A lot of reality in that comment. I would have handled the same exact way. Please report back once the block has some hours operating with your repair.
Tig welding 4047 on block
- Thread starter shushikiary
- Start date
Will do, I'm working with Jake to get the tune going. I've done my first 3rd gear pull at waste gate spring pressure (14 psi). I already have 500 miles on the block with the fix now, plus many hours idling making sure it wasn't spewing oil on me. I went through half a gas tank just idling.
Calculated MAF at redline was 335 g/s, which is about 450 theoretical hp using a MAF calculator.
The high lift cams cause some transient issues currently, low RPM high load (under 2.5k rpm) it can be a little picky, and on throttle lean in it stutters for half a second and then it's fine. Jake says he can tune that out with vanos but I'm fine leaving it if the vanos changes would hurt spool.
Going to target 24.5 psi for a theoretical 700whp. I'm going for a reliable 100k miles at 700whp.
Calculated MAF at redline was 335 g/s, which is about 450 theoretical hp using a MAF calculator.
The high lift cams cause some transient issues currently, low RPM high load (under 2.5k rpm) it can be a little picky, and on throttle lean in it stutters for half a second and then it's fine. Jake says he can tune that out with vanos but I'm fine leaving it if the vanos changes would hurt spool.
Going to target 24.5 psi for a theoretical 700whp. I'm going for a reliable 100k miles at 700whp.
Increase timing, increase separation angle on the exhaust, richen fuel.
Yea the issue seems to be lean low load ignition... which E85 is not friendly to. I bet running a larger spark gap might help but then hurt you high load.
So everything you said would help.
So everything you said would help.
Well tried to do my first run at 19 PSI today but something is up with the reflex. Looking at the tables I suspect the default setup for the EBC is having issues. Waste gate only worked fine, but it even set at a target pressure of 19 it basically held the waste gate shut and I had a large throttle closure from 3500-6krpm which lead to I think a timeout because my STFT's went to 0 and she ran super lean. I felt that happen and took my foot out of it immediately. The backend flash was set to 25 PSI, so I ended up running 25 PSI with the massive throttle closure instead, lol.
Everything is fine, just need to get the reflex EBC setup figured out... lots of questions there. Like is the feed forward a % value or is it just the PID error, or is it the straight up MAP pressure. Is the PID output inverted given they are running an open collector setup to control the valve (reflex sources the ground, aka 0 duty cycle is no ground applied it seems, as the comment in the xdf for the aux outputs seems to hint that is "high", but 100% is "low").
So just have to get the reflex EBC figured out.
Oh I should add Jake changed the vanos and a few other things which really cleaned up the low RPM response and the throttle lean in response.
I should also add that I talked to a welder in the area who has tried to weld several Alusil blocks. He has had no luck, attempting to TIG weld the block resulted in cracks in the HAZ after the weld cooled, just as I've read other people say happens because the silicone in the block concentrates on the outside of the weld puddle making it very brittle. They did not try a 4047 filler rod though, so I suggested it to them and they said they'd try it next time someone brings an Alusil block in....
Aka my non-willingness to just take it back to them and have it weld it is justified.
Everything is fine, just need to get the reflex EBC setup figured out... lots of questions there. Like is the feed forward a % value or is it just the PID error, or is it the straight up MAP pressure. Is the PID output inverted given they are running an open collector setup to control the valve (reflex sources the ground, aka 0 duty cycle is no ground applied it seems, as the comment in the xdf for the aux outputs seems to hint that is "high", but 100% is "low").
So just have to get the reflex EBC figured out.
Oh I should add Jake changed the vanos and a few other things which really cleaned up the low RPM response and the throttle lean in response.
I should also add that I talked to a welder in the area who has tried to weld several Alusil blocks. He has had no luck, attempting to TIG weld the block resulted in cracks in the HAZ after the weld cooled, just as I've read other people say happens because the silicone in the block concentrates on the outside of the weld puddle making it very brittle. They did not try a 4047 filler rod though, so I suggested it to them and they said they'd try it next time someone brings an Alusil block in....
Aka my non-willingness to just take it back to them and have it weld it is justified.
Last edited:
Played a little with the Reflex EBC. Don't have answers to all my questions, but halving the feed forward table and upping the min PID value to -20 and the max to +20 got me 16.5 PSI. I'm just playing with it my self because I always get frustrated with how fast tuners respond... doesn't matter who the tuner is, they pretty much all respond pretty slow because they are all just so back logged I think so I've kind of learned that's just the name of the game.
I'd love to have a tuner give me just a starter tune to save me time and then tune it all my self, which this kind of is at that point here in regards to the EBC and PI as I can now adjust those my self with the Reflex (like I could with the AIC 6 for PI).
It's just nice to know how things are actually calculated because then I can get pretty close first or second try instead of having to play with it 20 revs to get it right.... so without knowing that at this point, I'm having to just play with it.
Once I get the feed forward dialed in we'll see if I have to play with any of the PID constants. It would be really nice if the CAN integration for the N54 was done so the PI would shut off if the engine decides to stop fueling, but it is what it is until that is working.
Doing more reading on the welding, hobart makes a proprietary 4943 filler rod that might even be better than 4047, but for welding A356 (a lower % silicone than Alusil, but is used in many transmissions and other engine blocks today) for sure works better using 4047 than the normal 4043 welding rod used for a lot of aluminum. The added silicone in the 4047 rod reduces porosity, helps prevent contaminant pick up, and helps fill better than a 4043 rod. 4047 is typically used in thin metal welds where no leaking can be tolerated after the weld (aka the lower porosity). No idea when I'll hear back from the welder on their attempt with 4047, but my block won't be the guinea pig.
I'd love to have a tuner give me just a starter tune to save me time and then tune it all my self, which this kind of is at that point here in regards to the EBC and PI as I can now adjust those my self with the Reflex (like I could with the AIC 6 for PI).
It's just nice to know how things are actually calculated because then I can get pretty close first or second try instead of having to play with it 20 revs to get it right.... so without knowing that at this point, I'm having to just play with it.
Once I get the feed forward dialed in we'll see if I have to play with any of the PID constants. It would be really nice if the CAN integration for the N54 was done so the PI would shut off if the engine decides to stop fueling, but it is what it is until that is working.
Doing more reading on the welding, hobart makes a proprietary 4943 filler rod that might even be better than 4047, but for welding A356 (a lower % silicone than Alusil, but is used in many transmissions and other engine blocks today) for sure works better using 4047 than the normal 4043 welding rod used for a lot of aluminum. The added silicone in the 4047 rod reduces porosity, helps prevent contaminant pick up, and helps fill better than a 4043 rod. 4047 is typically used in thin metal welds where no leaking can be tolerated after the weld (aka the lower porosity). No idea when I'll hear back from the welder on their attempt with 4047, but my block won't be the guinea pig.
Materials and process conflate to a high skill learned through significant repetition. It's not a "welding job" with a new rod material. IMHO ... and to your point on not being a guinea pig ... if it was me I would not let anyone weld that block that has not welded that kind of block a number of times.
It's like finish work. Never go to a guy that is a really good painter, who shoots Dupont, and ask him to shoot your car with Spies Hecker. The process and materials are tightly interwoven in the experience expertise.
It's like finish work. Never go to a guy that is a really good painter, who shoots Dupont, and ask him to shoot your car with Spies Hecker. The process and materials are tightly interwoven in the experience expertise.
Ugh, I really need to know more about the Reflex.
The value for the N20 sensor is "patched" but the sensor voltage to absolute pressure is wrong for an N20.
This makes sense as then my PI values are making sense, but that then also makes me thinking my boost target is going to be silly. Which then also makes me realize that the Reflex does not know ambient pressure unless it has a pressure sensor inside the box.
Given the PID target boost pressure table says "boost" pressure, is it assuming a 14.7 PSI ambient pressure? I assume it's also affected by the pressure sensor reference voltage being wrong.
If the pressure sensor reference voltage is right, then the PI table is set to give me 100% PI duty cycle at only 18 psi of boost pressure (at this altitude)... which doesn't seem right as that's A LOT of PI.
So I suspect that the reverence voltage table is being used, which means that 17.7 PSI of boost at my altitude means it reports as 16 PSI absolute in the PI table, which would put my PI injection at around 45% at that boost and from around 5000-7000 RPM. Which makes WAY more sense.
This however means my boost target of 19 PSI.... is not at all 19 PSI, as 19 PSI boost (assuming again 14.7 ambient) in the table would put me at an INSANE amount of actual boost, it will target 33.7 PSI of absolute pressure, but the table as 4.85 volts as 38.24 absolute. Thus that means it's going to target a MAP voltage of 4.4197 volts, which would be 46.94 absolute, which is 34.7 PSI of boost at my altitude.
Fun times. Really need to get this whole Reflex setup figured out. Jake seems busy and often ignores my questions when he responds, so if someone knows the answer to any of this I'd love to know.
The value for the N20 sensor is "patched" but the sensor voltage to absolute pressure is wrong for an N20.
This makes sense as then my PI values are making sense, but that then also makes me thinking my boost target is going to be silly. Which then also makes me realize that the Reflex does not know ambient pressure unless it has a pressure sensor inside the box.
Given the PID target boost pressure table says "boost" pressure, is it assuming a 14.7 PSI ambient pressure? I assume it's also affected by the pressure sensor reference voltage being wrong.
If the pressure sensor reference voltage is right, then the PI table is set to give me 100% PI duty cycle at only 18 psi of boost pressure (at this altitude)... which doesn't seem right as that's A LOT of PI.
So I suspect that the reverence voltage table is being used, which means that 17.7 PSI of boost at my altitude means it reports as 16 PSI absolute in the PI table, which would put my PI injection at around 45% at that boost and from around 5000-7000 RPM. Which makes WAY more sense.
This however means my boost target of 19 PSI.... is not at all 19 PSI, as 19 PSI boost (assuming again 14.7 ambient) in the table would put me at an INSANE amount of actual boost, it will target 33.7 PSI of absolute pressure, but the table as 4.85 volts as 38.24 absolute. Thus that means it's going to target a MAP voltage of 4.4197 volts, which would be 46.94 absolute, which is 34.7 PSI of boost at my altitude.
Fun times. Really need to get this whole Reflex setup figured out. Jake seems busy and often ignores my questions when he responds, so if someone knows the answer to any of this I'd love to know.
Built my self a spread sheet to convert between the N20, stock, and the scaling given in the reflex.
Low and behold the answer popped out to me.
The scaling in the reflex (volt, psi) (1.21, 0), (2.18,10), (4.85,38.24), makes it so that the voltage from the N20 TMAP gives you a relative boost PSI reported to the reflex (at sea level).
That would make tuning pretty easy at sea level as it will target an absolute pressure to give you that boost at sea level.
Little bit cheeky that is.
So I just have to adjust the values based on boost pressure at this altitude and all the tables in the reflex that reference "boost" or "MAP" are actually referencing relative boost pressure at sea level due to the scaling....
Low and behold the answer popped out to me.
The scaling in the reflex (volt, psi) (1.21, 0), (2.18,10), (4.85,38.24), makes it so that the voltage from the N20 TMAP gives you a relative boost PSI reported to the reflex (at sea level).
That would make tuning pretty easy at sea level as it will target an absolute pressure to give you that boost at sea level.
Little bit cheeky that is.
So I just have to adjust the values based on boost pressure at this altitude and all the tables in the reflex that reference "boost" or "MAP" are actually referencing relative boost pressure at sea level due to the scaling....
Update:
Hit 26 PSI by 4k RPM today, tapered to 23.8 at redline, thing hauls arse. I have a turbo speed sensor hooked up, she's hitting 140k RPM. I'd hate to see what someone pushing 30 PSI on this turbo would be pulling for RPM. So for the sake of the turbo's life I won't push it any further.
I keep getting super knocking codes that are very transient. Sometimes they show up, sometimes they don't, usually cylinders 1 and 5, but sometimes also 2.
I have not been running desensitized knock tables or any pot's. Tried the desensitized knock tables and they did nothing. Checked the spark plugs and did a bore scope, no hints of true knocking.
So I ordered some 100k ohm pots and will be putting them on the knock sensors to desensitize them like Terry (BMS) has done in the past on his built motors.
Once the false knock is figured out things should get even faster.
The numbers are 3:1 pressure ratio ( likely a little higher due to intake vacuum), which on the GTX3582r Gen II compressor map (1.06 turbine AR) suggests I'm flowing over 80 lbs/min of low, which would put me right around 700 wheel hp and a little over 800 brake hp.
Hit 26 PSI by 4k RPM today, tapered to 23.8 at redline, thing hauls arse. I have a turbo speed sensor hooked up, she's hitting 140k RPM. I'd hate to see what someone pushing 30 PSI on this turbo would be pulling for RPM. So for the sake of the turbo's life I won't push it any further.
I keep getting super knocking codes that are very transient. Sometimes they show up, sometimes they don't, usually cylinders 1 and 5, but sometimes also 2.
I have not been running desensitized knock tables or any pot's. Tried the desensitized knock tables and they did nothing. Checked the spark plugs and did a bore scope, no hints of true knocking.
So I ordered some 100k ohm pots and will be putting them on the knock sensors to desensitize them like Terry (BMS) has done in the past on his built motors.
Once the false knock is figured out things should get even faster.
The numbers are 3:1 pressure ratio ( likely a little higher due to intake vacuum), which on the GTX3582r Gen II compressor map (1.06 turbine AR) suggests I'm flowing over 80 lbs/min of low, which would put me right around 700 wheel hp and a little over 800 brake hp.
Last edited:
Sounds like you're making some good progress! Wow, that's spinning pretty good. Maybe I missed it, did you say what turbo you're running?
You don't hear or feel the knock, what you feel is the ECU shutting that cylinder down, so now you're on 5 cylinders for half a second until the ECU turns the cylinder back on.
It only ever happens at high boost high RPM, never before 6k rpm. Car feels great then you feel the cylinder shut off, no sound out the exhaust, no change in vibration, nothing.
Note I have solid motor mounts and I made my own solid transfer case mount, so you KNOW when it even pulls timing as you can literally hear it.
It only ever happens at high boost high RPM, never before 6k rpm. Car feels great then you feel the cylinder shut off, no sound out the exhaust, no change in vibration, nothing.
Note I have solid motor mounts and I made my own solid transfer case mount, so you KNOW when it even pulls timing as you can literally hear it.
Interesting. I have had superknock codes. Everytime it felt like ignition break up or about to misfire. Logs showed a small lean spike similar to when you get a misfire.
Cylinders looked fine and plugs also showed no signs of knock.
I did some digging and found very little. I did find someone that also believed it was a misfire and not actual knock on his car.
The logs show no spike in the AFR before the event. AFR is great every time, then the moment you get the super knock event reported by the ECU the STFT's go to 0 as you enter open loop, it shuts the cylinder that it thinks super knocked down and so that bank gets a lean AFR as fuel injection is turned off, then after maybe a quarter or half a second it will go back into closed loop and re-enable the cylinder which then brings the AFR back in line.
Show all this to Jake he didn't seem worried at all and just immediately said in the past that they have had to put POT's on closed deck motors and even motors with just forged pistons. Given Jake has dealt with a lot of built motors, if he's not worried, I'm not worried. He said that if the plugs didn't show anything then I'm very likely fine.
I've had the super knocking codes before the built motor a LONG time ago, and I found it was due to a weak coil and it would exhibit exactly what you are describing where the AFR would spike at the event. After going to the precision raceworks ignition coil setup I've never had issues with miss fires ever again ( note I change spark plugs every 25k miles and only use NDK 2 step colder iridiums gapped to 22 thousands).
Show all this to Jake he didn't seem worried at all and just immediately said in the past that they have had to put POT's on closed deck motors and even motors with just forged pistons. Given Jake has dealt with a lot of built motors, if he's not worried, I'm not worried. He said that if the plugs didn't show anything then I'm very likely fine.
I've had the super knocking codes before the built motor a LONG time ago, and I found it was due to a weak coil and it would exhibit exactly what you are describing where the AFR would spike at the event. After going to the precision raceworks ignition coil setup I've never had issues with miss fires ever again ( note I change spark plugs every 25k miles and only use NDK 2 step colder iridiums gapped to 22 thousands).
Interesting. Im stock motor but have been through this same thing.
No limp mode, it happens, I lift, it smooths out, I keep driving.
Haha, you quoted me before I finished the edit.
I described having a similar issue above in the edit and how I fixed it.
I described having a similar issue above in the edit and how I fixed it.
Mine went away after I swapped out an index 12. Coincidence ? Or maybe a bad injector. Weird for sure. I also use 97506s at .019. Either way, DME is strange that MANY have superknock codes, but plugs and pistons show no signs of knock.Im also Full E85 so knock is unlikely for a mild tune.
