[electronics people get in here] 0-5v in, adjusted 0-5v out?

[Roman]

Hey people,

 

I've found a few instances where it would be handy to have a programmable magic box which can take an input 0-5v signal, and output a modified one.

 

In the instance of installing a larger throttle body, since more air comes in at lower throttle openings you would want to change the linear 0-5v signal to start showing higher voltages quicker, so the ECU tips in the correct amount of fuel prior to the MAF or MAP sensor catching up.

 

Also for my airflow meter, I would want to slightly bump up the signal when it's reading very low load, as it's advancing the ignition timing too far.

 

I figure the easiest way to 'tune' such a thing would be to have two graphs side by side on a GUI on a laptop or something which could plug in via USB, and you can just drag a line on a graph for how you'd like both the input and output voltages to relate to each other.
Or maybe select if the input vs output is linear, exponential, etc.

 

Does anything similar to this, exist off the shelf already?

 

Before anyone says "Get an aftermarket ECU you cheapskate" I could fix up minor issues with such a device for pretty damn cheap, compared to $3000+ for an installed and tuned ECU that acheives not much over the factory ECU apart from getting an extra 500rpm before my engine runs out of puff.

 

Suggestions go!

[Yowzer]

Jaycar used to do a programmable voltage modifier kit that does exactly that, but dunno if it's still available.

I think they might sell a preassembled module that does it, but can't remember.

[ta63-1uzze]

I think you are thinking about it a little bit wrong, it's all done via scale from .5 to 4.5 . Anything out side of that is how the computer sees and recognizes a fault , so tickling it up at lower loads would make it over compensate at higher loads or openings making it go above the threshold and into fault.

Is there something a little more precise that you are talking about, what is the issue that you are trying to get around?

Can I ask why you want a bigger throttle body? Do you believe it's a restriction at the moment?

I am intrigued as to where this is going, throttle bodies are definitely a volumetric designed device and unless you have substantially increased the air induction of the engine there is not allot of good that comes from increasing its size

[sentra]

use an arduino it will take 5 seconds

hide it in one if your empty mcdonalds cups

[Roman]

I think you are thinking about it a little bit wrong, it's all done via scale from .5 to 4.5 . Anything out side of that is how the computer sees and recognizes a fault , so tickling it up at lower loads would make it over compensate at higher loads or openings making it go above the threshold and into fault.

Is there something a little more precise that you are talking about, what is the issue that you are trying to get around?

Can I ask why you want a bigger throttle body? Do you believe it's a restriction at the moment?

I am intrigued as to where this is going, throttle bodies are definitely a volumetric designed device and unless you have substantially increased the air induction of the engine there is not allot of good that comes from increasing its size

I dont want a bigger throttle body - but I have a bigger than standard one fitted, as it's a requirement of changing to a RWD plenum for my engine.

 

The TPS on this throttle body runs counter clockwise (or the otherway compared to other Toyota TPS's) and seems to run across a lower voltage/resistance range than the factory one.

 

As according to my OBD scanner the ECU thinks the throttle is only opening between 8% and 50%, checking with a multimeter confirms it looks to have a different resistance range. So the easiest solution seems to be to 'spread out' the voltage a bit so the OBD reads closer to correct.

 

With a larger throttle body the accelleration enrichment isnt quite correct anymore, so ideally you need to add more air (voltage)at lower throttle angles.

So changing from a linear angle vs voltage to non linear so it ramps up the voltage quicker (But stops at the same peak) would effectively simulate the standard throttle body as far as the ECU is concerned.

Not a huge deal for the TPS and accelleration enrichment, but the first scenario, I dont know any other ways around it apart from bumping up the low load voltage.

At the moment when my car is at very low load, (say cruising 80kph) the motor starts pinking.

Also the car doesnt idle well currently, after cleaning and inspecting the ISCV it's all fine. But if I check in my OBD program, the ECU thinks that less than HALF the amount of air is coming in at idle, compared to what my Toyota Echo idles at. (5grams/sec in Echo, 2grams/sec in Carina)

I've checked for post MAF air leaks and done some troubleshooting, but cannot seem to pinpoint it to anything.

So whatever the MAF output voltage is when there's that miniscule amount of air coming through, I'd like to bump it up a bit so it's reading the correct signal which would correct the idle issues and also reduce ignition advance at low load to stop the pinking.

Normally I'd expect the knock sensor to take care of that, but perhaps it's a harmonic deadzone knock sensor wise on account of some modifications to the engine etc. And it seems to point back to the MAF being the cause of the issue regardless.

 

[UTERUS]

Can you fit a beams tps so the ecu knows wassup?

[Roman]

The throttle body that I am using is from an Altezza beams engine, which normally runs the electronic throttle body.

 

I've disabled that so it's just completely actuated by the cable. But on account of the wizard magic that's normally in the throttle body housing, the TPS needs to be on the other side so it rotates in the opposite direction to usual.

 

The FWD beams motor throttle body (to suit my ECU) is a little bit smaller and wont bolt to my plenum, and I cant swap the TPS over as it rotates in the opposite direction. (As per every other Toyota TPS that I've found...)
 

[Roman]

 

[Celica RA45]

dave afew of my mates bought the altezza manifold off me ,when they did the red top to rwd and made a plate so the std red top throttle bolted up to the altezza plenum

you might have to do the same

[ta63-1uzze]

I haven't had the time to completely read everything. Just a quick scroll , but what about making a adapter plate and just the video correct throttle body? Nice and simple. And if done well will look good too. No magic box required?

[ta63-1uzze]

Ah phone fail,was mentioned to say, use the correct throttle body and sensors,

[coldturkey]

TLDR - but there are many ways to skin a cat. It sound like you want a custom relationship between input signal and output? Could be hard to do with analogue electronics. If it was somewhat linear or exponential that you wanted then it could be achieved with transistors or amplifier circuits.

But the way I would do it is with a microprocessor, then it is infinitely customisable and you could set up any relationship you want.

I'll happily make up a circuit for you if you can decide decide what you want, piece of cake.

[downtrail]

Haven't really played around with obd scanners but are you able to use it to callibrate tps values? ie when closed look and burn that value then jam it to wot and look and burn that value? if not then use the correct tps for that ecu. dosen't matter if it works in the reverse rotation just wire it up backwards insted just swap the +5v and earth cables and your away.

[Llama]

use an arduino it will take 5 seconds

hide it in one if your empty mcdonalds cups

 

By far the easiest way if you can't be bothered messing around with a PIC or something.

[Ned]

TLDR - but there are many ways to skin a cat. It sound like you want a custom relationship between input signal and output? Could be hard to do with analogue electronics. If it was somewhat linear or exponential that you wanted then it could be achieved with transistors or amplifier circuits.

But the way I would do it is with a microprocessor, then it is infinitely customisable and you could set up any relationship you want.

I'll happily make up a circuit for you if you can decide decide what you want, piece of cake.

Cant think of an off the shelf solution either...

Happy to help also.

Simplest way that comes to mind is using a Mega AVR with 10 bit ADC and a USART. use the ADC to read in your 0-5V signal and have a lookup table with 1024 PWM values in it that you program into the EEPROM using a USB-Serial cable (the 5V FTDI one). Then use the PWM values and put them through an RC filter to get a smooth voltage for the ECU

Only questions are, what resolution do you need, what speed do you need and hopefully the input to the ECU is high impedance.

Resolution = how many steps do you need between 0 and 5V? an AVR can reliably do 256 but it can do up to 1024 well enough and with some averaging it will be pretty good. 5V over 1024 steps is ~5mV per ADC step.

speed = how many times a sec do you need to update the ECU? if its too fast, then the simple RC filter might not keep up at the resolution required etc. Plus the faster you need it, the faster the ADC needs to be sampled and if you need high resolution, you also need to average the input to try get rid of the noise on the bottom couple of bits.

If the ECU just samples the input with its own ADC and uses it to compute whats going on, then that will be fine, but if it has a load on the input (which it might to do combat noise) then a simple RC filter might not cut the mustard and you'll need to put in a rail-to-rail op-amp to drive the output from the RC filtered micro output (which is a good idea regardless really)

[h4nd]

> I've disabled that so it's just completely actuated by the cable. But on account of the wizard magic that's normally in the throttle body housing, the TPS needs to be on the other side so it rotates in the opposite direction to usual.

 

IF the sensor is a potentiometer,

   AND IF it is linear,

   AND IF the extremities will still work,

   AND IF it has a 3 wire connection,

then you can reverse the + and - feed, and keep the slider the same, to reverse the sense direction.

[Testament]

maybe a TPS off something else that gives the correct signal and rotates in the correct direction? somehow goto pick a part and test them all?

[slacker.cam]

The voltage modifier that Silicon Chip did a few years ago is pretty much the only 'off the shelf' product I can think of: http://archive.siliconchip.com.au/cms/A_111774/article.html

I think it would take care of your two problems fairly simply.

 

As already mentioned by H4ND you should be able to swap two of the three wires on the TPS to reverse it's action. If it's also giving you the wrong range I can think of a couple of dirty hacks that could help to scale the output. Basically your TPS is acting as a variable voltage divider between GND and some reference voltage (probably 5V, 8V or 12V). So you have GND on one terminal, the reference voltage on another, and the variable voltage output on the third. If you want to double (for example) the output then all you need to do is double the reference voltage. You need to be careful that the maximum voltage that the ECU sees is less than the reference voltage otherwise you're likely to see smoke or possibly start relying on the protection circuitry in the ECU which is not great. Building a small high accuracy reference voltage supply is 100x easier than any other option I can think of (including the voltage modifier above). 

[Ned]

+1 on what cam said.

if you dont need to do fancy pants log or exponential curves etc, then an hour with some resistors, a multimeter and maybe a voltage reg will get it to real 0-100% surely...

[slacker.cam]

Another couple of thoughts.

 

Does the ECU have a TPS calibration mode? Most do. 

 

Maybe the ISC valve is working sweet but it's fully opened when running and still not enough air is flowing. What happens if you crack the throttle stop a small amount to allow some more 'base' idle flow. I can imagine this could have been messed up through the process of swapping throttle bodies? Bit of a long shot and maybe you've already checked...