Leaking CSF radiator!

Dumaurier7

Lieutenant
May 19, 2020
532
208
0
As part of my single turbo build I purchased an "upgraded" CSF full aluminum radiator in anticipation of cooling problems and to also allow for possible repositioning of the thermostat if required. As the build progressed I did eventually have to cut and re-weld the thermostat fitting in order to get it (the thermostat) and the huge 4" turbo inlet/suction pipe to fit in the cramped space. At the end of the build I decided to pressure test the cooling system to ensure the integrity of my work and avoid any issues down the road, to my surprise and dismay the only thing found leaking was the BRAND NEW fully welded "upgraded" radiator which never once saw coolant ! I noticed that the leak occurred at the tube sheet adjacent to repositioned thermostat nipple, could it be the breach was caused by heat of the weld? since its a relatively small leak (seeping coolant and slow pressure drop) I plan to purchase an aluminum soft solder repair kit to address this issue, wish me luck!
 

Asbjorn

Lieutenant
Mar 10, 2018
854
602
0
European, based in China
Ride
Z4 N54 DCT
Just use your oem radiator. The CSF does not offer any increase in cooling performance as the area is no larger than the oem radiator. If you want better cooling performance you need to add more cooler area, ie an aux cooler. Increasing radiator thickness wont help much.
 

martymil

Major General
Sep 6, 2017
3,331
1,906
0
Down Under
Ride
S65 1m
sorry I disagree, we saw much more stable temps for longer with the csf on our n54 and the temps dropped much quicker but we also ran the auxiliary
radiator along side of this which helps a lot.

it holds a lot more water so it has some benefit.
 
  • Like
Reactions: F87Source

Asbjorn

Lieutenant
Mar 10, 2018
854
602
0
European, based in China
Ride
Z4 N54 DCT
sorry I disagree, we saw much more stable temps for longer with the csf on our n54 and the temps dropped much quicker but we also ran the auxiliary
radiator along side of this which helps a lot.

it holds a lot more water so it has some benefit.
We agree an aux radiator helps alot.
If you do not mind, I would love to see data proving that the csf offers more cooling performance compared to the oem unit. More water is not the same as more cooling performance, because it just slows down how fast the temperature changes. I went all the way up to 8cm thickness on my car with a custom radiator, and all it did was delay how long it took for the temps to go through the roof. Going larger radiator area (rather than thickness) was a much better path according to my testing. You can see they did the same on the S55.
 

martymil

Major General
Sep 6, 2017
3,331
1,906
0
Down Under
Ride
S65 1m
We agree an aux radiator helps alot.
If you do not mind, I would love to see data proving that the csf offers more cooling performance compared to the oem unit. More water is not the same as more cooling performance, because it just slows down how fast the temperature changes. I went all the way up to 8cm thickness on my car with a custom radiator, and all it did was delay how long it took for the temps to go through the roof. Going larger radiator area (rather than thickness) was a much better path according to my testing. You can see they did the same on the S55.
You answered it yourself if more stable temps and delayed temp rise is not a benefit I don't know what is, along side a bigger 850 watt fan from a 1m or late model lci and its a winning combination.
 

F87Source

Sergeant
Oct 14, 2019
332
151
0
Canada
Ride
2017 Bmw M2
Just use your oem radiator. The CSF does not offer any increase in cooling performance as the area is no larger than the oem radiator. If you want better cooling performance you need to add more cooler area, ie an aux cooler. Increasing radiator thickness wont help much.
That's actually not correct, the CSF radiator has a higher fin density and on top of that more fin surface area due to their louvred design compared to the OEM radiator. So cooling performance is indeed better. The D088 is the same and there has been extensive track testing data on that and the CSF radiator on the m2 forum.

Also thickness does help in cooling because you are also adding more fins via additional layers that comes with the additional thickness along with more volume, which is why Tyspeed with his pikes peak m2 (probably the fastest track bmw built on the forums) was able to cool his car with the AC delete CSF radiator which is only adds additional thickness, compared to the thinner street spec radiator which was not sufficient.
 
  • Like
Reactions: martymil

F87Source

Sergeant
Oct 14, 2019
332
151
0
Canada
Ride
2017 Bmw M2
That's actually not correct, the CSF radiator has a higher fin density and on top of that more fin surface area due to their louvred design compared to the OEM radiator. So cooling performance is indeed better. The D088 is the same and there has been extensive track testing data on that and the CSF radiator on the m2 forum.

Also thickness does help in cooling because you are also adding more fins via additional layers that comes with the additional thickness along with more volume, which is why Tyspeed with his pikes peak m2 (probably the fastest track bmw built on the forums) was able to cool his car with the AC delete CSF radiator which is only adds additional thickness, compared to the thinner street spec radiator which was not sufficient.
I wanted to also note that although a thicker radiator will help with cooling it's not as efficient as more surface area because the aie gets hotter as it passes through the core so the fins at the back of the core won't be able to exchange as much heat. However with the thickness we see on these csf radiators even the race spec ac delete versions it isn't thick enough to prevent effective hest exchange and we still see a performance gain.
 

Dumaurier7

Lieutenant
May 19, 2020
532
208
0
Just use your oem radiator. The CSF does not offer any increase in cooling performance as the area is no larger than the oem radiator. If you want better cooling performance you need to add more cooler area, ie an aux cooler. Increasing radiator thickness wont help much.
As stated at the beginning of the thread, the OEM cannot work anymore because the thermostat mounting point had to be changed, I don't know of any way to weld the plastic used for the end tanks, thanks anyway for your response. I'm surprised to see so many responses to this thread! lots of good information!
 

Jeffman

Major
Jan 7, 2017
1,618
1
628
0
Don’t mean to thread jack, but will a thicker CSF radiator cause the fan to push aftermarket inlets into the belt on an N54?
 

Asbjorn

Lieutenant
Mar 10, 2018
854
602
0
European, based in China
Ride
Z4 N54 DCT
That's actually not correct, the CSF radiator has a higher fin density and on top of that more fin surface area due to their louvred design compared to the OEM radiator. So cooling performance is indeed better. The D088 is the same and there has been extensive track testing data on that and the CSF radiator on the m2 forum.

Also thickness does help in cooling because you are also adding more fins via additional layers that comes with the additional thickness along with more volume, which is why Tyspeed with his pikes peak m2 (probably the fastest track bmw built on the forums) was able to cool his car with the AC delete CSF radiator which is only adds additional thickness, compared to the thinner street spec radiator which was not sufficient.

Alright alright, so lets say that translates to a 5% larger radiator area? In other words it will delay the time it takes to hit limp mode by a few corners on a hot day on track. Compare that to what is actually needed, which is more like 50-60% additional radiator area depending on the weather, power and track. This based on what BMW did with the S55. My only point is that the performance to cost ratio is super low with that CSF radiator, and it is not the first time we hear about it leaking.

And btw, isnt Tyspeed sponsored by CSF? Do you have data that his car is no longer overheating? I know he did a lot of cool aero stuff, but when I checked around two years ago his CSF + 2x s55 aux cooler setup was still hitting above 110C on track.
 
  • Agree
Reactions: Torgus

F87Source

Sergeant
Oct 14, 2019
332
151
0
Canada
Ride
2017 Bmw M2
Alright alright, so lets say that translates to a 5% larger radiator area? In other words it will delay the time it takes to hit limp mode by a few corners on a hot day on track. Compare that to what is actually needed, which is more like 50-60% additional radiator area depending on the weather, power and track. This based on what BMW did with the S55. My only point is that the performance to cost ratio is super low with that CSF radiator, and it is not the first time we hear about it leaking.

And btw, isnt Tyspeed sponsored by CSF? Do you have data that his car is no longer overheating? I know he did a lot of cool aero stuff, but when I checked around two years ago his CSF + 2x s55 aux cooler setup was still hitting above 110C on track.
It is more than 5% because every fin on the radiator has been louvered and fin density has been increased, so that + the fact that the radiator has fins all the way through its thickness surface area increase is likely in the ~20% likely greater range. Also any surface area or heat exchange area increase helps with total heat transfer effectively meaning peak temperatures will be lower at every situation, the increased volume is what "delays" limp mode for a few extra corners.


Yes tyspeed is sponsored by CSF leading to the development of their race spec radiator (a thicker radiator) which was the only thing capable of solving his over heating issue. Yes temperatures are still near limp mode, but keep in mind he is pushing 500 whp. The data is from his track videos from his youtube and instagram, he used to have alot more data before deleting alot of his videos. With the street spec it would be at limp mode temperatures of over heating, compared to the stock radiator which he used before hand there was an improvement.


But the final iteration to push temperatures even further below limp mode was water sprayers and that was so effective it helped even in super high elevations of pikes peak. Overall yes the CSF radiator helped but was still near limp mode until the water sprayers.



Now non tyspeed m2's:

1) Paradoxicalm2 (the owner of radenso radar - a super brilliant guy): His car was used to develop the CSF cooling suite, literally everything was based off of his car. His data showed on a modded m2 it would hit limp mode on a hot summer day, after the csf radiator the coolant temperatures would now be under control. And this is the street spec radiator (the race spec was designed off of tyspeed's car). he has since sold his m2 so it is no more.

2) Zm2 the current benchmark for a fast driver who is reaching the thermal limits of his m2. With his CSF radiator and new tuning techniques he is able to push 450 whp and 500 wtq for endless laps at a maximum temperature of 250F which is limp mode. He is still working on it and we believe with the new high ethanol content (keeps cylinder temps cool) tune and intercooler iats will be low enough to keep egts under controlled to the point where he will have coolant temperatures under limp mode for the entirety of the track session. So yes his CSF radiator seems to be working to keep coolant temps under control as well. Note before hand even with a dinan stage 4 tune he was over heating the stock radiator, and dinan stage 4 is only 375 whp.


Note the m2 and all F series n55's has an oil to coolant heat exchanger which means the coolant temperatures are sacrificed to cool oil temps, so this version of the n55 will always struggle with coolant temperatures while this heat exchanger remains.



That's what I see so far, and it shows the CSF radiator does work. In terms of leaks I have never seen any one on the m2 platform report that.
 

Dumaurier7

Lieutenant
May 19, 2020
532
208
0
My F22 which is a European (RHD) version does not have the oil to coolant heat exchanger "my" version was not meant for hot climates as stipulated on Real OEM, in order to supplement this I am in the process of installing a second oil cooler. Experience gained from my years of owning an RX7 has thought me that oil cooling is an important part of engine function and directly impacts overall coolant temperatures since the oil cools as well as lubricates.
 

Asbjorn

Lieutenant
Mar 10, 2018
854
602
0
European, based in China
Ride
Z4 N54 DCT
It is more than 5% because every fin on the radiator has been louvered and fin density has been increased, so that + the fact that the radiator has fins all the way through its thickness surface area increase is likely in the ~20% likely greater range. Also any surface area or heat exchange area increase helps with total heat transfer effectively meaning peak temperatures will be lower at every situation, the increased volume is what "delays" limp mode for a few extra corners.


Yes tyspeed is sponsored by CSF leading to the development of their race spec radiator (a thicker radiator) which was the only thing capable of solving his over heating issue. Yes temperatures are still near limp mode, but keep in mind he is pushing 500 whp. The data is from his track videos from his youtube and instagram, he used to have alot more data before deleting alot of his videos. With the street spec it would be at limp mode temperatures of over heating, compared to the stock radiator which he used before hand there was an improvement.


But the final iteration to push temperatures even further below limp mode was water sprayers and that was so effective it helped even in super high elevations of pikes peak. Overall yes the CSF radiator helped but was still near limp mode until the water sprayers.



Now non tyspeed m2's:

1) Paradoxicalm2 (the owner of radenso radar - a super brilliant guy): His car was used to develop the CSF cooling suite, literally everything was based off of his car. His data showed on a modded m2 it would hit limp mode on a hot summer day, after the csf radiator the coolant temperatures would now be under control. And this is the street spec radiator (the race spec was designed off of tyspeed's car). he has since sold his m2 so it is no more.

2) Zm2 the current benchmark for a fast driver who is reaching the thermal limits of his m2. With his CSF radiator and new tuning techniques he is able to push 450 whp and 500 wtq for endless laps at a maximum temperature of 250F which is limp mode. He is still working on it and we believe with the new high ethanol content (keeps cylinder temps cool) tune and intercooler iats will be low enough to keep egts under controlled to the point where he will have coolant temperatures under limp mode for the entirety of the track session. So yes his CSF radiator seems to be working to keep coolant temps under control as well. Note before hand even with a dinan stage 4 tune he was over heating the stock radiator, and dinan stage 4 is only 375 whp.


Note the m2 and all F series n55's has an oil to coolant heat exchanger which means the coolant temperatures are sacrificed to cool oil temps, so this version of the n55 will always struggle with coolant temperatures while this heat exchanger remains.



That's what I see so far, and it shows the CSF radiator does work. In terms of leaks I have never seen any one on the m2 platform report that.

Do you happen to know the nozzle size he uses for external water spray?
 

F87Source

Sergeant
Oct 14, 2019
332
151
0
Canada
Ride
2017 Bmw M2
My F22 which is a European (RHD) version does not have the oil to coolant heat exchanger "my" version was not meant for hot climates as stipulated on Real OEM, in order to supplement this I am in the process of installing a second oil cooler. Experience gained from my years of owning an RX7 has thought me that oil cooling is an important part of engine function and directly impacts overall coolant temperatures since the oil cools as well as lubricates.
Interesting because my Canadian m2 and m235i both had that heat exchanger.
 

Asbjorn

Lieutenant
Mar 10, 2018
854
602
0
European, based in China
Ride
Z4 N54 DCT
No clue but he sprayed methanol too which is a huge cooling boost due to evaporation but it's also a fire risk.

Are you saying he sprayed with methanol outside the coolers? Holy moly...

I tested with 2 x 60cc water spray in front of my radiator(s) and it didn't help at all. I know people say I installed the nozzles in the wrong locations + directions etc, but I have a feeling the nozzle size is the real issue. However running 2x60cc/min x 25min track sessions is already 3L of water. So for me it is not super feasible to use larger nozzles. For hill climbing you probably only need to spray for 12-13mins so you can probably go up in nozzle size for that application. Anyhow it is super easy to turn the system on and off and make A-B testing on track, so anyone who claims it works should be able to show us the coolant temp data before and after. You could even do a 30min track session and turn it on during the last 15mins while logging the entire time. Done.

Its a bit more difficult to do with the OEM cs CSF radiator as the weather etc will have changed between swapping them. So whether it helps 5% or 20% we will never know haha. Fins and thickness offers diminishing returns, and isn't enough no matter what. The core problem with the N54/N55 is the lack of radiator area, that I think we can all agree. At least BMW agrees...
 

F87Source

Sergeant
Oct 14, 2019
332
151
0
Canada
Ride
2017 Bmw M2
Are you saying he sprayed with methanol outside the coolers? Holy moly...

I tested with 2 x 60cc water spray in front of my radiator(s) and it didn't help at all. I know people say I installed the nozzles in the wrong locations + directions etc, but I have a feeling the nozzle size is the real issue. However running 2x60cc/min x 25min track sessions is already 3L of water. So for me it is not super feasible to use larger nozzles. For hill climbing you probably only need to spray for 12-13mins so you can probably go up in nozzle size for that application. Anyhow it is super easy to turn the system on and off and make A-B testing on track, so anyone who claims it works should be able to show us the coolant temp data before and after. You could even do a 30min track session and turn it on during the last 15mins while logging the entire time. Done.

Its a bit more difficult to do with the OEM cs CSF radiator as the weather etc will have changed between swapping them. So whether it helps 5% or 20% we will never know haha. Fins and thickness offers diminishing returns, and isn't enough no matter what. The core problem with the N54/N55 is the lack of radiator area, that I think we can all agree. At least BMW agrees...
I believe so because he only had one reservoir for the water sprayers and his meth injection so it must be the same fluid, but I could be wrong. But methanol mix will help evaporative cooling because alcohol evaporates faster and easier than water. If the mix is 50:50 it's not too bad to be spraying outside, because extreme winter windshield washer fluid is 50% meth as well.


Are you spraying the radiator directly or are you spraying the ac condenser in front of the radiator? Because there's only a tiny gap between the ac condenser and radiator for you to stick in the water jet.


Well so far the csf radiator has been shown to work, so their increase fin density and thickness is working. However it will be less effective than more surface area like with a larger radiator. It does frustrate me bmw is so stupid and they valued tucking in the intercooler under the radiator (and thus sacrificing radiator surface area) over radiator size. Then they crippled intercooler performance by throwing in this dinky POS that heat soaks even stock. They could've just made the intercooler a thinner but larger one that sits in front of the radiator (like VW) and thus allowing for a full sized radiator. But nope another dumb bmw idea.