Twincharging - calling Warpspeed! 3

[madmac666]

thread71-174690

Hi guys, i subscribed to this forum because i found the thread (referenced above) on twincharging while doing a Google search for information on this subject.

I have to say, the thread and contributions from forum users was very good and informative, which gave me a direction to follow for my own twincharger build.

Some background on this: I have a 1997 Mitsubishi Lancer Evolution 5 which i've used for sprints and hillclimbs in Scotland for a few years now. The state of tune of the car is such that we were producing around 650hp, but with the big turbo at 2.6bar, lag was always an issue. Twincharging seemed the most sensible option for the car to eliminate the lag and hopefully produce a torque curve with spool much lower in the rev range.

I'm using a Harrop HTV 1320 blower and the existing turbo which is a hybrid T04Z with 0.82 housing. This is the turbo that gave us 650hp previously, but on a 0,63 housing. The system is compound compression with the turbo feeding the blower. The blower has a 75mm pulley and the crank 150mm pulley so a ratio of 2:1.

I have the blower mounted and all pipework finished and finally had her on the rolling road a few days ago. We saw some impressive torque figures but had to cut the session short due to belt slip on the blower pulley when the turbo reached 1.5bar. The system i have uses a 32mm toothed belt but the length of the belt seems to be an issue as it's stretching and allowing it to jump over the pulley when the turbo spools up. We also had issues with controlling the turbo boost and decided i need to install another wastegate to allow better control of the boost. I'll have that done by next week RR session so we have control of the turbo boost and can carry on mapping.

It seems the belt starts to jump when the turbo reaches 1.5bar and the drive simply cannot cope with the power needed to increase boost by a factor of 2. I have a larger pulley (100mm) which i'll be fitting over the weekend and hope the longer wrap around of the belt on the larger pulley and the reduction of the drive ratio should sort out the issue of the drive. If it doesn't then i have another modification i can do to run a shorter belt on the blower and split the ancilliaries onto 2 separate belts but i'd like to try this again with the reduction in drive frst before i go to the trouble of a re-design of the belt drive. Space is at a premium so 32mm is the widest i can go and still have a chassis leg...

We never went above 4000rpm. Despite this, we saw 460ft/lbs torque and 360hp at 4000 and 2.9bar which confirms there's definatley power to be made! Obvously slowing down the blower will mean the turbo will need to be producing more to reach our goal of 2.5bar total. The outlet temp of 80c from the turbo at 1.5bar was logged, which is about the limit of the blower seals. I hope to keep the turbo around 1.5 bar so the outlet temps don't fry the blower seals. The combined air then goes through a very good intercooler and reduces the final temps to 16c (this is Scotland remember....!) at the plenum.

I think we're almost there and hope to have it mapped on the larger pulley with the extra wastegate fitted for turbo control. Looking at the RR graphs and data we collected the other day, we monitored the pressure from the turbo and combined total and can clearly see a multiply of the turbo by the blower very close to 2:1 from 1800rpm to 4000rpm but we couldn't control the turbo and it looked like it would have carried on climbing which would have killed the engine or the blower.

Because the engine already had the big turbo and supporting mods to make good power, we hope to see some good figures once the issues are ironed out. The engine itself is built to produce 1000hp 'reliably'.

I would like to hear any views and comments on my findings so far. I had considered swapping the turbo and blower order to reduce the power needed to drive the turbo but i felt the extra restriction on the exhaust (because the turbo would be working harder than it used to) would likely end up making less power, as the exhaust back pressure would then be much higher. It would certainly reduce the power needed by the blower


Thanks for posting the useful information, it really was a huge help and if you have anything to add that could help us, i'm all ears!

Donald

 

[Warpspeed]

About as far as is possible from where you are.

Melbourne Australia.

 

[70btdc]

What a great result! Do you have any data that you could share regarding temps and pressures at each stage?

Good luck with the sprinting, hopefully I will be able to see your beast in real life sometime soon...

 

[gruntguru]

madmac666. Are you running full sequential injection? If you are and the injection timing has been optimised, it is possible that you are scavenging a lot of air during valve overlap (a good thing) but your AFR meter (wideband lambda?) will be reading the average AFR, so with 12.4 average and a lot of scavenged air, the combustion AFR may be substantially richer than 12.4.

NB. If you need to reduce overlap, before you change cams, why not try opening up the lobe centres by retarding the inlet cam and/or advancing the exhaust cam a smidge?

 

[gruntguru]

Madmac. Good to see you using Aussie dyno's over there. Strange RPM/KPH scale - which one is correct? The torque curve looks surprisingly peaky - even turbo motors usually produce flatter curves (tho' without the instant response I grant you). Not sure where your cams are dialled, but I think if you advance the exhaust cam by 5-10 degrees, you will lose less boost to scavenging at low rpms. In addition the extra exhaust blowdown will spool the turbo earlier. Combining the two may produce more low-down torque and perhaps even reduce the smoke.

 

[Warpspeed]

It doesn't look at all peaky to me.

Between 4,000 and 7,000 there is 500 to 600 ft/Lb, only a 20% change.
That is rather a lot of torque !

But one thing that is rather less obvious, (but Donald will probably confirm), the acceleration potential, especially in the lower gears will now be far more fierce than it was originally with just the straight TO4-Z. It is not JUST the Hp and torque, but the ability of the engine to very quickly gain rpm under load in the lower gears.

There is a lot more to the responsiveness of twincharging, and how it actually feels to drive, than just eyeballing the dyno sheets may suggest.

Difficult to put into words really, it is something you have to experience yourself, but Don will know what I am trying to say.

 

[patprimmer]

With a roots blower, it comes onto boost even at idle almost as quick as the throttle opens.

With a turbo it needs both rpm and time to spool.

It is not the least bit surprising that a twin charger gives instant response and huge top end without sacrificing bottom end. That to me is the original driving force behind such a project.

What I had not considered (although once mentioned seems so obvious) was the extra pressure multiplication from the roots blower increases exhaust scavenging to such an extent.

On considering this, the cam strategy at overlap should be similar to that if only a roots blower where used. OK if only a roots blower with a somewhat restrictive exhaust where used.

If it where a SBC and presuming duration is measured at 0.050" I would say with 280 deg duration it needs about 114 deg lobe centres.

Being a 4 valve engine with smaller valves but bigger total curtain at the same lift and no shrouding between the valves at small lift, I would think 0.050 is a bigger %age of total lift and flow at lower lift is relatively better than a 2 valve engine at the same lift. Bottom line is significantly better scavenging, so even wider lobe centres might help. Some top fuel guys run 116 deg lobe centres, but they also run 300 deg or more duration.

With twin cam and adjustable cam gears I would open the lobe centres a stupidly big amount just to see, providing piston to valve allows it.

I agree with sequential timed EFI you are probably losing more air than fuel during scavenging.

It is my understanding that unburned air in the exhaust shows lean not rich or average as the O2 sensor senses the O2 but not the HCs, but black smoke indicates only partly burned fuel discharging from the exhaust.

Open the lobe centres but keep a close eye on exhaust valve tappet clearance to look for valve stretch. If valve stretch is even hinted at, close the lobe centres 2 degrees again.

Sorry if the post is a bit random as I typed as I thought and did not edit it for logical progression and concise structure.

Regards
Pat
See FAQ731-376 for tips on use of eng-tips by professional engineers &

http://eng-tips.com/market.cfm

for site rules

 

[patprimmer]

You now have me scratching my head how I can economically and efficiently get a small roots blower to fit my counter clockwise rotating Honda with minimal room between the head and the firewall in the CRX chassis and even tighter room constraints for the crank drive pulley and the sub frame.

If only one could get decent bearings and rotors for the Toyota SC14. They are dirt cheap and fairly compact and big enough I think for a D16 Honda even bored and stroked to 1.8 litre.

Hmmmm

If it could be turned over so the inlet port becomes the outlet it would not need the drive direction to be changed.

Sorry I will cease with the thread jack and start a new thread if I don't come to my senses.

Regards
Pat
See FAQ731-376 for tips on use of eng-tips by professional engineers &

http://eng-tips.com/market.cfm

for site rules

 

[Warpspeed]

I agree Pat.

Slightly less duration with wider lobe centres would definitely be the way forward for a more general purpose type of twincharged vehicle.

But then again, Donald is not the least bit concerned with fuel economy, just winning.
And the cooling effect of both extra air and extra fuel, with over scavenging, will add to power production, detonation resistance, and greatly aid thermal management.

A higher EGT would certainly benefit the turbo, but the blower more than makes up for that with the extra mass flow.

No doubt more tuning and more development lies ahead, but it is probably not such a bad thing to start out with it the way it is.

There is a very great reliability benefit in having a high detonation margin, and keeping peak EGTs down.

 

[Warpspeed]

Pat, don't know about Hondas, I have never twincharged one.....

But on a Ford Laser 4WD, I fitted an SC14 down beside the gearbox, ditched the SC14 blower clutch, and ran a long jack shaft, coupled to the SC14 input with a small chain coupler.

That ran along the side of the sump and left plenty of clearance for all the existing belt driven engine accessories, none of which needed to be moved.

It is the only space it would reasonably fit.
It also left plenty of room for a large turbo, mounted in the usual location.

 

[patprimmer]

OK

I will start a new thread

Regards
Pat
See FAQ731-376 for tips on use of eng-tips by professional engineers &

http://eng-tips.com/market.cfm

for site rules

 

[gruntguru]

With twin cam and adjustable cam gears I would open the lobe centres a stupidly big amount just to see, providing piston to valve allows it.

Opening lobe centres will improve piston to valve clearance (as I'm sure you know). Agree with everything else you said.

Warpspeed. Regardless of the vehicle purpose, I think it should run fairly wide lobe centres if it isn't already. Duration is probably OK as it is.

It is my understanding that unburned air in the exhaust shows lean not rich or average as the O2 sensor senses the O2 but not the HCs, but black smoke indicates only partly burned fuel discharging from the exhaust.[i/]

Unburned scavenged air is 100% lean so the "average" or overall AFR in the exhaust stream looks leaner than the portion that was in the cylinder during combustion. (I am/was talking about unburned air with no HCs). Result = WB lambda reads 12.4, cylinder sees a lower-number/richer-mix.

Black smoke is predominantly carbon particles. Available oxygen reacts easily with the outer hydrogen atoms leaving carbons with no oxygen left to burn them.

 

[madmac666]

Thanks for the input again! Because of the issues i had at the start (with the overheating blower), my attention has been to get the system useable before the end of the season so i haven't had much thoughts on cams and timing yet.

I'm just back from Loton Park Hillclimb near Shrewsbury which saw 100% reliability and the car performing faultlessly all weekend. This is a track i've only been to once, and the first time i was driving a Hayabusa engined Wesfield as the Evo was missing a SC at the time. My best time in the Westie was a 59.4s - this time i got the Evo to within 0.09s of the class record to finish with the class win and a personal best time of 54.80s (the class record being 54.71s set by Roger Banks in the mighty Audi A4 V8 twin turbo in 2007) To get that close without knowing the track well shows the car is very quick as it is.

I have a pair of S2 cams with duration of 272. These are timed to the manufacturers settings 117.5 degrees but to date, we haven't had a chance to play with the timing yet. The car is away to be stripped and lose another 100Kg over the winter although i still hope to keep it road legal! I have plans for building a new engine before next year as this one really is tired after 5 seasons of racing. I had always thought it best to use this old engine as a test bed for the compound charge system in case it went horribly wrong!

Food for thought on the cam timing tho. I'll pay attention to this during the spec and build of the new engine and hopefully we can have time to play with the cams at the pre-season dyno session next year. From the recent results at the hillclimbs, i have every confidence that the car is more than capable of taking most of the class records next year providing reliability remains.

I have one more event left this year - a sprint at Boyndie Kart track this weekend coming. This should be interesting as the circuit is so tight that i'd be having to brake for the next corner just as the turbo was spooling previously but now i think the car will hump the class record big time due to the instant response and mid range torque. I'll update with the video from Loton later today

Cheers for all the tips so far, i will of course keep you updated with progress!

 

[Warpspeed]

Wooo Hooo......

Very good going indeed, well done Donald.

And I am sure you will blitz the field at the cart track too.
The instant response and wide torque band should make your car a lot more drivable on such a really short tight track, than many of the other cars there.

I just wish I could be in Scotland to watch you do it.

 

[madmac666]

Warpspeed, have you had any dealings with Harrop Engineering? My contact there is Richard, he appears to be a petrolhead and has keen interest in my project.

Here's the video from loton park. Running 2 videos side-by-side of this run and another where i make it round the first corner cleanly, i lost 0.3s - more than the margin i missed the record by! Had i not been so late on the brakes at the first left hander, i'd have the record in the bag on that last run

http://www.youtube.com/watch?v=XI1CsX4S7hc

Ce'st la vie!

 

[bigjuicesr20]

Donald,

Wicked video!! Quick question, how is your SC pulley/belt system set up? You mentioned that you had to redesign a better system to eliminate belt slippage. Are you using a cog belt/pulley similar to this

http://www.fignoggle.com/components/gears-and-timing-pulleys-and-belts.htm

or running the typical multirib belt that comes with most of the SC's?

I have an Audi 5 cyl Coupe Quattro and have a couple SC's I can use for this project. I don't care about belt noise levels as I'm just going to rally/auto cross this car. I just don't want belt slippage.

 

[madmac666]

I wanted to eliminate slip from the outset so i chose an 8mm pitch 30mm HTD belt with alloy pulleys. Initially we had the belt jump because it was too long (1760mm) and the small percentage of stretch combined with the fact that the SC got it's torque after all the other ancilliaries, contributed to the belt jumping when the load was on the SC. I got round this by making pulleys at the right ratio to fit an available belt which runs directly from the crank to the blower, with a tensioner on the loose side. The other ancilliaries such as PAS and alternator are run from a slave belt system therefore allowing a lot more torque transfer. In hindsight, i should have had this system from the start but wanted to keep it simple and didn't think the belt would jump. The new pulley/belt system relieves the ancilliaries from the loading of the SC at full boost.

I didn't think i should risk the multi-V belt at the start. mainly because i was having pulleys custom made and didn't want to have to remake them if the belt couldn't cope. As it turned out, i had to get more made anyway! Development...

I was thinking that because i have a damaged SC here, i could loan it out to others wanting to do a test fit and see if that is the SC for their project. When i started, i ordered the SC, then while waiting for it to arrive, made a wooden mock-up of the unit to make manifolds and pulleys so when the SC arrived, it bolted straight on.

 

[madmac666]

These are the guys i used for the custom pulleys. Alan is very knowledgeable about belt drives and is located near the Gates belt factory in Dumfries, Scotland. He does a lot of development work with them i'm led to believe:

http://www.timingbeltpulleys.co.uk/pages/home.php

You would have to give them drawings for the parts to work from but the quality is very good and the prices are surprisingly cheap.

The best tip i can give you is to find out the lengths of available belts and design your drive around the nearest size.

 

[Warpspeed]

Harrop Engineering are not that far away from me here, but I have never done any business with them, or know anyone there.

Donald is right, use an automotive belt catalog to select a cam belt off a fairly common vehicle, and design your blower drive around that, right from the outset.

They always give tooth count, and the length will be tooth count x 8mm pitch. Curiously there are always an odd number of teeth on automotive belts to even out the wear, which is not a bad thing.

Using a readily available belt will save a lot of pain and frustration later on. The automotive belts are thinner and more flexible than the industrial belts, and are much better suited to small pulley diameters and the very high belt speeds we need.

Aluminium pulleys are ideal for a competition engine, but will show rapid wear in a long mileage street engine.
For the street, stick with steel pulleys, cast iron may be o/k, but mild steel is much safer at very high rpm.

Plastic idlers should be of large diameter, (to keep the idler rpm sane) and placed on the outside of the belt on the slack side, just before the blower pulley.

Working up a reliable blower drive is always one of the major frustrations of installing a custom blower.

 

[madmac666]

Have to admit in the early days the aluminium pulleys showed signs of excessive wear. This was partly due to road grit coming off the wheels of the tow van and getting in through the bumper vents to lie on the inside if the belts, made wors if towing any great distance or in the wet. Once started, the grit would stay in the belt and grind the aluminium away. I got round this by fitting a simple guard to stop 80% of the grit getting in there in the first place.

Future plans include having the pulleys hard anodised to reduce wear from the belt. For my application, steel pulleys were not even considered due to weight. The aluminium pulleys will eventually wear out but by then, i will have no doubt refined the design and need a new set made.

I also ended up using the standard sprung tensioner as it had the travel i needed to remove and refit belts easily. I modified it so it would spring into the belt and then i could lock it in place with the centre bolt. The beauty of this is it's so simple and strong and gives me a large range of movement as the belt inevitably stretches. I found that by not locking it in place, the inertia of the SC on the overrun would pull the tensioner open and cause the belt to jump off. The belt drive was one of the biggest challenges of the build - but looks easy now it's working properly!

 

[Warpspeed]

"The belt drive was one of the biggest challenges of the build - but looks easy now it's working properly!"

Yup, that is always the way.
Usually people doing this for the first time are fairly impatient to get to the initial road test stage, and the blower drive is usually the very first problem that shows up.

Aluminium is perfect for this project, but just cautioning people that aluminium is a less than perfect long term choice for an everyday road car. Hard anodising will help a lot, but that too will eventually flake off.

These pulleys are never cheap to buy new, so just a word of caution on material selection.