scheides

iTrader: (10)

Last year, I dyno'd my car on a dyno dynamics dyno with tephra v5 setup to do mapswitchng. The boost curve on my pump (base) map was certainly not flat, but it made power and it wwas my kick-around daily tune so it was no biggie:



Now, I was on the dyno yesterday with my car (96530006 - 05 evo8) as well as a friend's car (evo9), both with tephra v5.10, both running mrfred's direct boost control mod (same as before) with a GM BCS. I don't have my graphs with me because they were garbage, but I do have the IX's dyno. Long story short we ended up yanking out the GM BCS and putting in a MBC.

On the street, this car would boost to 25-26psi solid and run huge negative correction to try and bring it back down. When we got on the dyno, this is what happened (lower plot):

(upper plots are switched to MBC)

BOTH cars behaved this way on the dyno. A spike, followed by a huge drop-off in boost, then slowly ramp back up towards redline; power down the entire run.

Anyways, in my logs, boost would be waay down, but I would see zero error correction until around 5000rpms or so.

BOTH cars work awesome on the street, except for the overboosting issue. Anyone have any ideas what would cause this?

MR Turco

iTrader: (16)

I have been running 5.10 with averaging turned off without seeing this issue. Have you taken logs with averaging off and then on back to back? Also, what does your correction table look like?

derekste

iTrader: (9)

what do the boost control tables look like? is this psi-based or load-based target?

scheides

iTrader: (10)

The correction table is fine, pretty basic settings;similar if not identical to the ones posted in mrfred's thread for doing direct boost control.

I do not have before/after graphs with map averaging on/off, it did not don on me until this morning that it might have been the cause of the issue. I wish I would have thought of it yesterday on the dyno

Like I said, both cars behave perfectly on the street. MAP averaging does seem to help smooth out boost on the street, the boost does not fluctuate as much as it does with it off.

scheides

iTrader: (10)

Here's my tables, sorry I guess I should have included them in the orig post. Nothing fancy about them:

mplspilot

iTrader: (7)

Scheides, one thing i would suggest is fixing your boost desired table to follow your actual boost curve that you get with your WGDC settings. Right now you are requesting a flat line of 23psi starting at 2000 rpm. That's when your boost error correction probably kicks in and ups your WGDC up to whatever value you have in "max WGDC correction" — 10% if it's stock. So by the time you're at 3000-3500rpm your WGDC is way too high and you shoot over 23psi, up to 26psi, that's when your correction table takes out 11% of WGDC and maybe does it again, and maybe again because the checking interval is 1 (or 2, i dont think 1 actually works as 1), so you get a massive correction downward, and that huge dip.
By then you maybe fall out of the correction control loop, because you're below -3.1 psi from your requested 23psi...
It'd be easier to figure out with logs. I may be completely off here..
my 0.02

scheides

iTrader: (10)

Well the first few runs I had the top two columns set to zero for error correction, meaning it could not do much upwards error correction. Its still SO weird that it behaved so differently on the street than on the dyno, and furthermore different than it had behaved last time I was on teh dyno.

I think that's hardcore the biggest problem--I did not have aggressive enough upward correction set; I've never needed it on the street so never worried about it.

With that being said, I just discovered that I think I was on v5.8 (v5.08 if you will) still; I didn't have an option for 'map averaging' in my tephra mod list, and a little digging discovered old xml. When I loaded the v5.10 xml in, I get jibberish.

I think I'm just going to leave it until v6 gets realsed; I think I just missed/skipped the v5.10 boat, or backtracked at some point to v5.08.

*sigh* sorry guys.

MR Turco

iTrader: (16)

dont forget too that the car will load up completely differently on a dyno dynamics compared to the street.

scheides

iTrader: (10)

Yea I know; typically it's not that much different though. The thing that just is crazy to me is how different it is now from last year. It performed so perfectly then, and now just garbage.

MR Turco

iTrader: (16)

The weather, at least in New England has completely screwed my boost tune and it was PERFECT in the fall. Like you, i am waiting for V6 to come out. Between that and SD there is going to be a lot of tuning this spring.

scheides

iTrader: (10)

Yea, the sad thing about this is that I went for a baseline before I install my BBK. Since that top dyno plot, I had a HG failure, swapped in Kelford 272's in place of HKS 280s, GSC springs, perrin air filter, and a buschur quiet axleback (in place of my bullet). So much has changed, it seemed prudent to get a baseline and try and do a before/after comparison but looks like I will not be able to do that now

After the dyno I discovered a boost leak at the TB (loose t-bolt clamp of all things!) and air leaking past my turbo's bearings into the crank case so yea....yet again I have some work ahead of me.

I'm starting to wish I had left my car as stock as possible....sigh. No fun in that though

MR Turco

iTrader: (16)

Looks like a couple issues were due to stock parts failing

GST Motorsports

iTrader: (8)

You are targeting 23psi in your boost tables and the car hit 26psi on the dyno, why is it confusing that error correction pulled the boost?

Now since you said you had the "top two columns set to zero for error correction, meaning it could not do much upwards error correction." then it took the logged error to be less than -2.3 before it would add back in WGDC, and then only add in 4% (according to your table)

This is the problem with zeroing out the top two upwards error correction cells. You just found out why.

if you are doing that to reduce spool overshoot then I would look at zeroing out the Max Upwards correction VS TPS and filling back in normal upwards correction in the error correction table. I've been singing this virtue for awhile, because when tuned properly, it works VERY VERY well. Not many people seem to listen though. Send me a PM, I'd love to help if wanted.

btw, I've tuned a bunch of cars with map averaging via 5.10 with no issues.

fostytou

iTrader: (15)

Yeah, definitely looks like a 3+psi overshoot to me and a big reduction, then getting stuck under the curve for a bit. MAP averaging may have exhasperated your problem if you had higher correction intervals, but it looks like you are using 1.

Try not allowing upward correction and getting a duty cycle curve that fits the boost profile better. I'm guessing there was higher loading or a different gear was used than is on the street and caused your issue.

Worst case you could have just zeroed the correction table... to me that seems easier than going through the trouble of hooking up an MBC, and you'll see how close your curve is.

scheides

iTrader: (10)

Thanks a ton for the advice. I'll do some research and test it out before I bug you. The problem still bugs me several fold: it was not a problem the last time I was on this dyno, and it works perfectly on the street. Absolutely perfect.