twin-screw vs roots vs centrifugal vs turbo

Did I miss anything?

Anybody got any good links for understanding the pros and cons of each? As a bonus, good sources for fundamental understanding of each with more detail?

I know each of these things has been discussed here and there, but I need a grounding in anything I'm going to consider.

Turbo and centrifugal can have lag, but make more power up top than they can at the bottom(all depends on size too), and twin screw & roots type make good low-medium power(pending the size again)

Whipple claims their twin screw has better efficiency than the other forms though....which is odd to me.

Whipple's arguement

yamaha wrote: Whipple claims their twin screw has better efficiency than the other forms though....which is odd to me.

Yeah, I wonder how that works. But when it comes down to it, I don't even know how 'efficiency' is defined for a supercharger... The amount of energy needed to provide a given flow and pressure? Degree of heating of the charge? I know Whipple makes particular reference to needing an intercooler less often due to better efficiency, but now I'm stacking manufacturer claims based on other manufacturer claims...

The wikipedia article (cringe) on superchargers notes that roots-type require backflow of the pressurized area under the supercharger into the device to work, and that this is part of the inefficiency. Could the twin-screw's efficiency be due to being both positive-displacement and not requiring this backflow? (though honestly, I can't quite get my head around how the twin-screw's trapping a unit of air between the lobes and passing it through is all that different from the roots' trapping units between the lobes and case wall? Is it just the relative continuity of the volume with the twin-screw?

In reply to ransom:

Thats way above my understanding on this subject.....as far as I could tell, the biggest difference between the roots type and twin screw were the uneven number of lobes on the screws for the twin screws.

As far as the centrifugal goes, they're essentially just the cold side of a turbo with a pully added(Some are more advanced and such though)

I was discussing my need for FI on my Mustang with a Corvette buddy. My statement was that I wanted a roots or twin screw so I could build more power and torque lower in the RPM range. He asks "Can you smoke the back tires when you're starting off?" I sez, "sure, pretty much anytime I want below 30mph."

"So, how much more torque do you need down low?"

"............................................"

Which makes me think maybe I'd rather have centrifugal in the pony car.

In reply to ultraclyde:

The correct answer is you need tires that are worth a E36 M3 in the back.....then repeat this process after more power breaks them loose.

ultraclyde wrote: I was discussing my need for FI on my Mustang with a Corvette buddy. My statement was that I wanted a roots or twin screw so I could build more power and torque lower in the RPM range. He asks "Can you smoke the back tires when you're starting off?" I sez, "sure, pretty much anytime I want below 30mph." "So, how much more torque do you need down low?" "............................................" Which makes me think maybe I'd rather have centrifugal in the pony car.

Sounds like you need to spend some $$$ on getting the power down. I started a thread a while back about supercharging versus Turbocharging as I have been piecing together both systems for my car (automotive ADD is a bitch) Both have there advantages and draw backs you have to figure out what you are trying to achieve and then look at the options and see what gets you there.

I'd always go with a turbo except maybe for drag racing. The supercharger's only advantage is instantaneous power delivery*. In exchange for a little lag, turbos give much better efficiency (that is, more power since you're wasting less power to make the boost) and the ability to change boost levels not only without changing parts, but in an electronically adjustable manner.

*except centrifugal, which combines the lag of a turbo with the low efficiency and limited tuning options of a supercharger.

The biggest difference between turbos and centrifugals is the connection to the engine. Centrifugals are directly tied to engine speed, so the boost will rise with rpm. 3 psi at 3000 rpm, 6 psi at 6000 rpm, etc. This isn't lag - centrifugals don't suffer from "turbo lag".

Turbos are not tied to engine speed like this, so they'll make peak boost as soon as they physically can. If they can make a target 6 psi at 3000 rpm, they will. And they'll still make 6 psi at 6000 rpm. Makes for a very different driving experience.

There's a way around this with centrifugals but it's a bit goofy - you overspin the compressor and start venting boost. Without this sort of setup, a centrifugal kinda feels like a naturally aspirated engine with a cam set up for high power. It's fun, but it's not as fast as the turbo.

Twin screws and roots work pretty much the same. There are differences - one is pooping out compressed air whilst the other is packing uncompressed air into a limited volume - but you can consider them pretty much interchangeable from an implementation standpoint.

I like the idea of the instantaneous and predictable delivery of a positive-displacement, crank-driven supercharger.

I want to understand the whole mess better, but my own immediate use is likely to be street/autocross, where there is a premium on getting exactly what you ask for with the accelerator with a minimum of waiting and reacting to spool-up and so forth. This consideration would seem to trump peak power/efficiency, but if the turbo can be crisp enough, why not consider it?

I still haven't gotten to drive any turbos besides my stock WRX, and after other questions in other threads, I'm not sure that it's got lag, or goofy throttle mapping, or ??? It makes me unenthusiastic about turbos, but I don't want to write them off based on this one car. I know they can make useful boost at low engine speeds, but I have no faith in my impression about how quickly they can get there from off-throttle...

Stock WRX? Not the best example of boosted power delivery.

Open up the exhaust/DP and tune it...and be amazed at the difference in power delivery.

Keith Tanner wrote: The biggest difference between turbos and centrifugals is the connection to the engine. Centrifugals are directly tied to engine speed, so the boost will rise with rpm. 3 psi at 3000 rpm, 6 psi at 6000 rpm, etc. This isn't lag - centrifugals don't suffer from "turbo lag".

I thought the "lag" effect on centrifugal superchargers was caused not by any speed mismatch but by the nature of how centrifugal turbines work - the relationship between boost and turbine RPM isn't linear, and there isn't much boost at low RPM.

It's not technically the same thing as turbo lag but it feels similar.

What you're referring to is spoolup time, really. It's similar to a naturally aspirated car coming on cam. With the centrifugal, the only way to make more boost is to spin the engine faster. With a turbo, the turbo will spin as fast as it can - but there might not be enough exhaust energy at lower engine speeds to reach a given boost.

Lag is usually considered to be the delay between asking for power and getting it. Slap the throttle open on a turbo car at 4000 rpm, and there might be a short pause before you get the maximum boost the turbo can produce at 4000 rpm. The amount of lag depends how well the turbo is sized for what you're asking.

For autocross, positive displacement supers really do work well. For track work - well, there's a reason the turbo rules racing.

Driving feel to me - centrifugals feel like a highly tuned naturally aspirated engine. Positive displacement superchargers feel like a big displacement naturally aspirated engine. Turbos feel like the hand of god pushing your car down the road.

I guess I need to read more about turbos and what can be achieved in terms of lag reduction and what trades off.

I also definitely need to find out more about twin-screw superchargers (like did anything use them stock?)...

Nathan JansenvanDoorn wrote: Stock WRX? Not the best example of boosted power delivery. Open up the exhaust/DP and tune it...and be amazed at the difference in power delivery.

I'd love to, but this experiment in Factory Hot Rod ownership is doomed to replacement by whatever wagonish thing makes sense before long, so I'm keeping it bone, stone stock in hopes that the unicorn of an unabused, dealer-maintained, low-miles WRX will fetch a premium.

Unless I hit the lottery, in which case I'll keep it and try every experimental idea I have for making a modern car feel less heavy and insular

Rough guide to turbo: small = responsive but limited in power, big = laggy but more power. You can improve responsiveness by running a higher compression engine as well, but that means you hit the knock threshold sooner.

I have a little GT2554 turbo on my 1.8 Miata, smaller than most people would run. It limits me to 220 rwhp or so, but holy cow does it like to spool. The car sounds like it's powered by air, not combustion.

Twin-screws certainly aren't as common as Roots blowers. I think the Mazda Millenia used one. I'm pretty sure Mercedes used them, and wasn't there one on the Lightning pickup? The Mazda's probably the easiest one to find in a junkyard.

In reply to Keith Tanner:

Thanks!

I'd be pretty happy with 220ish hp out of my BMW M42 (same displacement, but I have no idea the relative potentials of those two engines).

I wonder how appropriately-sized the Millenia unit is... Presumably the Lightning unit would be massive for my application. Hey, the Millenia units are all over Ebay for ~$500. They look sort of long. this Autorotor unit is 1.5L to the Millenia unit's 2.3... I'll have see if I can dig up sizing info...

For the weight savings and packaging, I'd love to believe Whipple's claim of often avoiding intercooling, but I'm seriously dubious...

If for an M42, just use a smaller turbo.......the DASC guys proved supercharging is pointless on that engine. Turbocharging though, several guys have gotten good power from them.

In reply to yamaha:

Was it the engine, or the DASC setup?

I hate to write off the system most likely to give me the instant response I want because nobody got a particular system working well.

That's a non-intercooled Roots setup, isn't it?

Early WRXes were notorious for having really laggy turbo systems, due to restrictive exhaust with multiple cats, and even a cat before the turbo (which is a really bad idea as far as response goes). I think that got at least a bit better in the later ones, but I'm not sure by how much.

Most OEM turbos are on the small side, designed to produce lots of torque at low RPMs and with little lag. My old B5 S4 (2.7L, twin KKK K03s, stock), for example, made peak torque at like 2000 RPM, and it fell off a lot above 5000 (redline of 7000). Except for the slight whooshing noise, you almost couldn't tell that that car had turbos at all, and it made noticably better low-end torque than the 4.2L naturally aspirated V8 in the B6 S4 that I replaced it with.

Aftermarket turbo kits are usually sized larger than OEM applications, those K03s max out around 180-200 crank hp each, whereas the GT2554 that Keith mentioned (which is about as small as you'll find in any aftermarket kit) is good for more like 250. The GT2560 in my Miata is rated at 300, and in top gear on the freeway will go from vacuum to full boost in under half a second. In 2nd or 3rd gear (like at the autox) it goes to full boost in about as much time as it takes to get your foot to the floor.

One other important difference in forced induction systems is how they respond to altitude. Positive displacement superchargers (and I think most traditional centrifugal ones, too) behave like naturally aspirated cars -- as the altitude rides and ambient pressure drops, the torque curve drops proportionally but stays the same shape. Turbo cars, OTOH, will spin the turbos faster to make up for the altitude losses to some degree. Raising altitude on a turbo will raise the boost threshold (minimum RPM to make full boost) and increase lag, but the peak power will not drop anywhere close to as much as a supercharged or NA car. If it's got MAP-referenced electric boost control and enough headroom in the turbo, it may not even drop the peak power at all.

The downside that I encounter for a turbo at autox is that the torque at a given RPM/TPS point can change as the boost rises. This makes powering out of corners a bit harder to get right than with a naturally aspirated or supercharged car, but it's far from an insurmountable problem.

Oh, as for cars to look for rides in to see what a well-put-together turbo system is like, I suggest a Miata with a Flyin' Miata FM2 kit. Don't look at a Mazdaspeed Miata, though -- somehow Mazda managed to combine a turbo with lousy top end with one that had lousy spoolup characteristics and get the worst of both worlds on that car.

Keith Tanner wrote: There's a way around this with centrifugals but it's a bit goofy - you overspin the compressor and start venting boost.

The right way around it would be to use a planetary transmission or CVT to change the supercharger speed. Not very common as it's about the most complex boost control out there.

Positive displacement superchargers actually suffer worse than naturally aspirated cars with altitude gain, as the supercharger typically drops off in efficiency. We saw that with a couple of cars that were tested at our shop and again at sea level.

There are some of those Millenia compressors on eBay for $150 or less. 2.3 might be a decent size for a 2.2 engine, you wouldn't have to spin it too hard.

Whipples are less efficient at cruise and more efficient at WOT than Roots, although I'd still rather see an intercooler on all of them.

Which brings up another wrinkle. Positive displacement superchargers have the throttle body before the supercharger (usually). Turbos and centrifugals have it after the compressor. This means that an intercooler is throttled volume on the former, which can have driveability effects. You can put the throttle body post-SC, but holy cow does it make noise and you need a very big blow-off valve so you don't disassemble your intake piping when you change gears. I know there's been some Rube Goldberging going on to make this work with dual throttles and the like, but generally speaking it's not ideal.

In reply to Keith Tanner:

Very interesting. I wonder whether I'm just thinking wishfully to wonder whether the throttled volume of an intercooler and attendant plumbing on a supercharger setup might introduce temporal effects on the same order of magnitude as lag on a conservatively sized turbo.

My M42 is only a little 1.8, so still concerned the Millenia SC might be too large. Do the displacement ratings relate between twin-screw and Roots? Eaton seems to suggest (third-to-last QA) twin-screws need to be spun faster... But they would, wouldn't they?

In reply to ransom:

IIRC, the DASC was either an eaton M42 or M62 supercharger.....no intercooler AFAIK, but I think a few guys adapted liquid to air for them. The most powerful one I remember didn't get past 220hp. That mark should be very easy to reach with a short lag time on a t25-26 sized turbo. A rotrex centrifugal might even be a good choice.

There were several guys thinking the DASC was the best thing in the world on their ti's, but Jeebus they cost a lot for the kits for the power levels of a s50 swap.....