VNT control

[d-jimbo]

anybody want to share with me the principles of controlling a vnt? i've had a look but can really find anything. a point in the right direction would be appreciated.

cheers,

jim

[silverzx]

Multiple boost controllers would be a start I guess but ultimately you want it reading off a table calculating Throttle x RPM and then calculating an "ideal" boost. Think Ruans been programming something like this for a while.

[Ruan]

I'm not just giving the answers away - took me long enough to figure it out.

Essentially you need to devise a way of of being able to limit the boost until a certain RPM, not produce too much boost whilst cruising causing high temps and pumping losses, but still being the least laggy as possible and being able to control the boost at high RPMs - all the while remembering that more boost on a VNT turbo is only a measure of exhaust manifold restriction, you have to balance it quite finely or you'll end up running into a loop of more exhaust pressure = more inlet pressure = more exhaust pressure, overheating everything, but yeh mate i gotz 60psi of boost... I accidentally found out that yes, my GT2052V will happily pin a 60psi gauge, but still not go anywhere.

But yeah - Multiple ways of doing it...

[Mr Whippy]

Don't some of the VAG people who have done this use an actuator off the throttle arm on the pump too, since that essentially gets you your throttle input?

Then it's "just" a matter of mixing the boost actuation signal via the 'shaft' with the throttle signal via another pusher...

Blend together the signals with cammed pushers to get some curved response vs input and you might get a reasonable VNT pusher.


Obviously that is heavily simplified but iirc the VAG people were messing with that on mk3 diesel Golfs back in the mid 00's, and had a fairly refined system.
Surely it's just taking that and adjusting the variables with the appropriate non-linear action amounts (via cams or whatever else they used)



I'm surprised though Ruan, is the feedback loop really that bad? I'd imagined with just boost actuation you would still got ok performance for the total power output vs running similar power on a wastegated turbo. Or is the 'closed' restriction that significant at cruising/steady state conditions it is actually higher than a closed smaller wastegated turbo anyway?

Making a little ECU to control two boost actuators sounds like a cool idea. One look up table, two inputs, one output
Now go do that with I guess duty cycle operating vacuum control solenoids? Get the calibration of those into the mapping by simply driving and adjusting till it reads the right value (live mapping)
Crude but as effective as it'd need to be for Td use!

Hmmmm

Dave

[Ruan]

You seriously need that feedback, or yes, as you say the cruising pressures are way higher, at 70mph if you have it in at a point where it spools very quick, you're talking 18-20psi inlet and 25-30psi exhaust pressure, WAY too high really, you end up having to ease off the vane stop point, in which case you may as well have a Wastegate turbo...

But essentially, that's what I've done, got a Microcontroller to read MAP, TPS and RPM - you then have a table TPS vs RPM which then you have to set up with values from scratch, but it's not that difficult, depending how fine your table is - the finer the smoother... So essentially it's a boost control and limiter all in one... I actually have two maps it can run, one for cruising and one for ragging the ass out of it, purely because of the way the injection pumps work and it not being able to know the position of the LDA - the throttle percentage isn't actually really relative to IQ, so when you're cruising you'll find it's producing too much boost, but when ragging it, you'll find that it's a bit unresponsive under trailing throttle and some transients you tend to encounter.

Then use PID control to give it a target boost pressure with MAP as your feedback... It's then a case of fudging round with the PID variables to make it not hunt and overshoot too much for all conditions - in fairness that was actually the second hardest part...

The actual hardest part was essentially rewriting and refudging a load of the firmware for the controller I am using, most solenoid valves work on 140Hz PWM, typically the controller I use wants to use 100Hz PWM, therefore the solenoid valve used to go nuts... With adjusting the internal clocks, I've got it to 137.somethingHz now, which seems fine as long as you don't use the outer edges duty cycle, or it gets very jittery/jumpy...

When I did it I was sad enough to do an adaptation/calibration function which allows you to adjust the outer edges where the the solenoid valve needs to work, it actually only needs from about 35-70% duty cycle... Also have the same thing on the throttle pedal, so it reads a value for 0% and a value for 100% and works between that, makes life so much easier!

</geekrant>

[d-jimbo]

interesting stuff... cheers guys. i've done a bit more research and taken into account what uve said and due to the fact that i'd rather steer clear of microcontrollers and the like for now, im gonna try to control one mechanically just using throttle and boost pressure inputs.

[silverzx]

To be fair there are plenty of decent enough "non-VNT" turbo's out there suitable for our needs.

[ginge191]

d-jimbo, have a look and read behing the mechnically controlled VNT's that are controlled with simple rod linkages; as the throttle opens more, the rod linkages move the VNT vanes to create a wider internal.

This is a VERY simple way to control a VNT, but doesn't regulate anything else than "more fuel/more air" system, and doesn't take into account RPM etc.

you can run alongside the wastegate with the turbo alongside the rod linkages...

[Jonny81191]

I'm glad I've bought a wastegate turbo!

I still don't get VNT, admittedly I haven't read up on it though

[ginge191]

[Jonny81191]

oh! that's simple then!

[Ruan]

VNTs don't have a wastegate.

They essentially adjust the A/R...

As the vanes close up - they accelerate the gas towards the tips of the turbine blades through Variable Nozzles, as they open up, the gasses slow down and don't hit the tips at as sharp angle, causing less backpressure and causing the turbine wheel to slow down - or controlled right, will hold a set boost level.

[ginge191]

oh! that's simple then!

no, because as ruan and whippy have already mentioned, matching fuel to a variable turbo is not easy

@ Ruan, i'm sure there was a VW build i was reading where it had a matching wastegate AND rod linkage system :|

[Jonny81191]

now I don't get it. I'll stop trying lol

[Dum-Dum]

Excuse me potentially showing my ignorance but for basic mechanical control on a VNT isnt it just a case of having a actuator on the VNT leaver/rod so that as you push boost into the actuator it pushes the rod and opens up the nozzle, I know the there would be alot of fine tuning for that method but my head is saying how hard can it be.

[ginge191]

Excuse me potentially showing my ignorance but for basic mechanical control on a VNT isnt it just a case of having a actuator on the VNT leaver/rod so that as you push boost into the actuator it pushes the rod and opens up the nozzle, I know the there would be alot of fine tuning for that method but my head is saying how hard can it be.

Yeh, again, from what i've read on mechanical VNT vane control, using this method eliminates any different RPM, and you need differnt pressure at various rpm, if you have a full throttle with little RPM to overtake for example, the vanes are opened to full A/R potenital, but that's not ideal, you'd want cloesd vanes, working UP to larger vanes to increae the A/R in increments.

This is my knowledge... maybe a load of balony, but from what i understand from mecahnical vane control, feel free to shread this to pieces.

[Ruan]

The problem is that VNT Turbos will produce too much air for our engines to handle at lower RPMs...

But having it in full closed position to produce as much boost as possible means that cruising you'll end up producing like 15-20psi, getting things very hot due to high backpressure and causing quite high pumping losses...

You need to be able to reduce the backpressure at low throttle and restrict the boost levels so it doesn't end up surging.

[ginge191]

You need to be able to reduce the backpressure at low throttle and restrict the boost levels so it doesn't end up surging.

An optional wastegate, opening when linked to lwo rpm's and high throttle? :think:

I'm going to sleep on this one

[Ruan]

The whole point is that the turbo can do it all itself... The efficiency is completely wasted if you use a wastegate on it...

[Dum-Dum]

External Watergate with boost and vac to make it user controllable then a couple of solenoid to turn it on and off on q throttle switch. Could probably set it so you dont get more than 5psi below 1/4 throttle

[darrenjlobb]

External Watergate with boost and vac to make it user controllable then a couple of solenoid to turn it on and off on q throttle switch. Could probably set it so you dont get more than 5psi below 1/4 throttle

Lol and not use the vanes in the vnt at all? That would totally defeat the point of a VNT.....may as well run a wg turbo..

You can accuate it pretty well with a boost can, only issue is it will try to "overboost" at lower rpms ( the rpms an equiv wg turbo wouldnt have spooled by) and it will surge, so you have to watch your right foot, which means you dont get the full ammont of power from it as it would burn more fuel if you could stabilize the boost, but you really need an ecu as ruan said to do that, or a fancy mech linkage like the vw boys use...

An easy way to imagine a vnt / wg turbo is the wg turbo just "wastes" exhaust gas once its making boost, whereas the VNT is designed to always be able to flow it all, and "closes up" taking advantage of it lower in the revs, therefore being much more effcianct / less laggy.

[d-jimbo]

a fancy mech linkage like the vw guys have was more what i was thinking. obviously gonna need a lot of fine tuning but i dont mind that.

[Ruan]

The problem with a mechanical linkage is that it has no concept of time, it's not smooth in the slightest and will leave you with a car that is only good for one application... Either cruising or blatting it...

For example, the controller I use has a function that times the throttle, as in trailing throttle conditions you want it to be in full spool position constantly, trying to achieve as much boost as is available/allowed for a given RPM... But under that trailing throttle you might find that that's the exact throttle position for cruising, so it's adaptive to your driving style, if the RPMs stay within a certain window and the throttle position stays within a certain window for a certain time frame, it'll allow it to go into cruising mode...

Also I've got a condition where above 1500RPM and at 0% TPS it'll go into full closed mode, this massively helps engine braking, but I have the ability to adjust this so I can control the amount of engine braking... They're not just for quick spooling, there's so many advantages to them, by just having a rod/linkage setup, you may as well just grab a normal WG turbo... You've got to think of the exhaust side as well as the inlet side as it's variable, so can make your engine a totally different beast by just changing the vane position.

IMHO anyway.

[mark_airey]

A bit of reading

http://forums.mg-rover.org/showthread.php?t=466797

http://www.torquecars.com/forums/f61/vn ... sel-15815/

Two threads about the same setup on two different forums but looks like it has potential....

http://www.thesamba.com/vw/forum/viewto ... nt+control

http://www.vwdieselparts.com/forum/view ... 042#p76175

Thats should keep you busy for a while..................

[ginge191]

I'm very intrigued ruan as to how you're runnning this VNT, i know you've said you're not giving answers away out of pride, (fair play) but i'm also impressed you've managed to get an electrical/pressure controlled vane control AND taken into consideration RPM and not just "got a VNT working", but adapted it to a mechanical engine.

Your car going to be @ FCS ?

[Ruan]

It's really very simple, it's only the software that makes it special...

It's simply just the hall sensor reading the flywheel, picking up both marks and converting to RPM, a MAP sensor (off a Lancia Delta Integrale might I add!!!) and a throttle position sensor off a 406 ECU pump and a 140hz PWM solenoid valve and a small microcontroller to do the maths...

The only thing that I've got to do is do a solenoid valve driver - do it with a proper FET, the controller gets very hot actuating the valve, hence why it's not working atm, if I run it for long, it'll cook the controller board.

[ginge191]

It's really very simple, it's only the software that makes it special...

It's simply just the hall sensor reading the flywheel, picking up both marks and converting to RPM, a MAP sensor (off a Lancia Delta Integrale might I add!!!) and a throttle position sensor off a 406 ECU pump and a 140hz PWM solenoid valve and a small microcontroller to do the maths...

The only thing that I've got to do is do a solenoid valve driver - do it with a proper FET, the controller gets very hot actuating the valve, hence why it's not working atm, if I run it for long, it'll cook the controller board.

Ruan, with your profession, this is easy, but again, top marks for researching all of this.

Moving slightly aside, if a VNT turbo was to be used on a GTI, could there be an addition put onto a GTI ECU? :think: ... i've never looked at ECU's before, and looking at building a GTI-T project within this year, and continueing next year (got a shed being built solely for it :think: )

[Ruan]

Marelli IAW 1AP can't understand boost without significant modification... The only way the supercharged guys get away with it is because their chargers are just proportional to rpm... The volumetric efficiency increases as rpms do, so you just adjust the fuel map to suit...

You can't really apply a vnt to a petrol, exhaust gas temps are far too high and because you've got the option of high rpms, you just rev it to get the boost... There have been attempts, it does work, but the reliability of a vnt turbo on a petrol just isn't going to happen unless you're looking at one of the porker 997 borgwarner ceramic jobbies which from memory are well over 1700 quid a piece and are only a 60 or 62mm compressor...

A MODERN wastegate turbo will do everything and more, take a look at some cars with a Precision 6262 turbo o, you'll find people slating Precision, but they are darned good! Stupid quick spool and its a 400-500hp capable turbo... The problem lies when people put on a 'cossie t3 hybrid' and wonder why its a piece of shit!

[ginge191]

yeh you make a good point mate..
It was just a wonder, as said, i've never touched a GTI engine or ECU's for that matter, and yeh, i'll get reading behind it all before i go ahead with anything

[darrenjlobb]

Meh imo a mech vnt is still better than a wg turbo, as I said ealier, your taking advantage of the flow rather than just wasting it up high, tbh they work fine bar the surge issue, and having drove stock tdi's with vnt's tbh they fuel so little down there to keep torque down that its not much different than just using your foot to be carefull, besides unless you running a decent clutch, any bottom end torque will just ra*e the clutch anyways...

There a bit of a farce / pita to get working right, but a mech vnt will outrun a wg turbo anyday imo.

That said, theres nothing like a 80mm compressor lighting up at 4k