23 replies to this topic

[ausnotions]

The use of 2 stage supercharging on the pre WII GP racing cars has always fascinated me but I have never understood the thinking used to select the size and drive ratio for each Roots supercharger used. I post in the hope that someone else has been similarly challenged and will share their findings with me.

From photo's of the 1939 Mercedes M165 and the 1951 Alfa Romeo T158 the second supercharger appears to be half the size of the first. These are but two examples of using a larger first stage Roots supercharger blowing into a smaller second stage unit.

I can find no reference mentioning the actual cc capacity of each size supercharger, or the drive ratios used, on these 1500cc engines.

Does anyone know the formula - or perhaps advise me of a book reference or tech paper - that outlines how to size 2 Roots superchargers to a particular engine capacity to achieve this 2 stage supercharging?

Ted

[McGuire]

Lots of WWII-era aircraft used two-stage centrifugal blowers, so there is considerable technical lore about them in circulation. Of course the Roots blower is a constant-displacement machine while the centrifugal blower's output is roughly parabolic, and in aircraft engines the issue is altitude...but beyond that the principles are more or less the same. Here is the NACA paper by Dick Buck:

http://naca.larc.nas...41/naca-tn-794/

[ausnotions]

A good read and interesting that they actually tested a Roots supercharger on a Liberty V12 aircraft engine. Pity they did not progress the testing to include a second stage. Thanks for the response and link. Ted

[WPT]

A text book on thermodynamics may be of some help. To compress a given quanity of air, say at std atmosphere, to a given pressure takes work. The process which would use the least work is an isothermo (constant temp) one. It is difficult to cool the charge during the compression stage, so use a two stage setup with an intercooler between the stages. The specific volume of the air (inverse of the density, so units of ft^3/lbm) of the air at entry to the first stage is greater than the specific volume of the air at entry to the second stage, and the stages are in series, so the mass flow through both stages is the same, hence the second stage is smaller. This idea is well represented in a pressure (P) vs specific volume (v) diagram for a two stage reciprocating air compressor. WPT

[McGuire]

Originally posted by ausnotions
A good read and interesting that they actually tested a Roots supercharger on a Liberty V12 aircraft engine. Pity they did not progress the testing to include a second stage. Thanks for the response and link. Ted

Quite right. Here is the NACA paper detailing the supercharged Liberty project:

http://naca.larc.nas...aca-report-230/

The Roots blower basically went nowhere for aviation use as it produces no internal compression (and therefore is not a "true" supercharger, some sticklers will insist.) The centrifugal supercharger produces internal compression, while operating as to the square of its rotating speed -- in other words it produces boost in a relatively narrow band relative to the engine's rpm range. However, this is not a concern in aircraft engines since they operate at relatively constant speed and load.

So the Roots blower is capable of producing boost over a broader engine speed range, while the centrifugal blower is somewhat more efficient over its operating range.

Where I meant to point with the first link -- the paper underlines that beyond the speed and displacement of the blower(s), the biggest complication in ciphering out a multi-stage supercharger setup is the heat rise in the intake air between the stages. That is exactly WPT's point as well. The second link above provides more detail on the properties of Roots blowers themselves...which may help flesh things out further. The arithmetic is pretty straightforward and self-evident if you know the displacements, drive ratios and efficiencies of the blower(s). But for pre-WWII GP engines those specs can be pretty hard to come by.

If anyone cares... the Roots blower was invented in Connersville, Indiana (about halfway between Indianapolis and Dayton) in 1854 by two brothers, Philander and Francis Roots (not Rootes). They were attempting to devise a more effective water wheel to power their textile mill on the Whitewater river (which falls about six feet per mile) and came up with the twin-lobe configuration. However, water caused the wooden impellers to swell up and seize in the casing, naturally, but while trying to dry the thing out they discovered it made a mighty fine air blower.

[NTSOS]

In the April issue of Automovive Engineering, there is an interesting article, "BMW Beats the Lag"....about a fairly sophisticated 2-stage, large/small turbocharged application. It's a common pratice for diesel powered tractor pullers that use a simplified version to great effect.

John

[Engineguy]

I spent some time at the Connersville Roots plant about 12 years ago when I investigated a problem for them. The roots type blowers I was working with were pretty big; maybe 14-15" between lobe centers. The most fascinating thing to me was watching them hand fabricate large centrifugal blower impellers. Yes, it's ironic, and Roots Centrifugal Blower sounds like an oxymoron, but Roots manufactures centrifugal blowers too. Too big to cast with adequate strength I suppose, they use power hammers and Ironworkers to sculpt, freehand, those complex swooping individual blades out of 3/8" steel plate and weld them all into impellers about 36" diameter (I really dont want to know how many RPMs they turn ). They had wood/epoxy models they constantly laid the blades on to check the shape untill they got it right. The check models looked to be about 50 years old.

[BRIAN GLOVER]

This is a fascinating BB, isn't it?
Thanks guys.

Originally posted by Engineguy
I spent some time at the Connersville Roots plant about 12 years ago when I investigated a problem for them. The roots type blowers I was working with were pretty big; maybe 14-15" between lobe centers. The most fascinating thing to me was watching them hand fabricate large centrifugal blower impellers. Yes, it's ironic, and Roots Centrifugal Blower sounds like an oxymoron, but Roots manufactures centrifugal blowers too. Too big to cast with adequate strength I suppose, they use power hammers and Ironworkers to sculpt, freehand, those complex swooping individual blades out of 3/8" steel plate and weld them all into impellers about 36" diameter (I really dont want to know how many RPMs they turn ). They had wood/epoxy models they constantly laid the blades on to check the shape untill they got it right. The check models looked to be about 50 years old.

[Ray Bell]

Originally posted by BRIAN GLOVER
This is a fascinating BB, isn't it?

You can learn a lot here!

Connorsville? Roots? You could win bets on that information... put a quid on the bar in a pommie pub inhabited by ERA fans and see what I mean!

[McGuire]

If I recall the second NACA paper posted above will describe how to calculate the output of a Roots blower given the dimensions. Of course in the case of old GP cars, those dimensions will be hard to come by.

But I just remembered a convenient cheat: If you can find a cross-sectional end view of the blower and establish a scale for it, you can find the clearance volume and from that the blower's displacement. One of the impellers will invariably be shown in the vertical position (or if not you can change that with a little transposition). The area between the impeller and the case is the clearance volume and that, times the length of the case, times two, is the blower's displacement. Simply blow up the drawing on the photocopier to a convenient size, note the new scale, lay some graph paper over the area and count the squares.

[McGuire]

Originally posted by Engineguy
I spent some time at the Connersville Roots plant about 12 years ago when I investigated a problem for them.

Great story, thanks. I love walking into small factories and job shops and seeing the impossible performed by the few with practically nothing.

At one time Connersville was quite an automotive town. While the home office of Auburn and Cord was in Auburn, up north aways, the cars were actually built in Connersville.

[McGuire]

I seem to recall reading recently (for me that could be any time in the last three years) that the Roots plant in Connersville was being closed by Halliburton or Dresser Industries or whatever conglomerate owns them now. Inevitable I guess but still a shame...an entire technology, and a way of life even, is disappearing. I fear that eventually the USA will forget how to actually make stuff.

[Ray Bell]

Originally posted by McGuire
Great story, thanks. I love walking into small factories and job shops and seeing the impossible performed by the few with practically nothing.....

Maybe you can solve this mystery then... TNF couldn't...

http://forums.atlasf...s=&postid=17560

[Engineguy]

Originally posted by Ray Bell Maybe you can solve this mystery then... TNF couldn't...
http://forums.atlasf...s=&postid=17560

Regarding Anderson, Indiana...

Many other inventions were perfected in Anderson including: the gas regulator-Miron G. Reynolds, the stamp vending machine-Frank P. Dunn, clothes presser-H. Donald Forse, Irish Mail-Hugh Hill, flower car for funeral homes-Francis M. McClain, automatic gearshift-Von D. Polhemus, Sisson choke-Glenn Sisson, and the vulcanizing process that retreads tires-Charles E. Miller.

Anderson grew to such proportions that a Cincinnati Newspaper editor labeled the city "The Pittsburgh on White River." Other appellations were "Queen City of the Gas Belt" and "Puncture Proof City" (because of the vulcanizing done here and the rubber tires manufactured).

[Ray Bell]

Ah, thank you!

This has been bugging me for about six years. A guy sent me a copy of a page from the Dykes Cyclopedia advertising the Laurel 'roof' 4-valve head for T-models, which was made in (you guessed it!) Anderson Ind.

The line was at the bottom of the ad...

[McGuire]

Well whaddya know. I'd never heard of "the puncture-proof city," and I've been to Anderson dozens of times, and through there more times than I can count. (It's northeast of Indianapolis on I-69, about midway to Muncie. I'll be through there 10 times this May for the 500.)

It's definitely a racing and automotive town though. Ray Harroun lived there for many years, and it's the home of Anderson Speedway, home of the annual "Little 500." (500 laps with midgets on a high-banked 1/4 mile oval, totally insane. They run figure-8 racing there too. ) A number of cars were mfg'ed in Anderson, including the Nyberg, a big swell 60 hp touring car around 1912-15. A Nyberg ran in the 1912 500 I believe.

Anderson and Connersville were both in what was known as the "gas belt"... in the 1880's from Kansas to Findlay and Tiffin, Ohio they hit natural gas everywhere they drilled. They thought it would last forever -- some towns installed gas streetlights that burned night and day to show off and attract industry. One town in eastern Indiana even renamed itself "Gas City." But of course it didn't last.

[Engineguy]

Originally posted by McGuire
It's definitely a racing and automotive town though.... it's the home of Anderson Speedway, home of the annual "Little 500." (500 laps with midgets on a high-banked 1/4 mile oval, totally insane. They run figure-8 racing there too. )

Correction... 500 laps with 33 sprint cars, not midgets, on a 1/4 mile oval... 11 rows 3 abreast at the start, on a track that's about 4 cars wide if they're touching sidewalls. Like you said, totally insane. 33 cars would be heavy traffic on a 1/2 mile oval. It used to be more famous when it was by far the biggest purse in sprint car racing... then Earl Baltes started doing the gazillion dollar purses at Eldora, so you don't hear about the Little 500 as much. It's now the biggest purse in paved track sprint car racing. I've lived within 50 miles all my life, but have never got around to going to it. A lot of people attend religiously; claim it's the best racing action anywhere.

[McGuire]

Originally posted by Engineguy
Correction... 500 laps with 33 sprint cars, not midgets, on a 1/4 mile oval...

Thanks for the correction! Can't tell you why I typed "midgets," been to the Little 500 many times...maybe I was thinking of Night Before the 500 at IRP.

[Ray Bell]

This isn't going too badly for a thread about two-stage supercharging...

Very informative, a credit to all contributors!

What trivia is next?

[McGuire]

struggling back toward on topic....

...in the early applications of Roots compressors you will see some strange things by modern eyes...blow-through setups, which creates a large problem with a positive-displacement blower if the the throttle is suddenly slammed shut, unless a large bypass circuit is employed...also, very high drive ratios, sometimes close to 3:1. Roots blowers are generally not very efficient at high speeds, generating a lot of heat. Of course these compressors were invariably two-lobe, straight-rotor machines, not as efficient as the three-lobe, twisted impellers used today.

So the two-stage Roots blower setup was an attempt to increase efficiency. The typical setup used two blowers, a primary and a secondary, with identical drive ratios. These blowers would usually be of identical/similar case and rotor dimensions, except the secondary blower was of half length. (Technically speaking, slightly greater than half.)

Many of the twin-supercharger setups were not really "two-stage" but straight tandem setups: here two similar-sized blowers fed into a common intake manifold, or the supercharging duties were split in half, each of the two blowers taking half the engine as in the Alfa Tipo B straight 8.

[McGuire]

More on Roots blowers...

The impellers (which note: they counter rotate) take in air at the inlet (usually depicted at the top in illustrations, as shown below) pump it around the outside of the case and expel it at the outlet, at the bottom. The clearance volume is the area captured between the impellers and the case. So the blower's displacement or output per rotation is the clearance volume times the number of impellers (two) times the number of lobes on each impeller (two or three).

So the Roots blower is a constant-displacement pump. With every rotation, it takes in a fixed volume of air and delivers that fixed volume of air, pretty much regardless of speed. Thus the Roots blower can develop usable boost over a wide range of engine speed, even at very low rpm.

However....along with the clearance volume, the Roots blower also has a backwash or carryback volume, where the lobes mesh at the center of the case. Note that here the counter-rotating impellers will try to pump a small portion of the air back out through the intake at the top. Since in order to produce positive manifold pressure aka "boost," the blower must pump more air than the engine can take. The air backs up into the blower as it is compressed and in this backwash a considerable amount of heat will be created.

As a result, the Roots blower is not very thermodynamically efficient -- 55% to 60% at best. The efficiency can be improved by using three lobes on each impeller rather than two (Increasing the ratio of displacement to case volume, and reducing the ratio of backwash volume to clearance volume, among other things) or by twisting the impellers, so that each clearance volume is dumped into the manifold progressively rather than all at once. These two measures also make the Roots blower run quieter by smoothing the air pulses. The third path to greater efficiency, and the one mentioned here, is to use two small blowers rather than one large one, in either a two-stage system or a tandem arrangement.

Another way to control intake air temperature and thus improve blower efficiency is the intercooler, as noted by WPT. Though intercoolers were employed in the centrifugal blower setups used at Indy in the '20s and '30s, for whatever reason (weight? size? fragility?) they were not common practice in GP machinery. I have always wondered about that.

[McGuire]

Originally posted by McGuire


The efficiency can be improved by using three lobes on each impeller rather than two (Increasing the ratio of displacement to case volume, and reducing the ratio of backwash volume to clearance volume, among other things) or by twisting the impellers, so that each clearance volume is dumped into the manifold progressively rather than all at once. These two measures also make the Roots blower run quieter by smoothing the air pulses.

Well, you don't have to twist the impellers. Some bright spark thought of this clever idea:

[Nathan]

A few questions if I may in regards to this great thread...

- In a twin turbo set-up, would there be any benefit to running them in tandem the same way discussed with the belt blowers? That is, run the exit of the intake charge from the first turbo into the inlet for the next turbo?

- What are the differences in power loss in using a single blower versus using two?

- Has anyone tried using a CVT type drive on a centrifugal blower before?

Thank you kindly

[McGuire]

Originally posted by Nathan
A few questions if I may in regards to this great thread...

- In a twin turbo set-up, would there be any benefit to running them in tandem the same way discussed with the belt blowers? That is, run the exit of the intake charge from the first turbo into the inlet for the next turbo?

- What are the differences in power loss in using a single blower versus using two?

- Has anyone tried using a CVT type drive on a centrifugal blower before?

Thank you kindly

1. Yes, but then you have not only multiple compressor stages but multiple exhaust turbine stages. Or one can run multiple stages mixing superchargers and turbos. Very common on big diesels.

2. Depends, but usually multiple or tandem blowers are used to decrease losses aka increase efficiency.

3. Yes. To name one...the McCulloch VS57 (1950's) was a centrifugal supercharger employing a V-belt drive system in which the distance between the driven pulley sheaves was varied to adjust the drive ratio. Just like a traditional V-belt variable transmission, in other words. A later model, the VR57, used a centrifugal ball-race mechanism inside the blower to vary the drive ratio...3.5:1 to 5.5:1 at the impeller.

Also, stepped multiple-speed superchargers have been very common, RR Merlin for example. Through its development it went from a single-stage/single-speed to single-stage/two-speed to two-stage/two-speed supercharger system.