Mistakes were made, my R34 Story

[MBS206]

If you want cranky idle, go back to the old school firing order, out a carb on it, and a massive cam.

No matter how much you try to tune in a brake, it is never quite like an old, cranky as hell, carb V8.

Actually, you can build a hemi V8, and get a super cranky cam for it, and as much as you try and smooth the idle out, it will still sound cranky as f**k!

[Kinkstaah]

On 4/18/2025 at 9:28 PM, Dose Pipe Sutututu said:

That's why it's lurching, idling shit, and stalled 7 times.

Your approach is not ideal, it's literally idle hunting. Let's say your normal idle timing is 20° 

You would approach the idle timing table like this 

800rpm 20°

900rpm 20°

950rpm 0°

1000rpm 20°

So when your car hits the 950rpm cell, there isn't enough torque so it falls down to the 900rpm cell which has enough timing to produce the torque to reach the idle target again.

Then you'll need to dumb down the PID for the idle control motor to essentially ignore the 950rpm cell.

There's an art to great braps, when you don't have the option to advance an intake cam at idle.

@GTSBoy going to be real mad 🤣

Yuck. But it's kinda the same principle, moving timing around to make car run bad but sound good....

I went for a more spirited tuning drive to iron out a few things. Car no longer stalls, and it loved to dip RPM low when you clutch in. After tuning this (and the 4-5 tables that HPTuners/LS1 OEM ECU wants for idle air correction) it's now behaving somewhat normal.

All expected because there's a new TB, new Intake manifold and that new TB which is 102mm vs 78... has an entirely different IAC passage which is smaller because ?????

Unfortunately at this point I went to make further fine tuning changes to avoid it flaring up, you know... _tuning_ the HPTuners dongle died. Well to be more accurate - The USB cable died in a fashion that anyone who has ever charged a phone will recognize.

After the app demanding I resync it 50 or so times (which all 'succeeded' but all failed) the device does not want to sync and I've logged a ticket with support to see if it's fixable. US Support said it was a 'storage issue' but after removing the SD card inside it and formatting it/fixing it the issue does persist.

Unfortunately usually the fix is "Turn in your old dongle and pay $700 for the upgraded one" it's cheaper because I get some free licence credits I unfortunately don't need. However I'm 10 minutes down the road from HP Tuners AUS/VCM so at least I won't need to post it, and logged a new ticket for support over here.

Definitely drives different. My SOTP dyno believes it's probably making 310-320kw instead of the 280 before.  It scrambles for traction a little now whereas it previously different. It drives like a bigger cam car up top even though the cam is smaller, likely due to the cam not being advanced 6 degrees. The timing is deliberately low and the fuel is very rich so who knows if this will improve on the dyno. It may, imperceptably.

Also funny is removing the pineapples definitely makes the car squat more and axle tramp less. So this behavior of having more top end, squatting more, and scrabbling for traction more makes me think = more power. But I could have just been sitting on the threshold of that kind of behavior before.

Time will tell if my butt dyno is calibrated right. I need the exhaust leaks fixed before dyno tune for obvious AFR related reasons - I repositioned the pipes but I'm not confident it's fully sealed even if it is better.

At least the car does drive around while I cannot tweak/tune it for now. And I have aircon again.
medium success!

[GTSBoy]

Noice. Mostly.

[Dose Pipe Sutututu]

21 hours ago, Kinkstaah said:

Yuck. But it's kinda the same principle, moving timing around to make car run bad but sound good....

What I suggest doesn't have a 40° delta and won't have all your lurching, idle hunting issues.

Not to mention it's how most shops do it too.

So yes on theory your method "works" however once it goes into that hunt it takes much longer to settle as the PID idle strategy is working off it's tits to try target the set RPM.

[Kinkstaah]

My idle is set at 950rpm though - Moving the timing around 20 degrees is not really what I'd call a calm idle.

That said... neither is chop, by definition. The LS ECU likes to adjust timing to hold idle as opposed to air. It'd work, but generally speaking there'd be a discrepancy in the base idle and the IACV would want to move the timing around anyway to maintain said idle.

I think I'm just going to keep the timing steady anyway. Preserve my engine mounts. 

My aircon is now officially regassed.
As the guy was reversing I noticed my reverse lights do not operate, along with my reverse cam.

This is a bit distressing, because 100% of guides talk about which wire to connect to backup cams as "the goes with the [other color] wire". Often when doing conversions.

Unfortunately the R34 colour wires aren't documented
Unfortunately I had a T56 Magnum gearbox with it's reverse switch, which also isn't documented.
Unfortunately there's definitely not documentation for people with both of these in the one car.
Unfortunately I forgot.

After many hours of this, I have a reverse cam and reverse lights again.


The wire going through the trans tunnel to the reverse switch had broken. Upon inspection, it looks like this one wire had about 7 spade terminals and extensions in it.. for reasons I cannot possibly comprehend.

I also spent the 750 hours required to clean up the wiring behind my head unit which now looks like this:



This is a monumental improvement relative to what used to be there WRT triple gauges, head unit, traction control, wideband controller, and whatever the f**k OEM stuff still exists there in various states of connectivity/needed.

Next step is to check in at the Exhaust shop to see/confirm how much clearance I have, to decide what mid mufflers or 'resonators' (which are just straight through, narrower mufflers) I can add and hopefully cut out a lot of exhaust leaks, pinhole, v-band or otherwise.

But first step will be to 'take a look' before the next step.

[Kinkstaah]

Alright.

Exhaust has been looked at, and booked in 'soon'. I'm not 100% convinced it's going to be as good as possible but I'm going to go with it anyway. If I get a reasonable thing that works for 10 years I mean that's twice as long as I've lived with the current one.

I have a more pressing issue.

I have fixed my MPVI3 (by buying a new one).

Excellently, one can wire in analog 0-5v inputs to the ECU itself. I had wideband already via Serial so I also wired it in via the direct input. The idea being I can use the standalone logging without a laptop and have Wideband data in it.

The other wire I thought I'd use oil temp. This is where I've gone crazy.

HPTuners requires you to implement a formula so you know how much volts = how much temp. This seems relatively simple to me. However I cannot find the scale for this anywhere on the internet, nor decipher how to figure it out without removing the sensor from the car.

All I know is that voltage actually goes up as temperature goes down. I am using the actual gauge, so I can see what the temp is. The signal wire has been branched off into the MPVI3.

EXAMPLE:

2.68v = 58C
2.7 = 57C
2.74v = 56C
2.8V = 54C
3.06V = 49C
3.18V = 47C
3.37V = 43C

I think the gauge is 50-150C. It may be more. It may be less, because I can't find it for the love of f**k.
It appears all the information about the gauges I have has been scoured from the internet, but the sender is VDO 320.021

I believe the resistance chart is this

How the f**k do I convert this to a voltage?
Once I have a voltage... I then have to perform this transform of said voltage to show it in the scanner:

https://www.hptuners.com/help/VCM-Scanner/Content/vcm_scanner/defining_a_transform.htm

@GTSBoy you're probably my only hope here

TEST YOUR MIGHT

[Duncan]

*that* is a question for chatgpt (or someone that understands maths....but I'd try chatty).

Drop them the table and ask for the formula

[GTSBoy]

Um....

This is not easy. I will have a fiddle with it, but make no promises. I think, given the explicit resistances vs temperature, and your ECU measured voltages, it should be possible to match them up. I will do that by assuming a 5V supply from ECU to sensor, find out how much current would have to flow to produce that voltage, then use that current as a .... nah f**k. That won't work. The supply voltage is fixed, but there's nothing to say that the ECU regulates the current too. Um...

I dunno.

I guess there's a voltage divider setup in the ECU. As in, there's another resistance between the ECU's sense terminal and ground. That way, as the sensor resistance changes, that's how they get a floating voltage that varies with the temperature, without having to control the current.

Give me a minute or 3.

[GTSBoy]

Nah, it's not a simple voltage divider. I'm not enough of an electronics guru to know how they make these circuits work.

If I had a better idea of how the ECU's temperature measuring is done, I could then actually do as you want, which is turn that resistance chart into a voltage chart. But my approach has not worked.

What I did was interpolate the sensor ohms values for the temperatures you listed, as you did not have any of them on a temperature ending in zero or 5. These are:

°C	ECU V	sensor ohms (interpolated)
58	2.68	11.85
57	2.7	11.89
56	2.74	11.93
54	2.8	12.01
49	3.06	12.208
47	3.18	12.284
43	3.37	12.42

I then assumed 5V supply to the resister and calculated the voltage drop across the sensor for each of those, which is just 5 - the above voltages, and then calculated the current that must be flowing through the sensor. So you get:

						Values in sensor
°C	ECU V	sensor ohms (interpolated)			Supply volts	Volt drop	Current
58	2.68	11.85			5	2.32	0.195781
57	2.7	11.89			5	2.3	0.19344
56	2.74	11.93			5	2.26	0.189438
54	2.8	12.01			5	2.2	0.183181
49	3.06	12.208			5	1.94	0.158912
47	3.18	12.284			5	1.82	0.14816
43	3.37	12.42			5	1.63	0.13124

And then use that current and the ECU's sensed voltage (which must be the voltage drop across the in ECU resister is there is one) to calculate the resistance of that in ECU resistor. You get:

						Values in sensor			Other resistor
°C	ECU V	sensor ohms (interpolated)			Supply volts	Volt drop	Current		Volt Drop	Resistance
58	2.68	11.85			5	2.32	0.195781		2.68	13.68879
57	2.7	11.89			5	2.3	0.19344		2.7	13.95783
56	2.74	11.93			5	2.26	0.189438		2.74	14.46381
54	2.8	12.01			5	2.2	0.183181		2.8	15.28545
49	3.06	12.208			5	1.94	0.158912		3.06	19.25592
47	3.18	12.284			5	1.82	0.14816		3.18	21.46325
43	3.37	12.42			5	1.63	0.13124		3.37	25.67816

And that's where it falls apart, because the resulting resistance would need to be the same for all of those temperatures, and it is not. So clearly the physical model is not correct.

Anyway, you or someone else can use that information to go forward if someone has a better physical model. I can also show you how to interpolate for temperatures between those in the resistance chart. It's not fun because you've got to either do it like I did it for every 5°C range separately, or check to see if the slope remains constant over a wide range, then you can just work up a single formula. I'm just showing how to do it for a single 5° span.

For the 58°C temperature, resistance = 11.77+2*(11.97-11.77)/5

The calc is a little arse backwards because the resistance is NTC (negative temperature coefficient), so the slope is negative, but I'm lazy, so I just treated 58 as if it was 2 degrees away from 60, not 3 degrees away from 55, and so on.

[MBS206]

The circuit if not a resistor divider is using an opamp to deliver a specific current normally. By maintaining the current as resistance changes, the voltage does too.

Add to that, thermistors are normally non linear too,which can make creating a function impossible. Most uses of a thermistor people will utilise a lookup table to get the corresponding temperature.

[MBS206]

Can you measure resistance between the pin you're wiring to, and the sensor ground?

[Kinkstaah]

So to further complicate things or hopefully explain better:

This is a VDO sensor. It clearly does this math inside the gauge itself... because it displays a temperature in C to me instead of a voltage. The signal wire to the gauge is in Volts, not Ohms.

I have just teed into this and sent it to the ECU.

ChatGPT spat this out:

Temperature (°C)	Thermistor Resistance (Ω)	Voltage Output (V)
-40	36563.56	4.973962698
-35	26284.63	4.963854097
-30	19149	4.950518086
-25	14127.68	4.933166097
-20	10504.68	4.910527969
-15	7721.35	4.879055954
-10	5720.88	4.838133512
-5	4284.03	4.786165799
0	3240.18	4.721119717
5	2473.6	4.640900563
10	1905.87	4.543692515
15	1486.65	4.429695182
20	1168.64	4.296344225
25	926.71	4.144091368
30	739.98	3.972492431
35	594.9	3.782907287
40	481.53	3.577860996
45	392.57	3.361217186
50	322.17	3.136573398
55	266.19	2.908608143
60	221.17	2.68039363
65	184.72	2.455599277
70	155.29	2.239608872
75	131.38	2.035132288
80	112.08	1.846579676
85	96.4	1.674774149
90	82.96	1.511882199
95	71.44	1.359001674
100	61.92	1.222169588
105	54.01	1.100403407
110	47.24	0.989775394
115	41.42	0.889528391
120	36.51	0.800974069
125	32.38	0.723478416
130	28.81	0.654148313
135	25.7	0.591893137
140	23	0.536380597
145	20.68	0.487551867
150	18.59	0.442640126
155	16.74	0.40213318
160	15.11	0.365841848
165	13.66	0.333073247
170	12.38	0.303758956
175	11.25	0.277572169
180	10.24	0.253917873

[MBS206]

Yeah, it looks like it's being ran as a constant current setup is my guess.

Get some more data points, reading the degrees C on the gauge, and what the voltage is you're seeing out of it.

The sensor itself is not linear, even in smaller ranges of like 20 - 60 it's non linear. But if we can get some good plots, particularly at those 5volt steps, it will be very easy to work out, and then I can give you a logarithmic formula if its you need an actual formula.

Also, very likely ChatGPT will lead you astray with some of this stuff, as it totally depends on how the circuit is built.

[Kinkstaah]

Yeah - I tried throwing the formulas from ChatGPT and it tried it's best as it told me that the Formula in HPTuners as requested is a Linear function, but the original data I provided it (resistance to temp) is not linear and thus it kept trying to suggest formulas that would be more accurate in regions I actually want it to be accurate in.

But I didn't quite understand it at the time. I have now thrown the data into graphs and can clearly see they're all different shapes. Given it's an oil temp sensor I probably want it to be most accurate between 80 and 120c - So I don't really see much alternative other than driving the hell out of the car and letting it cool down from 120C, noting the voltage on the sensor as it chills out.

But at that point if I write it down, well, I'll be able to know this from the voltage. Would look pretty stupid on a datalog or a video showing me going down the straight at 0.265V Oil temp 

[MBS206]

8 minutes ago, Kinkstaah said:

Yeah - I tried throwing the formulas from ChatGPT and it tried it's best as it told me that the Formula in HPTuners as requested is a Linear function, but the original data I provided it (resistance to temp) is not linear and thus it kept trying to suggest formulas that would be more accurate in regions I actually want it to be accurate in.

But I didn't quite understand it at the time. I have now thrown the data into graphs and can clearly see they're all different shapes. Given it's an oil temp sensor I probably want it to be most accurate between 80 and 120c - So I don't really see much alternative other than driving the hell out of the car and letting it cool down from 120C, noting the voltage on the sensor as it chills out.

But at that point if I write it down, well, I'll be able to know this from the voltage. Would look pretty stupid on a datalog or a video showing me going down the straight at 0.265V Oil temp 

Extend the sensor out of the car, put it in a pot of oil, heat oil up.

You'll get to 80-120 then 😛

[Kinkstaah]

I should have prefaced all of this with "I'd really like to not pull anything out of the car for this" 

[The Bogan]

5 minutes ago, MBS206 said:

Extend the sensor out of the car, put it in a pot of oil, heat oil up.

You'll get to 80-120 then 😛

You could check/calibrate that with a thermometer, we have long glass thermometers at work which go from -10 to +150°c, they are pretty cheap too I believe,  as monkey fisted soldiers tend to break them, and getting new ones don't seem to send red flags to the people that hold the purse strings

Edit: after a google they apparently cost about $80, but, digital ones seem to be much cheaper at around $40

[GTSBoy]

Do not ever trust ChatGPT with anything math related. They can't do math. They have no idea what it is.

With enough data we can fit a decent equation in Excel, or if the available fits in Excel aren't good enough, a Matlab clone.

[MBS206]

1 hour ago, GTSBoy said:

Do not ever trust ChatGPT with anything math related. They can't do math. They have no idea what it is.

With enough data we can fit a decent equation in Excel, or if the available fits in Excel aren't good enough, a Matlab clone.

Put in excel, graph it, fit a trend line, display formula.

I totally am not reinstalling MATLAB. Hopefully you have it installed if you want to use it 😛

But if HPTuners can only fit a linear equation, no need for any of the above with some simple mafs...

[MBS206]

2 hours ago, Kinkstaah said:

I should have prefaced all of this with "I'd really like to not pull anything out of the car for this" 

Unplug the thermofans and let it idle... Watch the temp gauge as it all goes up... Record the values.

And done.