Let's talk about doing a triple turbo setup
03-11-2008, 01:51 PM
#46
Registered User
Thread Starter
Join Date: Apr 2003
Location: Cummins Technical Center, IN
Posts: 6,564
Likes: 0
Received 6 Likes
on
5 Posts
Shawn, that's good work on SolidWorks concept.
Of pivotal importance in this whole concept is keeping the exhaust headers/manifolds absolutely as short and compact as can reasonably be done. This will reduce lag and improve efficiency considerably. Remember, the fewer pounds of steel there are between the turbo and the exhaust port, the less heat gets wasted trying to heat that up.
Thermally, it's like the difference between a lightweight sports car and a big CTD racing around a road course---which will be faster, even with proportionally the same power? Of course the car will be faster, because it will corner and brake so much better.
There's very much a "Goldilocks" methodology used in performance-- things have to be PROPERLY sized-- not too big and not too small. Bigger is not universally better, and racers who win do "as much as it takes and no more".
For example, if two crankshafts can both complete a couple races without failure, but one is 5# lighter, which one will the racer use? Same thing for pistons, rods, you name it.
Applying this methodology to a CTD is useful in a "performance" mindset, but we define performance quite differently--- we're not too keen on rebuilding our engines regularly and downtime is nobody's friend.
All that just to say, that the exhaust needs to be just strong enough to where failure isn't likely, and just long enough to allow fitment.
Bracing will be required, but the increase in two small charger over one larger one will be minimal, and the decrease in leverage of the fabricated manifold will probably offset any weight increase entirely, if not overcompensate.
12Valve-- your broken head was due to manifold shrinkage-- fairly common on 12V trucks, and one of the reasons ATS developed the replacement manifold to begin with. It was not for performance, per se-- but for shrinkage control and longevity.
A fabricated manifold would have much less shrinkage to deal with, and properly designed, the loading on the manifold bolts would be completely with the design spec of the factory turbo/manifold weight.
Keep in mind that you're dropping a lot of weight by swapping heavy cast manifolds out for fabbed tubular headers.
This is something a skilled fabricator could easily pull off. I'm thinking firstly of Dave Cross of Passenger Performance in Aldergrove, BC.
After all, look at the CTD manifold he fabbed here:
Of pivotal importance in this whole concept is keeping the exhaust headers/manifolds absolutely as short and compact as can reasonably be done. This will reduce lag and improve efficiency considerably. Remember, the fewer pounds of steel there are between the turbo and the exhaust port, the less heat gets wasted trying to heat that up.
Thermally, it's like the difference between a lightweight sports car and a big CTD racing around a road course---which will be faster, even with proportionally the same power? Of course the car will be faster, because it will corner and brake so much better.
There's very much a "Goldilocks" methodology used in performance-- things have to be PROPERLY sized-- not too big and not too small. Bigger is not universally better, and racers who win do "as much as it takes and no more".
For example, if two crankshafts can both complete a couple races without failure, but one is 5# lighter, which one will the racer use? Same thing for pistons, rods, you name it.
Applying this methodology to a CTD is useful in a "performance" mindset, but we define performance quite differently--- we're not too keen on rebuilding our engines regularly and downtime is nobody's friend.
All that just to say, that the exhaust needs to be just strong enough to where failure isn't likely, and just long enough to allow fitment.
Bracing will be required, but the increase in two small charger over one larger one will be minimal, and the decrease in leverage of the fabricated manifold will probably offset any weight increase entirely, if not overcompensate.
12Valve-- your broken head was due to manifold shrinkage-- fairly common on 12V trucks, and one of the reasons ATS developed the replacement manifold to begin with. It was not for performance, per se-- but for shrinkage control and longevity.
A fabricated manifold would have much less shrinkage to deal with, and properly designed, the loading on the manifold bolts would be completely with the design spec of the factory turbo/manifold weight.
Keep in mind that you're dropping a lot of weight by swapping heavy cast manifolds out for fabbed tubular headers.
This is something a skilled fabricator could easily pull off. I'm thinking firstly of Dave Cross of Passenger Performance in Aldergrove, BC.
After all, look at the CTD manifold he fabbed here:
03-13-2008, 06:33 PM
#49
Registered User
Thread Starter
Join Date: Apr 2003
Location: Cummins Technical Center, IN
Posts: 6,564
Likes: 0
Received 6 Likes
on
5 Posts
I thought you guys might find this interesting-- it's an email I received from Dave of Passenger Performance (the fab who made the manifold pictured in my post above).
Email:
Email:
Justin, you have certainly piqued my interest.
For the secondary turbos I would face the turbine discharges towards each other, both heading to the center of the engine, than use the Y pipe from the secondaries discharges' to the primary turbine entrance. So the secondaries would be "bottom mount" and the primary would be a "top mount".
I would use 3 to 1 radial style collectors feeding the t25 turbine inlets, certainly easier than the 6 to 3 divorced t4 collectors.
For the primary turbine inlet the fab time between divorced and open would not be much, I would utilize divorced for best performance.
This is definitely something I can tackle, and would be a great performing setup, especially if we milled the intake manifold off the other side of the head. I just did one this week for a customer in hawaii, we are porting it right now. As for cost of this setup, I would guess there is about 3500-5000 dollars worth of work involved, something like this would have to be charged hourly, I am sure we could come up with a more accurate quote if you are serious about the build.
Also we would definitely need a truck here for the fitting, can you supply a truck?
For the secondary turbos I would face the turbine discharges towards each other, both heading to the center of the engine, than use the Y pipe from the secondaries discharges' to the primary turbine entrance. So the secondaries would be "bottom mount" and the primary would be a "top mount".
I would use 3 to 1 radial style collectors feeding the t25 turbine inlets, certainly easier than the 6 to 3 divorced t4 collectors.
For the primary turbine inlet the fab time between divorced and open would not be much, I would utilize divorced for best performance.
This is definitely something I can tackle, and would be a great performing setup, especially if we milled the intake manifold off the other side of the head. I just did one this week for a customer in hawaii, we are porting it right now. As for cost of this setup, I would guess there is about 3500-5000 dollars worth of work involved, something like this would have to be charged hourly, I am sure we could come up with a more accurate quote if you are serious about the build.
Also we would definitely need a truck here for the fitting, can you supply a truck?
03-13-2008, 08:00 PM
#51
The Guru
Join Date: Jan 2004
Location: Airdrie Canada
Posts: 6,589
Likes: 0
Received 0 Likes
on
0 Posts
I vote for 2xGT3582R setup as secondard (three cylinders into each turbo), then S400 as your primary. The exhaust from the two GT35 would discharge into the S400. Tons of air, wicked spool-up. Expensive yes. Where would everything fit, I have no idea.
03-13-2008, 09:45 PM
#52
Registered User
Thread Starter
Join Date: Apr 2003
Location: Cummins Technical Center, IN
Posts: 6,564
Likes: 0
Received 6 Likes
on
5 Posts
Mike, I wouldn't think that combo of turbos would work well.
First, since each GT35 is being fed HALF the drive energy, it would take forever to spool. Those turbos are way too big They'd spool way worse than a K31 as a single
A pair of GT3582s is the equivalent of a single 86mm inducer-- ON TOP! Even Tim Barber doesn't need that much seconcary, LOL!
Second, that much capacity needs a WAY bigger primary. Even an S480 is way too small to match with those turbos as secondaries. IMO, you'll be in S500 territory with small chargers that big, and even a S500 is still probably too small.
The primary that would properly feed two GT3582R turbos would be over 200lb/min-- that's darn huge.
Fitting everything is, of course, the long pole in the tent.
First, since each GT35 is being fed HALF the drive energy, it would take forever to spool. Those turbos are way too big They'd spool way worse than a K31 as a single
A pair of GT3582s is the equivalent of a single 86mm inducer-- ON TOP! Even Tim Barber doesn't need that much seconcary, LOL!Second, that much capacity needs a WAY bigger primary. Even an S480 is way too small to match with those turbos as secondaries. IMO, you'll be in S500 territory with small chargers that big, and even a S500 is still probably too small.
The primary that would properly feed two GT3582R turbos would be over 200lb/min-- that's darn huge.
Fitting everything is, of course, the long pole in the tent.
03-14-2008, 07:09 PM
#53
The Guru
Join Date: Jan 2004
Location: Airdrie Canada
Posts: 6,589
Likes: 0
Received 0 Likes
on
0 Posts
Hohn, they use them for twin turbo gasoline car that pump out 600 to 800 ponies. They spool wicked, trust me. Run a tight turbine housing pipe on the GT35's, then run the first turbo off the rear three cylinders and pipe the second turbo off the front three cylinders. I would just run a 2 long rectangle channel/tubing and hang the turbo off that. It would work. Massive amount of steel under the hood. The S400 is big yes but people have been putting those turbos on twins forever, and they work, they are cheap. The best part once everything is spooled up as you pointed out big big air. Why would you run three turbos? I bet that you would lose too much heat with the extra lenght of tubes you would have to run. Next would be loss from friction from just spinning the turbos. You would get better gains from putting in an intercooler between two stages on twin turbos. Your twin garrett turbo set-up works pretty good and to improvement over that would really difficult with the only real gains would be from variable turbine housing and variable compressor vanes. very expensive, mix in computer control on your wastegates. I would bet that it would be one cool set-up.
03-14-2008, 08:18 PM
#54
Registered User
Thread Starter
Join Date: Apr 2003
Location: Cummins Technical Center, IN
Posts: 6,564
Likes: 0
Received 6 Likes
on
5 Posts
MIke, the twin GT35s on gassers is an apples-to-oranges comparison.
A GT 35 will spool just fine if fed all 6 cylinders (much like a stock HX35 would spool), but with just 3 cylinders it will be very laggy.
Moreover, ANY s400 is still waaaaay too small to feed a pair of GT35s.
It would almost like building a traditional twins setup using a HTT64 over an HTT 66-- the 66 is clearly too small. Only in your proposed setup, you'd have the equivalent of a K31 on top! How would would twins work with a K31 over an S400? Now do you see the problem? There's no point in having twins if there's no compounding, and to have any compounding, the larger charger must significantly flow more air than the smaller one-- I personally use a 100% larger as the minimum, but there's pros and cons of slightly bigger or smaller.
To help visualize the lag of your proposed setup with the paired GT35s, let's look at the turbine maps, which will tell all for comparative purposes.
First, as a reference, is the turbine map of the Stg 1/2/3 Garretts based off the GT37:
Now let's look at the turbine maps for the GT35s:
At a PR of 2, the GT 35 with it's tightest housing wants ~21lb/min of gas flow. But we're driving two of these turbos, so we need double that-- 42lb/min.
Compare that to the GT3788 at 30lb/min on the button.
What this means is simply that the paired gt35s will spool over 30% later than the single GT37. That's huge, imo. If you put rpm to it, then if the single secondary spooled at 1800rpm hard, you wouldn't hit that point with the paired secondaries until 2400rpm! That difference in lag might be the difference between towable and track only.
Granted once you DID hit 2400 rpm, your primary would light a little bit later and you'd have MASSIVE airflow after about 3K. But huge airflow after 3K means nothing to me if I have little to no boost below 2K rpm where I spend most of my time. I suspect others would feel similarly.
Paired GT35s work superbly on a twin turbo gas engine because you have a lot more rpm available, and for that application I'd think them superb. But we're dealing with an rpm-limited CTD of somewhat modest displacement and miserable breathing potential as stock, and we need something that will deliver airflow somewhat early.
I'm reminded of good career advice I once received: "If you can't do the work at the bottom, you'll never make it to the top."
The same can be said of airflow and RPM range.
I'm not saying that paired GT35s won't work-- they will. They will just be way laggy--more than all but the most hardcore are willing to accept.
An S400 is simply out of the question though with that much secondary. You'd need a MINIMUM of a "Phat fish-poop" turbo to feed it (fish=bass, turd=poop)
Jh
A GT 35 will spool just fine if fed all 6 cylinders (much like a stock HX35 would spool), but with just 3 cylinders it will be very laggy.
Moreover, ANY s400 is still waaaaay too small to feed a pair of GT35s.
It would almost like building a traditional twins setup using a HTT64 over an HTT 66-- the 66 is clearly too small. Only in your proposed setup, you'd have the equivalent of a K31 on top! How would would twins work with a K31 over an S400? Now do you see the problem? There's no point in having twins if there's no compounding, and to have any compounding, the larger charger must significantly flow more air than the smaller one-- I personally use a 100% larger as the minimum, but there's pros and cons of slightly bigger or smaller.
To help visualize the lag of your proposed setup with the paired GT35s, let's look at the turbine maps, which will tell all for comparative purposes.
First, as a reference, is the turbine map of the Stg 1/2/3 Garretts based off the GT37:
Now let's look at the turbine maps for the GT35s:
At a PR of 2, the GT 35 with it's tightest housing wants ~21lb/min of gas flow. But we're driving two of these turbos, so we need double that-- 42lb/min.
Compare that to the GT3788 at 30lb/min on the button.
What this means is simply that the paired gt35s will spool over 30% later than the single GT37. That's huge, imo. If you put rpm to it, then if the single secondary spooled at 1800rpm hard, you wouldn't hit that point with the paired secondaries until 2400rpm! That difference in lag might be the difference between towable and track only.
Granted once you DID hit 2400 rpm, your primary would light a little bit later and you'd have MASSIVE airflow after about 3K. But huge airflow after 3K means nothing to me if I have little to no boost below 2K rpm where I spend most of my time. I suspect others would feel similarly.
Paired GT35s work superbly on a twin turbo gas engine because you have a lot more rpm available, and for that application I'd think them superb. But we're dealing with an rpm-limited CTD of somewhat modest displacement and miserable breathing potential as stock, and we need something that will deliver airflow somewhat early.
I'm reminded of good career advice I once received: "If you can't do the work at the bottom, you'll never make it to the top."
The same can be said of airflow and RPM range.
I'm not saying that paired GT35s won't work-- they will. They will just be way laggy--more than all but the most hardcore are willing to accept.
An S400 is simply out of the question though with that much secondary. You'd need a MINIMUM of a "Phat fish-poop" turbo to feed it (fish=bass, turd=poop)
Jh
03-14-2008, 09:05 PM
#55
Registered User
Join Date: Nov 2006
Location: Rochester, NY
Posts: 2,333
Likes: 0
Received 0 Likes
on
0 Posts
Had an hour or two of free time at work today. Ignore the lack of compressor outlet, I never made it that far the other day and forgot about it today 
Also, that rear manifold is a little hinky right now. I had to modify the splines a good amount to get the exhaust outlets to line-up, but I didn't spend as much time as I should have to get the smoother runners. For what it's worth, if we REALLY want it to look right, the front manifold needs longer runners to match the rear manifold even though the flanges line up as-is.
Also, that rear manifold is a little hinky right now. I had to modify the splines a good amount to get the exhaust outlets to line-up, but I didn't spend as much time as I should have to get the smoother runners. For what it's worth, if we REALLY want it to look right, the front manifold needs longer runners to match the rear manifold even though the flanges line up as-is.
03-14-2008, 09:37 PM
#56
Registered User
Thread Starter
Join Date: Apr 2003
Location: Cummins Technical Center, IN
Posts: 6,564
Likes: 0
Received 6 Likes
on
5 Posts
Killer, Shawn!
In real life, we wouldn't have to lengthen the runners in the back-- be can just alter the angle the runners take towards the turbine inlet. The front header would be tighter to the block to account for that.
Great representation-- it communicates the idea, even if the details aren't all there.
Next step-- model the GT4202 and the pair of Gt22s. Oh, and don't forget the water lines, oil lines, and multitude of inlet plumbing
Slacker!
In real life, we wouldn't have to lengthen the runners in the back-- be can just alter the angle the runners take towards the turbine inlet. The front header would be tighter to the block to account for that.
Great representation-- it communicates the idea, even if the details aren't all there.
Next step-- model the GT4202 and the pair of Gt22s. Oh, and don't forget the water lines, oil lines, and multitude of inlet plumbing
Slacker!
03-14-2008, 09:49 PM
#57
Registered User
Join Date: Aug 2005
Location: Airdrie, Alberta
Posts: 7,780
Likes: 0
Received 0 Likes
on
0 Posts
Slightly off topic, but the subject or waterlines. Stock 3/4" lines, inlets on my turbos is 18mm banjos (probably go to some -8 or -10 JIC fittings instead). Just run the water in paralell (what I'm thinking), or would in series have any advantage?
03-15-2008, 06:01 AM
#58
Registered User
I think that you should rotate the turbine section a little,
There is no need for the turbo axis being parallel to the crankshaft. I was thinking about using different length runners for the turbo because with a split housing you'd get the pulse of the shorter runner more towards the small side of the turbine. The Y where the exhausts meet would be very short, and with the secondaries being mounted higher up (like with flipping the stock manifold to get the secondary up) we could get the primary almost into stock position.
The piping would be extremely short on the hot side (less energy loss) but a little longer on the cold side.
(I don't have CAD on this 'puter...)
Just my 2c
AlpineRAM
There is no need for the turbo axis being parallel to the crankshaft. I was thinking about using different length runners for the turbo because with a split housing you'd get the pulse of the shorter runner more towards the small side of the turbine. The Y where the exhausts meet would be very short, and with the secondaries being mounted higher up (like with flipping the stock manifold to get the secondary up) we could get the primary almost into stock position.
The piping would be extremely short on the hot side (less energy loss) but a little longer on the cold side.
(I don't have CAD on this 'puter...)
Just my 2c
AlpineRAM
03-15-2008, 11:00 AM
#59
Registered User
Join Date: Aug 2003
Location: Jeffersonville, Ohio
Posts: 3,604
Likes: 0
Received 0 Likes
on
0 Posts
Chris
03-15-2008, 02:22 PM
#60
Registered User
Join Date: Aug 2005
Location: Airdrie, Alberta
Posts: 7,780
Likes: 0
Received 0 Likes
on
0 Posts
I think that you should rotate the turbine section a little,
There is no need for the turbo axis being parallel to the crankshaft. I was thinking about using different length runners for the turbo because with a split housing you'd get the pulse of the shorter runner more towards the small side of the turbine. The Y where the exhausts meet would be very short, and with the secondaries being mounted higher up (like with flipping the stock manifold to get the secondary up) we could get the primary almost into stock position.
The piping would be extremely short on the hot side (less energy loss) but a little longer on the cold side.
(I don't have CAD on this 'puter...)
Just my 2c
AlpineRAM
There is no need for the turbo axis being parallel to the crankshaft. I was thinking about using different length runners for the turbo because with a split housing you'd get the pulse of the shorter runner more towards the small side of the turbine. The Y where the exhausts meet would be very short, and with the secondaries being mounted higher up (like with flipping the stock manifold to get the secondary up) we could get the primary almost into stock position.
The piping would be extremely short on the hot side (less energy loss) but a little longer on the cold side.
(I don't have CAD on this 'puter...)
Just my 2c
AlpineRAM