bleucamaro

Wow, what a neat application. I'm looking at doing the same, or similar to the S2. I was originally planning on running an LED, but I don't think it would be as effective. Since the S2 DME does not have a KLR (knock control is done in the DME) can someone suggest a pin to take a look at wiring it to. My understanding is that going from raw data from the knock sensors is inadequate, and I need to look for a processed signal.

Can anyone suggest a pin to look at on the S2 DME?

Rogue_Ant

The KLR is what processes the signal from the knock sensor.... your S2 won't have either. You could do something like this though:

http://autospeed.drive.com.au/cms/article.html?&A=0348

I just built mine today


Rogue

Oddjob

I believe both 16v motors (S and S2s) actually have 2 knock sensors - so yes, he will be able to monitor the knock sensor(s) signal to the DME. But using a simple counter may not work as well as monitoring the KLR knock detection signal being sent to the DME on a 951, versus the raw signal from the sensor.

Pauerman

Rogue,

I'm surprised that your log show no knock activity at all. Have you configured your setup to only register "noise" above a certain threshold?

Since our cars use Bosch FLAT RESPONSE knock sensors, these sensors pick up noise in a broad frequency range - from 5 KHz to 20 KHz. The output signal it generates is low - around 30mV/g. With that said, if you haven't created any threshold parameters within your software, you should be registering a knock signal from at least the engine noise if you are taking a reading directly off the sensor.

Products like the J&S Safeguard and Link's KnockBlock use this ajustable knock threshold approach for determining knock with flat response sensors. They require that the car is driven throughout the RPM range without boost so that the engine's noise signature can be determined. Once this is established, the knock event output signal can be configured so that noise levels above the baseline value are recognized as knock.

Rogue_Ant

I currently have it hooked up to pin 15 on the KLR... it simply looks for voltage spikes from the KLR and counts them. I do need to verify that it is working properly, because the reading I'm getting (nothing) bothers me as well.
The AEM EMS uses voltage threshold for knock counting, and I could set mine up that way. But I don't know a 'safe' voltage output from the knock sensor opposed to a 'knocking' voltage.
I do have a J&S safeguard that I plan on using, but since I've sent it in to J&S for servicing, I'm having a hard time getting John (J&S) to respond to me


Rogue

Trucho-951

Sounds to me like you have a faulty cycling valve, it could failing "safe". That is, as a solenoid designed to close off the boost signal to the wastegate when activated, it is probably too old and can only close off the boost signal for those 3 seconds you experience, then it fails in open mode and consequently your boost level falls off. Do you have a spare cycling valve you can swap in?

samluke

I wish it was that simple. Cycling valve was one of the first things I tried, along with computers, knock sensors, temp sensors, MAF, TPS, and then wastegate.

Luckily I have my cycling valve relocated so that its easy to change. I have tried multiple cycling valves new and used trying to fix the problem. I tried another one, just the other day. Right now we have bad snow so its hard to test. I have a race coming up in April. Hopefully I can find something and fix it by then, but everytime I think I am on to something it leads to a dead end. I can't believe I am the first to have this problem.

Of course many cars now have MBC or EBC's which would likely mask the problem I am having.

Full load signal between KLR and DME is about the only thing I haven't had chance to check out yet.

Transaxle

The KLR is programmed to open the CV completely for 2-3 seconds if you move the throttle quickly from part throttle to full load. That is OK with your system.

Since you get desired boost with an open CV, the CV and the WG are looking OK. If it then reduces boost to a lower level constantly would indicate a problem with the pressure sensor in the KLR. If you already tested with another KLR, the sensor should be OK.

Another point to check ist the throttle sensor. Full load starts at ~60% open throttle, but from 60-100% are several boost curves stored in the KLR (see picture). Maybe your throttle sensor is somehow limited.

Another point is that there is no communication about boost levels or knocking between the KLR and DME. The KLR is a standalone system providing only the full load signal and a corrected ignition signal to the DME.

This advice has no basis more reliable than my own meandering experience...

samluke

Transaxle,

Thanks for the insight, it gives me some food for thought. I just tried a $20 manual boost controller to replace the cycling valve and the boost is rock steady, so its clearly a problem in the control system somewhere, I just have to find it.

I just changed the TPS again, and it made no difference, both were in spec.

It looks like I am back to trying computers again. I have tried my spares, maybe they are all bad. I need to hunt down some others to try. I have another KLR that I can try, but only 1 spare DME

Trucho-951

Samluke, I finally got around to installing the knock counter. Since I still had a little bit of 100 octane gas in the tank after I got back from Thunderhill, I decided to install the knock counter and do some data logging.

I know you are looking for knock data on a completely stock car, so I hope this info is still relevant. My car still has the stock k26/6, but also has some minor mods (MAF, 38mm Tial WG, black **** MBC, Testpipe).

Anyway, here are my results on a nice Napa Valley 85 deg F day:

100 Octane, 17 psig boost, results = 0 knock counts per WOT run in 4th gear.

I didn’t have enough race gas to continue so, I stopped at a local gas station for 10 gal. of 91 Octane, dropped the boost a little and continued data logging,

91 Octane, 16 psig boost = 10 knock counts per WOT run,

15 psig boost = 06 knock counts per WOT run,

13 psig boost = 04 knock counts per WOT run,

11 psig boost = 0 to 1 knock counts per WOT run.

In none of these runs was there ever any audible knock. If you’d like I can post the boost and AFR curves for each of the runs.

Thanks for your link to the $12.95 knock counter, I am happy with how easy it was to install and the valuable information it provides. I wish it had a reset button, would make it a lot easier to data log. Hope this info helps some.

951Boost

Guys, regarding the reset questions: I haven't received my unit so I'm just going by the Curtis manual which can be downloaded here: >>>LINK<<< go down to 700 Series (cased) and click on "manual".

If your unit has a fourth terminal (pin 4) it can be used for resets. See pages 14 (installation) and 17 (operation) on the manual.

Probably the easiest way to do it is to "T" from the 12+ wire to pin 1, install a normally open momentary pushbutton switch and run the wire to pin 4.

BTW- Model breakdown; 703= 3 wire counter, D= case style, etc. etc.

ross255

I want to get one of these counters, but they dont ship international.
Would anyone be willing to help me out and order one for me and send it on to the UK?

Oddjob

These particular models only have 3 terminal tabs on the back side. The reset tab is not included/installed, so its unclear if the function is available on this model, or not. There is a small non-protruding pin at the location of where the 4th connector tab would be. I have not tried to solder a connector tab to it yet, but I intend to try to set it up with a reset switch, if it is capable of that function. If you get yours and try it before I get around to it, let me know what you find out.

Thanks.

951Boost

Oddjob, I put in a call to one of the application engineers at Curtis to ask him about this. I was told they would be very shorthanded over the next couple of days but would get back to me asap. I'll post any info I get from them, but if you test your unit before they get back to me please let us know.

nize

i just ordered one.