Sam's 3.5 Disco MEGASQUIRT START/FINISH INSTALL THREAD! :)

[FridgeFreezer]

I would manually interpolate the results to something sensible, as you would anywhere in the map, so you don't have big jumps between cells which gives rough running and clearly not right

Let me and my anorak point out that MS interpolates anyway, I'm sure you know that, but just for any newcomers...

Sam - a tip for doing screenshots:

Hit the "print screen" key on your keyboard, open MS paint / paint.net / whatever and then either hit Ctrl+V or select "Paste" from the menu and it will paste a screen image, you can then crop it and save as a JPEG or whatever.

[Mr Noisy]

Ah ha!

That helps, thanks fridge!

I need to get a wireless card for my old steam powered laptop so that I can easily get online at work without having to take it out of the car and plug it in!!

[Hybrid_From_Hell]

OK

Sam sent me the MSQ, so I have made changes directly into it, and have taken a screen dump and pasted up the before table and the after table

This could be a long complex and confusing post, so I will try and keep it basic and clear(ish)

TS or MLV both are good, but both will only do so much.

When you have done some rough logs and updated the software VE trable things will improve, but, you will get to the next stage were you have a few things

that will start working against each other.

I will try to explain.

My "Rough" Map (VE) and Sams is ctaully not that far apart, I set the Rough MSQs to be rich a tad, this gives some protection, what Sam has done is good, very good

he has now tuned OUT the richness, and Tuned IN the VE table his engines Likes vs what it was getting

An observation from the Fuelling curves is I guess that this truck is either both Heavy / Overgeared, as the fuel "Curves" (look at the MAP from a distance ) are Flatter than no0rmal, showing that it needs more mid range richness / fuelling to cope with the throttle being depressed more, if the gearing was better / closer or less weight / or both then the curves would be more vertical a bit, this is not "Wrong" this is "His engine" Charactertisctics. A simailt curve is seen onn Auto fuelling vs Manual

Any

Sam has done well in that out of a possible 144 cells 14 have not been changed.

However, this does NOT mean that the other 130 cells have all been tunned, yes they have but to differing desgrees. ie Cell Number X may have been visted 2500 x but cell M only 6 that means that the value in cell X is more likley and I say MORE likley to be close to spot on than Cell M with only 6 visits

I say 'more likely' and not 'spot on' as the cell ALSO takes values from cells around it, and its algorithims then do clever stuff.

This you can find several things

"Odd" cells, or "Odd Clusters"

This could be a cell being changed but without enough real visits and info to get it right, it could be that this rogue cell them "influences" ones next to it as they are brought into the equaltions

Or blocks of rogue cells, or cells that have a big step between them, etc etc

This is were you need to "reset" the VE to give a smooth fuelling curve.

Blue is lean

Green richer yellow more so, red rich

In Sams table vs the one I have reset, there are only MINOR changes, but the effect is to give back fuelling curves.

This should be reloaded into the ECU and driven and compared.

Does is feel better in places ? which ones ? what RPM and MAP ? make a note

Where maybe does it feel worse, write down

Then relog.

When you have this relog look to see if the system has either taken what I have popped in and made minor changes, or, is the system really not liking my changes and resetting them ?

as an example of this, I have changed Bottom row from 3000 RPM

3000 3500 4000 4400 4900 5200 5400 (54 not entered or changed)

was

27 12 7 11 35 58 59

32 34 37 36 48 55 60

These are (or some are ) significant changes, personally I think this is injector closing, and they may reset, does it feel better worse in these cells, pops and bangs with one vs another ?

The cells Top right, mean that where sam has highlighted these have not been altered, but these cells HAVE palyed a part in the value of the cells near them, so reset these cells to closer be what you might EXPECT them to be HAD they been tunned, then retune, then see what happens to the cells that were reached and altered prior, but do these change EVEN IF THE SAME CELLS NOT ENTERED STILL ARE NOT and then review

Sam is at the stage now of having to botht une and interpret and force it to retune to see what happens, I would say and hope that the MSQ sent back to him possibly feels "Smoother" that the lumops bumps and steps taken out of smoothed, better in some areas, and maybe even better overall, or, poss more likley a tad less "exciting" in some places.

If this is the case its because the smoothing has flatted the cam coming on song, there is on most fyuelling maps a point where the cam suddenly "wakes Up" and wants a disproportionate amount of fuel to the cells a moment before, these should be learnt indicated and rememebred and these NOT then smoothed so much.

A session of testing what I have sent back, and a report back, followed by a relog and retune or 3, and then repost up of the amended maps will be interesting.

Bear in mind logging on a hot dat WILL in a perfect sceanario give differing results to a cold day, a +/- 1 maybe 2 on a cell vlaue is small (and smaller the bigger the value !!) so tunning will often see a cell move up a point and then another tune down a point.

What you are aiming for is a VE with no massive steps, no cells that look wrong (and have been reset and forced to retune to PROVE) there values, and one that whatever the table says "Feels and drives as you wnat it"

Off you go Sam, look forwards to the report back

Tables before and after

Nige

[Bowie69]

Let me and my anorak point out that MS interpolates anyway, I'm sure you know that, but just for any newcomers...

Yup, sorry, I probably wasn't quite clear enough, I meant for the unchanged cells.

I think in TS if you hover over a cell after tuning it gives you the number of visits in that session/log, at least I am pretty sure MLV did, so I assume TS does

[Mr Noisy]

Right!

Going to read that post a few more times!

Thankyou Nige for your help there, will get on with that on Saturday hopefully (night in the workshop tonight) and then see how it feels!

Much appreciated!!

[Hybrid_From_Hell]

Yup, sorry, I probably wasn't quite clear enough, I meant for the unchanged cells.

I think in TS if you hover over a cell after tuning it gives you the number of visits in that session/log, at least I am pretty sure MLV did, so I assume TS does

It do !

[Mr Noisy]

First observations are it runs and the cold flat spot is not as apparent, but I had to import my target AFR table again because it had reset to default.

Another thing I forgot to ask, after driving tomorrow and noting how it feels, if I VE Analyze again should I choose normal or hard??

Thanks!

[Hybrid_From_Hell]

HARD !

[Mr Noisy]

HARD !

LOL, thanks

[Hybrid_From_Hell]

You now have a good VE Table, so you wnat the system to have to work extar hard to convince you that cells need changing

[Mr Noisy]

You now have a good VE Table, so you wnat the system to have to work extar hard to convince you that cells need changing

sounds exciting!

[Mr Noisy]

So

Just got back from a fantastic day's offroading at Silverdale near Stoke-on-Trent.

Renowned as a car wrecker and an impossibly slippery site, we did indeed have a fun time sliding around in the mud

It was ALSO the first day where I have not been frantically looking at the laptop screen every 2 seconds and have been using the car for its intended purpose rather than a very expensive way to draw pretty coloured lines on a computer screen!

AND I'm most pleased to say my newly squirted V8 did not miss a beat and was truly wonderful throughout the day. Not too much fuel used either, not 100% sure if it has better economy now but time will tell. Enough of that though, today I had POWER, I had RELIABILITY, I had INSTANT STARTING, I had WATERPROOFING, I had THROTTLE RESPONSE, and I even had SMOOTHNESS.

All this gave me many many smiles per gallon for sure.

Backtracking just a couple of days, *ahhh, just opened a beer*

Nige very kindly looked at an msq that I had produced from his base map using a couple hours on MLV and a couple hours on TS.

I thought I had it pretty damn good after TS because it certainly made some changes to the MLV version!

However, Nige had a look for me and made a few tweaks.

He said it might feel different, and it did in places.

Perhaps I did notice it wasn't quite as defined when it came on cam, more of a smooth increase in power. Perhaps not, but boy was it smooth!

Couple of jerky bits before, on his tweaked msq it was lush!

Yes, maybe not quite as raw sounding at certain ranges, but very smooth all over, felt like OEM quality tbh!

So then I opened up VE Analyze Live in TS and let it do its thing, this time with the cell change set to HARD, to restrict TS making as many changes to nige's 'human brain tweak' (what I mean by this is that although TS had done lots and lots of clever stuff, it just takes the human brain to take the msq to the next step, sorting out the areas that TS cannot 'see' basically)

After around an hour of mixed driving, I had hit 123 cells and was struggling to reach any more. My cell weighting was not bad, had visited most cells many times, the harder to reach ones less so.

I was aware of the fact that I had to get results that night because i was offroading today!!

The cell change table in TS indicated that the majority of the TS tuning operation was leaning the table. Some areas such as midrange high MAP had been richened, but the overall appearance was leaning out.

I stopped at the lights, burned the final tune to the ECU (TS updates the tune every 15 seconds anyway but does not burn it so it would be lost if you switched off the ECU) and then stopped and restarted VE Analyze Live.

This clears the cell change table, so you can see again where alterations are made.

I drove back to the garage and only three cells were altered, low MAP around 2000rpm, all three were richened by 1 point.

I was therefore happy that TS was now happy with the tune as it had not chosen to change any more cells.

Pulled into the garage, turned off the car and switched ign back on, then opened my VE table. Spent a few minutes looking through the numbers trying to spot any oddments. Using the 3D view also helped here. There were 6 or 8 cells probably that I though were a few points out, a few I changed by two points, a few by 3 and the other couple 4 points. Maybe just one by 4 points. Anyway, looking at my new 3D view there were no dodgy high or low spots, nice and smooth looking throughout, which I thought mist have bee a decent place to start seeing as it don't have any variable valvetiming or lift, there should be no reason for any major jumps really, just a decent smooth curve (quite a flat curve as Nige noticed of course!)

I then burned the tune, switched off ECU, saved tune to desktop and closed TS.

Then I swapped back to NB lambda, in light of today's ScrapHeap Challenge I thought i better stick my cheapo NB underneath rather than all the posh WB gubbins.

So, swapped the sensor, then opened TS and project properties (same as the configurator thingy in megatune) and switched back to NB sensor.

Then in exhaust had settings reverted to narrowband, set 0.510 switch point and 10% authority.

I also switched off AFR target table for VE table 1 in lambda AFR settings.

Fired it up and it ran, always a relief

And it's been blummin marvellous today!

Will get on lappytoppy shortly and take a screenshot of my new VE table so that you can laugh at my tweaks and tell me I've done it all wrong

[Mr Noisy]

right so heres nige's tweaked table

and here's my tweak on that

[Mr Noisy]

So, after a somewhat unfortunate delay in the Christmas post, finally my Tachfix box arrived from Shaun at Extra EFi!

After a moment taken to admire the build quality and time taken to make it pleasant to look at:

I opened it up to see what was inside:

Seems we have a relay coil, a transistor and a resistor

Cheers!

[CwazyWabbit]

Quite a good mark up they make on those then......

1K Resistor is pennies

2n5551 NPN transistor £0.80

Any old relay that you rip off the contactor out of ...... £1 ?

So by the time you add box and wires about a fiver at a guess.

Then about 10 minutes to make + P&P

sell for £35

I've seen less sell for more though

[Mr Noisy]

Oh yeah, fully aware of the markup, but the numbers confuzzled me a bit and I DID NOT WANT to blow up my ECU so I thought if take the plunge for the purposes of the build thread and show people what a finished item looks like!

Also I get my rev counter working without needing a new ECU

[CwazyWabbit]

As mark ups go it isn't that outlandish in all honesty! However it was very public spirited of you to donate your money for the education of others ........ and it does look pretty with it's clothes on

EDIT: also if you don't happen to have the bits to hand the postage/minimum order values can soon make the diy build considerably more expensive

[Mr Noisy]

Yeah I expect there is about £7 in components from RS, a relay should be easy, so could be done for a tenner.

Anyway as above with all the diagrams I was confused but now we have nice pics of a finished item.

Will see if it works later!

Tbh I wish Shaun hasn't fitted the other half of the Superseal connector to the wiring because I would have pinched that and used it elsewhere

[Mr Noisy]

Not too difficult! That fast idle valve relay socket which is now going spare has proved very handy so far for extra power and earth connections!

Yayyyyy!

[Mr Noisy]

And before you mumble and grumble, I specifically used crimpers here so that if I want to remove the tachfix I can put the tacho wiring back to original in a jiffy!

[robp]

Good result there Sam, glad you finally have a working tacho

I know it might seem a big mark up for those few components but the way I look at it you are also paying for the research that has gone into solving the issue and producing a working solution you can easily plug in.

PS where are the off road pics?? did I miss a page

[Mr Noisy]

Yeah so lets break it down:

33.50 on eBay.

2.00 ish to post.

3.35 to eBay.

1.34 to PayPal

Components and hardware £10 (remember the Superseal connnector!)

Sticker 50p

Bit of cable and heatshrink 50p

Approx profit 15 quid?!

[Mr Noisy]

Regards off road pics, yeah urm, the off road picture novelty has kind of worn off, don't tend to take shots anymore but just imagine an abused disco in a hellishly muddy site

[Bull Bar Cowboy]

So, after a somewhat unfortunate delay in the Christmas post, finally my Tachfix box arrived from Shaun at Extra EFi!

After a moment taken to admire the build quality and time taken to make it pleasant to look at:

I opened it up to see what was inside:

Seems we have a relay coil, a transistor and a resistor

Cheers!

I could say .................... I told you so .....................

[Mr Noisy]

I could say .................... I told you so .....................

I know, you were exactly right in your instructions, but it confuzzled me so I thought best option here for a noob is to pay the man and get the product, plus it could be used as demonstration for the thread!