"Standback Tuning For Dummies Guide" The Official Q&A...

[spnkr21]

Nice work outta you Jackson. Sucks you had to take matters into your own hands after spending so much money on a product. I'll have to give this program a try.
Thanks again dude!

[MattJackson86]

So now that I have been able to log with the SB i have been dialing in my tune. AFR looks very good now and I have spent some time trying to dial in the boost with the PID settings. I have settled with the values 29, .075, 250 and the boost seems pretty good. I think its as good as its really going to get. My tune is for 18.5 psi til 5500 where I start to ramp down.

My MAF definitely seems to be maxed out or very close to it. I consistently see >4.75 volts in my logs and 4.95's are not uncommon. Is this really that common on a stock turbo? Its driving fine but any cause for concern? I was thinking about going for the bigger MAF/inlet but I dont think I will gain anything from it right now, especially for the cost.


Attached is the DH log and the SB log in CSV format. Any comments on boost, timing, or anything else would be appreciated. I plan on going to the dyno to see how it performs and to fine tune anything if necessary.

[jhershorin]

Matt - can you send me a log of you reving out - timing seems a bit conservative.

[Cue]

Originally Posted by jhershorin Matt - can you send me a log of you reving out - timing seems a bit conservative.

Agreed.... you should be hitting over 12/13 around 5300rpm. Plus it looks a little rich IMO. Tapering down, you could easily do 12.5 around 4500rpm to high 11's at 5700rpm... and that's still a bit conservative.

[MattJackson86]

Yes sorry the 4th gear ended early. Ill get one tomorrow revving out to redline. My timing may seem low as I am pulling 4-5 degrees of timing to keep knock down. I attached a pic of my timing chart for reference.

As for the AFR. I shoot for around 11.7-11.9. This gives me room for when the SB wants to act weird and go 12.2ish or when it gets cooler. My 3rd gear pulls dont often mimic the 4th gear and may run leaner.

Thanks for looking btw.

[RichieRichness]

Its looks like you need to start adding more timing at around 4k rmp...?

Matt...are you going to get the CP-E Flashes?

[MattJackson86]

I have them already.

Will get those logs today btw. How can I add timing though if I am pulling 3+ degrees to rid me of knock? I know my timing my look low, but I dont see myself adding any without the onset of KR. You can see in my above log that I still get the occasional blip of .7 kr around 5k.

[RichieRichness]

Your always going to have a little KR (.4 -.7). Log timing with your DH since you can see how it interacts with KR and rpm...imo it makes things easier to read. Always take 2 logs back to back to check for consistancy. Post some new logs after that.

[MattJackson86]

Yea I still expect kr and am ok if its under 1.0. But if I increase timing and not pull it so much, I would expect my .7 kr blips to become larger. Ill get some logs this afternoon.

[MattJackson86]

Here are my logs. 3 from the SB, 2 from the DH which match up with two from the SB. My AFR needs making a little smoother imo, but otherwise im happy. Just looking for 2nd opinions. Thanks.

[RichieRichness]

Matt,

Thanks for making the mazda tuner app....I used it yesterday and it's awesome for logging! You da man!

A couple suggestions when logging. Scale your AFR's down so we can see spikes and dips better....it takes some playing around with the DH software and graph options to make your logs reader friendly (try right clicking objects on the y axis to make those changes). If you go back a couple pages in this post you will see how I scaled everything in the DH software.

Your AFR's look okay right now but your timing is way off...I can't imagine that your car pulls very hard with that timing curve. I put together a chart to show you where you want to be as far as timing goes. This is based off my tune that was done by Jake.

Your probably going to have to make a bunch of adjustments to get your timing right. Keep an eye on your afr's after making adjustments...post up some SS of your map. I want to see your MAF and fueling settings.

jackson's timing.GIF

[MattJackson86]

Thanks for the info. Thats comparison chart definitely is interesting. I will play with my timing again. I will start by not pulling 4-5 degrees and only pull 3-4 degrees and see how the kr is. I expect it to return, but we shall see.

Glad you like the app. It works way better for logging imo. I am working on getting OBD information included in the logs, so no more separate DH/SB logs. Still a work in progress and no ETA yet.

[pebenito]

Originally Posted by MattJackson86 Glad you like the app. It works way better for logging imo. I am working on getting OBD information included in the logs, so no more separate DH/SB logs. Still a work in progress and no ETA yet.

So are you planning on merging in the DH logs, or actually connecting over OBD/CAN?

[MattJackson86]

Connecting to the OBD port directly. It can supposedly be done through the SB, but I have not found how. So it will require a OBD-USB adapter, but will make logging easier as you can start/stop in one location (my software) and have it all work.

It will work, i just want to also make sure that it does not slow down my sample rate to much. I get about 6 samples a second right now from the SB. Thats pretty good and I dont want it to drop much lower by now logging OBD data too. We shall see. Merging would be easy to do though so that could be another option if this does not work out, but I like the idea of one software to log it all.

[RichieRichness]

Thanks Matt. Also wanted to let you know that the auto logging option didn't work for me. I had it set to the default 3 volts. Tried it a couple times with no luck.

[MattJackson86]

Really? Odd. Works great for me. Hmm...let me think and Ill pm you.

[Cue]

Originally Posted by MattJackson86 Thanks for the info. Thats comparison chart definitely is interesting. I will play with my timing again. I will start by not pulling 4-5 degrees and only pull 3-4 degrees and see how the kr is. I expect it to return, but we shall see. Glad you like the app. It works way better for logging imo. I am working on getting OBD information included in the logs, so no more separate DH/SB logs. Still a work in progress and no ETA yet.

Dont forget to try other gas brands. I had a similar issue with KR just to find out that the 91 I was using just couldnt be tuned. Played around with different gases and it turned out that shell (for me) was a night and day difference.

[MattJackson86]

I already get crap from the fiancee from only going to Chevron. She's gonna think something is seriously wrong if I switch to Shell and now only go there

Its a good point though. I will play with try that once I confirm the KR returns. Ill start slow on timing so I don't get too much KR.

QQ: Advancing timing, which is what I will be doing going from -4.5 to -3.5, will that also richen or lean me out a tad?

[RichieRichness]

Originally Posted by MattJackson86 QQ: Advancing timing, which is what I will be doing going from -4.5 to -3.5, will that also richen or lean me out a tad?

You will see that your AFR will lean out/increase as you add timing...this is because your allowing more time for the fuel to burn....which means less fuel in the air mixture being read by your wideband.

[MattJackson86]

Ok. I will get my AFR looking good at about 11.5 with no spikes too high. Then when I advance timing it will lean out a bit and I should move back towards to my target of 11.7ish. Maybe I was getting kr from being too lean before, so I pulled 4.5 degrees and this made me richer and thus less kr? Ill try it.

[pebenito]

Originally Posted by MattJackson86 Connecting to the OBD port directly. It can supposedly be done through the SB, but I have not found how. So it will require a OBD-USB adapter, but will make logging easier as you can start/stop in one location (my software) and have it all work. It will work, i just want to also make sure that it does not slow down my sample rate to much. I get about 6 samples a second right now from the SB. Thats pretty good and I dont want it to drop much lower by now logging OBD data too. We shall see. Merging would be easy to do though so that could be another option if this does not work out, but I like the idea of one software to log it all.

I look forward to this, though I'm hoping my the bluetooth OBD adapter I'm using with the Torque app on my phone will work.

[MattJackson86]

Im sure it will. It probably is just a bluetooth serial port, so to my app it wont look any different. And as long as it uses the ELM327 protocol then it should be fine.

The hardest part is getting the Mazda specific PID information decoded. I have AFR and KR done, so the important ones imo. Would still like the others though (BAT, etc).

[pebenito]

Originally Posted by MattJackson86 Im sure it will. It probably is just a bluetooth serial port, so to my app it wont look any different. And as long as it uses the ELM327 protocol then it should be fine.

Yeah, I figured that was the case (bluetooth/usb bridged to serial).

Originally Posted by MattJackson86 The hardest part is getting the Mazda specific PID information decoded. I have AFR and KR done, so the important ones imo. Would still like the others though (BAT, etc).

Would you care to share KR (and any other future ones you figure out), I know people that use Torque would be interested.
http://www.mazdaspeedforum.org/forum/foru...29/#post716381

[MattJackson86]

Maybe. I have all the PID information already for everything, its the decoding that is difficult and time consuming. As in, taking the value 0BAC and converting it to 11.75 AFR. Just figuring out the formula is not so straight forward. Ill get it done though.

[RichieRichness]

Matt,

I just did a couple logs today and found that the timing in your program doesn't reflect that of what I see in my DH. Here are my logs. You can see the timing in the .csv files are lower than the DH reads....?

2-11-2011_3rd_Log1.GIF
2-11-2011_4th_Log1.GIF

[RichieRichness]

Ive also noticed I cant hit my peak boost target of 18 psi. It's maxing out at high 16s. Anyone know why this is? Do I need an aftermarket BPV?

[MattJackson86]

Anything the SB modifies will be skewed compared to the DH. So I believe that this is the reason your timing looks different. Same will go for MAF volts etc. Someone please correct me if this is not the case.

As for the boost, its your PID settings probably. They can be a PITA to get right but it makes the world of difference in boost control (good and bad). I use 29, .075, 250 right now.

FYI, new version of my software was posted online.

[spnkr21]

The only thing you log with the dh is bat+afr. Everything else is skewed including boost.

[RichieRichness]

My current settings are Kp=30 Ki=.09 kd=150. After studying the logs I took it looks like Kp, Ki and Kd have a direct effect on WGDC. It looks like as these values increase/decrease it has the same effect respectively on Wastegate DC. It looks like my waste gate is dumping too much boost, preventing me from hitting my boost targets. This makes me think to decrease Kp/Ki/Kd to decrease WGDC thus increasing my boost.

I'm not sure how kp, ki and kd interact with each other to adjust wastegate DC and don't feel like experimenting with these values until I have a better understanding. I've also heard it is wise adjust the boost curve to match the torque curve...I'm assuming I need to dyno my car to know this information? Can someone point me in the right direction here?

Attached are a couple of my latest logs.

[MattJackson86]

It does take a lot of testing, but i used this to help me understand the effects of changing p,i, or d. Scroll down to "Manual Tuning": https://secure.wikimedia.org/wikiped...PID_controller

I just studied that and kept referencing it as I tuned. It seems to have worked out. My boost is still sitting about .5-.8psi lower than my target, but I think thats as good as it gets. Every now and then it will jump over my target and work its way back down.

As for the boost curve, I just set it to gradually come on with the RPM's and with the TPS volts.

FYI, my timing changes are going well. I guess when I originally tuned I was too lean and got KR, so I pulled timing and this rid me of kr but probably by making me richer. By making myself richer in the first place my timing is now normal. The tune is quite different now but so far seems way better. My timing changes are set to 0 except for a couple lines at 5400rpms where I pull 1.5 degrees. Still working on that section but 0KR for the most part.

[RichieRichness]

Originally Posted by MattJackson86 As for the boost curve, I just set it to gradually come on with the RPM's and with the TPS volts.

What about trying to match your tq curve?

[MattJackson86]

Originally Posted by RichieRichness What about trying to match your tq curve?

Have not done that. Not been to a dyno to do so and this is the first I have heard of such a method of tuning boost. I just set it so it comes on full (18.5psi) at 3800rpms and then keep it on til 5800rpms where I taper it down slightly.

[the_caruch]

Originally Posted by RichieRichness My current settings are Kp=30 Ki=.09 kd=150. After studying the logs I took it looks like Kp, Ki and Kd have a direct effect on WGDC. It looks like as these values increase/decrease it has the same effect respectively on Wastegate DC. It looks like my waste gate is dumping too much boost, preventing me from hitting my boost targets. I'm not sure how kp, ki and kd interact with each other to adjust wastegate DC and don't feel like experimenting with these values until I have a better understanding. I've also heard it is wise adjust the boost curve to match the torque curve...I'm assuming I need to dyno my car to know this information? Can someone point me in the right direction here? Attached are a couple of my latest logs.

increase your Ki. try .2

In my experience the stock turbo isn't affected too much by small changes to the PID loop. So personally if i were you i would put in .5 or .7 into your Ki and see where you are at. It's up to you though

[MattJackson86]

See from my testing I found its somewhat a crap shoot as what to change. Ki to .2 may help you, but changing kp or kd may also help. I am currently at: 29, .075, 250, so you can see my kd is a lot higher than yours. Just start slow and see what changes, then based upon that link determine what to change next.

[RichieRichness]

I read this part about correcting droop...it looks like Kp should be the starting point.

DroopA pure proportional controller will not always settle at its target value, but may retain a steady-state error. Specifically, the process gain - drift in the absence of control, such as cooling of a furnace towards room temperature, biases a pure proportional controller. If the process gain is down, as in cooling, then the bias will be below the set point, hence the term "droop".

Droop is proportional to process gain and inversely proportional to proportional gain. Specifically the steady-state error is given by:

e = G / Kp
Droop is an inherent defect of purely proportional control. Droop may be mitigated by adding a compensating bias term (setting the setpoint above the true desired value), or corrected by adding an integration term (in a PI or PID controller), which effectively computes a bias adaptively.

Despite droop, both tuning theory and industrial practice indicate that it is the proportional term that should contribute the bulk of the output change.

[MattJackson86]

And if you look at that chart I linked you too you will see how it mentions that a high P will benefit some areas and hurt others. I found you wnat the lowest Kp and Ki possible that still works. Raising those values seems to introduce some negative effects. Then if it doesn't get up fast enough increase you Kd.

[www.cp-e.com]

For those that are further interested in PID algorithm explanations. This MoTeC technical document has an excellent explanation of these functions.

P value or gain, always carries the most weight. I have always found that the ones that you will ever really find any use with making more significant changes are only P and I. D is more for long term, kinda.

Oops, can't host it. Email me and I'll send it to you.

[RichieRichness]

I can hit my boost targets now (much closer at least). Only change I made was to Ki...set to .25

Thanks caruch for the tip with Ki settings.

This is a graph showing my boost compaired to boost target.

boost.GIF
Matt,
Your boost is all over the place...why?
jackson's boost.GIF
Sent from my iPhone using Tapatalk

[MattJackson86]

I was trying to get closer to my target. I used to have a graph just like yours but I dont like the fact it settles 1 psi below target. Mine is a little better in that it keeps trying to obtain the target, but it does look erratic. It does look though that I have a lot more data points than you do on the horizontal.

Sander, what are you suggestions with our PID settings?

[RichieRichness]

Originally Posted by MattJackson86 I was trying to get closer to my target. I used to have a graph just like yours but I dont like the fact it settles 1 psi below target. Mine is a little better in that it keeps trying to obtain the target, but it does look erratic. It does look though that I have a lot more data points than you do on the horizontal. Sander, what are you suggestions with our PID settings?

I'll take some 4th gear logs this afternoon. As far as being 1 psi off target, I'm on average about .5 psi below.

boost change.GIF