Before I started this repair, the combination gauge with the compass and acceleration didn’t do anything. The arrows and backlight didn’t even light up. I wanted to give a quick overview of what I did to get them up and running.
Backlights
The backlights were a pretty easy fix. 1 or 2 bulbs were dead while the others just had some corrosion on the contacts. I removed all of the bulbs and swapped them for LEDs (4x 74-GHP3: Green from Super Bright LEDs). I also took some high grit (maybe 600) sandpaper and cleaned the contacts until they had a slight shine. After that, all good! You can power up the unit by applying 12V where shown below. The LED lighting is definitely less evenly distributed than the incandescent but it also runs so much cooler and requires very little current. There were actually some melt/burn marks on some of the plastic in there due to the incandescents.
Firing up the VFDs
So at this point the VFDs didn’t turn on at all so I opened up the case and started poking around on the PCBs. I quickly realized that the two big caps had leaked all over and were long dead and dried out. I didn’t take a good before picture but you can see the residue in the after picture below. This is after some scraping and cleaning with IPA.
Quick tip on desoldering those. Grab the old caps with pliers and rotate one way and then the other and you’ll rip the capacitors off the leads. Then use your soldering iron to desolder the leads. The alternative is desoldering them from the bottom but that is very difficult since you then need to separate the two boards enough to access the underside of the top board.
You can solder in the new caps from the top and it will be fine. Just apply heat long enough that the solder can flow down into the through hole and to the other side. You can purchase any electrolytic capacitor that is 470uF and 35V+ since that is what was there originally. I got these at microcenter for like $4 total.
Success! The VFDs now light up. The picture below is without the backlights. Either I had them taken out at the time or I didn’t apply 12V to one of the leads.
I ended up putting the gauge back in the car at this point at giving it a shot. The compass worked! However, the acceleration and mpg did not…. The FSM describes how to check the signals coming into the gauge and I went ahead and did that – all looked good. The gauges were getting the right data, they just weren’t using it. Back to the PCBs we go.
I needed a way to simulate the car’s speed sensor so that I could work on the gauge inside and know when things are working. I think the speed sensor is an LED and encoder so you supply 5V for the LED, and the output is just a digital wave. The frequency / period of the wave is directly related to the car’s speed. So I wired up an arduino so that I could output a 5V digital signal to the gauge.
You want the arduino to simulate acceleration and deceleration from -0.5->+0.5 g’s. So I headed on over to excel to figure out what speeds I need to trick the gauge into seeing. The FSM mentions that the encoder is divided into 24. So each revolution of the encoder gives 24 waves. So once you calculate a speed you then have to convert that speed to a wave’s period by using the wheel’s circumference and the 24 factor.
Time (s) | Accel (g) | Accel (fps/s) | Speed (fps) | Speed (mph) | Tire Revs/Sec | Freq of Wave (hz) | Period of wave (ms) |
0.1 | 0.01 | 0.3 | 0.0 | 0.0 | 0.01 | 0.1 | 7046 |
1.0 | 0.10 | 3.2 | 1.8 | 1.2 | 0.33 | 7.8 | 128 |
2.0 | 0.20 | 6.4 | 6.8 | 4.6 | 1.24 | 29.8 | 34 |
3.0 | 0.30 | 9.7 | 15.0 | 10.2 | 2.75 | 66.0 | 15 |
4.0 | 0.40 | 12.9 | 26.4 | 18.0 | 4.85 | 116.4 | 9 |
5.0 | 0.50 | 16.1 | 41.1 | 28.0 | 7.54 | 180.9 | 6 |
So you can see we want the arduino to initially send very slow waves and over the course of 5s it will speed those up until they are 6 ms long. So I wrote the little bit of code below that uses digital pin 13 to send the appropriate digital signals to the gauge.
[gist https://gist.github.com/derglow/f926ab8df2d38c2f61ee1bde8bf1e1e3]
So now I could start troubleshooting the acceleration gauge. I recently bought a little $25 pocket oscilloscope off amazon and it was super useful here. Here’s a look at what the arduino is sending:
[youtube https://www.youtube.com/watch?v=g2W8VQpPehA]
Next I googled the chip nearest to where the speedo signal connects to the board. It’s a binary counter, so it takes a digital waveform and converts it into a running count. The datasheet also shows which pin is the input pin for the digital signal. I measured there with the o-scope and saw nothing. At this point I didn’t know what to do. There are about 20 components in this area and all looked visually ok. So my last ditch idea was to flip the board over and resolder all of the contacts. I think this is what people do to fix the dash power supply a lot of the time. You have to desolder a few things to get to a point where you can flip this PCB over. This is a medium difficulty job and not very beginner friendly.
It worked! I put it all back together and powered things up again and applied the arduino signal and voila!
And here it is in action:
To put things in perspective, these things run about $75 used. So you have to decide if this is worth the trouble. This one came from a turbo and I couldn’t find the turbo’s gauges for sale at the moment on ebay or 300zxpartsforyou. So in my case definitely worth it.
Hey I have the same cluster and for some reason the backlighting is completely “swapped”. the lights are completely off when I use the dimmer switch and put everything else to full and vise versa. The lights won’t turn on unless I turn on all the lights in the car. Any idea why?
Man, mine works when it wants to: most of the time ALL THE SEGMENTS in the VFDs light ups, if I turn it off and on it will sometimes work as it should or some times random segments of the VFDs light up. I swapped the same capacitors and resoldered a few connectors, but I don’t know what else to do.