5/21/22 I removed the Nixie dash system from the Corvair and am building the NDCA board for the Volvo. The board has SMT parts that I sourced from my engineering stock. I needed to steal a male RA DE-9P from the office. Done. I am also missing the Zener diodes. I'll omit them for now. I hope to make the oil display work. I'll need to run it in the house to test that, to find the serial data problems using the oscilloscope. The Car cable colors are: 5 Red Power12V 9 Blk PowerGnd 3 Wht Points 8 Blk PointsGnd 7 Red Speedo5V 2 Wht SpeedoSig 6 Brn SpeedoGnd 1 Yel OilPressure The Disp cable colors are: Board Pigtail DE9 Cable Signal Pin color Pin Color name 1 brn 2 gn/wh SCLK+ 2 red 6 wh/gn SCLK- 3 org 3 bn/wh MOSI+ 4 yel 7 wh/bn MOSI- 5 grn 4 bu/wh CS+ 6 blu 8 wh/bu CS- 7 vio 5 or/wh 9V 8 gry 9 wh/or Gnd See NDCACables.xlsx for the correct version. Volvo uses much different sensors than does Chevy. https://www.mnrelectronics.com/250-4165-universal-vss-signal-generator-magnet-kit-with-pick-up-coil/ The speedometer cable isn't the same, so the Auto Meter sensor isn't usable. I can move it to the '58 wagon. The engine temp sensor is a 5/8 nut with sealer on the sender. I need to find something similar or just not do temp, since my temp gauge works fine. I will see about getting the Rostra coil shown above. https://www.rostra.com/universal-aftermarket-cruise-control-by-rostra.php#2504160 $40, no magnets. https://www.ebay.com/itm/144536617938 I am thinking about making a Hall sensor to mount under the car with a couple rare earth magnets strapped to the drive shaft. I have some analog chips. I can try one, with a magnet on a hole saw mounted in my drill, for a test. Done. It moves from center of Vcc up or down to rail at 1/2" spacing. I'll try it again mounted on a steel bracket to see how that affects it. Same behavior. I made a sensor on the old cable using a second chip and another 0603 0.1uF capacitor. It works with the drill, at 1/2" distance from magnets. I'm going to encase it in that adhesive shrink tubing and mount it with a P clamp. I cut the first piece off too close to the chip, so it needed another tube around the chip. Done. I'm going to work on the oil pressure code now. Auto Meter 2242 Oil sender calibration data This was inflating it the first time. I noticed some hysteresis when letting the pressure out slowly. See OilGauge.xlsx It calcaulates the PSI from the gauge ohms and some other parameters. I need to get gauge ohms from the ADC which is 4095 for 3.3V There's a 330 ohm resistor in series with the gauge to 3.3V. ADC = 4096 * R / (R+330) R = 330 psi ohm 0 250 20 200 30 158 40 140 50 120 60 100 70 83 80 69 The spreadsheet calculates the PSI from the ADC counts. This needs testing. int oilOffset = 132; int oilScale = 133; oilPsi = oilOffset - (10 * oilADC / oilScale); TEsting with air tank. Tank nixie 0 1 133,132 80 71 75 68 70 70 146, 119 55 57 It reads a couple PSI high midscale. So what. The display glitches are likely due to insufficient hold time on serial data out of some shift registers realtive to the serial clock, which has different delays on different parts of the board. I added a jumper wire on ser clock from oil to temp chips to make it faster for those. I ordered a remote hose for the oil sender from Ebay. Autometer 3227. 3 ft braided AN4 hose with 1/8 NPT fittings. I am going to get the speedo working first since it's the thing I need to be legal. I will use a 3" long angle bracket under the driveshaft for the Hall sensor. There's a piece of sheet metal below it to make installation easy. The code I'm using is very old! It was in Arduino The current code is in CorvairDash. I made DashCodeVluv.ino for the Volvo. 5/29/22 The tach and speedometer are really sucky, so I am changing from the frequency measuring trick to a more analog approach. I plan to read the tach and speed inputs and treat them as analog inputs. I'll use 0 for LOW and 1000 for HIGH. I'll apply the same smoothing as I have been using. The result should be a trapezoid wave. I'll read it by having a state variable tachWasHigh and speedWasHigh and change the threshold accordingly, like a Schmitt trigger. I updated the tach code to use this methond. It seems to work, but I'm still getting false triggers. I need to make all the smoothing a lot smoother, it should be close to 170 for some effective smoothing. Time to do the same for the speedometer. I also need to add a second smoother for each of the RPMs and the MPH. Done. I made the speedometer and odometer use the new method. I need to make the speeo/odo scale be based on reality, not a number. 5/30/22 I rewrote the code to use the actual units for calculating the conversion facotrs for the odometer and the speedometer. It worked correctly the first try. I discovered that the tach was producing gibberish without the diagnostic print statement after I removed the periodic reading of TachPin, so I added a 100 microsecond delay in the main loop. The speed and odometer were correct, other than a slight adjustment to tire diameter. The tach is on 50 count resolution, which is less distracting than 10 counts. I decreased the odometer EEPROM write period to 0.2 miles from 1 mile. I like it now. It's much more well behaved. I had to replace my Teensy because I had left out D2, which was needed to reduce the ignition signal from 12V to 5V. I installed D2 and also added a 4.7K resistor across it, to act as a voltage divider. We'll see how long the Teensy lasts. It should be fine now.