Taking a journey through pain and a misfortune. A speed6 saga.

So where did we leave off?? Right.. a bad rod bearing on the 2.5L, bad rod bearings and worn out cylinder bores on the "good used" 2.3L... The 2.5L rod had enough hot spotting, plus my feeling that there was issues with the chamfer job I did on the crank shaft oil galleries.. I pulled the 2.5L out of the car and tore it down... To my surprise I found some scoring on the cylinder walls.. It wasn't bad.. but with the amount of carbon on the pistons.. I figure I was running a little to rich, causing some issues with washing down the cylinder walls... Forged pistons also require proper warm-up and some other special driving habits that I don't think I was really paying attention to...

I took both blocks to a couple of machine shops to get some opinions on machine work and costs.. There's no one locally who really deals with these motors, so the idea of getting into machine work just didn't sit well with me... I happen to come across a person on FB market place parting out a speed6... Again.. video's of it running, good compression numbers... it should be a good motor to drop in right??

Well god dam it!! Another fucking worn out running motor.. This one had a broken VVT actuator and the timing chain was hitting the valve cover..



Lets just say moral around this vehicle was shot... Once again I abandoned the car outside of the shop I worked at.. This time around for about 6 months while I sorted out some other life stuff... One of the things I sorted out was renting a shop for doing my own personal work..
In the back, my 93 Mazda 323 nick named princess.. A very nicely restored 1st gen Bronco getting a sniper efi install.. my buddies Camaro getting some regular maintenance work.. A lovely 2005 Mazdaspeed Miata with 25,000km's on it.. All original..


September of 2023 I melted down... The living situation I moved into after selling my house was a disaster.. said fuck it and bought a motorhome to live out of... Tired of waiting for promises to be full filled with work.... tired of trying to my own projects at home when all my good tools were at work... I quit my job... after 24 years of being a professional automotive tech, I had enough... Time to get all my tools into one location and focus more on my own dreams...

October of 2023 I dragged the speed6 to my own shop.. With all my tools now in one place.. It was time to get this cursed project fixed.. I went through all of the part I had... I decided to ditch the 2.5L, go back to a stock 2.3L block... I picked the best 2.3L block I had.. both of them worn, but one had a little meat left before it would be over the max oem specs.. I got a set of OEM rings, OEM bearings, decided to reuse the Gen 2 pistons and rods from the first motor and basically did a low buck stock rebuild... I was joking around with a buddy of mine, saying the 2.5L failed cause I didn't dress it up and make it look "pretty"... So this motor got some pained work.. cause why not??

My 323 has full polly mounts.. My buddies Speed3 has polly mounts.. I am old and can't stand how a car feels with polly mounts.. In my Mx3, I custom made solid rubber mounts using leaf spring bushings.. They work amazing... My limited time spent beating on the 2.5L in the speed6 left me really worried about the stock mounts.. For good reason, it's known that they are shit... I happened to have a left over leaf spring bushing.. So I did a thing...

Some modifications were made to the subframe for something to get bolted to..


Teaching my son the "binky" ways of CAD... We built a 4th engine mount... The design and concept is good.. but my execution isn't.. Measure once, cut twice?? eye ball it good enough and re-do it later?? I made the part that bolts to the frame with the bushing in it to wide... I've got a diagnosed disability that includes struggles with dyslexia.. when I work alone I am ok, but I am sure at some point I was giving my son measurements and I said the wrong numbers... I thought I'd run it, rather than redoing it... I might as well have polly mounts in the car.. this mount is grounding out causing a far bit of noise in cabin over 3500 rpm... If I narrow the bushing and bracket it'll fix the issue..




A friend of mine came for a visit.. She's an aspiring artist, never played with spray paint before.. So we played around a bit.. Motor's going back together nicely..


Almost ready for it's first Road test... You'll notice the big fancy stainless steel piece... I completed the cover for the cold air box I started building with the 2.5L.. The clearances are tight, when I get into some of the data logs you'll see... This cold air box makes a difference...

Well she looks fairly pretty now... That means she's going to hold up and stay together now right??


Only if I don't do stupid things... My first road test went pretty well.. I brought the car back in went to jack it up to check for leaks and I did something stupid... Using a block of wood and jacking the car up on the subframe from the front of the vehicle I made a big mistake..


No I didn't jack the car on the rad support.. I was using a block of wood because my jack doesn't go high enough to get the car on my good sized jack stands.. As a mechanic you can get lazy when you do things cause you've done it soooo many times.. Well I didn't pay close enough attention to how the block of wood and my jack were sitting on the subframe, as I was jacking the car right when it got to the top it started shifting... Just as I attempted to twist the handle on the jack to put the car down and re-set it properly it slipped.. My jack at full height punched through the rad support and as you can see destroyed the rad, cooling fans, ac condenser and rad support... In complete and utter disgust at myself... I threw some kitty litter on the mess, turned the lights off, locked the shop up and left...

If you ignore the busted rad support, you can see that I cut down the bumper bar during this build.. I also drilled out all the blocked off passages in the grill to match what I cut out of the support... This was to help increase air flow across more of the intercooler..


The car almost got stripped and torn down at this point for good... Almost.. I got lucky and FB market place to the rescue.. a guy locally was selling a V6 2006 Mazda6 with a blown motor for cheap... I almost bought the whole car just for the rad support and to see if I could make the V6 rad and stuff work because this was the ONLY parts available locally... The guy was super cool, he let me strip the front end of the car in his drive way while he was at work... I was really diligent about not making a mess of coolant and stuff like that... The V6 rad requires a remote reservoir that's pressurized.. I forgot to grab the reservoir from the car.. I've got a vibrant piece for making an aluminum coolant overflow tank that's pressurized... I thought about making my own tank.. But decided to spent the money and ordered a proper stock rad from rock auto..

Got the car all back together just in time for some fun in the snow... Lets just say, I quite enjoy a speed6's power split in the snow... It drives more like a rwd car, letting you hang the ass end out nicely under power..

 

Sooo... The car's running great.. It feels great.. While I get my tuning notes/discussion stuff together, Lets take a break on a good note and leave you with some recent actions shots of me teaching my son how to properly drive a car... sshhhh, don't tell his mom... Look at those smiles!

 

Tuning.. isn't this why any of you are here?? I talked about how this car was to be a tuning experiment for me.. So lets start.. My back ground isn't in ECU tuning. I barely graduated high school and had a knack for fixing cars.. Got a job at a local shop and started my career.. I am a horrible parts changer and would loose my ass if I worked flat rate... I fix the stuff, other guys don't want to, can't or don't understand how to work on..

Tuning and racing is a hobby for me that I am slowly developing into something more.. Before my speed6.. I was playing with HPtuners, Gen3 and Gen4 LS stuff.. I've played with some Ford 5.4L stuff.. Dug around in Audi/VW and Mercedes stuff.. There's been some Megasquirt, AEM and Link stand alone ecu stuff thrown in there..

Things I've learned over the years... If you want to get good at tuning a vehicle, go get good at diagnosing and fixing OEM vehicles... Focus on learning how to use a scanner to read data and develop a diagnostic path.. Don't rely on code to tell you what part to change, use codes to guide you, so that you can use data to verify and confirm the problem... If you're way of solving a problem is asking people what part you need to change in order to fix a vehicle, you shouldn't tune a vehicle.. You need to understand how components work, what they are doing and how to use that information to diagnose what part you need to change.. 90% of the "help me fix my car posts" can be narrowed down to 2 or 3 specific part tests with in 10 to 15 mins using scan tool data from the drivers seat..

Here's my first bit of contradictory information... Cleaning your MAF sensor should be done a part of maintenance or to help assist in diagnosing a problem... If you clean your MAF sensor and if fixes your problem, you still need to replace the MAF sensor.. Cleaning the MAF sensor is NOT a way to fix a problem... A dirty MAF sensor that is causing a problem has built up a "crust" on it.. You can not clean that crust completely off.. You can gently clean half the sensor element, but you can't get to the back side of it.. This "crust does 2 things... It insulates the MAF sensor element, which causes the MAF to read more air at idle.. So you end up with lower, usually negative fuel trims at idle.. under heavy load, but staying in closed loop (So not full throttle), the insulating effect of the crust will cause the MAF to read less air.. That means you'll end up with positive fuel trims.. At WOT in open loop, you'll see an increase in knock issues/leaner then commanded AFR.. There are some very very easy trends to look at with scan data that can diagnose this problem. MAF sensor issues are really easy to diagnose if your car has been properly tuned.. When you clean the sensor and it helps your problem, you are cleaning a layer off, or part of the crust which is why it can look like it fixes your problem.. The reason you should still replace your sensor, is because of response time... This can only be seen with a lab scope... It'll effect you're "accell enrichment" during throttle transitions... It'll cause the change in timing command to be slow...This happens to fast for it to show up on scan data, but essentially you're created a "lean" spot because the sensor doesn't react to the change in airflow fast enough.. You'll run to much ignition timing for a split second leading to tip in throttle knock issues... Cleaning your MAF as part of maintenance helps to ensure the "crust" doesn't form... So for all the quick responses to help me threads saying have you "cleaned your MAF sensor" my response is simple.. whats your scan data/logs looking like.. It's really really easy to see if this part of your problem...

Ok with that out of the way.. Lets talk more about tuning stuff.. Airflow, airflow, airflow airflow... If you are talking about "boost pressure" while measuring power gains.. Or looking at people asking "how much boost" they are running... Using that as your metric to think about power output, you are LOST!! Boost is how you measure the restriction in AIRFLOW above barometric pressure!!! Boost pressure plays an important roll in tuning, but not as a metric of how much power you are making..

OBD2 regulations... huh?? what? what do we need to talk about that?? Well it's important, CARB and OBD standards are dictating how manufactures can do certain things.. this was done to standardize things for emissions testing purposes and diagnostics.. spark tables, fuel, EGR scheduling... The data pids you're reading and what information you are getting from it... If you're using a scan tool that doesn't have OEM data capabilities than OBD2 regulations are dictating the data you're looking at... In order to tune a car... you need to look at data and understand the data you're looking at...

Calculated Load... Most of you are familiar with this if you are data logging with COBB... well, their version of calculated load, isn't the OBD2 standard of Calculated load, it's actually Absolute load... Calculated load, is a 0 - 100% data pid... Absolute Load is 0 - 400%... Calculated load follows manifold pressure.. Absolute Load follows air flow... If you are tuning with Versatuner, then you'll be used to see Absolute Load.. Absolute load is the measurement of air mass per cylinder, per intake stroke, normalized as a percentage.. So Load tuning isn't about targeting torque, or power directly... It's about being able to repeatedly measure and target how much air enters each cylinder, every intake stroke, under all kinds of different conditions.. By doing that, in theory if you make 270hp in 100 deg summer weather.. you'll make 270hp in 40 deg winter weather.. You'll make that same power at sea level as you will when climbing the mountains.. air mass per cylinder calculation is meant to account for changes in barometric pressure and changes in air density due to temperature.. I've seen some of this get talked about, but not really focused on in discussions..

Airflow and air mass are technically 2 different things.. Air mass takes into consideration how dense the air is... Airflow is a measurement of it's volume.. If you take a cup of air and compress it, you make it more dense.. The more dense it is the more fuel you need, the bigger the bang, the more power you make... When you are tuning this is important to understand.. The mass airflow sensor job is to measure air mass... That air mass is not the same air mass that's entering the cylinder.. The computer has to account for the fact the what it measures at the sensor changes.. It changes as it's pulled through the turbo in let pipe.. It becomes denser and hotter as the turbo compresses it.. It's density changes again as it flows through the inter-cooler.. One last time that density will change as the air gets heated up by the intake manifold and cylinder head before it enters the cylinder.. The absolute load reading you see on a scanner is the computers final calculated measurement of the air density entering the cylinder... following me so far??? I know my brain hurts.. time to take a break and gather my thoughts.. I'll let you sit in this for a bit..

 

This is taken directly from OBD2 regulations...

"Absolute Load Value (PID 43)
The absolute load value has some different characteristics than the LOAD_PCT defined in PID
04 This definition, although restrictive, will standardise the calculation. LOAD_ABS is the
normalised value of air mass per intake stroke displayed as a percent.
LOAD_ABS = [air mass (g / intake stroke)] / [1.184 (g / intake stroke) * cylinder displacement
in litres]
Derivation:
• air mass (g / intake stroke) = [total engine air mass (g/sec)] / [rpm (revs/min)* (1 min / 60
sec) * (1/2 # of cylinders (strokes / rev)],
• LOAD_ABS = [air mass (g)/intake stroke] / [maximum air mass (g)/intake stroke at
WOT@STP at 100% volumetric efficiency]
• * 100%. Where:
• — STP = Standard Temperature and Pressure = 25 °C, 29.92 in Hg (101.3 kPa) BARO,
WOT = wide open throttle.
The quantity (maximum air mass (g)/intake stroke at WOT@STP at 100% volumetric efficiency)
is a constant for a given cylinder swept volume. The constant is 1.184 (g/litre 3) * cylinder
displacement (litre 3/intake stroke) based on air density at STP.
Characteristics of LOAD_ABS are:
• Ranges from 0 to approximately 0.95 for naturally aspirated engines, 0 – 4 for boosted
engines,
• Linearly correlated with engine indicated and brake torque,
• Often used to schedule spark and EGR rates,
• Peak value of LOAD_ABS correlates with volumetric efficiency at WOT.,
• Indicates the pumping efficiency of the engine for diagnostic purposes."

Let's take a look at the stock tune how this correlates to engine power... We need to do some math, and for that we'll need some formulas.. I am going to work with imperial measurements.. Lets start with calculating mass air flow... MAF(lb/min) =((Displacement * RPM/2)/1728) * 2.7* (P/T).. Displacement is in cubic inches.. P= Map Pressure in PSI(This is measured in Absolute pressure, not Gauge pressure).. T= Temperature on the Rankine scale...

Next thing we'll need is a way to calculate HP... Let's take this simple formula from Garret Turbo's advice on choosing a turbo.. MAF = HP * A/F * BSFC/60 )... MAF = measured in lb/min.. A/F = air fuel ratio... BSFC = Brake specific fuel consumption, since we are using pump gas that'll be approximately 0.46.. Since we are going to calculate airflow using the Absolute Load values from our tune, we'll do some rearranging of this formula to solve for HP....That'll look like this.. HP= MAF/(A/F * BSFC/60)

Lets start here... Stock max absolute load targets for 4th gear..


We'll teach our selves some basic excel functions and build ourselves some tables.. While we are doing this, we've gotta do a little bit more work with formulas in order to calculate air flow from the absolute load readings... Absolute load is normalized to a naturally aspirated version of the motor.. It also uses standardized Baro and Temp values.. This is what our formula will look like MAF = Absolute Load * (((Displacement * RPM/2)/1728) * 2.7 * (P(Standard)/ T(Standard)))

Humm... 313hp at 5500rpm... That's not right.. or is it?? Technically this is assuming the engine 's running at 100% Volumetric Efficiency.. This isn't some kind of well built performance motor that's going to operate over 100% Volumetric Efficiency.. So if we want real world number we need to figure out what the Volumetric Efficiency is.. Let's take a close look at some of the data we can get from our Tune file...


Hummm interesting.. Both COBB and Versatuner list Volumetric Efficiency Maps... If you are familiar with tuning stand alone ecu's, you'll understand what a VE map looks like... The numbers here are a little strange, until you multiply them by 100 and take some other things into consideration like EGR scheduling, PCV, EVAP, VVT..

We can see there's 2 table.. Lets have a look at the VVT Modifier table and the VVT command tables side by side.. The overall lay out of these 2 tables over lap pretty nicely... So we can assume that the VE table is based off no Variable Valve timing... The VVT Modifier table, takes the base VE table and modifies it to account for variable valve timing..

When we put the VE and VVT tables side by side we can see some interesting things going on..
To me this didn't make sense by them selves until I combined the two tables by simply adding them together.. That looked better and made more sense.. If you are used to looking at the VE map in a stand alone ECU, then this map still wont look right to you.. From what I've seen from other manufactures.. OEM VE maps can look pretty messy... They don't follow traditional smoothed out rules that would follow your torque curve.. OEM manufactures will have EGR scheduling going on, PCV calculations, EVAP commands.. depending on how manufactures account for these.. It can show up in the VE table..


I'll talk more about this VE table stuff later, for now lets teach our selves how to do more complicated things within excel.. We'll build a look up formula with the ability to interpolate between values in order to best estimate what the engine's volumetric efficiency might be... We are going to multiply that by the air flow we calculated earlier in order to best represent the actual/real world engines air flow properties.. So now our mass air flow formula looks like this..
MAF = (Absolute Load * (((Displacement * RPM/2)/1728) * 2.7 * (P(Standard)/ T(Standard)))) * Volumetric Efficiency

Look at that... 270hp at 5500 RPM... The math isn't perfect, since I am converting things from metric into imperial, there's some rounding errors.. My look up formula for the VE table is using linear interpolation, it might be a little off.. I am using basic formulas as more of an estimation then an exact calculation... So to be 5hp off the advertised technical data... I'd say we are on to something here..

 

Sweet . Thank you, info is always good!

 

So we've looked at some airflow stuff.. Maybe we should switch and take a look at Data Logging.. After all need to be able to collect and go over data in order to tune...

First step... go here, spend the few buck and unlock the full version.. You can play around with the free version, but you'll be limited and likely not capable of doing all of the stuff I talk about doing..
https://www.efianalytics.com/MegaLogViewerHD/

Megalogviewer can look at any CSV files from any data logging programs.. So you can import Cobb logs directly.. You can export Versatuner Logs as CSV files... or as you'll see use other programs to data log.. Personally I do all of my data logging through a program called ForScan...

When logging, this is what your logs should look like... This is a 45 min drive.. If all you are doing is hitting record on your logger when you are doing a full throttle run, you are missing a LOT! of important data.. This goes for simple diagnostics.. got a fuel trim code?? wonky afr's?? This is how you can narrow down your problem and figure out what to test..


The program takes some setup work and getting used to.. It has some quirks to it if you are using it to look at logs from multiple platforms.. You'll need to setup custom profiles for the different types of platforms and build a field standardization list..


Data logging it's self.. In the stand alone world you'll see data logging rates used as a way of advertising how powerful their product is.. This is because time is soooo important.. At 3500 rpm, 29 engine cycles take place.. 29 times there's a suck, squeeze bang, blow event... In order for the data you are using to be accurate, you need to collect it at as a high of rate as possible.. OEM stuff is really slow compared to stand alone ecu's... This is a big down fall to doing OEM tuning work.. Lets take a look at some CSV files because not all logging platforms are equal either..

First up, you need to know how data is collected... The logged sends out a request along the network line it's communicating on.. It requests the current data from the pids you've selected.. That ECU collects that data and sends it to the logger.. This is done in a specific order.. Her's a COBB log. Things start at the time stamp.. from there data is collected across from left to right... once completed the next request takes place... This keeps repeating.. What's high lighted in blue, is the time current time stamp - the previous time stamp.. You can see it take 0.152 second to gather up 22 pids, report that data a start a new request cycle... The data pids start to skew at this kind of RPM.. You can see how the line before is 4630 RPMs, the line I've high lighted is 5102 RPM, the line after is 5357RPM.. If we look at the MAF voltage.. it goes from 4.02V and 4.25Vs.. with 4.22Vs in the middle.. The pids along the the green line don't line up 100%.. So any AFR or fuel trim corrections we come up with on that green line will actually fall somewhere between the line before and the line after.. If you look at the MAF curve voltages, that's 7 different cells in which the correction could fall in...


to be continued....

 

I realized while I was putting together my last post that I never finished my data logging rate comparison testing.. So I didn't have 3 data logs to show case... I was hungry for some pizza tonight and looking for an excuse to go for a drive... So here we go.. Lets continue...

I set the logs up with the same amount of pids because how many pids you have will effect the data rate.. In a minuet here you'll see why COBB and Versatuner requires logging to be done with a very narrow pid list.. which means you need to break your tuning up into different sections...

Here's tonights COBB log... I've gone from 22 pids down to 18 pids.. look at the change in time between data lines.. it's gone from 0.152sec to 0.144 seconds...

Here's the Versatuner log... This logger is the worst!... If you are trying to data log off generic OBD2 data, this is the kind of sample rate you can expect.. it's horribly slow!!! 0.354 sec...


ForScan has OEM level coding and re-programming abilities if you pay for it.. this means it's communicating on a different protocal.. It's data rate is much much faster.. First thing to notice, is that COBB and VersaTuner time stamps are in Seconds... ForScan is in Milliseconds.. This means we've gotta do a little math and convert the time stamp to seconds.. One thing I've done is high lighted the time difference between multiple lines.. if you average this together is about 0.0320 seconds between lines.. This much faster than COBB or Versatuner...COBB and Versatuner both have a steady data rate.. you wont see much change between lines..



Now the reason I high lighted the time between multiple line is due to OEM data rates.. Not every data pid refreshes at the same rate.. Something you Engine Coolant Temp that slowly changes over time might up date once every 100ms... where as engine RPM will be up dated as fast a possible because of how quickly it changes.. This is hard coded into the ECU... It's done because of the limitations around the ECU's processing power... Just like your desk top computer, the newer it is, the more processing power, the faster the data rates.. The more pids you can collect and look at.. here's a good example of what I am talking about from an HPtuners data log.. I zoomed out so that you can see all the gaps between the data pids and lines.. These pids are collected at different data rates.. Some loggers will interpolate between these data gaps in order to give you a smooth looking graph.. This is something else to watch out for if you are tuning other platforms, or using different loggers.. If you notice there's 2 pids that don't really have gaps.. These aren't computer generated pids.. HPtuners lets you input 2 external sensors... In this example these are inputs from the afr gauges I have wired in..

 

All right, looks like I've bored everyone.. So what do we need to talk about next to bring some life back into this topic.. Open loop vs Closed loop fuelling.. There's enough google information out there, so I wont get into breaking these terms down and what they mean... BUUT! did you know the Mazdaspeed ECU has a 3rd fuelling state? That's right, if you data log your fuel system status pid you'll see it cycle through 3 different states... Closed Loop, OL-Drive and Open Loop...

"Closed Loop" is self explanatory... "OL-Drive" is your traditional open loop status.. Anytime you are WOT, or under heavy acceleration you'll be in OL-Drive.. "Open Loop" is more of a middle ground transition status.. For a brief time during cold start up before the O2 sensor comes on line you'll be in this mode.. When you are slowing down in decel fuel cut, then start to accelerate again. You'll transition from OL-Drive, to Open Loop and then into Closed Loop... Basically anytime you are switching between Closed Loop and OL-Drive you'll transition through Open Loop...

This why there are 3 copies of some of the maps.. Closed Loop Exit Delay, for example. This is a counting value.. It determines how long different driving conditions need to be met before it'll command a status change... The higher the value, the longer the ECU will take to transition between each status.. I've tried doing some testing with different values in order to attempt to figure out which one relates to which fuelling status, but I haven't been successful... The counting value is not "time" based... It's not a measurement of 10ms for example... If I spent time on a dyno I might be able to control enough variables in order to figure this out, but that's not something I have access to or the money to pay for.. So for now we'll leave it as a hypotheses..


When I first started tuning my speed6 on the stock motor I did some testing with the HPFP tables... I need to re-do the testing in order to gather the data again, but I can tell you that each of these tables are referenced to each of the fuel system status.. It's been a couple of years and once I repeat the testing I'll confirm this information... From what I remember "A" is use during Closed Loop... "B" is used during cold start up/Open Loop... and "C" is used during OL-Drive.... The was the easiest to setup and test..