308 Speedo, Odo, Sender System Fix

Hi Gang,

Sorry this is so long but I hope it has some useful info.


I have a 1978 308 GTS with a Veglia Borletti speedometer. When I first purchased the car, the speedo worked but the odo did not so I set out to find out why. I dismantled the unit by first pulling the whole instrument cluster out (lots of threads on how to do this). I’ve got to hand it to the Italians for the ease of removal here. I did not have to remove the steering wheel, I just maneuvered the entire unit forward and then down and to the right, and it comes right out. Just three big connectors to unplug. The speedometer is easily removed from the cluster with two thumbscrews. This makes me think that they actually knew how often this thing would need to be repaired.

Once the unit is separated from the cluster, you rotate the outer steel band like a large screw and it pops off with a plastic ring and front window. Look for the gap in the ring flange and get this “threaded” on the outside rim of the plastic housing. This step while seeming simple is important so that you don’t scratch anything. Set these pieces aside. There are three screws on the rear that come out but more importantly, you must gently (I mean gently… this is old dry plastic) pry the plastic plug that is the strain relief for the three wires going in to the instrument part of the speedo. This piece comes straight out away from the housing. The three wires are +12V (Green), Ground (Black), Sender (Red) Signal. Once the little plug comes out, put your hand on the front of the unit and gently ease the speedo unit out the front of the housing into your hand while feeding the three wires into the back. Try not to smudge up the front face of the speedo. You should now see the speedometer and odometer in all of its naked glory.

Now I can only speak for this specific unit but I bet they didn’t change it too significantly for a number of years. Here is how it works. There are two sections to the unit and they are functionally separate. The speedo needle is attached to a D’Arsonval meter. This is simply a moving coil over a permanent magnet. When the coil gets a certain current, the meter moves up. The “sender” unit puts out a certain number of +12V pulses based on the rotation of the rear axle. There is about 530 ohms to ground and then the signal is fed into the SAY115Y Integrated circuit. This darned chip is obsolete. My Canadian friend, Gerrit, actually got the U.S. version of the datasheet and sent it my way – Thanks Gerrit! Anyway, the pulses are integrated into a current that is supplied by an on-chip current source. This current directly drives the meter. So, the faster you go, the faster the pulses are generated and the higher the needle goes. The duty cycle or shape of the pulses is not important, just that they are periodic. The SAY has a pulse re-shaper that makes sure the needle is integrating the same shaped pulse. There is a variable resistor on the board which will move the needle up and down for the same pulse rate. This is the calibration potentiometer. There is also a small screw that holds a fixed adjustment for how hard the spring on the needle wants to drive it back down. I will get to this later during calibration.

The odo section consists of a whole bunch of plastic gears to increment miles in tenths. There is the main odo section and the resettable mile counter section both driven by the same initial gears. The mile counter section is reset with a mechanical linkage to the front of the instrument panel. When you turn this little knob, a cable turns inside which turns the plastic gears, resetting the counter to all zeros. Of course, this does not affect the main odometer mileage. This entire little two-part mechanical counter is driven by a worm gear that is driven by a step motor. The step motor is a single-phase unipolar drive type. This means that it needs a single energized pulse to bump it a step. The SAY115Y contains a divide-by-64 circuit. So, while the input pulses from the sender are happily converted to current to drive the speedo needle, they are also being divided down such that two out of every 64 pulses will step the motor. The motor gets stepped twice because it steps each on the high part of the dived clock and the low part. The IC has two darlington transistor pair outputs that are designed to drive motor coils. The designers chose to connect one of these to a little PNP transistor acting as an amplifier to provide the +12V drive to the motor. The other output was connected directly to the motor for the Ground drive. This is the one that shorted to ground (inside the SAY IC) in my speedo.

I used the remaining signal that was still driving the PNP transistor and placed a small NPN transistor to provide the Ground or “low side” drive to the motor. This is shown as the “dotted line circuit” in the schematic. It seems to work fine and does not get warm.

The Speedo and odo are separate but do run through the same IC and use the same sender signal. One can be busted and the other working. If this is the case, your sender is probably fine because neither can work without those pulses. When speedos are reading high, they could simply need to be adjusted for wear or spring tension or the IC is failing. When the odo is not working, the stepper is probably ok (I don’t think these itty bitty parts can take out the coils in the stepper) but your drive circuit has failed. This type of failure is not that unusual for electronics driving inductive loads like a motor.

I reverse engineered the schematic and it is attached. Yes, I spent a couple of hours with a meter and a bright light (to see through the board to the copper). I hope this helps you guys. I wouldn’t work on one of these unless you understand electronics because you can break something pretty easily. Gerrit tells me that he is designing a board replacement (you can look him up on Fchat). If Gerrti does this, the board should only have a few connections and should be pretty easy to install.

Practical Facts:

1) The sender rotates 2 times for each revolution of the axle. This was measured by removing the sender, jacking up the back end, putting a little flag with a toothpick on the spindle sticking out of the axle and rotating the wheel one complete revolution.

2) The sender puts out 4 cycles per rotation, therefore the sender puts out 8 cycles for each rotation of the axle. This was measured by connecting the sender up to +12V, using a 560 ohm resistor from the sender output to ground (pulldown). The resultant signal is a square wave with about 6 V peak to peak. This was checked on an oscilloscope.

3) Doing the math yields 109.8 cycles per second at 60mph for a 24.5 inch tire diameter.

(((5280 ft/mi. * 12 in./ft. * 60 mi/hr.)/3600 s/hr.)/(Pi * 24.5in.)) * 8 cycles/rev.= 109.8 Hz

4) The odometer contains a divide by 64 circuit in the SAY115Y.

5) The odometer requires exactly 200 steps of the step motor to increment 1 full mile.

6) The odometer steps the motor on EACH half cycle. I point this out because up until now, every thing was based on a single cycle causing a single event vs. in this step, a single cycle causes two events.

7) Doing the math for 60mph yields:

((200 steps/mi. / 2 steps/cycle) * 64 divider)/60 seconds/minute = 106.7 Hz

Note: What the odo requires as an input is a better use to calibrate the unit vs. the calculation based on wheel size. This is because the tire diameter is a squishy variable.

8) If oversized wheels are used, the stock odo will not read correctly without some sort of circuit addition i.e. there is no adjustment for it. The speedo, however, can probably be compensated by the variable resistor.


Easy sender checker - Connect a +9V battery plus side to the green wire of the sender. Connect the minus side to the housing of the sender. Connect a 560 ohm resistor from the red wire on the sender to the housing (minus). The resistor is in a pulldown configuration. Connect a voltmeter negative lead oto the sender housing (minus) and the positive lead to the red wire of the sender. Rotate the sender and you should see the voltage change from 0.5-1V (Low) to 5-6V (High). I don’t know if this works on all senders but it works fine on this one.

When you reinstall your sender, make sure you get the small shaft on the sender into the cup receiver shaft sticking up from the axle. If you miss, the little coupler will miss the axle turning and you will think your sender is faulty. If you jack the back end up and get on your back looking up, you can see these two mate before screwing the sender down.

Calibration

I thought I could nail this thing and really tune it in tight but I cannot. The design of this unit is such that there is no way to get it perfect, just close. Since the odometer is pulse driven and essentially fixed all the way back to the wheel, I thought would start there. The calculations show that the odometer needs to have a 106.7 Hz frequency to simulate 60mph. A frequency generator with +12 amplitude was connected to the red input wire (sender input). The green wire into the speedo was connected to +12V and the black wire to ground. The generator was set up to output a 50% duty cycle at first.

The setting of the odo was recorded with the starting time and the unit was left to run for 5 minutes. At this frequency, there was exactly 5 miles shown on the odo. Step one complete. There is no adjustment for this, just verification.

The speedo jumped up to about 55mph. The variable resistor was adjusted to bring it up to 60mph at 106.7 Hz. Doubling the frequency and halving the frequency should have yielded 120mph and 30mph respectively, but it did not. This is where the second adjustment comes in. There is a small screw on the underside of the meter portion of the speedo. When loose, a plastic arm can slide to the left or right by a couple of centimeters. This arm simply winds the little resistance spring on the meter up or down providing more or less tension to resist the magnetic field that is positioning the needle. Since 60mph worked but the others did not, I would call this somewhat of a linearity adjustment. It also serves to pull the needle to peg at zero when the car is stopped, otherwise it floats a “few miles” above zero. Anyway, I messed with this in many positions and could get close but not perfect. I could get one or maybe two of the three readings to match the input frequency but not all three. I decided that I wanted the best accuracy below 85mph and concentrated on making 30mph and 60mph work. Because of this, when I put 213.4 Hz into the meter (double the 106.7 to get 120mph simulation), I got 114mph. This is the best I could do. It is interesting to me that all this talk of high speeds can be off by as much as 5-10 percent or more!

So, my speedo/odo works now and only so-so accurate. I don’t want to worry about it anymore. It’s getting to be time to drive.

-RJay

P.S. If anyone wants the schematic and layout for the board, please PM me because it is too big for the site.

 

Nate,

I can only answer questions for the 78 308 speedo - the limit of my experience here. Anyway, the speedo and odo are separate, in that, the odo is fixed by the pulses from the sender but the speedo can read high or low depending on a single potentiometer on the main board. So, you can read 10% high and have an accurate odo. You would have to pull the speedo, supply +12V, Ground and simulated sender pulses (frequency of 106.7 Hz to represent 60mph). Adjust the pot until the needle is at 60 and that will be fairly accurate since most driving in this range. This simulator can be a garden variety function generator but it must be able to drive 540 ohms on it's output or there won't be enough amplitude for the IC to count the pulses. Since you are not comfortable with the electronics, you should consider taking it to a speedo place.

Note: For all the DIYs out there, this speedo/odo is actually a pretty robust unit that is difficult to harm if you stay within +12 V, unless you start shorting things.