about the knock sensor this time? on the dyno runs you made last june, you were getting more power in the midrange with the knock disconnected; it is safe, do the gain remain the same in the long runs?

So one thing that is established is that it gives a fraction better low end and sacrifices a little bit at the top. That makes sense for better gas mileage, since we do normally drive around town in the lower rpms. The rpms below 3k are still inconclusive as you had stated, but do you think that it may suggest more power, regardless of the exact quantity?

what about the knock sensor this time? on the dyno runs you made last june, you were getting more power in the midrange with the knock disconnected; it is safe, do the gain remain the same in the long runs?
The knock sensor is still dummied with a resistor. Is it safe? I don't know. Are my head gaskets bad because I'm getting detonation, or are my head gaskets bad because it's a DOHC EJ25 that is known to have HG problems? I certainly wouldn't tell anyone else to bypass their knock sensor, but I'm willing to take the risk. In my particular case, the car runs substantially better - it's a night and day difference.

So one thing that is established is that it gives a fraction better low end and sacrifices a little bit at the top. That makes sense for better gas mileage, since we do normally drive around town in the lower rpms. The rpms below 3k are still inconclusive as you had stated, but do you think that it may suggest more power, regardless of the exact quantity?
I'm assuming the "it" you're referring to is the throttle body spacer. As I said, the data simply doesn't support drawing any conclusions. Here are the raw runs - how much torque is the TBS making from 2000-4000rpm?
http://www.submariner.org/thepno95/Pictures/Subaru/Dyno%20curves/TBS%20runs.jpg
There's enough variability in those 3 runs to cover the difference between the TBS and non-TBS runs.

Not to mention that what the horsepower and torque curves look like at WOT on a dyno don't really tell you anything about what sort of mileage the car is going to get at part throttle.

Pat

That dip, from what I understand is normal. I'm suprised it disappeared with the spacer. It might be from a hiccup in the flow or wave tuning of the intake. I'm not sure. I know I-Speed found this hiccup early in their RS reflash tuning, and they just tuned it out of the ECU. They think it's part of the ECU programming, maybe something it does between closed/open loop or something goofy. You always get a noticable dip and then your power comes on hard after that. I did get a RS reflash from I-Speed and dynoed the car afterwards on a local Dyno Dynamics up in MN with the reflash as the only mod. I never got a fully stock dyno, but the I-Speed makes more power everywhere and smooths out the overall behavior. One thing it does is smooth out that bump around 3000rpm. In my dyno, there is no bump, rather it just keeps on climbing and pulling till it tops out around 4000rpm then starts dropping off. It makes me wonder why that mechanically disappears. Interesting.

Get some butt-dyno time and a couple tanks through the car and see how you like the spacer. They will soon be available through TiC, so a lot of hands-on experience/info is nice for the potential buyers. It's good to see some dyno runs, although, the low rpm stuff needs to be sorted out. I plan on getting a spacer when TiC and geddesk2 gets some available. My car should hopefully be up on the dyno again by Jan for some Open ECU fun. I wouldn't mind adding another set of TBS/NonTBS results to this thread and some butt dyno comments as well.

Your top end certainly breaths well. What work have you done? My virtually stock setup just plumets after peaking at 134ft-lb around 4000 rpm. :lol: I'm hoping a shiny new full exhaust will fix it and make 'er keep pulling till redline. If I could get a curve like yours, I'd be happy.

Is the midrange torque that different from the DOHC to the newer SOHC engines? I would have expected it to be more peaky around 3500-4000 rpm. Just curious...if you happen to be familiar with both types.

OK, you guys are killing me.

That dip, from what I understand is normal. I'm suprised it disappeared with the spacer.
There is no dip. Look at the thread I linked to with my Dynojet runs. There is no dip. There's a hump up at ~3800rpm or so, but there is no dip. I have graphs from stock runs in Jan 2002, no dip, MRT exhaust in Jan '02, no dip, Cobb heads and cams in mid '02, no dip.

And did the TBS really take the dip away? Did you look at the graph of the runs with the TBS? Look at the greenish torque curve - omigod, a dip at 3000rpm! But it's not there on the yellow curve. And obviously the third curve is no help whatsoever. So, once again, at the low end the graph is inconclusive. There is not enough repeatability in those three curves to draw any useful data from.

Get some butt-dyno time and a couple tanks through the car and see how you like the spacer.
I've already reported my butt dyno results in the TBS thread. I felt zero difference installing or removing the spacer. I don't intend to put it back on the car.

Your top end certainly breaths well. What work have you done?
Read the linked dyno threads.

Is the midrange torque that different from the DOHC to the newer SOHC engines? I would have expected it to be more peaky around 3500-4000 rpm. Just curious...if you happen to be familiar with both types.
It's in the FAQ/Threads of note sticky. http://forums.nasioc.com/forums/showthread.php?t=195694

Pat

Thats some inconsistant data there... Cant do much of anything with it.

Sorry, I'm an idiot. :rolleyes:

I completely missed the link as it is normal, same color text, so it was non-existant to me unless I went over it with my cursor.

I've actually seen your MRT/Cobb thread a while ago, but I didn't link a name to it so didn't know it was you.

I haven't been keeping up in the TBS thread, so I missed your comments there.

The dip? I'll leave that up to debate. It's noticable in stock form just through normal driving. The engine has a little hiccup there where it's switching modes or something and feels like a slight lag time. It really bugged me how goofy the stock ECU behaved, and it's one of the main reasons why I got an I-Speed reflash. I know I-Speed commented about the dip early on in their development as well. I know the '98 ECU is different and doesn't "learn" things, and that may be why it's different than the newer products. Maybe the newer ECUs are set up different or "learn" the dip. I don't know. The SOHC vs DOHC doesn't really show any dip, so ... :confused:

Sorry to take a couple years off your life. :lol:

No harm done. :) I have to say that I'm not a big fan of the hyperlinks on this forum, as they're not very obvious.

As for the dip, I've never seen any dyno results of a stock "new" 2.5L, so I'm not sure what the story is. I know that my car never displayed any sort of dip when new, and the stock or nearly stock '99-01 SOHCs I've seen dyno'd didn't display anything like that, either. I have seen the I-Speed dyno graph somewhere, but that's the only one I've seen like that. I can't see Subaru going backwards with with ECU tuning, so it doesn't make sense to me that the '02+ SOHCs would have issues.

And, just like any other EFI ECU, the DOHC ECUs learn. I see comments on here pretty frequently implying that someone the "old" ECUs don't learn, or that the "new" ECUs learn in some fundamentally different way that makes modifications worthless. I'm not sure what the basis of these types of comments is.

Pat

I've seen the same too with the old and new. I know I-Speed doesn't want to flash pre '99 ECUs because they say it's a real pain to do. It's different in some way, but I don't know how. As well, I've seen the comments that you simply piggyback the older ECUs and tune as you like and the ECU won't learn and mess up the piggyback setup. The whole ECU tuning thing is pretty foreign to me, so I take it as others say.

I'm not certain about the dip either. On a dyno, I'm not sure how it would look, but the car does behave goofily around that range as it switches programming in some way. I've owned 2 '02 Foresters, and both drove the same. I've never liked the noticable changes in power nor its inability to cope with quick throttle changes. I can't really say if there's an actual dip in the dyno plot. Rather, I'll just say there are noticable fluctiations that occur. I-Speed fixed all that, so I'm happy.

Back on topic...

I've noticed two conflicting stories with the TBS.

First, most people comment on the improved low end torque with the TBS(when you look through old treads). Initially, I would assume this is atributed to the longer runner length. It would adjust wave tuning in the same sence a header with longer primaries would lower the peak torque range, wave tuning and scavenging. This would also justify the milage gain as well. At least, this is my thinking, along with some ability to improve flow linearity by forcing the air to flow straight for a short distance. There's been the cooling the intake air temp debate too, but I can't see too much of a change over 1" with air traveling at quite a few feet per second.

Second is TiC's definition of what the TBS does. They mention it increasing the...well I'll just copy it:

"Ok, so the question remains - what does a throttle body spacer do? It's actually really simple. It moves the throttlebody a set distance away from the intake "spider" to increase intake plenum volume. See, for top end you want as big a plenum as you can fit in there, BUT there's a trade off. A huge plenum will have an adverse affect on idle, low end and mid-range power. Therefore there's a compromise. That compromise is this - ideal plenum size for a street driven vehicle that needs power all though the rev band is 50-70% of your engine's displacement."

So, they're saying it has an opposite effect, increasing plenum volume and improving top end performance but sacrifices low end/idle performance/behavior.

So what's correct? Or are these two completely different aspects of the intake? ...kind of like an exhaust header having primary and secondary/collector sections(as well as the full exhaust) that work similarly but independently as well for a net effect. Sorry about the exhaust references, but it's kind of how I picture the intake in my mind as they both work in a similar fasion controlling/optimizing flow speed and linearity as well as pulse wave tuning.

I thought I had posted this in the big TBS thread here, but I guess I only posted it over on RS25.com. IMO, TiC's description is correct. Here's what I had posted on RS25:

The EZ30 manifold has an Iris valve that splits the intake plenum in half. The SVX's EG33 had the same sort of thing. From what I've learned on the SVX forums, the Iris valve is shut <4000rpms, which splits the plenum in two - smaller plenum at low RPMs. Once you get above 4000rpm, the Iris valve opens, giving you a larger plenum for high RPMs and creating a 2nd torque peak.

Based on that, I think the TBS will move the torque peak higher since it'll create a larger plenum. Make sense?

Pat

So thats like around 190-200chp? what kinda engine mods? ITs got a nice steady, even torque and power band. I'm assuming you have cams cause you are gettin peak power over stock redline haha.

So thats like around 190-200chp?
I would say more like 200-210. The average of 4 or 5 stock dyno runs 4 years back was ~103whp.

what kinda engine mods? ITs got a nice steady, even torque and power band. I'm assuming you have cams cause you are gettin peak power over stock redline haha.
Originally Posted by Back Road Runner
Your top end certainly breaths well. What work have you done?

Read the linked dyno threads.
Pat

Today I did a compression check, something I've been meaning to do for a while just to see how healthy (or not so healthy) the engine is. I rolled 179,000 miles over the weekend, and I'm still cruisin' around on blown head gaskets. (Actually, maybe just one is blown based on the compression test results).

Numbers:
1 - 180#
2 - 180#
3 - 120-135# :eek:
4 - 195#

For each one I let the car crank 5 times, then hopped out to look at the gauge. 1, 2, and 4 were at the posted value at the first glance. #3 was at 120#, so I cranked it over 5 or 6 more times, at which point it was up to 135#.

Imagine how much power the car would put to the rollers if all 4 cylinders were doing their share! :)

Pat

Rebuild time!

Complete spare engine (oil pan to throttle body) is sitting in my garage already, as is a spare Phase II shortblock (which the Cobb heads and cams will find their way onto).

Pat

That's about 45-48 over stock? Good job.

I have all this data in Excel, and I've created various graphs, but I hadn't done just a "now vs. stock" graph. This really drives home what all the work has done to this engine:
http://www.submariner.org/thepno95/Pictures/Subaru/Dyno%20curves/stock%20vs%20Cobb-Injen-OBX%20built%20EJ25.jpg
I think a better run, with more reliable low end data, would make the low end and mid-range look better, but even with the questionable numbers it's almost 10ft-lb even at the low end. +45ft-lb and +52hp at 6400rpm! The difference in the graph from about 5000rpm on up amazes even me.

Pat

She's definately pulling hard all the way up.:D

Complete spare engine (oil pan to throttle body) is sitting in my garage already, as is a spare Phase II shortblock (which the Cobb heads and cams will find their way onto).
I put the stock engine into the car 3+ weeks ago. Still using the Injen CAI, Fidanza flywheel, and full exhaust, so it's not really stock, of course. I'm impressed by how the low end and mid-range feels - doesn't feel much slower at all. However, from 5000rpm on up you can definitely feel the stock cams and heads just aren't up to the task. The engine really seems to be straining to redline instead of racing to redline like it did with the Cobb stuff.

I plan to dyno the current setup soon. I have the next couple Fridays off, so I just need to set up the dyno time.

Pat

I plan to dyno the current setup soon. I have the next couple Fridays off, so I just need to set up the dyno time.

Pat

Have you had a chance to setup dyno time? Really nice numbers you have.

I put the stock engine into the car 3+ weeks ago. However, from 5000rpm on up you can definitely feel the stock cams and heads just aren't up to the task. The engine really seems to be straining to redline instead of racing to redline like it did with the Cobb stuff.

Ditto. I am back to never revving past 5000rpms. The car is being picked up this weekend, so I can now focus on other things...

Have you had a chance to setup dyno time? Really nice numbers you have.
No, I've been busy with other stuff and haven't gotten around to it. I definitely still plan to get the car to the dyno again, I just need to find a free day. The good news is that there's another shop with a Dyno Dynamics that's about 45min closer to me.

Pat