Cld12pk2go's ATP GTX3071 Build

[cld12pk2go]

Originally Posted by BlueStreak @cld12pk2go. Looking at your boost profile, I'm seeing roughly 20PSI by 3200-3300. The compressor map indicates this is wicked close to the surge line. Not worried?

I have never heard a hint of surge, and per my math I am comfortably away from the surge line...





I am actually running quite a bit more flow than that plot now (more like low 60's lb/min flow)...

[atvfreek]

I love that power band!

[BlueStreak]

Ah, I see. The second VD plot shows you hitting 20PSI alot later than the other one. Was this done on purpose?

Here is my math. The LBS/MIN value assumes 20PSI across the board. I then look at the RPM when you hit 20PSI on your VD plot and match it to the values on the table.



So, 20PSI @ 3200 RPMS is ~21LBS/MIN and 20PSI @ 3700RPMS is ~23.5LBS/MIN. Looking at the GTX3071 compressor map at a 2.5 pressure ratio puts it very close to the surge line.

Do you have a table where you worked out your math? I'd like to see your calculations.

I'm asking because I have a 21PSI spring on my EWG setup and can hit 20PSI at 3200-3300 on spring pressure alone (which produces awesome response). My calculations have me a little worried on my GT3071R setup. Unfortunately, I can't hear any compressor surge (if it is surging) since the EWG is damn loud. I know on extremely high surge situations, boost can oscillate but my boost curve is very smooth and I'm unsure how the boost profile would react under mild surge.

[SleeperHatch]

How are you hitting 20psi on a 30r at 3200rpm

Sent from my EVO using Tapatalk 2

[BlueStreak]

Originally Posted by SleeperHatch How are you hitting 20psi on a 30r at 3200rpm Sent from my EVO using Tapatalk 2

Awesomeness? Haha.

I used to have a GT3071R with IWG and hit 30PSI by 3300RPMs because I was a little overzealous with the WGDC. It died a few months ago...

The only thing I can comment on w.r.t. differences is that my engine is forged with a whole lot of headwork done.

[cld12pk2go]

Originally Posted by BlueStreak Ah, I see. The second VD plot shows you hitting 20PSI alot later than the other one. Was this done on purpose? Here is my math. The LBS/MIN value assumes 20PSI across the board. I then look at the RPM when you hit 20PSI on your VD plot and match it to the values on the table. So, 20PSI @ 3200 RPMS is ~21LBS/MIN and 20PSI @ 3700RPMS is ~23.5LBS/MIN. Looking at the GTX3071 compressor map at a 2.5 pressure ratio puts it very close to the surge line. Do you have a table where you worked out your math? I'd like to see your calculations. I'm asking because I have a 21PSI spring on my EWG setup and can hit 20PSI at 3200-3300 on spring pressure alone (which produces awesome response). My calculations have me a little worried on my GT3071R setup. Unfortunately, I can't hear any compressor surge (if it is surging) since the EWG is damn loud. I know on extremely high surge situations, boost can oscillate but my boost curve is very smooth and I'm unsure how the boost profile would react under mild surge.

Basically, I use my datalog to get the Boost Pressure and MAF airflow for each 500 RPM breakpoint and then do the math to convert these to the proper units of lb/min on the x-axis and pressure ratio for the y-axis to allow me to plot where I am on the compressor map using the good ole Cartesian Coordinate system...

Since I am using a significant amount of meth, I need to apply a meth offset to my MAF g/s to get a more true airflow reading.

I have previously calculated my meth flow here at enough to shift my MAF down by 66 g/s:

Doing the math: Methanol Fueling Calcs



I have a progressive controller that kicks in at 3.2 volts on the MAF (around 145 g/s), so I just applied a best guess progressive flow profile to my meth offset.

I also have measured a little over a peak of 3 PSI delta between my IM and turbo discharge pressure at WOT up top, so I applied a similar progressive profile here for the Pressure Ratio calc.

I am assuming with my free flowing intake that my depression through it shouldn't be more than 0.5 PSI. I haven't tried running the fluid mechanics calcs to see if more or less is justified. This is just an old rule of thumb for the DP with a free flowing intake.

None of these three profiled corrections are going to be exact, but they should be plenty good enough to get a decent compressor map plot.


So here are the results:










The one thing we need to keep in mind is that the compressor maps are generated at ~68°F conditions. This pull was at 31°F, which means the initial air density was much higher allowing the compressor to move more massflow. So if we had a ~31°F compressor map, I would not be nearly as off the right side. Off course, this means I am actually a little closer to the surge line on the left; however, I believe I am still comfortably away from it due to the above plot coupled with the lack of any surge symptoms.

[BlueStreak]

Thanks! I'll study this as soon as I get behind my laptop.


Sent from my iPhone using Tapatalk

[rfinkle2]

Originally Posted by BlueStreak Thanks! I'll study this as soon as I get behind my laptop. Sent from my iPhone using Tapatalk

I was a little concerned about this as well. The difference in our calculations vs. cld12pk's is that we are using Garrett's method for determining appropriate turbo size for a particular engine displacement whereas cld12pk is using his maf sensor readings and converting to lb/min.

Interesting thing to note is that the GTX series supposedly has anti-surge machining built into the housing, but it looks nothing like the anti-surge machining option for the regular gt model:

[BlueStreak]

Originally Posted by rfinkle2 I was a little concerned about this as well. The difference in our calculations vs. cld12pk's is that we are using Garrett's method for determining appropriate turbo size for a particular engine displacement whereas cld12pk is using his maf sensor readings and converting to lb/min.

Precisely! I took the mass airflow output from the calculator at face value for all RPM breakpoints.

Though it is bang on for the upper RPMs, things aren't as *constant* across the board. At first thought, VE is assumed to be 90% across the entire rev range but in reality we know that not to be true. Assuming VE is the culprit in the calculations, actual VE would have to be higher in order to shift flow values to the right to ensure we're not surging. Perhaps VVT comes into play as well and we can't forget about cams...

I do like cld12pk2go's comparo tho. Using actual values to plot does make sense instead of using theoretical values since we can get them.

I'm going to have a look at a datalog when I have a chance to see what's what. Theoretically, I should be *safer* with my compressor map lying slightly to the left vs the GTX3071.






Sent from my iPhone using Tapatalk

[cld12pk2go]

Originally Posted by BlueStreak Precisely! I took the mass airflow output from the calculator at face value for all RPM breakpoints. Though it is bang on for the upper RPMs, things aren't as *constant* across the board. At first thought, VE is assumed to be 90% across the entire rev range but in reality we know that not to be true. Assuming VE is the culprit in the calculations, actual VE would have to be higher in order to shift flow values to the right to ensure we're not surging. Perhaps VVT comes into play as well and we can't forget about cams... I do like cld12pk2go's comparo tho. Using actual values to plot does make sense instead of using theoretical values since we can get them. I'm going to have a look at a datalog when I have a chance to see what's what. Theoretically, I should be *safer* with my compressor map lying slightly to the left vs the GTX3071. Sent from my iPhone using Tapatalk

The stock VE table clearly shows a ~10-15% higher VE in the lower RPM ranges compared to up top...

As an engineer, I love theory; however, actual process data >>>>> theory.

[cld12pk2go]

Final power from this setup:

http://www.mazdaspeedforum.org/forum/foru...0-7wtq-133471/