Two stage Superchargers

[Edgar Brooks]

Better to stick to proper nomenclature and not dumb down everything.

Better not be too rude to others, either; if Rolls-Royce refer to compression ratio, in their engines, then who are we, following on behind, to put ourselves above them? Note that compression ratio and pressure ratio are not the same thing, either. CR measures the decrease in volume, while PR measures the increase in pressure; the two figures are never the same.

Then that engine in the diagram has an aftercooler

And it was called an intercooler (again by the manufacturers) because it was interposed between the compressor and the engine.

Edgar

 

[PWR4360-59B]

OK guys, enough already. The combos are one stage with single, two or three speed drives. Two stages single or two speed drives some with intercooler. The correct answer on the boost or compression is additive with these aircraft engines. The merlin two stage will generate 5 ata of boost before the drag of the unit drops the shaft HP. see reno engines

I think what everyone is talking about is: say stage 1 is a 2:1 step up, stage 2 is say a 3:1 step up. In the first stage (all with zero losses at 14.7 atm) is 29.4 psi, then stage 2 multiplys that to 88.4 psi.. Then even when the set cylinder compression ratio see's the new stepped up number it multiplys that. Ratios are a thing that involves multiplication and division if your going a different direction of course. There is nothing ADDITIVE?????

 

[vanir]

Hey shortround you'd know, the 3-speed blowers on US navy birds are actually a two speed two stage with neutral aren't they? I mean the gear in the pit says hi, lo, neutral and those are the three speeds, so that's a two stage with two gears and a clutch on the second stage, right?

 

[Shortround6]

That is correct. It can also be seen in the power outputs. 2000hp to the prop in neutral at low altitude, 1800hp at 15,500ft in low gear (200hp to drive the supercharger) and 1650hp at 22,500ft in high gear ( 350hp to drive the supercharger). The engine was turning the same rpm and had the same manifold pressure. perhaps a bit of power was lost due to charge heating inspite of the inter cooler. The 1st stage (engine supercharger/main blower) had one gear and was always turning.

 

[vanir]

I know I'd be pushing it but do you think the same kind of nomenclature relates to the Jumo 213E/F? There's one blower, then a second stage blower on multiple gears with a neutral? Or do the multiple gears function in each pressurisation stage?

Might be worth reminding people that centrifugal blowers don't force induction to the manifold like roots blowers do, so aero blowers in series aren't like many imagine, it doesn't raise all this positive manifold pressure. It's just about altitude and nothing more. Centrifugal blowers compress the air charge inside the casing, it is drawn into the manifold with normal induction vacuum, but in a pressurised or compressed state, hence the german word kompressor for this type of blower. Doesn't push the air. It just compresses it.

So when you have thin air you compress it and it's still thin and doesn't make much power, because you're very high, you get the compressed air and run it through another blower in series and compress it again, whammo it now acts like sea level air and makes tons of power. Roots blowers on drag cars blow the air into the manifold, they're much less efficient. Most of what people imagine about blowers actually holds true for roots blowers but not centrifugals, which function precisely like multiple stage turbos rather than classical (diesel) blowers.

I'm rambling a bit now but what the hey; this is also all why centrifugals (and turbos) love intercoolers and aftercoolers, but doesn't do much of anything for a roots blower on a drag car. What does those babies is water injection at the inlet, the vanes heat up, so blower temperature is a major horsepower drag. In centrifugals and turbos the blower is oil cooled and the charge itself heats through compression, so you need cooling at the blower exhaust of preferrably each stage, but at least one blower exhaust before it hits the manifold.
These points tell you things like how the Daimler and BMW approach to overboosting are very different.

 

[Shortround6]

I don't know about the Jumo engine, we know the Merlin two stage engine had both impellers on a common shaft and both turned at the same speed. There was a Griffon with a 2 stage 3 speed supercharger which may have had the impellers on a common shaft as the two stage two speed Griffons did.

I believe you are incorrect about the centrifugal blowers, while they are not positive displacement like the Roots, a Single stage centrifugal blowers they will certainly make positive pressure in an intake manifold. A big difference, which is immaterial for aircraft use, is that any particular centrifugal compressor has an effective upper limit on the pressure it can create. Say, for arguments sake, a particular centrifugal compressor has a 3:1 pressure ratio. It can deliver 3 times the input pressure and that is it. Even if you block the output duct/s you are not going to get much more than 3 times the input pressure becasue not only does it depend on moving air (imparting velocity to the air) to make pressure there is enough extra space in the compressor housing for excess pressure to bleed back to the intake side.
The Roots compressor moves a "slug" ( a fixed volume) of air through the compressor for every revolution of the compressor. The faster you turn it the more air you move, the pressure rise doesn't actually occur until the air is going out the discharge. If you turn the Roots supercharger slower than the engine needs air it doesn't really develop any pressure, If you turn it faster you than the engine can use air normally you will get positive pressure in the intake manifold. Here is the real difference. Take both compressors and turn them with electric motors with the internal combustion engine turned off. The Centrifugal will hit it's limit and then there will be little or no change in maximum pressure. The Roots will continue to build pressure, at least to a much, much higher point.
A big difference between the Centrifugal and the Roots is that while the Roots will double the amount of air it moves if you double the speed of it's impellers, the Centrifugal will move four times the air if you double it's speed, at least in theory.

 

[vanir]

Believe me I was having a real hard time trying to explain what I was trying to visualise.

I know their application, I can fit them and build an engine depending how we're going to set it up, roots or centrifugal or turbo. For me the big difference is the "band" in which centrifugals operate which is very much like a turbo, and graphing them you do it on a turbo chart. For roots you use a different system and it's right across the rpm range. So you have to size blower outer casings with different impeller sizes for various stages to overlap your curves.

This makes them perfect for aero engines because they're much more efficient, but only within a finite band of rpm based on flow rates and relative casing/impeller sizing, boost a separate issue entirely but indirectly associated being about turn speed. This isn't a problem for aero engines because they use a small band of rpm anyway.

So for me planning an engine build one is like a fan forced induction and the other is like a compressed air bottle maybe, they're different to work with in what you imagine is being done with the air charge but it's hard to find the right words.

Say for example, manifolding and head machining has almost no effect with a roots, some but so very fractional, but with a centrifugal the marked difference is shocking, a little head machining and you need to up the casing size and rematch an impeller, you can really fine tune with them. Roots are much too brutal for pansy things like that, just feed me is all they want to say, throw me on and feed me with a dirty big set of carbies.

 

[Shortround6]

Well, in theory, a Roots will match the engine across the rpm band. it will flow twice the air at 4000rpm as it does at 2000rpm and three times at 6000rpm. The Roots will supply boost at idle, subject to other restrictions. So, again in theory, with the right blower casing, impellers and drive ratio to crankshaft speed you could get a Roots compressor to give you, say, 6lbs of boost anywhere from idle to max engine speed. If you double the engine speed and thus double it's need for air the speed of the Roots will double and so supply twice the amount of air. Things are never so perfect in real life
The centrifugal on the other hand, increases flow with the square of the speed, so as and engine builder you have to make choices, a small centrifugal that provides no boost at low rpm (low air flow) but gets to the boost you want at max engine rpm, a larger centrifugal that gives the boost you want just of idle but provides way too much air at high rpm and needs blow off valves or other stunts to keep from over boosting and a middle of the road approach that still requires boost control. The big centrifugal is large, heavy, has lots of inertia (like adding a heavy flywheel to the engine) and sucks up power while venting compressed air to the outside.

As you say, aircraft engines operate over a smaller rpm band than car engines. I wonder, though, if the improvement in the centrifugal engine from better inlet flow (bad word but what do you call the the plumbing from the supercharger to the intake valve? it is still the engine intake but it is the supercharger discharge) is due to helping things along at less than full boost? I mean, if you have two engines and one has a roots giving say 6lbs boost at 2000rpm and going up a bit when it gets to 6000rpm vs the same engine using a centrifugal that gives 2lbs at 2000rpm but gives 6lbs somewhere short of 4000rpm and blow off valves keep it from going much over 6lbs the rest of the way (no carbies, fuel injection, venting fuel/mixture is a bad idea would more manifold/head work really give more power just to the centrifugal at 4000-6000rpm or would the advantage be in the 2000-4000rpm range?

I don't know, just going on theory not experience.

 

[vanir]

Oh no it's not like that at all, they've an efficiency curve like a turbo that's only about 3000rpm wide on modern ones. War era ones are just less efficient overall so it's subtler but important. A roots has no such thing, its sweet spot is just rev harder or change gear drives.
They have an island cfm rating and max pressure ratio and the boost limit for their impeller/casing ratio, with the rpm on a given engine this boost is available decided by blower speed (internal gear ratio and belt ratio). You size casing to engine cfm demands by how much boost you want, impeller and gear ratio to the rpm of that engine you want the maximum boost at. It only comes in that band of a few thousand rpm, outside of it the blower isn't doing much or not enough of it or something, plus this sweet spot is right up near maximum impeller speed so this is like a caged section of performance you have with centrifugals (and turbos), I presume this is why roots are rarely in series but centrifugals often are.
I'm not being patronising, I'm just trying to extrapolate how modular they are.

I've got some Merlin 25 specs and whilst the pressure ratio goes up (1.34@2800engine spd/19K blower speed, 1.5@3200/22K, 1.7@3600/24.5K, 1.86@4000/27K), the blower efficiency peaks and has its sweet spot right in the middle there, just like a turbo (55%@2800, 59.5%@3200, 62%@3600, 59.8%@4000), so you get to a point, in a Merlin normally operating at sea level on a dyno at about 3800rpm where it's just superfluous to rev higher. With ram air maximum efficiency pulls back down to 3000-3250rpm and that's right where you want it in flight, on the bench it's about 3600. Maximum volumetric efficiency is 127% at 3600rpm (125/3200, 122/4000) But boost keeps going up, at 4000 it's 12.1psi, at 4400 it's 14.7psi but that's graphed, it won't stop pinging past 3400rpm on 104ron fuel SA conditions and perfect spark.
Exhaust port pressure flies off the scale at about 3200rpm basically, its harmonics and flow dynamic gets all messed up.

So you see I start playing around with head machining and valve overlap for dynamic pressures we're talking some pretty detailed fine tuning I can do with that blower. But it's because of its tuning band, they have a little performance band, very unlike a roots. Roots don't even like headwork and they sure don't like overlap, they're just a big huffer and that's all.

 

[Shortround6]

Well, I did say theory

I Know that modern turbos are modular and I think I was getting at that the centrifugal could not supply the desired boost over the entire normal operating band of a car engine, you can (have to?) pick one part of the rpm and optimize it for that part and try ( by playing with the housings, impellers, etc) to get as wide a band as you can.

Something doesn't sound right about your Merlin example. We are talking about the R-R Merlin 25 airplane engine aren't we? The Merlin 25 supercharger could hit a pressure ratio of 3.1@3000 engine rpm/28.47K blower speed. 18lbs boost at 9250ft at 3000rpm. 9.49 drive gears to the supercharger. Merlins were limited to 3000rpm under power, at least from the factory, what some racers or tractor people do with them may be another story. They were allowed an over speed of 3600rpm in a dive but that is with a part closed throttle and the propeller driving the engine.

 

[wuzak]

I don't know much about Roots superchargers, but I believe they are generally less efficient, or maybe less effficient at high pressure ratios.

In any case, in 1939 Mercedes-Benz improved their W154 Grand Prix car by changing from the M154 V12 with two parallel Rootes type superchargers to the M163 with a two stage system with two Rootes type superchargers in series.

In 1950/51 the Alfa Romeo 158 with one Rootes type supercharger became the 159 with a two stage system of Rootes type superchargers.

 

[vanir]

The figures I have are dynamically modelled from factory specification with engineering software, designed to tell me things like thermal efficiency peaks at 2800rpm and even with perfect spark pinging starts in at 3200rpm with a piston speed of 3733fpm so despite the fact that calculated hp continues to rise with rpm to a peak of 1770 at 4000 as far as flow dynamics goes, there is overpressure in the exhaust port and high chance of engine failure causing an engine speed limitation of 3200-3600rpm absolute (assuming unbreakable conrods), or a 3000rpm operational guideline. This returns a 1610hp peak.
It also tells me head flowbenching is geared for precisely 3000rpm (residual exhaust % is lowest at this engine speed). It also tells me I could aim for about 1770hp in overboost at 3000-3200 if I wanted that facility.
It says something that Merlins are actually famous for, at least with SEAC because of poor fuel quality in Burma. That this engine would not at all like any reduction in fuel quality, british 100/1941 (100/130 or ammendment 5). Standard aviation gasoline would restrict ideal engine speed to 2800rpm and limit boost. Historically Hurricane II's and IV's sent to Burma had to be detuned like this so had shocking performance in the theatre until the late war when ground had been retaken and supplies brought up.

The figures are all calculated however.

 

[GregP]

Vanir,

What are you taling about? The thermal efficiency of WHAT? if you're going to have a technical discussion, shouldn't you identify what you're talking about?

Are you takling about an aircraft engine? if so, what engine? And what is the point you are trying to make in the post?

Just curious since I can't tell from the text.

 

[vanir]

Greg if you trace the conversation back a few posts it'll all make sense.

 

[Shortround6]

The figures I have are dynamically modelled from factory specification with engineering software, designed to tell me things like thermal efficiency peaks at 2800rpm and even with perfect spark pinging starts in at 3200rpm with a piston speed of 3733fpm ...................

The figures are all calculated however.

Some thing seems wrong with these figures? Merlins didn't run at over 3000rpm in service use and at 3200rpm it would have a piston speed of 3200fpm. and as noted before the pressure ratio of the supercharger and the supercharger speeds don't match up withe gear ratios the Merlin 25 used.

 

[vanir]

Engine modelling always exceeds real world limitations, that's the idea. If you could build a magical unbreakable Merlin an engine model will say what it can do at 3500 or 4000rpm given its specification. From the mapping you see the most likely cause of engine failures in service conditions by predicting fictional ones.

You might be right about the blower but, it looks like they've customised figures to use the software's automotive blower mapping shell for the big Merlin aero motor so that it turns out performance mapping which matches the factory figures. It's mentioned in the software help pages. This one had to have belt sizing, impeller ratio and island cfm ratings among other race engineering values entered to be represented using the software engine, so the blower figures are probably off.

 

[Shortround6]

I can understand using computer modeling to explore operating areas that were never reached in service. It is when the computer model has some rather large differences in what was actually done that I start to doubt it's accuracy. If it can model the engines as built and operated then it can use that data to predict performance. If it is using data that doesn't reflect what the engine could really do (or what it couldn't do) then any projection of that data doesn't seem to be much good. The piston speed thing is pretty basic, Merlin had a 6in stroke, 1 FPM for every RPM. to get 3733fpm at 3200rpm the engine would need a 6.999in stroke.

 

[vanir]

Got me there mate. It's automotive race software, they reliably match calculated with automotive race builds dyno figures, I know that much I've built engines using it (saves you a fortune), I dunno what the story is with them putting a Merlin 25 in the engine build library. Their program shell is really geared for 1-8000cc. The blower mapping engine is meant for automotive race types.

 

[Siegfried]

Roots type superchargers have a more linear output in proportion to RPM, likewise with the Lysholm types and the G loader scroll types seen on current Mercedes Kompressor models.
This is important in terms of drivabillity compared to the non linear output on centrifugal or axial types. I think the big Jumo 205 diesels used roots blowers. For turbo superchargers the waste gate linearizes the output.

 

[vanir]

Thanks Siegfried, that puts into perfect words what I was trying to say a little while back. I really overcomplicated things by introducing the modelling while trying to find the words, effectively trying to demonstrate what that is saying.

There's a high alt prototype that uses 2x 603 engines for thrust and a blank 605 engine is used just to drive a massive roots blower that second stages them. It was a ca.14,000m alt recon proposal I think.
Anyways it definitely appears the aero industry had roots blowers kicking around for specific applications, just that with the limited rpm band of aero engines in general, you get better efficiency out of the "non linear" centrifugals. That difference can easily be shown with computer modelling.