Turbo bandit

[luke2152]

The mk2 mazda mx5 ones are a perfect fit!

 

[luke2152]

Have done a few moderate pulls and just as the turbo starts to kick in its hitting an artificial limiter.  Datalog says it my overboost limiter set at 200kpa (absolute) but I've only been using 40% throttle and less than 6000rpm so that seems very implausable to me.  Besides my wastegate should open at 7psi which is 145kpa.  Before I blame the map sensor can someone confirm that one would not get anywhere near 200kpa (thats 15psi of boost!) without really giving it some?

Edited July 19, 2017 by luke2152

[Arttu]

Yep, sounds unlikely that you would got over 200kPa at that point. But not completely impossible depending your turbo sizing.

But it's slightly moot to keep guessing about this Just check the datalog, most likely it can tell what's going on there. My wild guess is that you are losing crank trigger sync at that point. Check "sync loss count" and "sync loss reason" values from log for that one.

[slingy1157]

Yep. Datalogs will tell you everything you need to answer this.  What does the map sensor read when static? And what map sensor do you have and is the right one selected to suit in the software.  If a gm style map sensor, alot advertised as 3 bar aren't.  You need to google the part number and check.  1 , 2 and 3 bar gm map sensor all look the same so be carefull.

[luke2152]

Datalog was what told me the map was so high.  I can't attach datalog on here though but have posted a question with datalog on the msextra forum: http://www.msextra.com/forums/viewtopic.php?f=101&t=66750&sid=8928f870c16ed61231d065f226369f4f

Map sensor is this one: http://www.efi-parts.co.uk/index.php?productID=321

And selected 3 bar GM sensor on the software.  It gives 95kpa when static.  I also tried it with software set to 2 and 1 bar gm sensor and they gave 65kpa and 35kpa readings when static so its definitely not one of those.

 

Btw the tiny digital boost gauge I mounted next to my clocks is about as useful as a potato and just shows random numbers

 

[Arttu]

Ok, if log says 200kPa and you have selected correct sensor in the settings then it's quite likely that you really have that much boost. 95kPa static and 60kPa idle sounds pretty much right. Typical idle values are between 60-75kPa with individual TBs.

I'll try to find time to take a look on your log, hopefully later today.

[Arttu]

All right, I checked briefly your log and settings. First of all, the MAP sensor readings look consistent so I think you are really having 200kPa here and there. So probably you should check your wastegate setup.

Then some other notes:

-You are running dangerously lean under boost.

-I noticed that you have set fueling algorithm as ITB. As far as I know it isn't supposed to work properly on boosted engines. Feel free to correct me if this has changed recently.

[luke2152]

I will check my wastegate setup.

The ITB algorithm has a check box 'multiply by MAP'.  I assumed this was for use on a boosted engine unless its just for an extra degree of control on an NA engine.  Maybe I should just use speed density

[Arttu]

It has been a while since I properly went through how the ITB mode actually works so I don't remember all the details. But I remember there was some catch why it wouldn't work properly on boosted engines. And I think also the code developers recommend it only for NA engines unless this has changed recently. So it might be possible to get it working but I wouldn't recommend trying it as your first tuning project

Speed Density doesn't work too well either on ITB engines. Problem is poor resolution on partial throttle so typically you can only get it properly tuned for partial or full throttle but not both.

My recommendation would be Alpha-N with that "multiply MAP" option. That has been working really well for me, on several engines. And if you want more tuning options for variable boost etc. you can add Speed Density as secondary load on top of it. Feel free to ask if you need more detailed instructions how to set this up.

[luke2152]

When using a secondary load table how do you tell it when to switch. Do you use table switching option and then switch at (for example) 110 kpa. Or is it done in one of the other settings. Tunerstudio is awesome but has so many different options!

[Arttu]

With secondary load table it doesn't switch between tables but continuously blends values from both. How I have been doing this is to use the Alpha-N table as main tuning table and Speed Density table for finer corrections on different boost level. So for starters I put some quesstimated values on Alpha-N table and Speed Density table full of value 100 which means multiply by 1 so no correction. So at this point all tuning is done by Alpha-N table. Then I set wastegate to minimum boost and start tuning the Alpha-N table. Once I'm happy with tuning I turn up the boost and check if AFR is still ok. If not, I'll do required corrections to Speed Density table. Then more boost and repeat corrections and so on... If tuning for only one fixed boost level usually there isn't need to touch the Speed Density table at all.

If needed I can post some screen shots of relevant settings.

[luke2152]

Ok thats makes sense!  Now when i use the table generator to make a base map it makes one with 15 - 170% TPS on the Y axis. Not sure why because you can't have more than 100% TPS. Is that correct or should i rescale it to max 100%. Do you have a VE table to get me roughly in the right area?

[Arttu]

Here is a settings file that you can use as an example. Probably not a good idea load in all the settings but you can export the tables and take a look rest of the settings to get an idea how to set them on yours.

Then some explanation about most interesting parts.

General settings: Even though the Alpha-N table is used as main tuning table I still set Speed Density as primary load. This doesn't matter for tuning since both tables are treated equally. But this way you get the MAP reading for "load" value which is needed to get some other functions like boost control working correctly.

VE table 2, Alpha-N: Note division of throttle position rows. You need more rovs for low throttle openings to get get them tuned properly.

 

VE table 1, Speed Density: This is used to just add some extra fuel for higher boost areas. Engine VE tends to increase when you get closer to peak efficiency area of the turbo so you need to compensate this if you use higher boost than what the Alpha-N table is originally tuned for.

[luke2152]

Excellent info there Arttu.  Especially little details like more rows on the low throttle positions.

Today I slapped myself because I had been an idiot.  I didn't realise the wastegate has ports for hoses on both sides of the diaphragm (I still have no idea why!) but I had my boost signal plumbed into the wrong side such that boost would hold it closed instead of push it open.  That explains the big map reading.

I briefly experimented with speed density alone - as you suggested it does not work well with small throttle openings.

Back to Alpha N and set all of the AFR table likely to see boost quite rich to be on the safe side and then ran VE autotune through the revs with gradually increasing throttle openings.  Then smoothed away out of place looking numbers of the new map.  Now it runs really well and most importantly is not going lean in boost.  I still haven't tuned full throttle at high revs or full boost but I feel like I'm getting somewhere at last.

[Gixer1460]

12 hours ago, luke2152 said:

 I didn't realise the wastegate has ports for hoses on both sides of the diaphragm (I still have no idea why!) but I had my boost signal plumbed into the wrong side such that boost would hold it closed instead of push it open.  That explains the big map reading.

  

The dual ports are for use with a decent boost controller that can add or subtract boost at any point that it is programmed to do so - like if using boost against gears it can add more boost depending on gear and reduce it if a particular gear / rpm is wheelie prone - all clever stuff!

[luke2152]

Bike is running great.  Have put in couple of hundred miles on it.  Mid range torque is great at 0.5bar boost.  Keeping out of the top end revs til its fully run in and I have a decent AFR gauge.  Think I have a bit of oil leaking from under alternator.  Hard to tell exactly but I'm guessing there's a big o ring where the alternator goes into the crankcase?  Also my header bolts keep coming loose.  I used copper doughnut style gaskets.  Will they settle down and stop coming loose after a few more heat cycles or should I use a pressed wire OEM type gasket.

[Tombola]

I have a feeling I'm going to have this same problem, did you buy the flanges laser cut from Eblag? They flex a little when being nipped up, I've already put spring washers on, so I was either going to drill through the cap head and wire them together or see if there's some sort of thread lock.

[slingy1157]

11 hours ago, luke2152 said:

Bike is running great.  Have put in couple of hundred miles on it.  Mid range torque is great at 0.5bar boost.  Keeping out of the top end revs til its fully run in and I have a decent AFR gauge.  Think I have a bit of oil leaking from under alternator.  Hard to tell exactly but I'm guessing there's a big o ring where the alternator goes into the crankcase?  Also my header bolts keep coming loose.  I used copper doughnut style gaskets.  Will they settle down and stop coming loose after a few more heat cycles or should I use a pressed wire OEM type gasket.

I never had problems with oem style gaskets

 

[vizman]

all my blow through suzukis loosened their header bolts....

[wombat258]

I use OEM pressed wire exhaust gaskets and have not had a problem with the bolts loosening. I did have the turbo flange nuts loosen, but loctite solved that problem.

[luke2152]

When I say loosening - if I don't tighten them every 50 miles they will fall out.  Its very annoying.  I'm going to try oem ones and if that doesn't work drill and lockwire.

[Gixer1460]

Simple - bin the bolts and put some studs in there + brass or aluminium nuts either self lockers or with spring washers - job done! Can't say i've ever had a significant loosening - with a blow through set-up (and without additional support) you are hanging considerable weight off 8 bolts not designed to carry anything like that weight! Used both types of gasket - both work - prefer the pressed wire type for more solid mounting!

[luke2152]

I'm still having problems with boost creep and its not the wastegate.  Wastegate opens at correct pressure as near as I can measure.  So I took it for a run with the wastegate completely removed and (apart from sounding like a spitfire and making the neighbours hate me) still gets into boost.  7.9 psi doing a long full throttle pull in 4th.  With wastegate fitted (it is 7psi gate remember) it went all the way to 12 and I think it would have kept climbing (based on the datalog).  Fortunately my AFR's are decent now.  Clearly my wastegate pipe is not flowing enough to control boost.  It exits at 90 degrees from the collector but looks difficult to make a smooth flowing pipe that doesn't interfere with the front wheel.  Also could do with cleaning the inside of the welds up so make it a bit smoother inside.  But is a 38mm gate big enough in pricipal for this setup.  I would have thought so since saab seems to manage ok with an internal gate on this TD04.

[Gixer1460]

Looks like (from the 2nd photo) a 25/30mm pipe between the collector and the WG flange? If it is - definitely too small, should be ideally about the same as the outlet dump pipe. A 90 deg connection isn't best.... a curving connection from higher up, curving down into the flange may help? And a 38mm Tial should be good for controlling near 300hp

[luke2152]

The pipe is the same size as the outlet dump pipe.   I think 38mm OD 35mm ID but didn't measure.  But I didn't realise til I took the photo that the hole for the pipe is a little smaller.  Although it isn't as bad as that when I get a finger in so might be a trick of the camera.  Either way I need to grind that out!   I'm thinking I could possibly bring it out to the same 90 degree position but use a 90 degree bend off the collector to improve flow.

Edited August 1, 2017 by luke2152