Is it worth doing parallel twins?
03-16-17, 01:51 AM
#3
I will go over some reasons why parallel twins aren't as good as a single turbo because this is a very misunderstood topic.
Parallel twins will have more lag than a single turbo of the same output potential.
There are generally two things that cause this increased turbo lag- the compressor surge lines and lack of exhaust energy.
--------------------------
Compressor surge lines-
When looking at a turbo compressor map there is a curving line on the Left that defines the turbo surge line.
This surge line shows how much boost the turbo can create per airflow- it is a limitation based on the specifications of the compressor.
If your motor is able to spin the turbo up fast with exhaust energy you will still be limited by the compressor surge line on how much boost you can run in the low rpms (low airflow).
An example is on the EFR 8374 and EFR 9180, if you try to run 20+ psi boost at 3,000rpm the turbo compressor will surge and the car will buck. The exhaust spools the EFRs so well the limit is the compressor surge line.
So, what about parallel twins?
To compare parallel twins surge lines you just take one of the compressor maps and halve the flow (the other turbo is providing the other half flow).
This almost always leaves you with a surge line worse than a single turbo of the same airflow limit as the two twins combined.
There are exceptions.
The EFR 7163 developed in the Indy Car program has a very good surge line. In parallel twin configuration over ~12psi boost and under ~20psi of boost it actually has a better surge line as twins and can flow more than a single EFR 8374.
So parallel twin 7163 would be laggier under 3,000rpm and then hit harder compared to a single EFR 8374.
--------------------
Exhaust energy-
Turbo exhaust housing A/R is pretty widely misunderstood. It is the ratio of exhaust Area (runner cross section) to the Radius (how far from center of turbo wheel).
If you have a single turbo with a divided T4 0.92 AR each scroll is actually 0.46AR.
Same applies to parallel twin turbos- you add together the AR of the two turbos. Twin 0.60AR exhaust housing turbos leaves you with 1.20AR total.
You can see that people that have done twin turbos in the past on rotaries probably didn't understand this.
Another tempting trap on parallel twin turbos on the rotary is to feed each turbo off a single exhaust runner. This hurts spool as well since you have half as many exhaust pulses.
The whole idea of equal length manifolds and divided turbos is to get as many pulses as possible as that smoothly accelerates the exhaust wheel.
Big low frequency pulses are bad for accelerating the exhaust wheel. You can just play with a child's pinwheel to figure this one out.
If you feed each parallel twin turbo with both rotors you have a complex manifold and the extra merges and surface area will cost you some exhaust energy over a simple single manifold.
Again, there are exceptions.
I have a design in mind that will feed each parallel turbo from both runners without complicated merges.
This is to simply use 2 1/2" to 1 1/2" schedule 80 reducers off the engine and 1 1/2" pipes joined for one turbo and then place another 1 1/2" schedule 80 45 degree bend inside the 2 1/2" section of the reducer and join these with pipe for the second turbo.
If you wanted to scale this for stock sized twins it works out to 2" to 1 1/4" schedule 10 reducer (thinner wall) to 1 1/4" schedule 80 pipes to the turbos to keep the same runner area on the inner and outer pipes.
Pipe within a pipe with equal area for the inner pipe and outer pipe flow areas.
I actually want to try this in a sequential twin set-up to get the advantage of low runner volume on the center primary turbo pipe and the velocity stack shape with the outer pipe flow area closed off.
---------------------
Compressor co-surge-
Another issue with parallel twins is compressor co-surge. This occurs when one compressor is flowing slightly less than the other for any reason and the second compressor completely takes over the total system flow actually causing the first turbo to surge/stall/reverse rotation.
The load of taking over the total system flow slows the second turbo and the first turbo spins back up and takes over total system flow while the second turbo surges/stalls/reverses.
This is why V engines with twin turbos work best with two completely separate intake tracts into the cylinders each feeds.
The FD stock twins in parallel mode do not suffer this co-surge as they are basically both fed off the same exhaust (there is plenty of communication between the two exhaust ports in the stock log manifold) and so accelerate at the same rate even if one compressor is contributing slightly more or less.
Parallel twins will have more lag than a single turbo of the same output potential.
There are generally two things that cause this increased turbo lag- the compressor surge lines and lack of exhaust energy.
--------------------------
Compressor surge lines-
When looking at a turbo compressor map there is a curving line on the Left that defines the turbo surge line.
This surge line shows how much boost the turbo can create per airflow- it is a limitation based on the specifications of the compressor.
If your motor is able to spin the turbo up fast with exhaust energy you will still be limited by the compressor surge line on how much boost you can run in the low rpms (low airflow).
An example is on the EFR 8374 and EFR 9180, if you try to run 20+ psi boost at 3,000rpm the turbo compressor will surge and the car will buck. The exhaust spools the EFRs so well the limit is the compressor surge line.
So, what about parallel twins?
To compare parallel twins surge lines you just take one of the compressor maps and halve the flow (the other turbo is providing the other half flow).
This almost always leaves you with a surge line worse than a single turbo of the same airflow limit as the two twins combined.
There are exceptions.
The EFR 7163 developed in the Indy Car program has a very good surge line. In parallel twin configuration over ~12psi boost and under ~20psi of boost it actually has a better surge line as twins and can flow more than a single EFR 8374.
So parallel twin 7163 would be laggier under 3,000rpm and then hit harder compared to a single EFR 8374.
--------------------
Exhaust energy-
Turbo exhaust housing A/R is pretty widely misunderstood. It is the ratio of exhaust Area (runner cross section) to the Radius (how far from center of turbo wheel).
If you have a single turbo with a divided T4 0.92 AR each scroll is actually 0.46AR.
Same applies to parallel twin turbos- you add together the AR of the two turbos. Twin 0.60AR exhaust housing turbos leaves you with 1.20AR total.
You can see that people that have done twin turbos in the past on rotaries probably didn't understand this.
Another tempting trap on parallel twin turbos on the rotary is to feed each turbo off a single exhaust runner. This hurts spool as well since you have half as many exhaust pulses.
The whole idea of equal length manifolds and divided turbos is to get as many pulses as possible as that smoothly accelerates the exhaust wheel.
Big low frequency pulses are bad for accelerating the exhaust wheel. You can just play with a child's pinwheel to figure this one out.
If you feed each parallel twin turbo with both rotors you have a complex manifold and the extra merges and surface area will cost you some exhaust energy over a simple single manifold.
Again, there are exceptions.
I have a design in mind that will feed each parallel turbo from both runners without complicated merges.
This is to simply use 2 1/2" to 1 1/2" schedule 80 reducers off the engine and 1 1/2" pipes joined for one turbo and then place another 1 1/2" schedule 80 45 degree bend inside the 2 1/2" section of the reducer and join these with pipe for the second turbo.
If you wanted to scale this for stock sized twins it works out to 2" to 1 1/4" schedule 10 reducer (thinner wall) to 1 1/4" schedule 80 pipes to the turbos to keep the same runner area on the inner and outer pipes.
Pipe within a pipe with equal area for the inner pipe and outer pipe flow areas.
I actually want to try this in a sequential twin set-up to get the advantage of low runner volume on the center primary turbo pipe and the velocity stack shape with the outer pipe flow area closed off.
---------------------
Compressor co-surge-
Another issue with parallel twins is compressor co-surge. This occurs when one compressor is flowing slightly less than the other for any reason and the second compressor completely takes over the total system flow actually causing the first turbo to surge/stall/reverse rotation.
The load of taking over the total system flow slows the second turbo and the first turbo spins back up and takes over total system flow while the second turbo surges/stalls/reverses.
This is why V engines with twin turbos work best with two completely separate intake tracts into the cylinders each feeds.
The FD stock twins in parallel mode do not suffer this co-surge as they are basically both fed off the same exhaust (there is plenty of communication between the two exhaust ports in the stock log manifold) and so accelerate at the same rate even if one compressor is contributing slightly more or less.
03-16-17, 10:10 AM
#4
Well written Blue TII. I agree, parallel on the stock twins doesn't really gain you much. You introduce a whole lot of lag and gain little up top. Pretty much the biggest benefit is simplicity in the engine bay. Keep the twins sequential, or go with a properly sized single for your application. If you want to stick with twins and make power, go with BNR. We have made 456 wheel hp on our Mustang dyno with BNR twins run sequentially.
Last edited by IRPerformance; 03-16-17 at 10:13 AM.