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 This darned chip is obsolete.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.

 

Fantastic post! This stuff has been talked about a number of times but I don't think I've seen anyone actually show how it all works, etc. Thanks!

 

Beyond Excellent !!!
Thanks for taking the time to do it in such great detail. (with great photos too)

 

Hi Bob
Excellent description. I hope to have a final version replacement circuit built by end of May.

Glad I coould help.

Gerrit
http://dino308gt4.com

 

lots of threads on how to do this
Could you point me to the thread on the Cluster removal. I tried a couple but
this article came up in both but no how to remove the instrument cluster.

Thanks in advance

stephen

Eventually I want to swap out my 85MPH for a REAL speedo. although I never go more than 85

 

I found a thread that covers instrumnet removal under speedo removal,
so I don't need the link.
Thanks

stephen

 

"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."

I have a question. My speedometer looks like it reads approximately 10% or more over what I am actually traveling. Does this mean that my odometer will be 10% or more over the distance that I really traveled? If so, who would I go to to get that adjusted like you did with yours. I am not THAT good with electronics to try it myself. Maybe I should adjust it 10% under what it was and take a long drive to catch up! hehehe

Thanks,
Nate

 

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.

 

Hi Bob,

Reverse polarity will fry the early speedo circuits promptly! I noticed on your schematic that there is better input voltage protection than on my 1976 version of the circuit. But I still wouldn't rely on the circuit to protect the chip from that reverse. For those who haven't guessed, the IC (SAY115X) is NLA.

Gerrit

 

Great info in this thread . . . thanks.

 

Great read. Thanks for all your documentation and hard work Bob!

 

Just plain AWESOME!!!!!

 

Good morning,
This chart is very interessant . So ,like this guy know his subject I explain mine.
I've got a 348 and the speedometre doesn't work.I've got the power on both green/white wires connectors (speedo and sensor)
I think the problem comes directly to the speedo.
What can I do to repair it.
Thank's lot.

 

Oh heck yes!!!

Well done RJ and Gerrit

Dave

 

Wow. This is one of those jaw drop threads, the kind where you see the level of knowledge someone has to understand this stuff and know how to repair it. Ive never been good at electricity, so once it goes much beyond simple circuits and switches my mind goeas blank. Though I did delve into the early Bosch fuel injection and had a pretty good idea how to make it function.

But I have a question to ask some of you guys, both Ray and Gerrit as well as anyone else reading. These parts are aging and becoming increasingly delicate as well as harder to repair. Some of the parts, the SAY115Y IC chip IIRC, for one, are becoming entirely unavailable. At some point these could become totally unrepairable, or only at very great cost.

There is a guy here in town who owns a 69 Dodge Charger 440 R/T that has the rally gauge package. These parts have in the last few years become gold, with prices exceeding thousands for good quality originals. When I was looking at his car I noticed all the gauges were Stewart Warner and mounted into a flat panel, not at all like the stock recessed originals. This was when he said he put them in to keep the originals on a shelf at home. Apparently a few of these cars have been stolen or had parts stolen off of them, and in his case he was just trying to keep the car a driver without risking someone taking his cluster or having one of the gauges fail. They will last pretty much indefinetly sitting on a shelf.

So here is the question. Is there a mechanical sender that could replace the electric sender on the 308, so we could route a cable forward and drive a standard speedometer head? I want to be able to drive the car, so its show worthiness is is not at all important, but this stuff needs to work. I would have no problem with a set of Stewart Warner or Smith gauges in the cluster as long as it worked well and was reliable, while keeping the original stuff under safe keeping. Any ideas? Or, is there perhaps an elecrical model speedo and sender available aftermarket that could be used?

 

Paul you could try calling Mo-Ma in NM. They repair all kinds of these units on daily basis. They've done regular mechanical ones for me and just did a tach for Doug328. I'd bet they could answer your question about parts or fixes.

 

Both cars speedos work (or were), so I dont have to fix anything (yet). I was just commenting that perhaps it wouldnt be so dumb to run something readily available that works good (better?) and leave the Vague Liah stuff on a shelf with the Air injection/charcoal evaporation, air pump, A/C parts, DOT 5mph bumpers, etc.. While someone might really go postal seeing an aftermarket ignition or a Tubi, I doubt anyone will freak out too bad if they dont see the stock gauges. A nice set of Smiths might actually look more "sporty"?

 

I don't understand half of what he said, but I also am very glad he said it!!!!

Very nice work!

 

We had a thread where a member rebuilt his dash for improved visibility, a 328, IIRC.

 

Sorry, I meant Mo-Ma may be able to tell you how available or easy to fix the circuit boards are today. I figure if parts are available or easy to fix, hey no big deal.

However, a nice set of Smiths ivory faced gauges would look cool in that black dash!

 

I guess I should have priced Smith gauges before I opened my mouth. Nevermind that idea.

So another question. Is there anything on these speedos to tune up or lubricate, something to extend thier life as long as its out and in our hand? Or is it just a wait until it quits scenario? And the SAY115Y IC, is there something else that will work, or are we screwed if it pops?

 

Hi
I have looked high and low for an alternative to the SAY115Y to no avail. I bought 10 last year and have been fixing speedos locally with that supply. The scariest part is removing and re-installing the bezel, the electronics are simple. The chips are a b***h to solder in because the tin plating has disappeared due to age.

There are similar chips but not pin compatible, e.g. UAF2115 (also out of production but seems plentiful).

But you have choices when originality is not an issue. The threads on the transmission end are for a standard (metric) speedometer cable. A Fiat speedometer head would work in that case. And as a bonus I believe those are the same diameter and angle as the Ferrari 308 ones. In fact the 1976 308GT4 parts manual shows a right angle attachment with cable, valid to SN 8198. It is a 1:1 ratio so going straight up instead of 90 degree would work as well. The angle drive is a mere $425 so I would try without it first
If you prefer aftermarket, then there are speedo senders (not compatible electrically with Ferrari heads) that will fit and you can match an appropriate head unit. These senders typically put out 8-32 pulses per rev, the standard Ferrari unit is 4 per rev.

And the third option is to replace the circuit in the speedo head with something 'modern'. I started up this path using the LM2917 but never added an odometer pulse generator. I might finish that project this winter.

As to tuning or lubricating, there is really nothing you can do to extend the life. The one act that is extermely high risk is boosting or reverse polarity on the battery. And alternator diode failures. If you need to trickle charge your batery all the time, get your alternator looked at. This is how my speedo went out the second time.

Gerrit
http://dino308gt4.com

 

Paul . . I looked at a lot of gauges . . didn't see "Smith" . . but went with some Autometer gauges .. my buddies that aren't purists think they look better . .I've only put the tach in so far but I have a speedo to mount and plan on changing all of 'em at some point . .. I started a thread in the 308 section about it with a pic or two . . let me know if you want any more info.

And guys I'll take a pic of the tools I made for peeling that bezel back . . it's not quite as user friendly as I hoped . . someone else mentioned they got the speedo and tach apart without having to deform the bezel 'cause the body is plastic on these 2 gauges . .. you might want to try that . . personally I don't see how that was possible but I can be impatient.

cheers

 

Smiths gauges: http://www.holden.co.uk/displayProductsByBrand.asp?brand=Smiths&brandCode=Z3

 

If I'm reading the stock 308 schematic correctly, the sender is fed by 12V ignition power, and then the ground side is connected to the speedo circuit.


On this new Autometer electronic speedo, they show if reading off a 2 wire sender the sensor is wired completely different . . . the speedo signal is hooked to one side of the sender (probably providing +12V) and the other wire is hooked to ground.

I can mount something on one of the drive axles . . . is there anything I can do to the stock sender to make it compatible? . . . called Autometer and they don't think the speedo will work with the stock sending unit if it is how I described.

thanks,

Sean