i think a clutch type or viscus (lsd or torque converter style) center section would be ideal. it would give you some “slip” while it engaged and disengaged.
heat would be a factor for both of those units and they change their “clutching” properties with heat and also the viscous unit would add “weight” or resistance with change in temp.
adding any rotating mass to the turbocharger itself is going to decrease efficiency and increase lag as well as make boost harder to control and reduce the responsiveness of the turbo in general. You would be best off to use a clutched supercharger in conjunction with a turbo that kicks out once the turbo is spooled, however these systems are massively complex and really not worth the time and money/effort as better results are usually had with a properly designed turbo system in the first place. Drive somthing like a new bmw 335i which has a nice twin turbo setup and a turbo/supercharger combo setup will become pointless; the responsiveness is good enough you really dont need to supplement it at all
A high speed electric motor will do a good job of spooling the turbo up faster. My masters paper was about research done in using electric assist to build boost faster. It’s an interesting idea to use in hybrid vehicles that have battery capacity. In diesels if you can get the boost up faster you can give it fuel faster without incomplete combustion (smoke). Allows for better performance while keeping emissions down.
more moving parts at high rpm is not always the best thing to think of.
there was a movie with chuck norris and he had a dodge ram with a clutched supercharger. you are about 20 years late
naysayers that arent engineers and dont understand his question… get lost. you havent got a clue.
he isnt thinking of using a standard a/c clutch at all. he is thinking of something like the freewheel setup on a bmx and 10-speed bike. you dont need to use a clutch at the turbo shaft. you probably need it elsewhere.
1st gear in a dodge 3-speed automatic uses a freewheel setup. its a bunch of rollers and a housing with ramps that “catch” a shaft when power is applied and allow it to spin freely when the power isnt applied. its called the overrun clutch.
its possible and will require a much wider custom dual-cartridge ball bearing center section and the turbine wheel/shaft assy has to have a longer shaft. you place a cartridge near the turbine and another near the compressor and leave the center of the shaft open for the overrun clutch assembly and a gearbox setup. youll also need to provide oil to the bearing cartridges and the overrun clutch setup.
the theory part of things. you calculate the airflow needed (lb/min) at say 1500-2000rpms for the engine, taking volumetric efficiency into account.
then look at the compressor map for the turbo that you are using and look at the airflow at the bottom and the pressure ratio on the left side that you want to have at your low rpm point. at that point (or near) should be another line telling you the shaft speed to make this pressure based on the airflow. set your gearbox that drives the turbo shaft to meet that compressor speed at the rpm point. you will want something like an ac clutch to turn off the system when at idle and when the full spool boost is reached. given the gearbox ratio that you need to get actual boost at such a low rpm, you will overspeed the turbo with the engine high rpms.
the system will work should work well though it will be difficult and expensive to manufacture for an individual and difficult to plumb.
more realistic options?
-small shot of nitrous on the bottom end and shut it off at full boost.
-VNT turbo. greatly increases the velocity of the nozzle aimed at the turbine. i think a newer powerstroke ford uses a VNT turbo. take the exhaust housing from that and a T4 or GT compressor and mate the two. should work fairly well and be reasonable for a custom turbo shop to put together.
-HX35 with a modified divided turbine housing (or any turbo with the divided housing like the evo). block off one side of the turbine housing with an inconel throttle plate in the low rpms till the boost comes up. 
Brian
^^You totally nailed what I was talking about.
dont they make variable vein turbos for that? if you want quick spool off the muscle y not just add a little shot of nitrous to get it spoolin?
i thought of this back in college. then discussed it at length with carnut. i think everyone has thought of it at one point. In theory, it works, but in practice it’s very difficult… it’s just more moving parts to break…
the idea of having spring loaded cogs cycling at 50K hertz is laughable. the people talking about the bearing loads don’t really know what they are talking about, though. There are a number of couplers that could be used to eliminate any tangential loading to the bearings.
gaseous fuel (propane) and oxygen injected into the hot side would probably work far better…
as far as current technology, it’s better from an engineering standpoint (as a company, taking into effect mass production and cost) to develop existing technology to better suit needs. for example, introducing Variable Vane technology and sequential turbo systems and tuning them to act like the ideal system, which is what you’re looking for with the freewheel idea.
i actually really like this idea, it could be something one day.
not really related, but it’s a cool idea nevertheless that’s been a little more developed.
you mean like a 10 speed bike? yeah that wont hold up lol. the overrun clutch i was talking about is different though. ill see if i can find a pic.
yup.
even a CVT style gearbox on a vortech/paxton/procharger etc supercharger would be easier to implement. a little mircoprocessor magic to control it and youd be all set. the CVT would only have to operate at 8k max so it should be fine. the blower gearbox would do the rest and that is already designed to work at the high speeds. full boost off idle and no spiking would be nice.
yea that one is pretty cool too and probably easier to control than my CVT idea. its less efficient than a belt though but that doesnt matter that much. 
Brian
Keep in mind it would take a large amount of energy to spin the turbine at a point where it would generate boost. And what you would end up with is basically this:
With possibly the benefit of a turbo assuming it worked as well as a conventional turbo.
Given the tolerances, and speeds one would be hard pressed to make it reliable or spin fast enough. Might as well look into twin-charging at that point.
Wouldn’t it have to be geared like 50:1 for the engine rotation to drive a turbo? Is that realistic?
basically what he wants to do is make a procharger that run and makes boost at low rpms and then once rpms start to climb, have the exhaust turbine take over as it exceeds the speed of pulley. Its really not a terrible idea in theory.
Lots of ideas sound good in theory, hence why I posted. There would have to be gear reduction to make it work. Twin charging sounds good, but a lot more parts(maybe not) and would require actuators and extensive piping for the system be efficent.
Admittedly I have skipped the second page, but would a “1-way bearing” utilizing sprags be more along the lines of what you are thinking Articzap? (Think 1 way, spring-loaded slip roller, no moving parts) I am positive I know exactly what you are talking about. I think you idea could be more constructively defined as a “lag eliminator”. While I cant see any type of gear driven Turbocharger style compressor brought to high enough speeds to make sizeable power mechanically, I can see it taking off quite a bit of the time needed to spool.
I am well aware of the rotational speeds that a Turbocharger can see, but I cant see how a setup such as the above would influence this negatively other than excessive complexity. There would be no increase to the shaft rotational weight, and bearings such as these are already widely used in applications that can easily see 40,000+ RPM. Not quite Turbine shaft speeds, however not so far out of the question that it couldnt be done.
Discuss?
The fact that adding anything to the rotating mass of the turbo charger will hinder it’s spool, the fact is if you want to accomplish this using some sort of mechanical aide, a CVT setup in my opinion would be the best method. Anything you do to add components to a simple machine will complicate the unit. Everything has its positives and negatives.
Dont know for sure if you are replying to me, but if you read my post, there would be NO change to the rotating assembly mass, and the most drag it would see during free-wheel would be about on par with adding a roller bearing to the center of the shaft with no race.
I am not saying the whole idea will or will not work, but the idea of driving a shaft with a 1-way sprag bearing is sound in my opinion, even at those rotational speeds.