Retrofitting an '88 Legend FE display.

[GasSavers_bobski]

Hey folks... I'm starting a new thread rather than further hijacking someone else's. I guess I'll start off with quoting the relevant posts:

Originally Posted by bobski

We've had the technology for decades. I've been working on retrofitting a display into my CRX:

It's the driver information display from an '88 Acura Legend... $30 from salvage. The only down side is that it's meant for a V6... The injector signal pulsewidth needs to be scaled 2/3 to work properly with my inline 4. It will still show changes in FE, just not an accurate measure in MPG or l/100km.

Originally Posted by skewbe

bobski, that's really cool to be reusing that component. I've been considering similiar 2/3 type circuits for other applications too.

Do you know the pinout on that thing or if the mpg/trip function will work with just volts and a vss/injector signal? Or is it integrated into the honda ecu?

Originally Posted by bobski

No, everything is thankfully external. I traced the single wire that runs to the ECU a few days ago... Inside the case, it runs to the no.1 injector pin via a 10k resistor. That's it.
The device actually has a whole bunch of functions: Clock (the unit takes the place of the clock in the legend), distance to empty indicator, trip odo, current fuel economy, average fuel economy, engine oil level, coolant level, washer fluid level, service and oil/filter inverval odos. It isn't installed yet, so there may be other functions lurking in the software that I haven't come across with simple power and ground.

Lets see... The wire from the lighting switch is 12v+ to dim the display, the pink wire from the ECU is the switched ground injector signal via 10k resistor I was talking about above.
The fuel level sending unit needs to output 2-5 ohms when full, 25.5-39.5 ohms at half tank and 105-110 at empty. Conveniently the same values as the CRX sender. I haven't looked at the low fuel sensor... Judging from the circuit diagram, it's the same design as found in other Hondas and Acuras. I can look up specs if you want.
The coolant temp sender needs to output 158 ohms at 54?C/129?F and 32-49 ohms at 85?-110?C/185?-212?F. Also stock CRX values.
The fuel filler door switch is part of the door release lever, so I assume it's triggered momentarily when you pull the release... I haven't investigated further.
The engine oil level switch is in the oil pan... Too much work as far as I'm concerned. The only odd thing I noticed is that the corresponding wire in my donor car showed about 50 ohms of resistance. Not no connection or good continuity, but 50 ohms. The dipstick showed way over full, so I assume water got in the crank case... The switch may have been damaged.
The coolant level switch is built into the lid of the radiator overflow bottle... Simple on/off operation via a reed switch and float with internal magnet. The washer fluid switch is installed in the bottom of the bottle, similar design to the coolant switch.
The speedo and pulser are cable driven, so I assume that's also typical Honda spec... Four 12v pulses per cable revolution or whatever it is.

If you're going junkyarding, the display is on the top of the dash above the radio, HVAC controls and such. Make sure you've got an info display equipped car before you start disassembling things... The standard issue clock looks very similar to the info display when unpowered.
The control unit is above the glove box. Once disconnected from the box, the wiring harness between the display and control box can be pulled back towards the display intact. That way you don't have to deal with resoldering the wad of wires between the two.


So do you have a diagram for a good 2/3 pulsewidth scaler? It strikes me as a problem that might necessitate a processor, but I don't really know. Simply dropping every 3rd pulse might work as well, but it's hard to say how the info center would react to that.

Originally Posted by skewbe

I don't have a solid schematic yet, but for an application that can be off a few pulses I'm thinking an analog frequency->voltage->voltage Controlled Oscillator circuit will git 'r done (with appropriate input and output voltage tweaking).


from:
http://www.uoguelph.ca/~antoon/gadgets/pll/pll.html

Here is a frequency to voltage chip that looks like it is built for a magnetic pickup, fyi:
http://www.national.com/pf/LM/LM2917.html

so basically we have
vss->LM2917->trim pot voltage divider to adjust ratio->VCO->voltage divider for instrument input.

Edit: oops, you are trying to divide the injector pulse, not the vss, you need to detect the duty cycle (read inductor) and make a new signal with the appropriate duty cycle...

edit: oops, looks like we are hikjacking a thread...

[jwxr7]

that is a great idea , Why haven't I thought of trying that? Saves alote of the hard stuff of DIY. Would it be better to stick with ones out of any pre obd2 car if it is going to go in an obd1 car? Seems like it would be simpler.
I'll have to do a walk around the junkyard and see what they have. What other cars would have a factory mpg read out? My buddy has a 95 newyorker with one. Most of the others I can think of are obd2.

I am geeked about trying this .

[CO ZX2]

Originally Posted by jwxr7

that is a great idea , Why haven't I thought of trying that? Saves alote of the hard stuff of DIY. Would it be better to stick with ones out of any pre obd2 car if it is going to go in an obd1 car? Seems like it would be simpler.
I'll have to do a walk around the junkyard and see what they have. What other cars would have a factory mpg read out? My buddy has a 95 newyorker with one. Most of the others I can think of are obd2.

I am geeked about trying this .

Won't you need a unit compatible with you throttle body injector? And if you could only find a V6 unit, halve the readings?

[mrmad]

I don't see the circuit skewbe showed would work for what you need. The LN2917 is a frequency to voltage IC, meaning it will take a certain frequency and convert it to 0-5V. The input frequency is adjustable by changing different capacitor and resistance values on its inputs.

The next circuit then takes voltage measurements and converts them back to frequency. The problem here is you need to be meaursing pulse widths, and the output would have the same pulse width. Plus, the Legend display would be assuming the Legend's injector size.

I would think the easiest thing to do would be to connect it up and see if you can get it to read mpg and use it as a relative measure of mpg.

[jwxr7]

Originally Posted by CO ZX2

Won't you need a unit compatible with you throttle body injector? And if you could only find a V6 unit, halve the readings?

It would probably be easier, but a multiport would work too. I would just have to come up with a multiplier that I can go by. It would be a useful instrument even if it is way off. I could see instant feedback which would lead to improved FE driving.
like mrmad said

I would think the easiest thing to do would be to connect it up and see if you can get it to read mpg and use it as a relative measure of mpg.

I remembered my ex-girlfriend's 89 pontiac grandprix had a trip computer, and a friends 1990 olds cutlass supreme had that as an option too.

[skewbe]

Originally Posted by mrmad

I don't see the circuit skewbe showed would work for what you need.

Nope, it won't. It might be useful for calibrating the vss signal (i.e. if you had different sized tires than the computer expects) though.

I think the "put it on the passenger seat and wire it up and see how it does" approach is a reasonable suggestion.

I would be tempted to do the hiway mile marker thing compared with the trip and see if there is a factor you need to multiply the miles by, then compare actual fillups to the mpg readings and see if there is a fuel consumption multiplier (divisor).

Once you know the correction factors, you should be able to calculate a displayed MPG to use as a target for your honda and know how to convert it to an accurate MPG number. So basically you drive around with a not-so-arbitrary figure in your head of what the display should say if you are trying to maintain, say, 50mpg. Throw a calculator in the glovebox if necessary along with the conversion factors but it would be good to see that it does work on it's own without investing in extra circuitry, heck it'll probably be "good enough" to be very helpful as-is.

[GasSavers_bobski]

Originally Posted by mrmad

Plus, the Legend display would be assuming the Legend's injector size.

The Legend and the Civic injectors in my CRX are the same flow rate - 240cc/min. It's just that the Legend has 6 injectors where my CRX has 4. The display only measures the pulsewidth from one fuel injector, so it must be effectively multiplying that by 6 internally to get the total fuel flow.
I'm gonna to pick up some alligator clip jumpers and such to make temporarily rigging this thing in place easier. With any luck, I'll have it spitting out a reading this evening.

[GasSavers_bobski]

It works!
The speed signal appears to be spot on. The stock mechanical trip odometer went from 172.4 to 191.1 miles over the course of the test drive, the info display trip odo read 18.7 miles.
The FE reading seems to average over the course of a few seconds. Coasting to a stop results in the MPG number gradually dropping to 0 a few seconds after you actually stop.
Cruising at around 50 mph.
It maxes out at 99 MPG. Too bad... The numerical display reads up to 199,999.
I'm not sure about the accuracy of the FE reading yet. It's definately displaying on the fly, but the average reading isn't right. I pulled into the garage, shut the engine off, fiddled with the cooling fans for a minute and then started trying to take pics of the trip milage reading.
When I gave up on the trip milage picture, I flipped to the average FE reading. It read 11.3 MPG... Seemed more than a little low, even with correction. As I stared at it, it ticked down to 10.8 MPG, then 10.4... This is with the engine off, mind you. Like a hypermiler's nightmare or something.
I think I know what's happening. The injectors are a switched ground system. When the ignition is on, one side of each injector is constantly powered. That means the wire between the ECU and injector has 12v+ until the ECU decides to fire the injector, at which point the injector's internal resistance limits current flow, the ECU sinks what current does make it through and the wire appears more or less grounded. So, the info unit takes 12v+ on the injector wire to mean the injector is closed, and ground to mean the injector is firing.
I was running the info display off a cigarette lighter outlet, which is switched on and off indirectly by the ACC contact of the ignition switch. With the ignition off and ACC power on, the injectors weren't getting any power, so the injector wire had 0 volts and maybe even a connection to ground. The info unit read this as the injector firing at 100% duty cycle while the car was standing still. The average FE plummeted.
So, I'll have to wire it up to ignition switched power, reset the average MPG meter and give it another try.

[skewbe]

cool beans, let us know if the mpg reading starts making sense (note my saturn reads .1gph with the engine off too ?!?)

[GasSavers_TomO]

Awesome! I'm going to my old junkyard and pick one of these up tomorrow. I'll get cracking on a How-to for the OBD1 guys.