Hypergear Turbochargers and High flow Services Development thread

[Lithium]

No - just gives you an idea (if you think about it a bit) of how much of a difference in shaft speed would be required going from full boost, full rpm to "no surge" with a closed throttle then back to full boost, mid rpm.

Going by this GT3071R compressor map a GT3071R on 17psi on an RB25 would normally operate at an absolute minimum of around 105000rpm to supply full boost at any point in the rev range - going up to >120,000rpm at redline... so maybe around 20,000rpm variance in shaft speed for the full boost operation range.

Shutting the throttle off will make it charge to well <60,000rpm and then when you get on the throttle again it will have to add at least (at a guess) 50,000rpm in shaft speed again. Not sure how the differences would be with a functioning blow off valve and without, might have to trawl around and see if anyone has tested it with a blow off valve and a turbine speed sensor - or even just tested boost recovery with and without a BOV.

[Lithium]

Oh yeah,

All theory, of course. GTScott - be very keen to hear the results if you do some testing, or at least if you guys do the brake loading test what your thoughts are.

[Hanaldo]

I have been intending to put my bov back on... Its just easier said than done. Especially with having to find a new fabricator.

[hypergear]

When the throttle shuts, compressed air will bounce back trys to rotate the compressor wheel backwards and escape from the induction pipe causing that Tuuuutuututu sound. This is not good for a turbocharger as the compressor end is not made as a driven component for the turbine wheel, and I occasionally get BA turbochargers with twisted shafts from owners whom wanted dosing noise. So in this case a BOV is recommended.

How ever I do find that boost take less to build up between gear shifts with bov blocked.

To me the best way to minimize the induction pressure drop is by shifting it quicker at high RPMs.

[GTSBoy]

Both surge (low flow at high PR) and throttle closed flutter are examples of aerodynamic stalling of the compressor. This is a classic aerodynamic stall, like when a conventional aerofoil is at too high an angle of attack and the air flow separates from the back of the leading edge. Stall causes the wing to stop producing lift.

It's obviously a little different in a compressor, as we're not using air flow to generate lift, we're doing the energy conversion the other way around (mechanical energy converted to airflow).

I think the easiest way for you guys to understand it is if I take an example from conventional centrifugal fans. These fans are related to compressors, being centrifugal devices that work on the same principles, it's just the pressure ratios and shaft speeds are drastically different. Take a look at the linked image. This is a fan curve for a conventional fan. The fan curve is somewhat similar to a compressor map. You have mass flow of air on the X axis and you have pressure rise on the Y axis. The performance curve itself is the line that starts on the Y axis at just over the 4HP / 80% efficiency line, curves up to the right and then down to the right, passing off the right side of the graph just below the 1 inch (of water) pressure line. You will notice that the selection range is highlighted. When sizing up a fan, you want the operating point (pressure and flow) to fall in that range. This is very important, and relates to the compressor stall issue.

So let's imagine that we have a design for an air flow system that requires 8000cfm of air delivered at 2 inches of water. This, conveniently, is right in the middle of the selection range for this fan. Such a system might be nothing more than a nozzle at the end of a pipe so that we can blow leaves around, or a snowmaker or provide the combustion air for a burner, or just about anything. If you run the fan at the shaft speed that matches the linked curve, with the nozzle on the outlet that will pass that much air at that pressure, then the fan will operate at that point on the curve.

Here's the best bit. Fan curves make sense. If you open up the nozzle, you reduce the restriction on the system and obviously enough, the flow will increase. If you opened up the nozzle so that the back pressure was only 1.5 inches, then the flow would increase to 10000 cfm. And the opposite is also true. Close the nozzle down and the pressure will rise and the flow will decrease. But there is a limit to how far you can close it down, because at some stage you will reach the top of the fan curve. Continue to close the nozzle down and the flow will continue to decrease - it can hardly increase now can it? There's not enough pressure available to make it flow more. But now that we are on an operating point to the left of the fan curve's peak, we are in a region that no longer makes sense. If you were to open the nozzle up so that the restriction is decreased then the back pressure would drop and the flow would decrease. That doesn't make any sense right? Right. Fan's can't handle it. The fan will want to jump rapidly back and forth between the two operating points on the curve that line up with the same pressure rise, but each time it jumps from the low to the high flow point, the air flow goes through a change that spits it back the other way. Horrible, destructive vibrations can ensue. This fundamental aerodynamic instability is stall.

It is exactly the same when you do something stupid to a centrifugal compressor, such as having it spinning fast, pumping lots of air at a high PR, then slamming a downstream valve shut, killing the flow. The pressure waves do bounce back and forth in the inlet tract and when they arrive at the compressor wheel they jack the operating point waaaaay up above the stall (surge) line. If this happened only once it would be bad enough, but it can happen 20 or 30 times in the gear change time or just dosing it up in front of Maccas. As Stao says, it puts enormous mechanical loads on the shaft.

Compressor bypass valves do have a purpose.

Edited April 8, 2013 by GTSBoy

[Roy]

When the throttle shuts, compressed air will bounce back and rotate the compressor wheel backwards and escape from the induction pipe causing that Tuuuutuututu sound.

Are you sure? Have you witnessed the compressor actually spin backwards? I would be amazed if that were true. You have god knows how much cfm pumping out the exhaust spinning the turbine which I think would stop the compressor ever stopping and reversing . There is nothing to actually cause the compressor to spin backwards

What will happen is it will be trying to compress air into a fixed volume and with a turbo effectively being a low pressure high flow pump the clearances will mean the air will be "chopped" vs compressed and you may get reversion of air pulsing back out the air filter but cant see the compressor wheel rotating backwards

[GTSBoy]

Roy, I don't know about the spinning backwards part, but have another think about what you just wrote. Throttle is closed. Engine has pretty much stopped passing air/gas. The only supply of lots of air at high pressure is the region between the throttle and the compressor.

But really, the wheel doesn't have to be spun backwards. Without some super duper mega exensive laser laboratory with picosecond pulse frequencies you probably couldn't visualise the wheel spin direction anyway. But ignore that, because the wheel doesn't need to spin backwards. All you have is a simple energy balance. You have a certain amount of energy tied up in the rotational inertia of the shaft and two wheels spinning in the correct direction, and you have a reservior of stored energy in the form of compressed air. The compressed air has to work backwards on the comp wheel as it flows back out through the clearance in the compressor housing, in through the vanes and out through the inducer. At the same time, the comp wheel is still trying to pump forwards (against a dead head, so losing that battle). It is no wonder that there is violent aerodynamic effects, and there is no wonder that there are sufficiently large maechanicl forces to twist the shaft. Under normal drive conditions the increase in drive force on the turbine is a gradual one. Much more gradual than the instant slam shut of the closed throttle on boost scenario.

[Roy]

Roy, I don't know about the spinning backwards part, but have another think about what you just wrote. Throttle is closed. Engine has pretty much stopped passing air/gas. The only supply of lots of air at high pressure is the region between the throttle and the compressor.

But really, the wheel doesn't have to be spun backwards. Without some super duper mega exensive laser laboratory with picosecond pulse frequencies you probably couldn't visualise the wheel spin direction anyway. But ignore that, because the wheel doesn't need to spin backwards. All you have is a simple energy balance. You have a certain amount of energy tied up in the rotational inertia of the shaft and two wheels spinning in the correct direction, and you have a reservior of stored energy in the form of compressed air. The compressed air has to work backwards on the comp wheel as it flows back out through the clearance in the compressor housing, in through the vanes and out through the inducer. At the same time, the comp wheel is still trying to pump forwards (against a dead head, so losing that battle). It is no wonder that there is violent aerodynamic effects, and there is no wonder that there are sufficiently large maechanicl forces to twist the shaft. Under normal drive conditions the increase in drive force on the turbine is a gradual one. Much more gradual than the instant slam shut of the closed throttle on boost scenario.

You are an engineers engineer He said the wheel spins backwards, it doesnt. As you point out it effectively chops the air and all sorts of harmonics are going on as the low pressure high flow (ie fans with big tolerances) compressor basically cavitates.

The KE in the wheels along with the engine running as an air pump means that the air on the hot side is still directional and going to keep spinning the turbine. Case in point put your foot over my tailpipe when I am running on the dyno and lift off at 6,000rpm.

[GTSBoy]

Yuh, but it's a tiny flow compared to when the engine is actually doing something. It's just the air that leaks past the throttle/idle valves and there's no combustion products to add lots of heat. That little flow would slip past the turbine pretty easily without putting much torque onto it.

[Roy]

Yep...and then there is the KE in the wheels.

[hypergear]

Once throttle is shut, engine no longer burst hot air into turbine housing. At that instance turbine shaft is rotated by inertial energy. The turbo compressor is a centrifugal pump, so the only way for majority of compressed air to escape is the same way they've been drawn, by forcing the compressor wheel in opposit direction.

Compressed air at that point slams back against the compressor wheel, depending on the size of impact (or how much boost is dosing backing) the size of the compressor housing and blades. while forces cancelling each other, the medium in transferring and absorbing this energy is the turbine shaft and thrust bearing.
o
Which is also why there is a two phase collar in side a bush bearing turbocharger as well as the opposing angular bearings inside a roller bearing turbocharger.

Edit:
Pardon me in the last post as I got the translation of wording wrong. the correct wording for the pump is centrifugal pump.

[Roy]

The compressor wheel is not a positive displacement pump. Its a centrifugal pump...air can pass through its tolerances. When idle a positive displacement pump generally acts as a check valve. A centrifugal pump can be blown through. A twin screw supercharger is a positive displacement pump. An engine is a positive displacement pump when there is no combustion as the turning wheels spin the crank that displaces fixed volumes of air

[hypergear]

Pardon me in the last post as I got the translation of wording wrong. the correct wording for the pump is centrifugal pump.

[Dose Pipe Sutututu]

The turbine can never spin backwards, if so it's just like blocking up your entire exhaust side, that would result in your motor just shutting off. I won't go into details because this would turn out to be a keyboard war LOL.

Just read this article:

http://www.gfb.com.au/images/stories/downloads/gfb_dp_compressorsurge.pdf

[GTSBoy]

That's simply not true. I'm not saying that the rotating assembly ever actually does spin backwards. I myself doubt that it would get the chance to. But it is critical to remember that the circumstances under which it is being proposed that it happens are when the engine is actually not passing a serious amount of gas. The throttle is closed, the injectors are not firing. The only air passing through is what is sucked around the throttle and through the idle control valve. Granted, if there is no BOV (which is the point of the discussion) then there is high pressure air upstream the throttle and the idle control valve and the vacuum in the inlet manifold would also be very high (high rpm throttle closed) so there would be more air flow than at idle....but there is no combustion, so the gas going through is colder than it would be at idle. It's hard to know without doing a bunch of calcs exactly how much gas we're talking about. But it is nowhere near as much as even when the car is just cruising along on light load. So the proposition that the small quantity of exhaust gas could not squeak through the turbine side if it were spinning backwards is not automatically true. Just like we are saying with the compressor side - there are clearances in between the turbine and the housing.

But as I maintained (and of course Roy also says) there is plenty of KE trapped in the spinning assembly. The light end is the compressor, the heavy end is the turbine. So if you are applying a force to the compressor though having trapped boost in the inlet tract, the turbine is a source of force in the opposite direction simply from its inertia and those counteracting forces may well be enough to twist a shaft as Stao maintains he has seen.

Doesn't mean anything ever spun backwards. But there is nothing in the physics of what we're talking about that says that the wheels simply cannot spin backwards either.

Edited April 9, 2013 by GTSBoy

[Lithium]

Given that the topic of momentum in the complete opposite direction to what we desire is being discussed as a result of compressor stall - I am guessing my suggestion that some of these "straight line" performance figures being lower than expected from the given dyno results being a result of the lack of blow off valves on the cars in question causing the turbos to lose WAY too much shaft speed on shifts and spending excessive (compared to if they had a functioning BOV) time building the correct momentum up again to provide the target boost level in the next gear is plausible?

[GTSBoy]

I would have thought so.

[jay-rod]

Ok to change up the turbo bov war here is me beating a WRX Sti.

Also, for the 1000m i got ~220 kph ( my Gps said 225 )

4th gear was still climbing and fast and as soon as i get into 5th its like the gear has no power, i guess the 5th gear is not a racing gear ( 1 to 1 ratio ?)

Anyway, below is my 400 rolling start against Sti, i also beat a 2011 Wrx STI aswell

Also i have video of me beating what i think was a slightly modified R32 GTR

( Sorry for the stupid laughs , me and my bro being idiots )

I can also put up my run Against Mclaren MP4 if you want? Pretty embarrassing though.

[Roy]

I never noticed a huge difference when I was changing cooler piping and didnt run a BOV with my bush bearing turbo that was borderline too big for a 2L. I think the trick is really how it is fuelled/tuned as the momentary rich/fouing fuel levels tend to cause the thing to bog down more then a drop in shaft speed.

But like all things its probably a number of small compounding causes that ends up in a measurable drop in ideal performance.

[Hanaldo]

Roy, I am using a Link G4 ECU with a 3 bar MAP sensor; so the lack of a bov wouldn't be causing any overfueling issues in my case. Also Luke (Landscribble) got similar results to me and as he said, he runs a bov.