Technical Turbo design discussion and opinions

MMP

Private
Nov 10, 2016
48
34
0
Houston, Texas
There are only so many ways to set up a shorty 3 into 1 manifold whilst maintaining stock location you know, but the devil is in the details... The MMP manifold starts off @ 32mm ID, tapers up to 36-39mm ID or thereabouts shortly afterwards, and then necks back down to to the equivalent of sub 29mm as it goes into the hotside, hardly ideal IMHO. i do agree however, that a high-performance TT setup is absolutely the way to go on our cars, which is why I went that route to begin with


you are right the devil is in the details and you are also right Twin Turbo is the way to go. We have owned many N54s twin turbo and also single turbo and also both types of turbos at the same time, and hands down the twin turbos are way funner to drive than the single turbos and with all the twin turbo options including 53mm ball bearing twin turbos we are starting to make now, any whp goal can be achieved with MMP twin turbos. Also we like generally what you are doing with the turbos although as you can see below we would make several key improvements like we have in our design but imitation is the best form of flattery and I must say your turbos are a striking resemblance to the MMP 1K turbos.


As far as manifold details….The MMP manifold perfectly matches the N54 head exhaust runner ID at 33mm and is a smooth transition to 45.1mm then merges to a 50.1mm collector. This smooth merge is only accomplished with a casting and we use the best investment casting process and highest-grade stainless-steel materials for our casting. Below you can see a pic of the cross section of our casting and how it matches the runners and merges for the best aerodynamic flow possible and least pressure losses. MMP manifolds combined with our turbine housings truly are the highest flow of any twin turbo setup for the N54 including the ones you present here.

manifold cross.JPG



I think you use a machined flange with machined insert welded in and cut and bent 1.25" sch 10 tubes welded together, correct me if I’m off please. You also would have to match the 33mm ID port size of the N54 with the counterbore for the insert, not sure what the N53 is but it is probably similar maybe you can measure it in the head just past the counterbore for the exhaust port insert and advise the size. the 1.25" sch 10 tube you use is 36mm ID. also from your flange to the ID of your pipe there is not a smooth transition as you can design with a casting, it is most likely a step change in diameter due to fabrication limitations that step change in diameter is very bad for flow as it causes flow separation and pressure waves and pressure losses add up there. My 45.1mm pipe ID size in my casting is 57% greater flow area than your 36mm ID pipe. also your collector is visibly smaller than mine. I would guess around 40mm ID but maybe you can clarify the size. If it is 40mm ID my 50.1mm ID is also 57% more flow area than your merged collector which again is less pressure losses, more flow, and more power from the manifold. I took a stab at modeling the cross section of your manifold based on the specs you provided to show what I think are the differences and posted the pic of the cross section below, but please let me know if I need to correct something for comparison purposes.

hydra manifold.JPG


The manifolds I designed I did as the highest flow area diameter possible based on the bolt pattern restrictions. If I could have done larger I would so I believe they are the largest possible with a bespoke casting design and using investment cast methods and not sand casting is absolutely required for the high end surface finish and precision needed since pushing to the largest size pipe not to interfere with bolt pattern so even .5mm tolerance matters for the casting otherwise will run into problems quick which is why our 1K turbos uses these investment cast stainless steel manifolds with our 1K turbos.

Also your cast turbine housing you mentioned you use 6.6cm^2 for the AR area, our turbine housings have a 7cm^2 AR area so again ours flow more and are optimized through CFD for our high flow manifolds, high flow collector size and large 4646 GT28 high flow 9 blade turbine wheel TD04 turbos with our high flow native TD04 compressor housings.


You are right the devil is in the details and this is the most details I have shared about our design anywhere but since you posted some info on our design as if you knew all the details I just wanted to make sure you had all the right information for comparison purposes.



Also nice work on the custom N53 head mating to the N54. I look forward to the flow bench numbers. Was the n53 head ported or is it as stock? The N54 head work we did over 2 years ago and posted about in another forum showed the N54 head did flow terribly and could be improved by over 40% at .5” lift with a porting job on the head as we did it. We saw flow increased from 167CFM at .5” lift 28inhg pressure drop to 235CFM which we were very happy with those results using the stock N54 valves and of course non of the headaches to solve when fitting a non standard n53 head to an N54 bottom end but we like to see the effort and new things being tried. Once we saw the great results 2+ years ago from the N54 head porting we started offering them for sale at $2000 which was a huge decrease from other options on the market at the time offering a ported head for $4000. Since then others now offer it and the market has come down more in line with reality.



We look forward to seeing more results from your testing, good luck.
 

fmorelli

Major General
Staff member
Aug 11, 2017
3,176
2,902
0
54
Virginia
www.morelliguitars.com
Ride
E89 Z4 35i, F10 535d
[...snip...] As far as manifold details….The MMP manifold perfectly matches the N54 head exhaust runner ID at 33mm and is a smooth transition to 45.1mm then merges to a 50.1mm collector. [...snip...]
I think it is awesome that you are sharing so much information from your experience and design work at @MMP. And as you noted, more than you ever have. I think it would be great if you created a separate thread on MMP design. I'm sure this is interesting to many of us. Your work certainly merits its own thread, rather than being buried in this thread of someone getting underway with their product development. I'm sure you would agree.

Filippo
 

[email protected]

Lieutenant
Dec 7, 2016
626
397
0
St. Louis, MO USA
www.rbturbo.com
Ride
'08 335i, '14 M6, '15 Tundra
you are right the devil is in the details and you are also right Twin Turbo is the way to go. We have owned many N54s twin turbo and also single turbo and also both types of turbos at the same time, and hands down the twin turbos are way funner to drive than the single turbos and with all the twin turbo options including 53mm ball bearing twin turbos we are starting to make now, any whp goal can be achieved with MMP twin turbos. Also we like generally what you are doing with the turbos although as you can see below we would make several key improvements like we have in our design but imitation is the best form of flattery and I must say your turbos are a striking resemblance to the MMP 1K turbos.


As far as manifold details….The MMP manifold perfectly matches the N54 head exhaust runner ID at 33mm and is a smooth transition to 45.1mm then merges to a 50.1mm collector. This smooth merge is only accomplished with a casting and we use the best investment casting process and highest-grade stainless-steel materials for our casting. Below you can see a pic of the cross section of our casting and how it matches the runners and merges for the best aerodynamic flow possible and least pressure losses. MMP manifolds combined with our turbine housings truly are the highest flow of any twin turbo setup for the N54 including the ones you present here.

View attachment 17921


I think you use a machined flange with machined insert welded in and cut and bent 1.25" sch 10 tubes welded together, correct me if I’m off please. You also would have to match the 33mm ID port size of the N54 with the counterbore for the insert, not sure what the N53 is but it is probably similar maybe you can measure it in the head just past the counterbore for the exhaust port insert and advise the size. the 1.25" sch 10 tube you use is 36mm ID. also from your flange to the ID of your pipe there is not a smooth transition as you can design with a casting, it is most likely a step change in diameter due to fabrication limitations that step change in diameter is very bad for flow as it causes flow separation and pressure waves and pressure losses add up there. My 45.1mm pipe ID size in my casting is 57% greater flow area than your 36mm ID pipe. also your collector is visibly smaller than mine. I would guess around 40mm ID but maybe you can clarify the size. If it is 40mm ID my 50.1mm ID is also 57% more flow area than your merged collector which again is less pressure losses, more flow, and more power from the manifold. I took a stab at modeling the cross section of your manifold based on the specs you provided to show what I think are the differences and posted the pic of the cross section below, but please let me know if I need to correct something for comparison purposes.

View attachment 17922

The manifolds I designed I did as the highest flow area diameter possible based on the bolt pattern restrictions. If I could have done larger I would so I believe they are the largest possible with a bespoke casting design and using investment cast methods and not sand casting is absolutely required for the high end surface finish and precision needed since pushing to the largest size pipe not to interfere with bolt pattern so even .5mm tolerance matters for the casting otherwise will run into problems quick which is why our 1K turbos uses these investment cast stainless steel manifolds with our 1K turbos.

Also your cast turbine housing you mentioned you use 6.6cm^2 for the AR area, our turbine housings have a 7cm^2 AR area so again ours flow more and are optimized through CFD for our high flow manifolds, high flow collector size and large 4646 GT28 high flow 9 blade turbine wheel TD04 turbos with our high flow native TD04 compressor housings.


You are right the devil is in the details and this is the most details I have shared about our design anywhere but since you posted some info on our design as if you knew all the details I just wanted to make sure you had all the right information for comparison purposes.



Also nice work on the custom N53 head mating to the N54. I look forward to the flow bench numbers. Was the n53 head ported or is it as stock? The N54 head work we did over 2 years ago and posted about in another forum showed the N54 head did flow terribly and could be improved by over 40% at .5” lift with a porting job on the head as we did it. We saw flow increased from 167CFM at .5” lift 28inhg pressure drop to 235CFM which we were very happy with those results using the stock N54 valves and of course non of the headaches to solve when fitting a non standard n53 head to an N54 bottom end but we like to see the effort and new things being tried. Once we saw the great results 2+ years ago from the N54 head porting we started offering them for sale at $2000 which was a huge decrease from other options on the market at the time offering a ported head for $4000. Since then others now offer it and the market has come down more in line with reality.



We look forward to seeing more results from your testing, good luck.

Have zero idea why you'd ever want to "taper up" on your primaries, especially on these ultra short primaries and to the level you have (33mm to 45.1mm!). We'd consider that more of a design defect than a benefit. This therein leads to your merger diameter of 50.1mm collector, once again, way too large for any breed of these turbos and likely another defect. Much like using 2" ID outlets per turbo, another overall unneeded and oversized defect.

Eventually you may realize it is NOT all about flow, it is also about velocity and overall function. You may want to start looking at some other turbos on other platforms, and while considering the power ranges and displacements they are supporting to help better design your turbos considering not only flow but also flow optimization.

This all seems to be backed up by some preliminary reviews of these units you have manufactured, in that they seem to have extremely poor spool/transient response and also do not seem to have much any benefit in power- to date anyway. Hopefully this gets better with time and what has been shown thus far is nothing but an unfortunate fluke after fluke. I'd say what Hydra has going on here is a MUCH better fit for 99.99% of the N54 community, while you'd be better off twin turbocharging a big block chevy.

Do agree however in that the twin setup are much "funner".:yum:

Rob
 
Last edited:

MMP

Private
Nov 10, 2016
48
34
0
Houston, Texas
Rob

Spool vs top end power (flow velocity vs flow loses), as you know always a trade off with fluid dynamic turbo systems. We are very happy with our design choices with spool being 15psi at 3300-3400rpm and we have the best design for top end power for twin turbos in this size.
 

[email protected]

Lieutenant
Dec 7, 2016
626
397
0
St. Louis, MO USA
www.rbturbo.com
Ride
'08 335i, '14 M6, '15 Tundra
Rob

Spool vs top end power (flow velocity vs flow loses), as you know always a trade off with fluid dynamic turbo systems. We are very happy with our design choices with spool being 15psi at 3300-3400rpm and we have the best design for top end power for twin turbos in this size.

Hey if you are happy guess that is all that matters. However for those customers out there looking to make the most ideal purchase they may appreciate making 15psi a thousand RPM's sooner AND still make the same top end power as well (on this displacement/RPM). There is NO doubt that this level of increased response would lead to a MUCH better transient response for everyday driving as well, no one likes a lazy turbo setup day in and day out.

As for the 7-8cm2 A/R, better for a 2.0-2.5L engines. 5-6cm2 are better for 1.5-2.0L engines, whereas per turbo we are utilizing 1.5L's per bank on the N54. Even the GC's, that everyone raves about, are only ~5cm2... and to further put it in perspective the OEMs are 4.9cm2. So if hydra is using a 6.6cm2, that is already fairly aggressive IMO, and "one-upping" that with "MEGA FLOW" commentary is doing nothing for you. As stated there are tasteful ways of accomplishing these things to get the most ideal overall results, and an all around better performing machine and "funner" everyday driver.

Thanks,
Rob
 
Last edited:

MMP

Private
Nov 10, 2016
48
34
0
Houston, Texas
Considering stock turbos spool 15psi at 2400rpm, stock frame “stage 2” turbos with 15T wheels spool 15psi at 2800-2900rpm, our stock frame stage 3 turbos with 19T wheels spool 15psi at 3100-3200rpm, we like the much higher extra power potential on the top end for a couple hundred rpm on spool.
 
  • Like
Reactions: 335is02

[email protected]

Lieutenant
Dec 7, 2016
626
397
0
St. Louis, MO USA
www.rbturbo.com
Ride
'08 335i, '14 M6, '15 Tundra
Considering stock turbos spool 15psi at 2400rpm, stock frame “stage 2” turbos with 15T wheels spool 15psi at 2800-2900rpm, our stock frame stage 3 turbos with 19T wheels spool 15psi at 3100-3200rpm, we like the much higher extra power potential on the top end for a couple hundred rpm on spool.

Think your figures are a bit pessimistic on your alternative examples much like they are likely optimistic on your benchmark. Despite that there are significant real world effects on transient response, that would be significantly apparent even with those deviations you've specified.

As such it is not always about putting her in 5th, finding a large hill to climb, stabbing the throttle at 1k rpm; to try to get a good psi per rpm boost climb to help with product marketing. There are times where one just wants to make a quick pass/etc and want their turbos to respond instantly as well. Your designs are best fit for twin turbocharging a much larger displacement than a 1.5L (7-8k rpm) cylinder bank IMO, and a poor choice for the vast majority of the performance N54's on the street.

Now back to Hydras regularly scheduled program.
 
  • Like
Reactions: Hydra Performance

Asbjorn

Lieutenant
Mar 10, 2018
811
552
0
European, based in China
Ride
Z4 N54 DCT
Rob

Spool vs top end power (flow velocity vs flow loses), as you know always a trade off with fluid dynamic turbo systems. We are very happy with our design choices with spool being 15psi at 3300-3400rpm and we have the best design for top end power for twin turbos in this size.

I wish there would be more competition in the spool segment (transient response really) that is not just OEM+ turbos though. Ideally we would have affordable fast spooling ball bearing turbos meant to produce no more than 500whp for racing on tracks. At least more competition in the "GC lites" segment would be nice. I love what @hydra is doing here, I just would not want to deal with supporting that much horsepower on the track.
 
  • Like
Reactions: frontside0815

MMP

Private
Nov 10, 2016
48
34
0
Houston, Texas
My numbers are testing I have done on my own car with 3 of those 4 turbos and I have logs to show for all except the stage 2 stock frame turbos which I have customer logs they have sent me to show the numbers.
 

MMP

Private
Nov 10, 2016
48
34
0
Houston, Texas
I wish there would be more competition in the spool segment that is not just OEM+ turbos though. Ideally we would have affordable fast spooling ball bearing turbos meant to produce no more than 500whp for racing on tracks. At least more competition in the "GC lites" segment would be nice. I love what @hydra is doing here, I just would not want to deal with supporting that much horsepower on the track.

We have some nice competition for that market segment coming out at a very competitive price (hopefully end of this year, was supposed to be end of summer but delays happen) and probably ball bearing variants of it some time next year as we will be soon making more and more ball bearing options. We have already done twin GTX3071r (5363) ball bearing wheels and bearings for bolt on to N54 as well as twin GTX2860r (4747) and they came out great.
 
Last edited:

MMP

Private
Nov 10, 2016
48
34
0
Houston, Texas
Would be cool to see a log from hydra To see at what rpm 15psi is achieved on his turbos to compare that aspect of it to the other numbers I posted.
 

[email protected]

Lieutenant
Dec 7, 2016
626
397
0
St. Louis, MO USA
www.rbturbo.com
Ride
'08 335i, '14 M6, '15 Tundra
Rob, can you show us how you designed your turbo manifolds?

Sure here is a bunch of info attached here from those much smarter than you/I/whoever else around here, and is of course what our current designs entail. Best bet for all is to entrust into those (ie. OE's) who are much larger, well funded, and much more talented than those wet behind the ears types who always default to "BIGGEST is BESTEST'EST for MOOO POWA".

Some deviation is at times warranted, but one must be tasteful and realistic on how much they push things in design. Proper matching of flow paths and components is something that is what is going to yield the best all around results, and using decades of studies on other platforms turbo designs as guides is often what is going to prevail in the real world over time. All said and was mentioned in prior posts, Hydra's design seems to be very tastefully thought out over anything else that is available at this time from an all around N54 performance perspective.
 

Attachments

  • MHIN54TurbochargerProductionDocument.pdf
    1 MB · Views: 155

[email protected]

Lieutenant
Dec 7, 2016
626
397
0
St. Louis, MO USA
www.rbturbo.com
Ride
'08 335i, '14 M6, '15 Tundra
So... You haven't designed or produced any is what you're saying.

The OE manifold is beyond satisfactory (more so absolutely excellent) for our current product line up from a flow AND performance perspective. It also has extremely awesome thermal characteristics, and has proven to be very reliable for over a decade now. So you are absolutely correct in that we have not designed any N54 manifolds, as there is no reason to design our own manifold (to date) and we are not alone in this regard either. And if we did rest assured it would be done to tasteful standards, which BTW would still be very similar to the OE manifold design especially in the sizing of the primary tubing and with only minor changes in the mergers cross-sectional area (which only would matter if the turbine housing has had a fair A/R increase to match).

As you seem to be one of those very vocal and self-proclaimed forum talents you probably are well versed to understood in that in general "smallish and restrictive" turbo manifold designs are less important as they still keep velocity up, and are usually much less of a restriction than what lies ultimately within the turbine housings/wheel assembly anyway. Or in other words an overly large and extremely mismatched exhaust manifold is NOT the answer for any turbo system, on any platform, at any time.
 
Last edited:

[email protected]

Lieutenant
Dec 7, 2016
626
397
0
St. Louis, MO USA
www.rbturbo.com
Ride
'08 335i, '14 M6, '15 Tundra
And your product line doesn't hit 900 at the wheels like your competitors.

You are correct we do not offer a true Stage 3 setup such as those hitting 900whp (ie. The one VTT Stage 3 twin setup and a few large ST setups). This is not our target audience as very few truly are looking for that kind of power on a N54, many do not want to enter into that pandoras box of $10s upon $10s of thousands of dollars into a platform that seems to be stuck in the high 10's/low 11's regardless... without extensive conversions outside of the BMW drivetrain anyway.

Nor do any want to have a very lazy turbo setup in everyday driving scenarios, most want to have a simple fast responding cost effective and reliable setups that makes up to and around 600whp with not much fuss. Much after this the platform seems to become a bottomless pit of problems, ever growing expenses, and lackluster support on many fronts.

Rob
 
Last edited:
  • Agree
Reactions: Torgus and Asbjorn

langsbr

Captain
Apr 5, 2017
1,176
673
0
Ride
07 335i 6MT e90
The OE manifold is beyond satisfactory (more so absolutely excellent) for our current product line up from a flow AND performance perspective.

People said the same thing about stock inlets for a decade as well. And Carbs, and flat tappet cams, and, and, and... Could it be that you believe it's "best" because it's what you use and you don't have an alternative offering?