nj1266

iTrader: (6)

I finally finished my testing on the ebay filter. I had some delays due to my favorite test ramp being closed by Caltrans I wanted to be consistent and do all my testing on that flat ramp.

My car has a TBE w/tp and stock boost. All charts below are a series of three per sesssion. I did 4 logging session of 3 back-to-back logs per session. I tried to keep the weather condition as follows: 20-25*C, 63-67% humidity, and 101 kpa baro pressure.

Here is what I found:

1. As mentioned slapping on the filter w/o tuning makes your car run rich. This happened despite the fact that the car did not hit higher load cells. I checked and double checked for leaks, but I consistently logged a rich mixture with the ebay filter untuned.

Here is a before and after chart. The before are always on top and the after are at the bottom:



So you will have to tune your AFR back to the way you had them before you installed the filter. It took me about two logging sessions to arrive at an AFR chart that somewhat resembles the one before I installed the ebay filter. It is not the same, but it is as close as I could get it to while keeping knock in the 1-2 counts here and there.

Here is the final AFR chart that I used with the ebay filter:



2. The filter also messes up your timing up top. My car lost 1* of timing from 6500-7500 rpm. At first I did not notice, but then I checked and found a consistent loss of timing by 1* up top.

Here is the before and after chart:



I had to go back to the timing tables and adjust them back up by 1* in the affected rows and columns. I was truely surprized by this and I have no explanation as to why the timing got pulled up top despite the lack of change in the load cells that the car hit.

3. I registered a boost increase in the 5500+ rpm by ~1 psi or less. The peak boost was not affected much. Infact it was a bit less, but that could be due to weather conditions even though I tried to keep the variations to a minimum. In the 5500+ rpm the ~1 psi gain was consistent and repeatable.

Here is a before and after chart:



4. The IATs are my biggest concern with this filter. With the stock air box the IATs dropped by an average of 9* F during the WOT run. With the ebay filter the drop was only an average of ~2* F. IMO, that is no good and the filter requires some sort of a shield. BTW, I did keep the snorkel for the stock box while testing the ebay filter

Here is a before and after chart:

nj1266

iTrader: (6)

But does the Ebay filter produce any power gain on a lightly modded car like mine? It did produce some gain, but not a lot. I did have one run when I recorded a 16 hp gain over the last run I did w/o a filter, but that was a HERO run and was not replicated in other runs. So I will ignore it.

Below is the best run I had with Logworks before and the best run I had with Logwroks after. I am going to compare results from the same logger to keep things consistent. The gain is 7 hp and most of it is concentrated above 6500 rpm. Are 7 hp worth the increase in the noise and possibly attracting the police. Only the end user can tell.

l2r99gst

iTrader: (2)

Great test, nj. I just added my review to the Apexi drop-in thread where we were talking about this filter: https://www.evolutionm.net/forums/sh...=277802&page=5

Our results and conclusions came out to be about the same.


Eric

nj1266

iTrader: (6)

Thanks Eric. It is good to know that you Bryan and myself arrived at similar results independently of each other. I like that a lot. What surprized me the most was the impact of the filter on the timing up top in my car.

Why would my car consistently and repeatedly lose 1* of timing from 6500-7500 with this filter?

Next up is to test the Apexi filter. A friend of mine wants to sell one used for about $75. I might buy it and test it.

I am really conflicted about keeping the ebay filter or going back to stock

l2r99gst

iTrader: (2)

I lost timing, too, more so in the midrange to upper RPMs. I'm pretty sure that I was at a slightly higher load, though, which would explain that. Unfortuantely, I didn't log 2-byte load.

I, too, am debating on whether to use the filter or go back to stock and test with a modified airbox and drop-in...that was my whole point of stating that Apexi drop-in thread. I know that the open cone filter will prove to be a much better option for higher boost and higher HP, but I don't like the increase in IAT. I'll have to go back and see what pressure drop jcsbanks saw when modifying his airbox.

I've been pretty busy and lazy lately, so I may just leave it as-is for a while. But, in the meantime, I will try to at least build some sort of heat shield for the ebay filter to see if it helps out with the IAT issues.

Definitely post up results if you test the Apexi. It has always been my favorite filter, but I decided to give the Ebay filter a shot from Bryan's recommendations and results. For $7, it can't be beat.


Eric

nj1266

iTrader: (6)

Here is the Hero run that I talked about in my prior post. It gave me 17 hp over the best run w/o the filter using logworks. If I can get consistent 15-17 hp from this filter, then I will definitely keep it on. Alas, this run was an orphan

C6C6CH3vo

iTrader: (9)

Did you happen to log the airflow as 2 byte by chance?

Wondering if timing was adjusted from the raise in temp and baro psi's. I guess with IX's you can be in the same load cells without knock but run less timing with same maps but with interprolation (w/o knock). The VIII FSM claims timing is advanced with elevation, I havent seen this yet though.

BTY - Thats heavey timing at peak if you ask me

RazorLab

iTrader: (8)

Nice testing NJ, again, pretty much the same results I have seen on about the same type of mods, but much higher boost.

I really honestly don't think it's worth it on a TBE 'stage 1' type evo, and I am honest about that with people that ask. With cams and more airflow mods it will help a bit more. It sure did on my personal EVO with much more mods than a typical 'stage 1' evo. By 'stage 1' I mean TBE / drop-in filter / fuel pump.

One thing it was nice to see from your DataLogLab powercurves is that although the peak numbers aren't much higher it does look like it holds power much better up top. Looks like almost 40whp at 7500 on this:



I did not see that type of gain on a real dyno but it could very well be like that on the street with more airflow.

nj1266

iTrader: (6)

No, I did not log the 2 byte airflow. I suspect that the solution to the richer mixture might be in the higher airflow due to the ebay filter. But I am speculating w/o any data to back this up.

The 1* timing pull is the biggest mystery to me. It happened in ALL THREE logging sessions with the filter prior to adjusting it on the last logging session. The timing pull was consistent and replicable.

Timing would be high @ peak if this was an Evo 8 OR if I was running higher than stock boost. Evo 8s usually like 2-3* timing @ peak on 91 pisstane. The 9 can run higher timing on 91 pisstane. If I was running 22 psi of boost, then I dial the timing back to 3-4* @ peak.

nj1266

iTrader: (6)

When results are replicate by three different testers independently, then they become more trustworthy. IMO, quality testing must be replicable and verifiable.

I very much agree with you. I am going to drive with this filter for a while and collect more data and then more than likely go back to the stock airbox with the K&N drop in that I have.

When I posted yesterday in the Apexi thread, I did not see the HP gain that I saw today. Yesterday, I had a Eureka moment right after I posted in the Apexi thread. I checked the before and after timing logs and in every log the timing was being pulled by 1* between 6500-7500 with the ebay filter installed. This is why the DLL power runs from yesterday had negligible power gains. When I set back the timing to the way it was before I installed the filter, the car gained the 7 hp you noticed in the DLL power run.

jrsimon27

iTrader: (1)

nice review but 7wph its not worth it imagine in a very hot day your intake air temps could go higher and produce more knock.
but great review totally different to what others claim like 15whp and up!

RazorLab

iTrader: (8)

It seems like everyone is just looking at peak power gains in this thread. If it did indeed gain 40whp at 7500 like depicted in your DLL powercurves that would be amazing. I doubt such a huge gain but even half that would be pretty damn good.

Would be interesting to get more data on this. I will try to do some street logs as well, instead of just dyno logs with less airflow.

The one funny thing is that your peak gain was exactly the same peak gain I have seen on the dyno.

nj1266

iTrader: (6)

The reason why I ignored the 7500 rpm hp gain is because I could not replicate it. If it is not replicable then I ignore it. What I could replicate is the 7-8 hp gain from the filter.

Check out this log of the same filter power run that I posted overlayed on a different pre-filter power run. This pre-filter power run is from the same batch that I posted before. You will note that there is almost no gain @ 7500 rpm. The problem with the pre-filter run, though, is the high torque numbers. These numbers are an anamoly. Usually the TQ figures are 272-275 ft-lb. Note that the power gain is 8 hp, almost the same as in the previous overlay.

RazorLab

iTrader: (8)

nj,

Ok, that looks more believable. Thanks!

nj1266

iTrader: (6)

What the vendors usually do is use their hero run and claim that this is the power that the filter produced. I could have done the same. I could have posted the hero run that gave me 16-17 hp as the typical gain from the filter. But that is incorrect and usually not replicable. Good data has to be replicable espeically when road tuing and dynoing where a lot of variables influence power output.