Project "Marginal gains..."
Turbo Mini Forums

My normal trickle charger keeps going past 15v? Why would I use that when all the specs explicitly say not to charge above 15v?

The very first thing (as in before the other tests) I did was to connect it (Li-Ion) to my daily, it was a cold day (-2.0 -2.5 Deg C), it didn't start, hardly turned over > doesn't get any less scientific than that. (It did start on another day that was over 5 Deg C.)

EDIT: Just to be clear about the chronology of events as the above is not clear;

Day 1 - fitted Li-ion during the afternoon, ambient 5 - 6 degrees, started OK, left to run for a while to check that it seemed to be charging from alternator > it did.

Day 2 - Early start, -2 to -2.5 Deg C (according to car display) not a chance, turned very very slowly, lots of relay clicking etc due to low voltage > swapped back to lead acid (480 CCA) which started straight away, (within 10 minutes of the Li-Ion failing to start it)

Had the daily for 4 years, same lead acid (Halfords battery) never had to charge it, all battery wiring is new as well are earths, has always started first turn of the key in that time.

All other checks / tests were done as a result of the daily not starting, to try and understand why.

I couldn't rule out that my normal lead acid charger hadn't contributed to the failure to start so I took it to a friend / ex colleague who had a Li-Ion specific charger > who offered to look at the discharge behaviour as it is his field of expertise as he was interested too.

I am not deliberately conducting cold temperature testing of it, just trying it in whatever conditions are present on the days I can try.

Edited by Aubrey_Boy on 21st Jan, 2015.

Surely something is wrong then? Because the same cell or smaller, same temps, started a fairly high compression 1.9 with a diesel starter?

It didn't start too well at first, then I kept going over the earth and feed and it was fine.

I'd get it on and fully charged, to a point where its starting the car, then see how long it works for.

Yep that's the plan as stated in my previous post, charge it with the Li-Ion charger and try again.

The 'condition' button on top of the Li-Ion battery shows it as fully charged still (as it did when I first tried it)

The fact remains that every time I put the lead acid back in with the same earth / feed leads it starts without a hint of turning over slowly so I am happy the car starter / wiring isn't the problem.

I can see you and others have used them and had success, all I can do is try it and state my findings, I have no axe to grind either way.

The daily has PAS / AC / High Amp alternator etc so not sure if / or how they might be clutched or disengaged so as not to cause extra drag on start up, so it might present a higher resistance / inertia (despite being low CR) to turn over than might be expected, dunno

Edited by Aubrey_Boy on 21st Jan, 2015.

interesting stuff, thank you for the data .i really like your common sense approach

Edited by robert on 18th Oct, 2017.

Medusa + injection = too much torque for the dyno ..https://youtu.be/qg5o0_tJxYM

turbo 16v k-series 11.9@118.9 :)

Denis O'Brien.

On 21st Jan, 2015 robert said:
interesting stuff, thank you for the data .i really like your common sense approach

Edited by Aubrey_Boy on 18th Oct, 2017.

So we have strayed into the scary world of Pectel SQ6 FBW setup...

We did this as a bench setup before comitting to the final loom modifications.

Now I haven't counted them myself but I read somewhere that an SQ6 has around 170 'fields' which have to be correctly populated for a FBW setup, these might be just single values and some full 3D tables.

Having spent the best part of a day just to get to the stage where you can depress the FBW pedal and see the FBW TB respond as you would expect I now appreciate this. I knew it wasn't going to be straight forward but there are so many things I hadn't even considered. This setup only required a total of seven channels to be up and running but still took all day:

Pedal position sensor 1 (PPS1)
Pedal position sensor 2 (PPS2)
Throttle position sensor 1 (TPSA1)
Throttle position sensor 2 (TPSA2)
Crank signal - "RPM" (We used a .wav file as an RPM source directly into the Crank signal path as the FBW shuts off after 30 seconds if it doesn't detect an RPM signal)
H Bridge 1 - Throttle body motor control
H Bridge 2 - Throttle body motor control

The PID controller commanded a fair amount of this time, especially the pedal demand versus TB response behaviour



I forgot to capture any of the live workspace data streams, but the above is a typical Pectel Caltool workspace, the 3D graph is part of the 'proportional' term tuning versus error and the windows above it would show the TPS and PPS (Pedal position) traces and numeric display values if this were a 'live' view.

I plan to set it up again this week for some further work and data logging and I'll video the Pedal controlling the FBW TB, again not ground breaking but it's my first time with 'aftermarket' FBW so it impressed me.

Still a lot more to look at;

Throttle blip on down change - FBW can be programmed to do this
Traction control / Launch control - FBW Specific strategies, the FBW throttle angle can be controlled as well as the usual ignition / fuel cuts.
Idle control - Again now controlled by the FBW and not an ICV solenoid valve

Cheers

Edited by Aubrey_Boy on 6th Jul, 2017.

Good work man. PID controllers take time to get close, but they can be very satisfying to calibrate.

fbw certainly opens a lot of opportunities, almost a shame its going in a mini :-

I use PID's pretty well daily for control system simulation work, mainly tuning the response so comfortable with them > but there not going to pitch me into a hedge if I get them wrong > unlike this stuff...

But I had my resident Pectel expert in attendance so as well as the specifics for calibrating with this pedal and TB the 'global' strategies implemented should be pretty robust too.

Spent a lot of time discussing the balance between making everything too sensitive versus response in a variety of scenarios, interesting chat about specific stuff for drift cars as regards pedal mapping.

It's really beneficial to me (not being engine calibration savvy) doing this as a bench exercise as you could see first hand how the TB responds to calibration changes, an example being how quickly the butterfly snaps shut on lift off, for ages we couldn't understand how some changes weren't behaving as expected but then remembered that below a certain RPM level (whatever RPM threshold you set it to) the FBW Idle control can over-ride the base FBW calibration so takes over to maintain idle and 'damps' the lift off behaviour (if you allow it to) when you are expecting it to snap shut.

Plenty that can still be got wrong though, tbh without his help I wouldn't have entertained FBW, but he's about to defect very soon so just making sure as much is done as possible before he leaves.

Also I've done quite a lot with the LiFePO4 stuff now as well, done over a 1000 miles / 25 hrs / 50+ starts with a 7Ah version, I've been datalogging charge current specifically / voltage drop on cranking etc, I'll put something up when I have finished the same tests on another unit for comparison.

But main problem is still cold weather starting (tested on a few different cars now) leading to sub 8 volt readings / ECU resetting etc (Ben is testing the 4Ah / 200+CCA (680g) version at the moment which starts my daily fine as long as it's warm)

This is the 4Ah version it's crazy small :)

684g



same footprint as an iPhone 5...






The 4 year old daily battery it replaced for some testing :)



Cheers

Edited by Aubrey_Boy on 6th Jul, 2017.

amazingly tiny , how cold does it have to be to not work?

Edited by robert on 18th Oct, 2017.

Medusa + injection = too much torque for the dyno ..https://youtu.be/qg5o0_tJxYM

Hey Robert,

The graph below is typical of many I have seen for LiFePO4:



It shows discharge curves at several temperatures, between 45 and 25 degrees there is little or no loss due to temperature until you reach about 70% capacity loss. But between 25 degrees C and 0 deg C there is a substantial drop.

I have noticed an audible drop in cranking speed below about 4 deg C ambient, by ambient I mean that the vehicle and battery have been out overnight at this temperature and everything is at this temperature when I try to start the car.

Below freezing on 2 or 3 occasions I have had it such that the engine barely turned over and the voltage fell to about 7.3 volts each time I tried to start, this resulted in everything resetting, clock, trip computer, radio, screwed electric window auto dropping etc...

I have repeated this behaviour in another similar vehicle and the internet is littered with similar.

It takes about 240A peak to turn the engine over properly, the LiFePO4 that I have done most testing with claims to be 400CCA but must be below this as it is a long way from turning over when cold, initially....

I say initially because through trial, error and what I have researched it is clear that each attempt to start 'loads' the battery and helps to self heat and start to bring the cranking performance back. It's a strange characteristic to get used to when you are used to lead acid.

The problem is that if the battery spec is marginal in 'warm' conditions, in cold conditions it is borderline between running out of capacity before the self heating is complete.

The video below is a good example of this behaviour in a bike albeit in much colder conditions, but the overall characteristic is similar

https://www.youtube.com/watch?feature=playe...d&v=OBfvR1EJJBk

But as I say I have run out of capacity on a few occassions before I reached the point there was enough self heat to start, but if I keep the battery inside overnight and refit it even when the car is covered in frost it turns over easily first try.

So depending on your circumstances (garaged etc...) / climate it could be just fine to use a 200+CCA / 4Ah unit but it's too marginal for what I want, the 'larger' units give you some margin when the cold temp effects kick in not to mention the ability to survive a little longer if you leave the ignition on whilst trying to work on / fix the car.

Cheers

Edited by Aubrey_Boy on 6th Jul, 2017.

fascinating data ,thank you .

Medusa + injection = too much torque for the dyno ..https://youtu.be/qg5o0_tJxYM

Very good stuff.

Are you planning to run the same tests with the larger battery to see how the behaviours differ between cell size? (&brand?)

Probably a stupid question.... how come the voltage is so low on the graph?

Yeah I have had 3 different LiFePO4 batteries on test now and 1 Odyssey, the largest LiFePO4 is currently in the daily and I plan to do a final 'group' data log of all the batteries as far as charging current, starting current, starting voltage etc. It was way too easy to overcharge the 4Ah in the daily (Typically they have a 12 - 20 Amp (3C - 5C) Max limit - however the cell manufacturers that I have seen say no more than 4C, so not sure where the battery manufacturers allowing 5C get the value from) and it was seeing 40 Amps (10C) peak regularly as the surface charge built up, not recommended as it's this which kills them.

The graph is just a random LiFePO4 cell, so it's just a single cell being measured. Any car / Motorcycle battery would be made up of at least 4 LiFePO4 Cells / prismatics in series each of nominally 3.3v

Edited by Aubrey_Boy on 24th Feb, 2015.

Right! that answers it.

Yeah, I think I overcharged mine a bit, it never did any damage but I would have liked to be a bit smarter with it. I didn't have control over it.

Can you write charging strategies with the SQ6M so you're not charging on accels if the battery is below x% and the coolant is below xdegC?

One of my bug bears with the SQ6 is that you can only use strategies which exist in the base metafile, you can't write your own.

It's not a problem as a supported Race team as you say what you want and the code appears but as a punter it costs plenty.

The attached screen grab shows what's available as standard as far as alternator control goes with the SQ6/M.

Edited by Aubrey_Boy on 6th Jul, 2017.

I use DBW but running the emerald all the hard bits are taken away. I'm still waiting for new firmware to be able to update more of the settings

Team www.sheepspeed.com Racing

Unfamiliar with pectel. But the alternator control relay functions look promising. Especially tps & hysteresis. Would be a nice function to cal in.

Is zerolag in that metafile yet then?...

This is that base metafile :)

So fully control & strategy capable

With the necessary amendments to make it "Luton XE" specific

All I need now is a rocket (Not just yet though)

EDIT: Screen grab



Hi Jim,

I didn't realise that Emerald had DBW, is it a cost option or does it come with the base ECU?

What pedal & TB are you using?

Cheers

Edited by Aubrey_Boy on 6th Jul, 2017.

It's a beta version of the firmware I'm testing for them along with TC, ALS should be with me soon too. I'm using the pedal and throttle body from the VAG 1.8 Turbo engine. Although plan to make new pedal to fit with planned floor mounted pedal box in the next few weeks

Team www.sheepspeed.com Racing

I minor Li-Ion evaluation setback...

An Italian registered Artic tried to spin me around in my daily and run me off the M6 near Spaghetti junction today

Apologies if you were delayed

Cheers

Edited by Aubrey_Boy on 25th Feb, 2015.

bet you'll sleep well!

Medusa + injection = too much torque for the dyno ..https://youtu.be/qg5o0_tJxYM

Oh shit, hope you're alright!

hope your ok buddy

Team www.sheepspeed.com Racing

All fine thank you, more inconvenient than anything really.