Inlet Duct PMD

[Anonymous]

I've just done a poor man's dyno  back-to-back test of the H&S inlet duct vs the OEM one. The inlet duct, by the way is the elbowed pipe that feeds the stock airbox - it runs just under the n/s rear light (in the UK anyway). Here are the results:

MR2 Roadster (UK) 2003 model
Mods: H&S twin exhaust; Corky's anti-flex plate
Test conditions: Temp - 16 degrees C; Fuel - Shell Optimax 98 RON; Car - fully fuelled and 1 person - 11.5 stone. Spare wheel in front - as stock
All tests performed in 3rd gear starting from 2K revs and starting timing from 3K revs with roof down and OEM air deflector up.

PMD Test 1 - OEM Duct

H & S  twin, std inlet duct, Corky's anti-flex plate 3k to 4. 5k - 5.06s

H & S  twin, std inlet duct, Corky's anti-flex plate 4. 5k to 6.5k - 7.65s

H & S  twin, std inlet duct, Corky's anti-flex plate  3k to 6.5k - 12.71s


PMD TEST 2 - H&S Duct

H & S  twin, H & S inlet duct, Corky's anti-flex plate  3k to 4.5k - 4.87s

H & S  twin, H & S inlet duct, Corky's anti-flex plate  4. 5k to 6.5k - 7.16s

H & S  twin, H & S inlet duct, Corky's anti-flex plate  3k to 6.5k  - 12.03


The results show that the inlet duct definitely makes a real difference. The bulk of the gains are at the top end of the rev range, which is what we would expect. The main thing that I'm happy about though is that this is not at the expense of low down tractability since there is also improvement at the low end of the rev range.

The results seem bear out the theory that the restriction in the OEM duct is for noise attenuation since I can't think of any other logical reason for it being there - it certainly doesn't do anything for low end performance.

I must admit that on the butt dyno I could feel an immediate benefit to the H&S duct - this was highlighted by the fact that I had swapped from the OEM to the H&S in the space of approx 5 minutes and then immediately gone back out for a drive whilst the feel of the OEM duct was still fresh in my mind.

[GSB]

Interesting results... good work...  s:wink: :wink: s:wink:

[MRMike]

Good work Rusty, that intersting admittedly, TBO I put more faith in your but dyno than the PMD methodology, and if you say you can tell a difference thats good news.  Can't wait to get mine..

On a bit of a side note, whilst I appreciate its only a guide, I really do not see the value in the PMD methodology.  There is IMO no way you can accurately account for 0.19 seconds difference when recording on a stopwatch, between 3 to 4.5k revs..I'm not sure why it's caught on really?

[Anonymous]

There is IMO no way you can accurately account for 0.19 seconds difference when recording on a stopwatch, between 3 to 4.5k revs..I'm not sure why it's caught on really?

agree with that mike but at the top end, there's nearly 7 TENTHS difference which would surely mean there is a gain.

Rusty, how about doing one more run but with no elbow pipe?
im running without it and dont really want to go back as it sound soooo much better!   s:wink: :wink: s:wink:

[MRMike]

agree with that mike but at the top end, there's nearly 7 TENTHS difference which would surely mean there is a gain.

Rusty, how about doing one more run but with no elbow pipe?
im running without it and dont really want to go back as it sound soooo much better!   s:wink: :wink: s:wink:

Yeah sorry didn't meant it to seem as though I was saying Rusty wasn't measuring accurately, I was merely stating that no-one will be able to stop a stopwatch accurately enough to gleam any meaninglful data from the PMD.

Based on Mark and Steve's thinking I don't think there is any doubt that this pipe will prove beneficial, I wouldn't have bought one otherwise   s:wink: :wink: s:wink:  

I agree maybe some runs without would be a good comparison

[Bongo]

I guess if it were averaged over a number of runs it could give an indicator?

[Anonymous]

Rusty, how about doing one more run but with no elbow pipe?
im running without it and dont really want to go back as it sound soooo much better!   s:wink: :wink: s:wink:

That could be interesting... your increasing the amount of air being sucked in but then again its a LOT hotter air, much like using a cone filter so it would probably balance out and you would either be equal power or maybe down a little.

I have to agree though, it does sound good with the pipe removed but i noticed a slight lag when i had mine removed.

[Anonymous]

Just for the record, results were averaged over six runs in order to reduce human error factor. Also all results were measured with the roof down and OEM wind deflector up. I must also add that out of all of the readings, none of the OEM duct ones were faster than the H&S inlet duct ones. Also, and I know it's a subjective opinion, but I reckon I got about 1 perfect run for each set up out of the six and even the worst run with the H&S was better than the best run with the OEM duct.

I'll fully own up that I was totally sceptical before carrying out the testing, as you'll probably have guessed from my previous postings. I was on a mission to prove that OEM was good and H&S was not but the results prove otherwise so I've changed my view and the H&S duct is still on the car.

Also, sorry to go on, Mike Coughlin form SC pm'ed me and suggested greater gains can be had by extending the end of the intake tube towards the side inlet to provide a beter cold air feed. Must admit I'd forgotten about doing that but now it's next on the list.

[MRMike]

I'm not sure how much benefit there is in having more airflow, into the engine. I've attached vacuum pipes to the airbox and it has done nothing on the GTECH. I cut into the base of one airbox two holes that would accept vacuume hoses, which I draped below, with the ends facing forward.  It did nothing.  Probably because of the arc of the vacuum making it literally impossible for a significant charge.  So that idea was scrapped. There's just no way I can see of feeding additional air into the stock filter location, beside the route Toyota have already identified.

I'm not even sure how much benefit there is in having a massively increased airflow, because on a WOT the ECU resorts to its pre identified maps anyways.  

I think its more important for the air to be colder rather than more volume..I guess we'll see when you try out Mikes idea.  I can see this working, as I said I think Toyota had the right idea, and I think with a little tinkering you could further improve it. The old MR2 had to have one of those scoops to aid cold air flow as I think they probably faced a similar predicament. I'm really surprised the TTE turbo doesn't appear to have any additional inlets.

[Anonymous]

Mike has already tried his idea and posted the results on SC - her ya go:

"After reading quite a discussion about air-inlet pipe options on the British MR2 Roadster forum I decided to try bypassing the venturi pipe, leading into the air filter box. I simply replaced that curved black plastic venturi-shaped tube, leading into the air-filter box with 3" flexible clothes dryer hose. `glad I did.

A half-dozen "poor-man's dyno" runs verified that the time has dropped 0.18 seconds from an average of 8.17 seconds to 7.99 seconds!

Furthermore, these new times were recorded with a Full tank; my records show that the previous times were recorded with 1/4 tank of gas. So, call it a 0.2 second improvement, which, I think, IS significant.

What's more, I could easily feel the difference ... it was definitely pulling quite noticeably stronger from 5,500 - 6,500 rpm. The rest of the range felt more torquey, too. And, although 6,500 - 7,000 rpm is not included in the "poor-man's dyno" run, mine now pulled just as strongly to the redline fuel cutout; it used to die noticeably after 6,500 rpm. So, that venturi pipe definitely costs some horsepower (in my mind) ... especially since these improvements were recorded with a showroom stock car (with no low-restriction muffler, header or cpu-chip diddling).

Last, but not least (?), the sound was quieter and deeper.

The above times were recorded at 35 degrees with a showroom stock 2003 (with ~35 pounds removed from no spare, jack kit or plastic tire cover). These times were recorded with the previous CAI (also 3" flexible clothes-dryer hose) from the left-side body vent to the stock inlet, below and in front of the fuse box.

These were, no doubt, the cheapest horsepower I'll get out of the beast!"

[Anonymous]

BTW - Nice one Markiii for sorting the GB out in the 1st place - respect!  s:P s:P

[Anonymous]

running without the duct, i can definatley feel the difference on a hot day. not sure if the temperature affects it as much with the pipe in though. thank god that heat shield is there!!   s:) s:)

[Anonymous]

You'll probably notice it the most when running in traffic - i.e little air movement through the engine bay = sluggish until you get up some decent speed and flush the warm air out of the engine bay. This was what it was like on my chipped 205 GTI 1.6 with a K&N 57i. It was nice at high revs / speed though due to cold air feed.

[MRMike]

You'll probably notice it the most when running in traffic - i.e little air movement through the engine bay = sluggish until you get up some decent speed and flush the warm air out of the engine bay. This was what it was like on my chipped 205 GTI 1.6 with a K&N 57i. It was nice at high revs / speed though due to cold air feed.

Yep I concur entirely. Its exacerbated in the 2 as I think it was Mark pointed out it s a low pressure area in the engine bay. Nice previous car as well!

[heathstimpson]

Good work Rusty; would be interesting to see a Dyno dun with both the exhaust and inlet pipe.

[Anonymous]

I really do not see the value in the PMD methodology.  There is IMO no way you can accurately account for 0.19 seconds difference when recording on a stopwatch, between 3 to 4.5k revs..I'm not sure why it's caught on really?

If you read through the PmD thread on SpyderChat, you'll see that a number of things were done to allow for consistent, accurate measurements.

1.  Even tach numbers were selected - to make it easy for one's eye to spot, when clicking the stopwatch.

2. An average of 6 runs (say 3 in each direction) is recommended - before making conclusions.  Furthermore, the PmD standard was specified to be done in 3rd gear - so that the pull would take long enough to make 100ths of a second irrelevant.  Tenths of a second are significant, especially when averaged over 6 runs.

You'll also see that the PmD definition is 3,000 - 6,500 rpm - not "3 to 4.5k revs".

The thread explains the methodology & its implications in considerable detail; I'm at a loss as to why so many have difficulty understanding it.

 m http://www.spyderchat.com/phpBB/viewtop ... s+dyno+pmd m

[Anonymous]

would be interesting to see a Dyno dun with both the exhaust and inlet pipe.

Note that a dyno run is not going to show the results of aerodynamic modifications, such as the results from taking cold air from a high-pressure area of the car ... because dyno runs are, obviously, done with the car sitting still.

I made measurements, running an inlet straight down from the air filter box to the rear anti-sway bar; there was no measureable benefit.  When switching to the driver's side inlet (in conjunction with by-passing the 3" to 2" venturi pipe between the fender inlet and air filter box), there was a 0.2 second difference on a showroom stock 2003 MR2.  It would be interesting to see/measure the difference on a car that had enhanced breathing (e.g. header and lower-restriction muffler).

[MRMike]

I really do not see the value in the PMD methodology.  There is IMO no way you can accurately account for 0.19 seconds difference when recording on a stopwatch, between 3 to 4.5k revs..I'm not sure why it's caught on really?

If you read through the PmD thread on SpyderChat, you'll see that a number of things were done to allow for consistent, accurate measurements.

1.  Even tach numbers were selected - to make it easy for one's eye to spot, when clicking the stopwatch.

2. An average of 6 runs (say 3 in each direction) is recommended - before making conclusions.  Furthermore, the PmD standard was specified to be done in 3rd gear - so that the pull would take long enough to make 100ths of a second irrelevant.  Tenths of a second are significant, especially when averaged over 6 runs.

You'll also see that the PmD definition is 3,000 - 6,500 rpm - not "3 to 4.5k revs".

The thread explains the methodology & its implications in considerable detail; I'm at a loss as to why so many have difficulty understanding it.

 m http://www.spyderchat.com/phpBB/viewtop ... s+dyno+pmd m

Fair points Mike, and I've read the Spyderchat thread since its inception.  I like the idea of timing intermediates as a more meaningful measure than 1/4 miles etc.  I think this methodology would work excellently with a GTECH.  However manually recording with a stopwatch the start, and end of the intervals you have a massive variance in the results already. Effectively you could be over/under the start/start point in one increment.  Fair enough averaging the results will reduce this variation, but in my experience if you applied standard deviation to these numbers they are significantly distanced from the mean, which calls into question the overall output

Its the best method anyone come up with as a guessstimate though.  I just don't follow how you could make directly comparable runs with this methodology.  Maybe i'm missing something!

[Tem]

I think this methodology would work excellently with a GTECH.

True, but that's why this is called poor man's dyno  s;) s;)  Everyone can measure time, but everyone doesn't have G-Tech or alikes...

This is far from excellent, but then again, a lot better than "it feels faster".

[MRMike]

I think this methodology would work excellently with a GTECH.

True, but that's why this is called poor man's dyno  s;) s;)  Everyone can measure time, but everyone doesn't have G-Tech or alikes...

This is far from excellent, but then again, a lot better than "it feels faster".

Very true Tem, very true. Perhaps I should stop being such a fussy bugger.  

But given that you can't put that much value into the the figures, how much more does it actually tell you other than "yeah its faster" which the Butt dyno told you anyways. I doubt a GTECH will even have the accuracy to 'see through' the different variables to indicate if the inlet pipe has made a difference or not, because as you know that too is not infallible. But it illiminates the main sample error associated with starting/stopping at the correct times

I'l shut up now as I'm starting to aggrevate myself.  s:lol: :lol: s:lol:

[Anonymous]

I just don't follow how you could make directly comparable runs with this methodology.  Maybe i'm missing something!

What specific problem?

1.  You stare at the tach with the stopwatch in hand and click when the tach reaches 3,000 rpm.
2.  You stare at the tach and click "off" when it reaches 6,500 rpm.

What is the problem?
   This is not rocket science!

As I've said many times, hundredths of seconds are not significant ... but tenths of seconds are, especially when averaged over a half-dozen runs.  But, it is a cheaper way of measuring before/after results.

If someone is claiming a 0.1 second difference from some modification, yes ... a 0.1 second difference is subject to doubt.  However, the significant differences, that we're looking for with modifications are usually greater than 0.1 second differences.

For instance, a showroom stock 2003 will run ~8.2 seconds.  Those with popular header, muffler & cpu mods are reporting 6.6 - 7.0 seconds; this is significantly faster.  Cars with mild turbo charging are reporting times down in the 5 seconds range; these cars are significantly faster than a car in the 8 seconds range or even 7 seconds range.

I wish those who put so much faith in dyno results were as careful testing (e.g. especially before, as well as after ... to say nothing of noting atmospheric conditions etc.) with a dyno as they are to disparage a cheaper alternative.  Dyno results are really not comparable from day to day ... or from dyno to dyno, either.  Yet, people accept those numbers as absolute truth.  Dyno results are always stated to be insignificant +/- 3%.  

Thus, the potential dyno errors within typical MR2 operating ranges are:
3% of 120 whp = 3.6 whp
3% of 140 whp = 4.2 whp

Short of installing a turbo, few single MR2 modifications result in much more than 4 whp gains ... which are within or close to the error range of dyno measurements within the MR2 operating range!  So, why is there such faith in dyno measurements ... like they came down from the mountain, etched in stone?

3% of 8.2 seconds for a stock PmD would result in a 0.25 second variation.  I don't think it's reasonable to feel that stopwatch results will always be "off" 0.25 seconds, especially when done carefully and averaged over 6 runs!

Sure, there are variations in PmD measurements; but, no worse than the variations in dyno measurements!

[Anonymous]

I really do not see the value in the PMD methodology.  There is IMO no way you can accurately account for 0.19 seconds difference when recording on a stopwatch, between 3 to 4.5k revs..I'm not sure why it's caught on really?

If you read through the PmD thread on SpyderChat, you'll see that a number of things were done to allow for consistent, accurate measurements.

1.  Even tach numbers were selected - to make it easy for one's eye to spot, when clicking the stopwatch.

2. An average of 6 runs (say 3 in each direction) is recommended - before making conclusions.  Furthermore, the PmD standard was specified to be done in 3rd gear - so that the pull would take long enough to make 100ths of a second irrelevant.  Tenths of a second are significant, especially when averaged over 6 runs.

You'll also see that the PmD definition is 3,000 - 6,500 rpm - not "3 to 4.5k revs".

The thread explains the methodology & its implications in considerable detail; I'm at a loss as to why so many have difficulty understanding it.

Have I made an error posting the results? I thought I posted 3-6.5K and also 3-4.5K & 4.5-6.5K. The reason for adding the 2 other splits was in an effort to isolate low end and top end results in order to see if the OEM pipe had better low down performance at the expense of top end performance. This couldn't have been verified from the std PMD testing. Fair enough, it may be questionable as to whether its possible to accurately time the shorter increments but surely it's better to have some indication rather than no indication. Furthermore, out of 6 runs the average of all of the H&S results are better than the average of the OEM results. As stated previously, all runs were in 3rd gear with the roof down and the OEM wind deflector up.

If people read the original info that I posted properley there would be no need for pointless discussions as to its validity. If you want to disregard the high and low increments then fine. If you think that PMD is a waste of time then fine - each to his own. I just thought I'd share my findings in case they were of interest to anybody else. IMO the consistency of the results is enough to convince me of their use as a simple performance measuring tool in the absence of a GTECH or other similar device.

[MRMike]

Sorry Rusty, this is my fault, and It's completely detracted from the information you kindly took the time to put up.

My bad!   s:oops: :oops: s:oops:

[Anonymous]

Hey look, no worries, at least people have an opinion. I'd rather that people stood up and said what they thought even if it's not my opinion it's still valid. After all how would we ever move things on if we all sat around agreeing with each other! I just want to improve the cold air feed now by extending the intake piping right into the side intake - this may make even less difference on the PMD but hey I'll take some measurements anyway when it's done. At least it may give some sort of idea as to whether it's a worthwhile improvement or not.

[Anonymous]

The reason for adding the 2 other splits was in an effort to isolate low end and top end results in order to see if the OEM pipe had better low down performance at the expense of top end performance. This couldn't have been verified from the std PMD testing.

Good use & choice of tools, Rusty.

There's nothing sacred or inherently "right" about the existing PmD procedures.  The only benefit of doing it any one way is comparability ... with others' test results.  Sometimes that's not what's required.

Fair enough, it may be questionable as to whether its possible to accurately time the shorter increments but surely it's better to have some indication rather than no indication.

Good job.  Your variations make it clear that the greatest benefit is at higher rpm, compared to mid-range rpm (as one might expect, intuitively, btw).

IMO the consistency of the results is enough to convince me of their use as a simple performance measuring tool in the absence of a GTECH or other similar device.

Agreed.  Precision, in the repeatability sense, is all that's required to measure improvement from a before measurement.  All that is really important is the difference in the measurements - before vs after.