Hey All sorry for the big post, trying to cram as much information into it as I can.

I am looking for some assistance in getting my SA table in shape.

Here are my engine specs.

• 383 Stroker HP & Torque: 430 HP / 450 FT LBS
  Compression Ratio: 10.0:1
  Aluminum Heads
  Hydraulic Roller Cam
  Cast Steel Crank
  Hypereutectic Pistons
  HSR 36lb/hr EBL Flash I

ROTATING ASSEMBLY:

• New BluePrint Cast steel crankshaft
  3.750" Stroke
  OEM type 5.700" connecting rods w/ 150,000 psi bolts
  Hypereutectic pistons
  External Front/ External Rear
  

CYLINDER HEADS:

• BluePrint Aluminum Cylinder Heads - H8002K
  64cc chamber
  2.02" intake/1.60" exhaust valves
  195cc intake/75cc exhaust runners
  

CAM SPECS:

• Cam Type: Roller
  .528 Intake .536 Exhaust
  221 Intake / 226 Exhaust duration
  @.050 - 110 degree lobe separation
  

IGNITION TIMING:

• 34 Degrees Total at 3500 RPM
  

Below is my current SA Table which was generated from the spreadsheet utility found here.


Current SA Table 12-21-19

That said, I do get knocks in various places. I am in the process of doing some VE learns.

Below is the most recent knock count from my last VE learn.


Here is the most recent corrections from the learn.




Lastly here is the total knock count over a 34 minute drive.




I am not sure how much of this is real knock, and how much is false knock. I do know at times I knock during downshift and other times I do not.

Any advice in getting this dialed in would be great.

Thanks!

 

Need to find out if any of that knock is real. The high counts at the lower loads most likely isn't real. The counts in the 95 KPa region, if real, need to be taken care of.

RBob.

 

That's precisely what I want to do.

Assuming it is real knock, from what I can tell, it is either fueling or timing related. I am leaning towards timing and my SA tables.

However that said. I am unsure how to correct my SA table, if needed, or what timing the engine really needs.

Additionally how audible should the knock be? Obviously the computer is detecting it, but as far as I can remember I do not hear anything different in the tone of my exhaust or from the engine itself when the knock occurs.

 

The spark plugs will show whether there is real knock or not. If real then lowering the SA by half the amount of peak retard will usually do the trick.

Getting false knock when the trans downshifts is normal.

RBob.

 

To Green89IROC30: What kind of knock sensor do you use? Is it a GM #10456549 for 350CID? Is the EBL Flash1 equipped with a ESC control module, integrated into the EPROM? Just curious since I could not find a knock sensor for my 383 Stroker set-up. Could not find a definitive answer on forums as what to use for a 383 stroker with FIRST® intake. Cheers!

 

Good question.

I don't know the P/N but it is a 100k ohm type for the '7165 ecm for a 350 87 IROC-Z

​​​​i also do have the ESC hooked up.

Thanks!

 

I see! Just noticed you have a 91 car powered by 87 power plant. So with the MAF you have an external spark control module with 100 Ohm KS and '7165 ECM. I have a speed density system with 3.9k Ohm KS and ESC module is part of the EPROM in the '7730 ECM. So in theory your KS P/N is 10456288. Interesting that the KS design for 350 works on your 383. Well, thanks for your time and quick reply.

 

I just got done pulling the plugs and had them evaluated at my performance shop and they found no evidence of knocking on the plugs themselves.

They suggested pulling some timing to see if it goes away, and if it does, then it was real knock, if not, false.

If it ends up being false knock they asked about a way to reduce the sensitivity of the knock detection on the EBL. I was gonna go digging into the bin to see how I might do that.

Thanks!


Yeah, my car is a bit...odd. I bought a 91 roller and used my 87 as a donor to get it running and wired up.

 

IMHO it could very well be false knock since the ESC was design for 350 OEM set-up. Not high performance rumbling cam headers motor lol. With my headers on the LB9 305, I was getting false knock all the time. Hell, I have the ESC disabled and sometimes I get knock counts which is beyond understanding. FYI, the SP383 Deluxe with fastburn aluminium heads calls for 32* SA max at 4000rpm (Updated 07-05-2020: 32* is indicated in the engine user manual, however on GM performance site they indicated 36*!) All the best!

 

Within the BIN you can only reduce/change the amount of retard relative to the counts. The knock reporting and counts are all in the hardware.

False knock can be caused by the knock sensor being too tight (11 - 14 ft/lbs is the spec). Exhaust rattles, a bad accessory bearing or cracked/loose accessory bracket or bolts can also cause false knock.

RBob.

 

Thanks, all good things I can check. Appreciate the responses.

 

Hey All,

In my quest to figure out my knock I was investigating the base timing of the engine.

With the ESC disconnected and the engine running and my Bosch timing lights set to zero advance.

I took the above video.

From what I can tell and please anyone correct me if I'm wrong this engine is timed at zero TDC.

Additionally if I advance the timing light a degree then the corresponding mark advances a degree as well.

The engine recommends 10 to 16 advance.

Is the above image enough to determine the base timing of my engine from those who are in the know.

Timing is not something I'm good at.

Thanks.

 

Looks like you are right. Your timing is at 0* which is weird. It should stall at 0*.
If your Tune is set at different then 0* you have a problem. You have to make sure what ever is set in your tune is what is set with your distributor.

 

Thats exactly what I thought.

I just created a tune with a base of zero to test for knock.

 

So my engine was indeed set to 0* initial timing.
The installer corrected it to 11 as the engine calls for 10-16 base.

I will investigate my sa tables more and let you guys know if I need anymore help.

Thanks!

 

Be sure to change the calibration to match: SA - Initial SA

RBob.

 

Thanks for the reminder, I made sure to do so.

Now for trouble shooting knock. I want to make sure my methodology is correct.

Knock can be caused by too much fuel OR too much SA, correct?

So for example, say my cAfr is 14.7 and my actual AFR is 14.7 and I get knock, I should retard the timing for the cell.
Say my cAFR is 12.5(PE mode) and my actual AFR is < 12 and I get a knock, I should lean out the mixture in that cell.

The above is what I am thinking of as a way to treat this knock.

Please let me know if I am approaching this wrong.

Additionally @RBob is there any way to get the EBL to only look at the Main/Extended SA tables and ignore any of the negative adjustment tables like CTS vs Vac or any of the other ones that can modify the SA?
I of course do not want to effect cranking SA and would want those tables to be left alone.

I ask because the engine manufacturer indicated that it would be beneficial to simulate a mechanical distributor curve as much as possible.

Thanks!

 

Usually a bit rich helps prevent knock as it also cools the chamber. Running at stoich is the most knock prone AFR.

I have no words (jaw drop) regarding that statement...

Note that during cranking the plugs fire at the base distributor timing.

RBob.

 

I don't know what to say. I thought I read in the EBL that something was used during cranking and I thought it was SA.
I am not sure why the engine manufacturer makes the recommendations they do.

 

You might be thinking of choke SA. That comes in only after cranking and the engine is running. Data logging should tell you what is going on. The dump log will tell you much more than the data analysis screen.

 

Have you verified with a timing light that the reported SA from the WUD and the actual timing are the same?

 

Can't say I have. I just had the installer reset the timing since they timed it at 0* when they put the engine in. It is now 11* and to be frank, I am not that great with a timing light.

 

So I am really at a loss as to what to do about the knock counts and I am afraid to damage the engine.

All I can do is ask for hand holding at this point.

I have reduced timing, enriched the mixture and still have knocks the same places.
Sometimes the knock counts for a given SA cell are high, other times they are low. I'll correct one area and get a different counts in another.

I am attaching to this post a collage of data as well as a zip of the logs and bin 24 in the hopes that something might be helpful to those willing to lend a hand here.

For the image of the results, there are three WUD knock readouts for tunes 22, 23 and 24. The top middle image is the difference in timing between tune 22 and tune 24.

For those who do look at the logs, log 24 should be paused @ 09:30 to keep the data between the three logs comparative

Shortly after that I WOT the car at highway speed which causes a downshift that spikes the knock count by ~400+ knocks.

I am still trying to sort out the WOT downshift knock issue too. I have been told that downshift knocks are likely false, but I am not ruling anything out at this point.

If anyone can shed some insight I would really appreciate it.

Thanks.

 

The spark plugs will show if there is real knock or not. Very important to check them. As you posted you have already done this. So it comes back to being false knock.

Can desensitize the knock sensor by mounting it on a 45° street elbow.

RBob.

 

Thanks RBob.

With the amount of knocks I have been seeing would you expect the spark plugs to show at this point?

I need to get a few things before I can desensitize the sensor, notably a proper torque wrench and the elbow.

That said before I commit to desensitizing it, I wanted to post a bit of my current data in Image form.

The top right table is the current difference in timing from my last run. Below that is the current table values. I took bits and pieces from EBL_F_3006 and my table to come up with this. It is probably wrong still though.

I seem to have made a knock count improvement in the lower RPM, I also uncapped the max retard at the high load side from 5 back to the default for EBL_F_3006

I did however hit some new RPM load ranges this run.

I will also be pulling the plugs once the engine cools down to double check the plugs again. Just want to be safe here.

Thanks for all the help!

 

Remember that your knock sensor was design for an OEM TPI 350. I would not rely too much on it. My 383 has none and it’s working fine. I use 91 pump gas ethanol free with BOOSTane to bring it around 98. I don’t have access to 93 here and fuel quality is so, so. I have a conservative SA curve at WOT for now with 12.8 AFR on the gauge. This is until I get to the dyno next summer.

Hope you have an AFR gauge? The commanded AFR in PE is not accurate. The gauge will tell you.

Cheers!

 

That is good to know, I think I feel more comfortable desensitizing it now, thanks. I do have an AFR gauge hooked up and am tuning against that for fueling.

I am curious what your SA table looks like. If you wouldn't mind I'd like to see the graph as a reference.

 

This is a graph at WOT for my 383 stroker (blue line). It is a copycat of OEM 305, TPI, 1991, Speed Density (7730), 5 speed, tune (ATNX). The main SA table is changing based on CT and PE. This example is at WOT at 176 Deg.F. I am waiting for the dyno next summer to tune the SA, so the graph most likely will change. The red line is S_AUJP v6 that circulates on Thirdgen.org. I believe it is for a 350CID with automatic transmission, but I have no other details. Basically, SA curve is specific to each engine. I have aluminium heads fast burn, so SA might not need as much advance.

***** Updated 07-05-2020****
The following chart is NOT ACCURATE of real SA measured by timing light. There is a difference of up to +6* between user input SA and real SA. That explains the dip between 2400-4800 rpm. Bottom line is, the SA table doesn't represent the real SA measured by timing light. For example, at 4400 rpm, the chart indicates ~30* but timing light measured ~36*. BE AWARE that knock could occur even if the user SA input appears to be low. This is due to the combination of ICM and SA latency table. PLEASE see following threads for more information:

https://www.thirdgen.org/forums/diy-...ncy-table.html
https://www.thirdgen.org/forums/diy-...-increase.html

********************

 

Thanks for posting all of that. I will take a closer look here to see if I might be able to adapt anything.

 

What kind of WB AFR do you have? Are you doing WB VE learns? Now I'm hooked. I won't sleep well until you figure this out. You better start saving money for rear tires when you do.

 

I bought this WB back in December 2018;
Yes I have been doing VE learns. I have since stopped recently though, to focus on just the knocks. I thought that I should have at least one constant in my quest against the knocks.

Monday my street elbow and torque wrench should arrive and I was planning to desensitize the knock sensor.

I have tons of data and logs I have been trying to analyze in Excel and google sheets as well I can share. If you are interested hit me up in PMs and I will send you a share link to google sheets.

 

Hey all.

I pulled the plugs got a new set.

I honestly can't tell if it's peppered.

I got a bunch of different magnifiers up to 30x and about all I can determine is that on some of the plugs it has cratering I guess you could call it.

There also seems to be some scoring of the porcelain.

I can't definitively see any specs because of the overall soot buildup on the porcelain.

I can try and get some pictures up if anyone would think they'd be helpful but I don't know how clear they'll be.

 

So after closer inspection, I do think there is peppering and that would mean that the knock is real as well.

Guess I have to spend more time looking at the SA and AFR.


plug 5

plug 7


plug 4

 

Just an update for those following along.

On my last run, I only had 18 knocks in areas I used to have 70+ so I think I am getting somewhere. Peak retard was 3* @1800/70 map and 2* @2000/50 map. Every where else was 1* retard




Weird that the manufacturer recommends 32-34 SA all in @3500 rpm, yet I knock at any SA close to that.

Totally gonna throw out everything they told me about this engine...

My average sa is in the 24-ish range now.

 

What kind of fuel do you use?
Mine use to knock at only 22* around 4000-4500 RPM. (updated 07-05-2020, real SA was actually 28* not 22*, see post #28) I could hear it!!! It was on 91 pump gas. Don’t know what kind of crappy gas they have around here. Even my Miata was pinging. By adding BOOSTane, it all went away. You are definitely leaving power on the table.

 

Yeah, that thought has been in the back of my mind the whole time.

I use 93 pump gas. As to it's quality, I couldn't hazard a guess.

 

You need a dyno session!!! :-)

 

I do indeed need to get back on the Dyno. It's 150 an hour for Dyno tuning at my local place.

 

Just an update for those following along and lending a hand.

I decided to check my SA at the crank and compare to the commanded SA as seen in WUD.

To do this, i set 4 ranges to a flat amount.
2000-2200 was set to 25* across all load ranges
2800-3200 was set to 28* across all load ranges
Using my timing light, which I finally figured out how to use semi-decently, I was able to confirm that the SA as seen in WUD was within between 0* TDC and 1* ATDC, which is pretty damn close.

After consulting a resource I decided to revert my changes to my main SA table and change the over all curve of the table. Additionally I zeroed out the two SA Bias scalers and associated tables. Whether this was the wrong move or not, I would like to know.
My goal was to be governed as much as possible by the Main SA table.

I set my all in SA to 32* @3500 and using the spreadsheet I mentioned earlier, calculated a new SA table.
I set my all in MAP value to be 80+ kPa only and my increment to 0.0650* per 1 kPa increase.



It generated the following curve: and table.


I got about 73 knocks total with this table scattered in a few places.


Looking at the data some of it seemed to be fueling related rather then SA related. However the higher end seemed SA related. I took a few degrees out and ended up with curve

After re-doing the VE learn with a higher start temp(172 vs 150) and altering Open Loop AFR Multiplier vs CTS a hair in the 32-44 C range. I made another run, I only got 13 knocks and none of them were above 2000 rpm or 60 kPa Map and of those my peak retard was 2* in the 60 kPa cell, everything else was 1*.


I'd say I am making progress. I will pull the plugs in the morning and inspect them.
That said, I have had low knock counts one run and higher knocks the next with nothing changing so only time will tell.

 

With that much SA now you must feel it in the seat of your pants!

 

aye, dyno is planned for this month.

Let me know how yours goes. I'll do the same.

 

I have yet another update for my weird random knock and for you @SbFormula

So my knock from 1400-2000 @MAP 20-60 truly does seem to be false.

I found a post here from @CORV3TT3 who basically had the same problem I was experiencing.

As it turns out, the donor for my build was originally an 87 IROC-Z 305 TPI car, that later had a 350 installed by the previous owner.
Due to this it had the 305 ESC, which I paired to a 350 KS.

Thanks to the linked post I found out the the 305 ESC is different then the 350 ESC and you need to pair the correct bore size KS to the right ESC for that KS.

In my case I needed LXE6 from standard electric since I already had KS7.

After replacing my 305 ESC(standard electric number LXE9) with the 350 ESC, I did not find any knock in the same range I have been fighting. In fact there were only 2 knock counts for my last run.

Now I need to make a couple more runs to validate the fix because one data point does not proof make. But it is waaay more promising then any other fix I have tried so far.

Given the Knock made no sense based on engine conditions, I started looking elsewhere.

Here is hoping!

I wish I had a way to bring attention to all the 350 swappers/383 builders using a TPI based system to make sure the KS and ESC are for a 350 before going forward.

 

Wow! Back to basic eh? Simple as that.

 

Yeah no kidding right? 👍

 

27 deg of timing on that motor at 4800 and above? That thing must be way down on power. Why is it tapering off so hard after peak?

i would expect that motor to make peak torque in the 3400-3800 rpm range. This is generally a place where lower timing is needed than your max hp rpm timing. 32-34 is common. Then max rpms usually increase to 34-38 depending for most sbc.

You really need a dyno to verify those values. If it knocks on the dyno you can sometimes see it more pronounced in the power curve or rpm log. I dont like factory knock sensors on non stock motors since i cant be sure its accurate all the time and they seem to be too sensitive and susceptible to false counts. Tuning a deg at a time on dyno untl power gain falls off to only few hp a deg then backing off 1-2 is the ideal way to hit the wot stuff imo

cruise can be more challenging but since pressure is low there its hard to really have knock related damage imo and it should tolerate 34-38 deg as well in most low kpa areas or low lv8 areas for maf cars. Using throttle position vs map vs mph at cruise can help find optimal timing. You want the lease throttle position to maintain a given cruise speed at a lower map value. Its time consuming tho

let the engine tell you what it wants but not necessarily the knock sensor. But atleast you found the correct knock sensor and it should start making more sense now

 

Thanks for the input.

Now that I've determined the cause of the false knock I'm going to be not only getting it on the dyno hopefully this month but also tweaking the table a little bit more.

I'm running a different table than I posted in the beginning of thread now.

I'm going to tweak it a little bit further before the dyno

 

I had some time today and yesterday to really digest what you had to say @Orr89RocZ , and yeah super happy I figured out the ESC was the problem with my knock counts.
With that in hand I created a test table to try once the weather clears(currently raining).
Below is said table.
If it goes well as a general driving table, I will use it as a base for my dyno tuning, which I hope to do this month, weather permitting, i.e no snow/salt.

 

Looks good to me. May find it can tolerate abit more in the cruise ranges like 1500-2500 rpm, 40-70 map but that is a good start imo

 

It's me again... Sorry all.

So I was testing my new table now that I have the correct ESC, when I found an area with sporadic knock counts. I also reverted to a previous lower sa table over all and this area persists.

Specifically ~1800 RPM, ~64 MPH, ~@50 Map, ~13% TPS,

By sporadic I mean, I can be in that range and experience no knock while other times I will get knock.

I am going to attach a xls of a snippet of the log while I was in this range. Hopefully someone can shed some light.
I did inspect the plugs and I saw no peppering specifically, but I did see some reflective specs, which could be aluminum I suppose. It wasn't much, just a spot or two on plug 1 which I just installed a new set yesterday.

I'm not sure what specifically is going on since the knock counts are intermittent but persistent in this range.

 

i wouldnt think there would be detonation signs on the plugs unless you can audibly hear it pinging, and by then the knock sensor would be going crazy.

That could be false or an area where the cam doesnt run well there. Maybe mild reversion effects, harmonics of some sort triggering a count. Not sure but i have similar things happen, at a certain rpm point. Above and below it was fine. May try different air fuel ratios there at that specific point and keep adjusting timing in the cells around that specific point