RB20 Turbo selection (twinscroll housing advice)

[Lithium]

1 hour ago, iruvyouskyrine said:

Lith I understand what you are saying and this is exactly what I want. Suggestions of things to look into that I would otherwise not consider. I understand your hesitation with the brand but I have confidence in them, my mates have used them with no issues for a while now. Like I said earlier I really don't know the finer details of turbos like the blade count and angle etc affects the performance. I thought comparing the sizes of the wheels would give me at least a basic indication of roughly what to expect from each turbo.
 

Cheers, glad you get what I'm trying to do - it's a bit hard sometimes online to put across where you're coming from.  I think any of the options presented will work well... and on your note of the EVOs spinning the wrong way, just for interests sake EVO 10 turbos spin "the right way" and are twin scroll with decent sized housings on both sides.

Good to know the Mamba turbos have held up, which ones are you talking about?  Have been waiting for results to start coming through for ages and the relative radio silence makes one think that people could be either avoiding them altogether, they are getting results not worth speaking of or they aren't surviving - or a combination of the latter.  

@discopotato03 I understand, I was just trying to clarify these are all things that have been long considered and not just kneejerk comments.  We are both pretty well researched.   

[iruvyouskyrine]

Didn't know that they changed back to a standard rotation on the Evo 10. Still seems like it will be difficult finding a T3 TS housing for them. My mates have been using TD05-18G and TD06SL-20G on SR20's making decent power for a while now, another mate has a GTX3076 on an RB25 same story. Basically identical to the Kando turbos (the owner of Mamba is the former owner/founder of Kando).

I think i'm almost at the point (unless something comes out of left field) where i just have to make my mind up between the GTX28060 and the 63 and see what the results are to be honest. R_34 has suggested a very rough 300-400 rpm difference between the 60/63 wheels, would you consider this roughly correct? Any idea what the power difference would be between the two as well?

[discopotato03]

I pondered this for an hour this morning and ended up deleting my thoughts because they all point towards ways of a trying to fix something that's avoidable/replaceable .I reckon I can see too much effort and money going into something that's dimensionally same as the 2.5L version .

I don't reckon I could live with a 20 even in a 32 GTST when things like Neos fall in and have way more potential .  

Edited April 19, 2017 by discopotato03

[iruvyouskyrine]

I agree that 25/neo is substantially better in every single way and there is no good reason to waste money on an RB20. I'm actually surprised it wasn't the first thing bought up after my original post 

Saying that, i'm sticking with the RB20. 100%. What are your thoughts on a GTX2860/63 with .83 A/R twinscroll housing on a gibson motorsport twinscroll manifold with external gate and E85?

[NickRB20]

What rpm are you hoping for your target boost?

[discopotato03]

I imagine the R31 GTSR engine was designed to make power at revs so low down torque wouldn't be its strong point . That manifold would have been part of a system to achieve a desired result . The mid 1980s was a learning time for all turbo road car manufacturers including Nissan . If you look at their other 80s engines eg L20a/28ETs/Z18ETs/FJ20ETs/RB30 ETs the pattern is similar . Lowish compression ratios and no intercooling except or the later FJ20 powered DR30s . They had to keep the exhaust side flow up and boost modest to avoid detonation . And whilst using pretty agricultural turbochargers . Power came at revs except for the larger L28 and RB30 engines . The tide turned with the later DR30 FJ20 where they reduced the inlet plenum volume / compression ratio up from 8 to 8.5:1 / turbine housing down from 0.63 to 0.48 . AND had smaller inlet tube plumbing with a side mount intercooler vs no IC . Maximum power was up and boost threshold came down , I had one of those cars around 15 years ago .

AFAIK the R32 RB20DET was the best all round version of that engine , it lost all the TVIS style BS and had a more modern BB turbo with better engine management . The things to keep IMO are the native manifolds because they are sized to keep the gas speeds up around where they need to be with those small 330 odd cc pots . If the cast iron RB20 and RB25 style manifolds were garbage there's no way they , esp the 25s , could make the power they do when modified . Just on this I did once see SKs cracked RB20 exhaust manifold and they are smaller bore than any 25 one I've ever seen . 

If you think you can make a twin scroll T25/T28 work go for it , all I know is that if the housing doesn't work with the turbine turbo performance suffers . It doesn't have to be out by much to push the boost threshold up . Any waste gate should have zip to do with turbo response because it will be shut until rated boost pressure opens it .

I'd be trying to keep this thing as simple as possible . RB20 manifold and turbine housing and fit a small Garrett or whatever BB cartridge into it . Or ask Hypertune to develop something that will come on sooner than the complete OE RB20 turbo does . Restricted Rally style turbos sometimes had quite small looking compressors but they still made strong low to mid rev torque . I'd be looking at Garretts GT2554Rs and GT2854Rs and crunching the numbers based on low to mid rev torque rather than high end numbers .

All this is attempts to work around what Nissan fixed with the RB25 . The RB20 really is an orphan based around the bore and stroke of the old L20a six . I don't think you can compare the RB20 with any other RB because of that small 78mm bore size . All other RBs were big bore and even Nissans poverty pak RB24 was a destroked 25 not a long stroke 20 . Really major works are not an economic reality with an RB20 because no matter what you do that small bore and lack of capacity will eventually paint you into a corner . The engineering solution is greater capacity and less extreme mods .

I would go so far as to say that a really well done RB25DET Neo could convince you that the eventual dirty 30 may not be needed . 

Your call I know what I'd do .

A .

[Lithium]

12 hours ago, iruvyouskyrine said:

Saying that, i'm sticking with the RB20. 100%. What are your thoughts on a GTX2860/63 with .83 A/R twinscroll housing on a gibson motorsport twinscroll manifold with external gate and E85?

What is the .83a/r twin scroll you are proposing to use?   I am quite apprehensive about the whole twin scroll GT28 idea because I am unaware of a native housing for them, and have seen aweful results from people modifying housings meant for other turbines in the past - not to say it won't work, but I'm just wary... especially when giving away the divider space on that small a turbine.  I'm not going to touch the "this will/will not work well" topic without turbine flow maps or any kind of evidence of performance on any other setup for fear of giving bad advice.  I'd say the most accurate input would be that you'll be a guinea pig if you try it, though it has potential to be pretty epic.

8 hours ago, discopotato03 said:

...and had a more modern BB turbo with better engine management .

I'd be trying to keep this thing as simple as possible . RB20 manifold and turbine housing and fit a small Garrett or whatever BB cartridge into it . Or ask Hypertune to develop something that will come on sooner than the complete OE RB20 turbo does . .

I would go so far as to say that a really well done RB25DET Neo could convince you that the eventual dirty 30 may not be needed . 

Really?  I honestly was pretty sure that the RB20DET had a journal bearing ceramic wheel unit on it?   If it is ball bearing then that REALLY highlights how much better the TD05s are, they are MUCH bigger wheels and flow and honestly weren't much laggier - definitely performed better than the ball bearing ceramic turbine RB25 turbos did on RB20s, which makes a mockery of Nissan/IHI's effort at turbo matching when you get a journal bearing unit designed for a 4pot behaving much better on their own engine.

To be fair it's been YEARS since I drove such a setup, but while my R33 was in for work with the GT3076R I borrowed my mates S13 Silvia running an RB20DET with a TD05 16G on it, road tuned to 20psi and if are happy with the lag of a GT3076R (or ATR43SS2 ) on an RB25 in an R33 then an RB20 in a lighter chassis with a 16G is surprisingly tolerable.  I'd love to know what a GTX2860R would be like, if they can spool even better and make the same power then I'd say they'd destroy a few ideas about what RB20s are like.   Bare in mind they'll never be glorious at real low rpm, but 20psi before 4000rpm on any 2litre - even an RB20... and you get a very solid push in the back.

 

 

 

 

[GTSBoy]

40 minutes ago, Lithium said:

Really?  I honestly was pretty sure that the RB20DET had a journal bearing ceramic wheel unit on it?   If it is ball bearing then that REALLY highlights how much better the TD05s are, they are MUCH bigger wheels and flow and honestly weren't much laggier - definitely performed better than the ball bearing ceramic turbine RB25 turbos did on RB20s, which makes a mockery of Nissan/IHI's effort at turbo matching when you get a journal bearing unit designed for a 4pot behaving much better on their own engine.

 

There's essentially no difference between R32, 3 & 4 CHRAs.  Same BB centre, same ceramic turbine.  Only the housings and compressor wheel changed, and not by much there either.

[Lithium]

1 hour ago, GTSBoy said:

 

There's essentially no difference between R32, 3 & 4 CHRAs.  Same BB centre, same ceramic turbine.  Only the housings and compressor wheel changed, and not by much there either.

Interesting!   Yeah I knew the turbine side was the same, which is HORRIBLE by the way but something got it in my head that the RB20 units were journal bearing too.  

Not too relevant, aside from the fact that it is dreadful how poor they were for how much money went into making them!    They should have been the EFRs of their time!  

[GTSBoy]

Red top RB20s and the RB30s had plain bearings.

[Kinkstaah]

Would a BW be more responsive than a TS Garrett?

Still think a 62/6758 will work better.
That said, yesterday drove a GTX3076R on a SR20 which was a TS setup and honestly the lag was really a non-issue, which was great for a 280KW setup let alone a 400+kw setup that this one was.

Is the RB20 really worse than the SR by that amount? I have a feeling any of the modern options in a twin scroll setup is going to be suprisingly good actually!

[R_34]

The twin scroll housing for the GT28 is one made for the gt28 wheel so I suppose it works OK.

A tuner in my country used one on a SR20/GTX2867/fullrace manifold/twin tial 38 gates/HKS step 2 cams.

371kw to the fly

660Nm

The boost was tapering badly hence the torque drops off.

 

Edited April 20, 2017 by R_34

[iruvyouskyrine]

Okay so RE the T3 Twinscroll rear housing;

Did a lot of googling and it seems that ATP Turbo are the ones who actually make this housing. I have sent them an email and are awaiting their response. I also emailed GCG and asked them if Garrett ever made such a housing and this is what they had to say. 

Quote

Hi Josh,

Garrett themselves never released these housings in that configuration, however, we do have access to an aftermarket 0.82a/r T3 split pulse turbine housing to suit the GTX28 series. You’d be looking at approx. $650.00 Inc GST, with a time frame or 7-10 business days from order.

and a follow up email when i asked for more info

Quote

Hi Josh,

I have attached an image for you. The housing is based on one of Garrett’s T3 split pulse housings, that are originally designed for a larger turbine wheel, so the housing is sleeved, and re-machined to suit the GT28/GTX28 series of turbine wheel. It has a 3” V-Band outlet, and we have the mating flange ($55.00) and clamp ($50.00) in stock.

The photo that was attatched was the exact same photos that are on the ATP Turbo website (link to it here)

Next up are some pics that i found online (which besides the ATP website is the only pictures that i could find anywhere online). The are from a german eBay add where a dude is selling what is advertised as "Garrett Gt28 Exhaust casing T3 Twin Scroll Garrett GTX28 V-Band" (link to auction)

I have sent the seller an email asking if he can provide me with more information about where he purchased it from. But it looks very similar to the ATP one and you can see it has been sleeved like the GCG rep suggested.

@R_34 can you provide any more pics/details on the housing that was on the car from that dyno graph? Thanks

[iruvyouskyrine]

Also, i had no idea that RB20/25 turbos were ball bearing. They really are staggeringly bad in that case, i just assumed they would have been journal bearings with heavy cast wheels.

@Kinkstaah Yes performance wise the EFR would most likely shit it in, but they are just so expensive for something thats going on a $20 engine. Saying that though, i will get some more prices on everything but this looks like its starting to add up pretty quickly. What EFR would you reccomend to make a punchy 220kw on 98 octane that has a T3 flange?

[Kinkstaah]

The truth about the EFR when I bought one, is I realised when you add a rear housing, and add a BOV that is decent, the EFR is actually as cheap/cheaper than a Garrett is. Not to mention the IWG is better still than the Garrett one is, etc etc. You even get a boost solenoid for your trouble.

Fabbing can be a thing though, only the 7064 and 7670 onward fit a T3 flange natively. These are big units and 300kw++
Given how nice that GTX3076 I drove the other day was though... it would probably be prettty damn good on a RB20 which can rev higher than a SR20..

Otherwise the 6258/6758 are T25 flange, but it's nothing an adaptor wouldn't fix. 6258 is a 450hp rated turbo and 6758 is 500. I'd probably go the 6758 for more headroom but the 6258 will definitely do 220kw atw..

But honestly if the target is a usable 220kw atw, the answer is a RB25 high flow from Hypergear or contact scotty nm35 and see what he's doing with the impulse turbo over there. That will be a bolt on solution and all thinking from there on can just stop and it'll hit the target with extreme ease

[R_34]

7 hours ago, iruvyouskyrine said:

 

@R_34 can you provide any more pics/details on the housing that was on the car from that dyno graph? Thanks

 

Nope there no pics of it unfortunately, well none showing that exhaust housing, they juste state GTX2867 twin scroll .82 for the turbo.

[Roy]

I have a new TD05-18G with no rear housing i will be happy to do a deal with. No twin scroll opyion but smallish turbo that i bought for my old RB20 before the head gasket went

[iruvyouskyrine]

Right so have done some more investigation and cost comparison into the EFR thing. It's starting to look really good besides the lack of BW using T3 flanges.

So my question is keeping in mind this is an RB20, would a T4 flange be too large? It seems very unconventianal to use such a large flange on such a little engine but hey if it's going to work then i am happy to run it. If i was going to use an EFR i would most likely be going with a 6758 and a .80 A/R T4 winscroll IWG housing. I know @Kinkstaah has been saying 6758 all along so keen to hear his thoughts on it.

With the EFR's is the 6758 the best choice for a punchy 220-230kw on 98 octane or should i be considering the 6258? Keeping in mind that that power is my lower goal, the more power it makes on E85 the better!

[Kinkstaah]

People like Havoc fabrication can weld you on a T4 TS to stock RB manifold... to which aforementioned turbo would bolt right on to.......

I only say 6758 after driving a 2L with a GTX3076R on TS which I figured would be utter laggy shit for me to realise no, no it wasn't at all. 6258 would be closer to 220kw for sure, but I have this feeling the 6758 would be responsive enough and also provide more fun happy times. These things do not like being overspun and I'd go a size up to minimize that risk and let things run lazier than being right on the edge.

Note: Never owned a RB20! Results may vary! But I still would expect using a BW6758 is going to be the best answer anyway, strongly, strongly doubt you'd regret it on there, it's about as good as a RB20 is going to ever get.

[iruvyouskyrine]

First off, I will not be using a standard RB manifold. It's just not happening.

In the old EFR thread there was this post

Quote

I don't mean to take away from all this big numbers talk, but what are peoples opinions/thoughts on a 6258 on a RB20? I'm currently running a 2530, would the 6258 be similar in response/spool?ike this

And Jeff's response to that

Quote

an RB20 would absolutely rock with the EFR6258. it would feel like an rb26 with good turbos ha

 I agree with you Greg, i think realistically a 6258 is my best choice for exactly what i want. BUT my gut feeling is as you said, a 6758 would still be responsive enough but have more power capacity and stop the overspinning of the turbine wheel. The twinscroll manifold and housing should help with spool as well. Of course all these assumptions are just from my gut feelings and reading forums/results on the web. I don't really know enough about turbo flow rates etc to make calculated guesses or use BW Matchbot. Hopefully that's where people like @Lithium can help me out