mx5cartalk.com

Hi All

I should have the Jaycar DFA by tomorrow and should have assembled, tested and installed sometime soon after, but before any tuning wonder if I should back down the ignition timing, mine's sitting on 14degree at the moment running 98ron. How far should I back down?

Cheers
Minh

why?

just leave it at 14 ?

other than that, when building it i strongly suggest that you omit the relay from the board. just use 2 wire links in place of the relay.

the relay is only there because a smalll number of cars will throw a fault code upon startup with the dfa connected. the mx5 is not one of these cars, so there is no need for it. so it's best to leav it out as it just one more mechanical thing that could possibly break down.

also, it is a good idea to add some additional heat sinking to the supplied heatsink. a strip of alumium should do it. it is not entirely neccesary, but mine sometimes overheats after about 20 minutes of having the hand controller connected, because the backlit screen on the controller draws a lot of current.

this will not affect running at all, it just means you cannot make tuning changes until it cools down again. (it still works fine tho). only takes 5 minutes to cool down.

i emialed you today with my tuning map. did u get it ?

Ok, got the DFA, assembled, installed but not yet tuned. I've also got a Smart Mixture Display setup show mixture. What would be a good approach to tune the DFA on the MX5? The instruction came with the kit recomends removing the closed-loop (disconnecting EGO sensor) then tune for smooth idle and off to the dyno for high load tuning.

How did you tune your DFA?

Fatty wrote:why?

just leave it at 14 ?

doesn't the timing changes when we're modifying the AFM signal?

Fatty wrote:other than that, when building it i strongly suggest that you omit the relay from the board. just use 2 wire links in place of the relay.

the relay is only there because a smalll number of cars will throw a fault code upon startup with the dfa connected. the mx5 is not one of these cars, so there is no need for it. so it's best to leav it out as it just one more mechanical thing that could possibly break down.

I decided to keep the relay incase anything happen to the DFA (blown fuse etc), in this case the relay returns to normal state which connects input to output (bypassing DFA).

Fatty wrote:also, it is a good idea to add some additional heat sinking to the supplied heatsink. a strip of alumium should do it. it is not entirely neccesary, but mine sometimes overheats after about 20 minutes of having the hand controller connected, because the backlit screen on the controller draws a lot of current.

this will not affect running at all, it just means you cannot make tuning changes until it cools down again. (it still works fine tho). only takes 5 minutes to cool down.

Luckily the kits been updated with the heatsink supplied.

Fatty wrote:i emialed you today with my tuning map. did u get it ?

Got the tuning map, thanks heaps James. My load range is slightly different to yours but still gives a good indication of where I should make the changes. I tune mine to give the widest load range to suit the signal response from the MXs AFM, currently idles at load ~95 and max AFM opening at 0, I could widen up the range at bit more to make full use of the 0-128 load range but this means when ACC-ON the AFM signal will exceed the ADC threshold (beyond the 5V).

Mr Starlet wrote:Ok, got the DFA, assembled, installed but not yet tuned. I've also got a Smart Mixture Display setup show mixture. What would be a good approach to tune the DFA on the MX5? The instruction came with the kit recomends removing the closed-loop (disconnecting EGO sensor) then tune for smooth idle and off to the dyno for high load tuning.

How did you tune your DFA?

i explained this in another thread so i'll just paste it in here:
i just tuned mine myself by watching my air / fuel meter (edit-same one as yours i think minh) and doing loops of the on ramps at edgars rd and plenty rd, on the western ring rd . give it a blast at wot on the on ramp at edgars rd, turn off freeway at next exit, adjust dfa. give it a blast at wot on the on ramp at plenty rd, turn off freeway again. adjust dfa. give it a blast at the on ramp at edgars rd again...so on and so forth around a few times til i had it sorted.
but yeah a dyno would give much better results i'm sure. i didn't bother with a dyno tune tho as i figure it would be a waste of money as this is a stop gap measure til i get my turbo sorted.

but using this dodgy method i took my mixtures from being pretty much constantly in the "rich" range on my afm (while under high load) down to about midway on the "normal" range. i did not try to get it too far towards "lean" because my afm is simply not accurate enough , and i figured it wasn't worth the risk. but yeah, i certainly noticed a good increase in power (based on my "seat of pants dyno" Mr. ) and fuel economy (based on litres per 100km - which i did measure before and after).

Mr Starlet wrote:doesn't the timing changes when we're modifying the AFM signal?

brain not working properly today, but i don't think so. i have some ideas but i'll leave it to others to comment. but my gut says that you should be ok, i just can't figure out how to explain myself at the moment. i'm a bit drained from a very busy weekend.

Mr Starlet wrote:I decided to keep the relay incase anything happen to the DFA (blown fuse etc), in this case the relay returns to normal state which connects input to output (bypassing DFA).

that's good thinking. but in my case returing to normal state would be no good either, as my car won't run wiith these big injectors.

Mr Starlet wrote:Got the tuning map, thanks heaps James.

no worries

Mr Starlet wrote:I tune mine to give the widest load range to suit the signal response from the MXs AFM, currently idles at load ~95 and max AFM opening at 0, I could widen up the range at bit more to make full use of the 0-128 load range but this means when ACC-ON the AFM signal will exceed the ADC threshold (beyond the 5V).

i'm not understanding this at all. my car got nowhere near load point 95. it idles at about 65.
or did you use "fine" mode ? so you get more points and finer control, but i think this mode reduces the voltage range down to 2.5 v or something ? i can't remember the details anymore.

Fatty wrote:

Mr Starlet wrote:I tune mine to give the widest load range to suit the signal response from the MXs AFM, currently idles at load ~95 and max AFM opening at 0, I could widen up the range at bit more to make full use of the 0-128 load range but this means when ACC-ON the AFM signal will exceed the ADC threshold (beyond the 5V).

i'm not understanding this at all. my car got nowhere near load point 95. it idles at about 65.
or did you use "fine" mode ? so you get more points and finer control, but i think this mode reduces the voltage range down to 2.5 v or something ? i can't remember the details anymore.

The circuit setup instruction were pretty crappy actually, if you were to follow exactly what it says then you most like won't get as wide a load range as you should.

The input circuitry is basically a buffer IC follow by an amp with adjustable gain (VR1) signal then fed to an Analogue-Digital Converter, the ADC has input range of 0-5V, I then probe the operating range of the MX5's AFM (~300mv to 3.96V), adjust VR1 so that 5V is presented at the ADC when there's an input of 3.96V, this means that load 128 = 3.96V. At idle the AFM opens slightly so the voltage reading drops slightly (can't remember how far).

After the ADC digitised signal then fed to the main micro-controller for processing. The output from the MCU goes to a Digital-Analogue converter then a final amplifier stage with gain get by VR2 (or is it VR3 - can't remember) then out to the car's ECU, all you need to do is adjust so that the output shows the same as input (3.96V).

yep, what you did sounds a bit screwy to me minh. i agree with david (carsokay).

from what you did, i don't think your input will equal your output with no tuning changes.

and yeah 65 tuning points is heaps more than enough. you really don't need to use the whole 128 points.

carsokay wrote:You do not calibrate the DFA based on the actual max voltage output of the AFM. The 1.6NA AFM is what you would call a 5V output type even though at sea level, it outputs 3.96V with the ignition on but without the engine turning.

Yes that is true, it would be classified as the type that works within the 0 to 5V range, but that doesn't mean its actual output is precisely 0 - 5.00V. When AC-On my AFM output 3.96V (@ECU), when engine on it's around the 2.94V, maximum vane opening at ~340mV (I got this by physically open the vane (flap)); output voltage inversely proportional to load. As far as I can see, 3.96V is the highest the sensor will ever output to the ECU, that's the range that the DFA will see and need to cater for, why waste load resolution to cover for voltage range that it'll never see.

carsokay wrote:You must calibrate your DFA using a 5V input that you generate using the potentiometer supplied with the Jaycar kit. You do not calibrate it for 3.96V that the AFM outputs. The load points that you see represent the max and min output of the AFM as applied to the DFA's 5V scale. You will not get the full 128 to 0. Anyways 65 to 0 is more than enough steps especially when you consider that the SAFC-II by Apexi has only 12 "steps".

What you see / interpret as load points are simply what the ADC (IC4 - TLC548) tells the MCU (IC3). To get the ADCs full conversion range hence the maximum number of loads, you need to tune the DFA to within operation range of your AFM, in my case my range was 0 - 3.96V and I see no point to cover more than that. As for the number of load sites, yes you can get by with 12 or 65 or more, but to me the more resolution the better fine tuning you can achieve.

For those interested in tuning for more load sites; do your basic tune that's described in the article to get your offset (VR3) right, then apply these tuning steps...



My DFA is tune for when the ACC is ON but eventually go one step further and adjust the kick-in voltage so that the DFA turns on when the car start (alternator generates ~13.8V). Doing so I could tune the DFA to run on 0-2.94V "engine running range" to take full advantage of the 0-128 load sites.

I hope this helps for those looking to do something similar.

Cheers
Minh

EDIT: Even with 0-95 load range, I find that the 60 or so load site happens at normal street driving; idles at 95 and ~36 at about 3kRPM (normal driving gear change), that leaves 36 load sites from about 3kRPM to redline. So then if you have 60 load sites then I would imagine your first two-thirds are gone in normal driving...that leaves ~20 for the more critical stuff.

Interesting note: during tuning I notice the ECU goes open loop at 4.6kRPM. It also goes open loop on rapid acceleration, not sure whats been monitor to make this change...maybe change in AFM signal, rapid rise in RPM?, definately not throttle change coz the manual TPS is an on-off affair.

i think it's dependant on engine load, not sure how that is determined tho

carsokay wrote:You may be on to something Mr Starlet. I have a second DFA that is now a spare and will set it up for a max input of 4V and see how this goes. We are of course logically assuming that whatever changes we make with the DFA, the resulting modified voltage output need not and will not exceed 3.96V and that the ECU will not respond to signals more than this.

Give it a try David and you'll soon realise how misleading the instructions are. Basically you turn up the volume of the signal to feed the processing units then turn down the volume after the processing unit by an equivalent amount so that effectively the input is the same as the output.

took 10min from lunch and tune the DFA to work ONLY after car starts (increase switch-on voltage to 13.6V), recalibrated VR2 and VR4 for 0-3.00V operation so now I'd be taking the full 128 load sites when engine runs, sweet