I have been from time to time been having a problem with my JDM h22a where every once in a while right at about 3,000 rpms, all hp is lost and it feels like my car has 10hp trying to make it up a hill. I will floor it and the car just bogs and bogs, and ever so slowly climbs rpms, and sometimes lasts even into higher RPMs. Then all of a sudden (its like 5 seconds later) the power returns and I am on my way. It only happens once in a while but its really annoying. I have search all over this site and honda-tech with many people experiencing the problem, but no one really has an answer. http://www.honda-tech.com/zerothread?id=1028177&page=1 Here is a 7 page thread on H-t with no real solutions. Now here is what I have theorized... A lot of people are pointing towards knock sensor as the issue, and alot of people are screaming get a new one, but I think that the stock one is doing its job. The knock sensor is simiply a chunk of peizo electric material which basicly converts shock/vibrations into current. When there is detonation, it cause a soundwave that triggers a large current spike in the knock sensor. My question to many of you, how many of you kept your balancing shafts? I did not. This could be a reason why I am sometimes having this problem. The frequency is smaller so the peaks in the sine wave of vibrations are spaced further apart. As RPM's rise, they become close and closer together to the point where the the knock sensor sees it as one continuous peak. The purpose of the balancing shafts is to cause a vibration that will cancel out the main vibrations cause from the crank and other moving parts (destructive interference) When functioning, these cause the amplitude of the vibrational waves to be smaller. Now because they are not functioning, these waves are back to their original size. Because it is at a lower RPM, as I said, these peaks are spaced further apart and because the amplitudes are bigger then what it was programmed from stock, on occasion, the amplitude of the the wave is large enough to trigger a spike in current high enough for the ECU to think that a knock has occurred. So the knock sensor is NOT faulty, it is just oversensitive. Now how do you solve this. The knock sensor sends low voltage to the ECU at lower RPMs and linearly increases the amount of current as RPMs climb. When its a steady increase, the ECU acts normal. When there is a very sharp jump in voltage, the ECU interprets this as a knock. Now, this is for all you EE's out there (I'm a materials engineering student with limited circuits background) what size resistor would bring the voltage levels down a bit. Another solution could possibly putting a rubber grommet as a washer for the knock sensor to dampen some vibrations, thus reducing the amplitudes of the vibrational peaks.