Where is my horsepower?
#1
Where is my horsepower?
O Queso here we go. I've been running this car for about 5K miles on the new engine now, and everything seems to be working well enough. Reliability has largely been good aside from the turbo failure, which I am assuming is just due to the age of that turbo (unknown, but the rest of the kit is ~20 years old from what I gather).
One thing that has bothered me since some of the first pulls on this engine have been Virtual Dyno's estimates for my car. I understand that virtual dyno isn't perfect, but even at 14 PSI on a 2560 I'm showing less than 200 HP.
I've read a handful of threads like this, but figured I'd make one to - hopefully - document the process as I attempt to figure out what's going on.
Things I'm planning to check/address as of now.
Details about the car/build can be found in my build thread and signature, but I'll try to post most of the relevant stuff here.
BP4W, 9.0:1 pistons, VICS intake, Stock NB1 Throttle body
Very Old Flyin Miata Manifold and intercooler kit, 2.5" downpipe to full 3" catless Kraken exhaust, Genuine 2560R Turbo
91 Octane Fuel (Idaho if that matters), This is as good as is readily available around here.
I did compression test the engine recently, and the cylinders were 145, 145, 140, 152. This seems low to me, but the numbers are consistent enough. This was my first time building an engine. I have not done a leakdown test.
Spark Table as of today. Again, this seems fairly retarded to me. I was looking at Brain's basemap for a 97 (same compression ratio as my car) which was running ~8 degrees more timing all over. I don't know what octane that was targeted at. I have reduced this based on what I've heard from the knock sensor.
I will likely go out tonight and take a few datalogs of pulls. I will post virtual dyno results and datalogs when those are available.
If anyone has any thoughts, tips, or other things to check please let me know.
My current tune is attached and can be viewed online at Simba Tune
One thing that has bothered me since some of the first pulls on this engine have been Virtual Dyno's estimates for my car. I understand that virtual dyno isn't perfect, but even at 14 PSI on a 2560 I'm showing less than 200 HP.
I've read a handful of threads like this, but figured I'd make one to - hopefully - document the process as I attempt to figure out what's going on.
Things I'm planning to check/address as of now.
- Getting the actual weight of the car.
- I've borrowed some scales from a friend, so I should have an accurate weight of my car soon. I believe I have a rough idea of what the car weighs, but I haven't weighed it since I put the turbo kit on, rebuilt the engine, added the wing, etc...
- Check mechanical timing.
- It seems to be pretty common cause for issues like this. When I rebuilt the engine I know I checked the cam timing, and I believe I checked the timing again after about 50 miles, but will be redoing this to check my boxes.
- Spark tuning.
- I've road tuned the car myself and consulted others, but I want to revisit my spark table. I'm on 91 octane and do have a way to listen for and datalog my knock sensor. I have "tuned" this table before while listening for knock, but the numbers I've ended up at seem low. I also know someone with a similar timing map and car setup who was making 250+ on Cali 91.
- I have only tune this car, and only installed my ECU about 1.5 years ago. I am comfortable with TunerStudio and the ECU, but still have a way to go.
- If possible, I'll try to get some audio recordings of the Knock Detective as well.
Details about the car/build can be found in my build thread and signature, but I'll try to post most of the relevant stuff here.
BP4W, 9.0:1 pistons, VICS intake, Stock NB1 Throttle body
Very Old Flyin Miata Manifold and intercooler kit, 2.5" downpipe to full 3" catless Kraken exhaust, Genuine 2560R Turbo
91 Octane Fuel (Idaho if that matters), This is as good as is readily available around here.
I did compression test the engine recently, and the cylinders were 145, 145, 140, 152. This seems low to me, but the numbers are consistent enough. This was my first time building an engine. I have not done a leakdown test.
Spark Table as of today. Again, this seems fairly retarded to me. I was looking at Brain's basemap for a 97 (same compression ratio as my car) which was running ~8 degrees more timing all over. I don't know what octane that was targeted at. I have reduced this based on what I've heard from the knock sensor.
I will likely go out tonight and take a few datalogs of pulls. I will post virtual dyno results and datalogs when those are available.
If anyone has any thoughts, tips, or other things to check please let me know.
My current tune is attached and can be viewed online at Simba Tune
#2
Your 120 row is around 3 degrees retarded, your 140 row is about 5 degrees retarded, and it gets worse from there. Even 100kpa row is 2-4 degrees retarded.
I always end my 100kpa row at 28 degrees, and take 1 degree out for every 1 psi of boost, that's where I'm getting my numbers. At 14psi, 28-14=14 degrees, and you're at 10/11.
One pull isn't going to break anything, and if it does, it was going to break anyways. Case in point, I've done a few dyno pulls on a stock engine at 300hp. Do a datalog with your current table, then add 4 degrees, then do another pull. If your VD numbers don't increase, you've got another problem going on.
Your compression is ~20-30psi low for a stock engine, but that could be gauge-to-gauge differences. Is that some harbor freight gauge or a decent OTC or similar tool?
I always end my 100kpa row at 28 degrees, and take 1 degree out for every 1 psi of boost, that's where I'm getting my numbers. At 14psi, 28-14=14 degrees, and you're at 10/11.
One pull isn't going to break anything, and if it does, it was going to break anyways. Case in point, I've done a few dyno pulls on a stock engine at 300hp. Do a datalog with your current table, then add 4 degrees, then do another pull. If your VD numbers don't increase, you've got another problem going on.
Your compression is ~20-30psi low for a stock engine, but that could be gauge-to-gauge differences. Is that some harbor freight gauge or a decent OTC or similar tool?
#3
Your 120 row is around 3 degrees retarded, your 140 row is about 5 degrees retarded, and it gets worse from there. Even 100kpa row is 2-4 degrees retarded.
I always end my 100kpa row at 28 degrees, and take 1 degree out for every 1 psi of boost, that's where I'm getting my numbers. At 14psi, 28-14=14 degrees, and you're at 10/11.
One pull isn't going to break anything, and if it does, it was going to break anyways. Case in point, I've done a few dyno pulls on a stock engine at 300hp. Do a datalog with your current table, then add 4 degrees, then do another pull. If your VD numbers don't increase, you've got another problem going on.
Your compression is ~20-30psi low for a stock engine, but that could be gauge-to-gauge differences. Is that some harbor freight gauge or a decent OTC or similar tool?
I always end my 100kpa row at 28 degrees, and take 1 degree out for every 1 psi of boost, that's where I'm getting my numbers. At 14psi, 28-14=14 degrees, and you're at 10/11.
One pull isn't going to break anything, and if it does, it was going to break anyways. Case in point, I've done a few dyno pulls on a stock engine at 300hp. Do a datalog with your current table, then add 4 degrees, then do another pull. If your VD numbers don't increase, you've got another problem going on.
Your compression is ~20-30psi low for a stock engine, but that could be gauge-to-gauge differences. Is that some harbor freight gauge or a decent OTC or similar tool?
I'll try to get those pulls done tonight once it cools down a bit.
Purely theoretical, but lets say I increase timing and see a horsepower gain, and then add 2 more degrees and see another gain, is it safe to say that because the engine has continued to gain power that it is not knocking or that any knocking is very mild/light? I read another thread last night where someone said that the engine won't knock AND make more power with an increase in timing (assuming the timing increase is not something dramatic). That makes enough sense to me because I don't imagine the engine is going to run very well if you're combustion process isn't smooth/efficient.
I understand that the 91 octane will likely prevent the engine from making it's full potential. I guess what I'm asking is if knock would show a similar curve to MBT, where the peak of the curve would be where knock begins to occur, or if knock would occur before that point.
In the illustration below, I'm assuming the green line is what being knock limited looks like and the red line is what the power/timing relationship would look like with something like E85. Around the time power falls off, is that where more severe knock is occurring? Prior to that point I'm wondering if there would be light or no knock occurring? Maybe that's simplifying things too much.
I drew this crummy picture to try to illustrate what I'm theorizing
Regarding the compression, the brand is "Shankly" which looks to be generic Amazon. It was what we had in the garage. My only time using it has been on this engine, so I don't have any data points to compare to, but I'll ask my roommate what his experience with it has been.
I'm not sure if it matters for a compression test, but I am using lower compression forged pistons @ 9.0:1
I really appreciate your time. Don't feel the need to go into crazy detail if you don't want to.
#4
Also, after writing that last post I got the car on the scales. 1 Pound off of the settings I've been using in virtual dyno. I figured it would be about 50 pounds off.
2515 with me in the car. Somehow only 1 pound off from the settings I've been using in virtual dyno
Mechanical timing looked good to me. Feel free to double check me.
Extension is showing TDC on piston 1 and 19 teeth between the marks
2515 with me in the car. Somehow only 1 pound off from the settings I've been using in virtual dyno
Mechanical timing looked good to me. Feel free to double check me.
Extension is showing TDC on piston 1 and 19 teeth between the marks
#6
So the timing indeed seemed to help, but I'm still a long way from where I'd expect to be.
I did a baseline pull, added 4 degrees to the majority of the boosted area, then did another pull with another 2 degrees added in. I did smooth the map out a bit.
I am using a FWLR Tuning Knock Detective which has a headphone jack to work as electronic det cans. I didn't perceive knock on any of the runs, although I do believe I heard a couple pings at lower RPM after running the +6 setup with a bit more boost thrown in. That is not shown in the photo below.
I have also already corrected the lean area at the beginning of the pulls. My fuel map is pretty dialed in, but that's part of the map that I don't hit often.
These three pulls were taken on the same stretch of road, pretty flat, pretty smooth, not perfect though.
I drove around and did some more pulls. Everything sounded clean. Here is the table I ended up with. I plan to smooth this out and do some more pulls in the future to get it dialed in.
Regardless, 190 HP at 10-12 PSI still seems a bit low to me. I guess it's not totally out of the range though. At this point do I just turn the boost up to 15 PSI, fine tune the timing and see where it gets me?
I did a baseline pull, added 4 degrees to the majority of the boosted area, then did another pull with another 2 degrees added in. I did smooth the map out a bit.
I am using a FWLR Tuning Knock Detective which has a headphone jack to work as electronic det cans. I didn't perceive knock on any of the runs, although I do believe I heard a couple pings at lower RPM after running the +6 setup with a bit more boost thrown in. That is not shown in the photo below.
I have also already corrected the lean area at the beginning of the pulls. My fuel map is pretty dialed in, but that's part of the map that I don't hit often.
These three pulls were taken on the same stretch of road, pretty flat, pretty smooth, not perfect though.
I drove around and did some more pulls. Everything sounded clean. Here is the table I ended up with. I plan to smooth this out and do some more pulls in the future to get it dialed in.
Regardless, 190 HP at 10-12 PSI still seems a bit low to me. I guess it's not totally out of the range though. At this point do I just turn the boost up to 15 PSI, fine tune the timing and see where it gets me?
#7
Verify/adjust base spark timing with a timing light and check what the offset is set to in TS.
Also, boost is a measure of restriction, not power. You could run 14psi with a stock muffler and only make 120whp if you did it wrong enough. Different trim exhaust wheels and housings flow very differently, so a 2560 is not always comparable to a 2560. Same on the compressor side.
Also, boost is a measure of restriction, not power. You could run 14psi with a stock muffler and only make 120whp if you did it wrong enough. Different trim exhaust wheels and housings flow very differently, so a 2560 is not always comparable to a 2560. Same on the compressor side.
#8
The dyno graphs are indicating you're headed in the right direction. Double check your base timing as Six mentioned, but it looks like you gained power while boost lowered, which is a great start. You won't (shouldn't) knock before hitting MBT. Typically I keep increasing timing as long as gains are substantial enough to justify it. So far 4 degrees advance got you 20hp. That's a huge gain, one I would expect would continue. My 28@100 then -1/psi recommendation is conservative, but you're still ~2 degrees below that.
Can you attach your .ini here? I can't open your tune without it.
Can you attach your .ini here? I can't open your tune without it.
#9
Ted, I missed your post yesterday, but I will adjust VD for the elevation and temperature. I suspect that will make a small difference. I also have an APR wing and aftermarket front bumper so I'll see how much effect the coefficient of drag has.
Will do on the timing, and good info on the boost levels. Makes sense enough, but that's a new way of thinking about it for me. If anything, I'm suspicious of my intercooler/piping, but my VE table seems to increase pretty linearly once I'm in boost, so I would assume it's flowing OK. Does that seem logical?
Curly, you don't think the 91 Octane will be a limiting factor here? I'm curious because I've seen people rag on Californian 91 octane so heavily, but I haven't been able to find any real information on why Cali 91 would be any worse than Idaho 91.
I attached the latest tune (haven't driven with that map yet), the datalog from the +4 pull and the ini. It said a .ini file is not supported when I tried to upload it so I changed it to a .txt file (SgtSlowConfig.txt), but I assume you could open it in notepad and Save As to change it back to a .ini without issue.
Verify/adjust base spark timing with a timing light and check what the offset is set to in TS.
Also, boost is a measure of restriction, not power. You could run 14psi with a stock muffler and only make 120whp if you did it wrong enough. Different trim exhaust wheels and housings flow very differently, so a 2560 is not always comparable to a 2560. Same on the compressor side.
Also, boost is a measure of restriction, not power. You could run 14psi with a stock muffler and only make 120whp if you did it wrong enough. Different trim exhaust wheels and housings flow very differently, so a 2560 is not always comparable to a 2560. Same on the compressor side.
The dyno graphs are indicating you're headed in the right direction. Double check your base timing as Six mentioned, but it looks like you gained power while boost lowered, which is a great start. You won't (shouldn't) knock before hitting MBT. Typically I keep increasing timing as long as gains are substantial enough to justify it. So far 4 degrees advance got you 20hp. That's a huge gain, one I would expect would continue. My 28@100 then -1/psi recommendation is conservative, but you're still ~2 degrees below that.
Can you attach your .ini here? I can't open your tune without it.
Can you attach your .ini here? I can't open your tune without it.
I attached the latest tune (haven't driven with that map yet), the datalog from the +4 pull and the ini. It said a .ini file is not supported when I tried to upload it so I changed it to a .txt file (SgtSlowConfig.txt), but I assume you could open it in notepad and Save As to change it back to a .ini without issue.
#13
Similar to most things in this tune, I did dial in the VICS at one point, although I can't remember if that was before or after the engine build. Virtual Dyno actually helped me realized that I had the activation backwards early on (IE, I turned VICS off when I should have been turning it on).
I did 3 pulls with the tune you posted Curly. One on the road I was using yesterday and 2 on other roads that seemed pretty flat. All three showed very similar results. I didn't change the VD settings from the original ones I was using.
Still need to get the VE dialed in. I lost some of the changes I made when switching the tunes around.
I believe I was hearing some light/intermittent knock around 2500-3500 when accelerating through gears around town. I've heard that before which was part of what led me to pull so much timing. I need to get a different style of aux cable to record the knock sensor output to my phone. I'd be interested to hear what people think of the sounds. I've listened to knock recordings and I feel pretty confident that I'd be able to pick it out of a recording, but confirmation would be great.
I did 3 pulls with the tune you posted Curly. One on the road I was using yesterday and 2 on other roads that seemed pretty flat. All three showed very similar results. I didn't change the VD settings from the original ones I was using.
Still need to get the VE dialed in. I lost some of the changes I made when switching the tunes around.
I believe I was hearing some light/intermittent knock around 2500-3500 when accelerating through gears around town. I've heard that before which was part of what led me to pull so much timing. I need to get a different style of aux cable to record the knock sensor output to my phone. I'd be interested to hear what people think of the sounds. I've listened to knock recordings and I feel pretty confident that I'd be able to pick it out of a recording, but confirmation would be great.
#14
If I were trying to tune this car, with that data, I would be looking for a dyno. If you look at the RPMs where your max hp or tq is, they're changing drastically from run to run. I wouldn't be sure if it was run variation or my changes making that difference.
While doing 2-3 pulls without changes is good to verify consistency (which is one of the problems you're having), you also want to compare a change to a previous revision. Often I add 2 degrees across the entire timing map, but it only makes an improvement at a certain rev range. I keep that timing and remove it everywhere else.
Boise is only 6 hours away from my work's dyno...
While doing 2-3 pulls without changes is good to verify consistency (which is one of the problems you're having), you also want to compare a change to a previous revision. Often I add 2 degrees across the entire timing map, but it only makes an improvement at a certain rev range. I keep that timing and remove it everywhere else.
Boise is only 6 hours away from my work's dyno...
#15
To curly's point, VD can be very good, but it has to be used properly, which few know how to do.
outside of all the common knowledge stuff (flat ground, repeatable conditions...) probably the most important thing is maximizing the number of data records in a pull. This means using the highest gear you can, and setting the logging rate as high as possible. From my perspective, tunerstudio based logging just isn't good enough, but I suspect you don't have high speed logging capabilities and are stuck with it. You should absolutely be doing pulls in 3rd gear or higher if possible.
not to burst your bubble either, but the "Dynojet" view is a ficticious 9% adder. Also, "boost" is meaningless, put absolute on your graphs.
outside of all the common knowledge stuff (flat ground, repeatable conditions...) probably the most important thing is maximizing the number of data records in a pull. This means using the highest gear you can, and setting the logging rate as high as possible. From my perspective, tunerstudio based logging just isn't good enough, but I suspect you don't have high speed logging capabilities and are stuck with it. You should absolutely be doing pulls in 3rd gear or higher if possible.
not to burst your bubble either, but the "Dynojet" view is a ficticious 9% adder. Also, "boost" is meaningless, put absolute on your graphs.
#16
I agree. While I'm trying to be consistent with the data it doesn't seem like what I've collected is the best. I need to find a road/area that doesn't have houses nearby.
We do have a shop in town that rents a dyno, and I've been meaning to reach out to them because I'd like to get this done instead of fiddling around and guessing.
This seems like a big step in the right direction though.
This is the dyno graph of all the runs on the same road. Baseline, +4, +6 and with Curly's tune. I purposely left the smoothing low.
Also, I did check my timing with a light and it was very slightly off. I changed the trigger angle about 4 degrees. 207->211
We do have a shop in town that rents a dyno, and I've been meaning to reach out to them because I'd like to get this done instead of fiddling around and guessing.
This seems like a big step in the right direction though.
This is the dyno graph of all the runs on the same road. Baseline, +4, +6 and with Curly's tune. I purposely left the smoothing low.
Also, I did check my timing with a light and it was very slightly off. I changed the trigger angle about 4 degrees. 207->211
#19
Thanks for all the input/tips guys. I'll try to get some logs in higher gears soon, and I got a cable yesterday that should allow me to record the knock sensor output to my phone. Not sure how much that will help, but I was curious to see if it was possible.
The timing light we have is adjustable, but I wish it was static, it would give me a bit more confidence in the accuracy of the base timing. Regardless, looks pretty spot on now.
The timing light we have is adjustable, but I wish it was static, it would give me a bit more confidence in the accuracy of the base timing. Regardless, looks pretty spot on now.
#20
Alright, so check it out. Pretty happy with how this turned out. I got the right aux cable and ran the headphone output from the @FWLRtuning Knock Detective into my phone. I was able to use that as an "external mic" when recording.
I believe my boost controller might be contributing some noise to the audio because it's mounted on my valve cover, so I plan to try to isolate/relocate that.
I'd be curious to hear people's thoughts. To me, it sounds like there's slight knock at about 3k RPM. Example around 36 seconds into the recording.
I believe my boost controller might be contributing some noise to the audio because it's mounted on my valve cover, so I plan to try to isolate/relocate that.
I'd be curious to hear people's thoughts. To me, it sounds like there's slight knock at about 3k RPM. Example around 36 seconds into the recording.