Modifying automotive tachometer for fewer cylinders...

[65ShelbyClone]

Out of cheapness and spite I have decided to investigate this. I would like to modify a car tachometer that only has options for 4, 6, and 8 cylinders so that it can be used on engines with fewer than 4. Switching between 1, 2, and 4 cylinders would be ideal.

It has a 3-position selector switch with different resistors on each pin, so I assume their resistance determines the signal pulse division. It has an IC, but the ID is obscured by the needle mechanism. Anyone know how I could modify it the way I need?

 

[Theatronics]

Need to know more.
What are the values of the resistors you mentioned.
What kind of display does it have, sounds like an old moving needle?

Can you measure the signal going into the needle display?
(is it a DC voltage that varies or maybe PWM signal?)

Normally the way a tac works is it measures the number of firings that occur on the main coil.  The distributor cap then diverts each spark to a different sparkplug. So if there are 8 sparkplugs, divide the number of sparks by 8 and you have 1 rotation of the engine.

For a pure analogue method, I suppose you could take the number of pulses coming in, amplify them by a fixed amount (controlled by a resistor on a n op-amp) and then feed the result into a capacitor.  Much like a switching power supply. the stronger and for frequent the pulses are, the higher the output voltage from the cap will be.

Feed that signal to a calibrated voltmeter and presto you have a tac.

I am guessing that the resistors are the feedback resistors on a OP-AMP.  You might try installing a variable resistor in place of one of them and then hook the unit up to a 4 cyl car.  Using some other device measure the rpms of the engine and then tune the variable resistor down to get the correct reading. 
Disconnect the resistor and measure it to see what the resistance is.

Then match that as closely as you can with standard values.  If you want, you can just leave the VR in but use a 15 turn PCB unit and glue the screw when you get it right. Otherwise it may turn and come out of calibration.

I could be totally wrong because we don't know enough about the unit you are using yet.

-Mike

 

[65ShelbyClone]

Need to know more.
1.) What are the values of the resistors you mentioned.
2.) What kind of display does it have, sounds like an old moving needle?

3.) Can you measure the signal going into the needle display?
(is it a DC voltage that varies or maybe PWM signal?)

Normally the way a tac works is it measures the number of firings that occur on the main coil. The distributor cap then diverts each spark to a different sparkplug. So if there are 8 sparkplugs, divide the number of sparks by 8 and you have 1 rotation of the engine.

1.) Ah, jeez, my notes are at home and I'm at work. One was ~24k I think for 8cyl, I forget the second, and the third was ~350-360k for 4 cylinders.

2.)Tach is analog. It's a Sun Super Tach II 0-8000rpm.

3.) You mean like with an ocilloscope? Not unless I have a signal generator that will allow me to run the tach on a bench at work......where the oscilloscopes are.

I figure the switch and resistors is just for some kind of divider. If it will help, I'll try yo get some detailed pictures of the PCB and components. I'm thinking it might just be easier to make a signal modifier with an optical switch and trigger wheel(to mimic a 4/6/8 cyl engine).

 

[ZomBiE]

Normally the way a tac works is it measures the number of firings that occur on the main coil. The distributor cap then diverts each spark to a different sparkplug. So if there are 8 sparkplugs, divide the number of sparks by 8 and you have 1 rotation of the engine.

What? 2-stroke engine? In four-sroke, it needs 2 revs for cranckshaft for one firing spark in single cylinder. CAMSHAFT does only one rev while cranckshaft does two. in V-8 you must divide strokes by 4 if you want actual engine revs.

 

[Theatronics]

Hey , I'm an electronics geek, not a motor head.  Sheshh.

Yes I forgot that in ICE's that there needs to be an exhaust cycle for each piston.
That would equate to 2 revs for 8 firings in a 8 cylinder engine.

That just means we need to divide (or multiply) by 2.  Eh. Not so hard.

What is the pickup or signal input method on the tachometer?  Inductive pickup, Direct wiring?

Is there a sensor on the crankshaft?

-Mike

 

[65ShelbyClone]

This tachometer is built so that the signal connects to the negative side of the ignition coil. The points or ignition signal on a car engine come the cam(50% of crankshaft speed).....My single cylinder is triggered off the crankshaft itself and thus will fire twice per 720* of rotation. A 4cyl will fire fire four times, so my tach will read 50% of actual rpm if I don't change anything.

Ignore the little black "4", "6", and "8" written on the board next to the resistors. I'm not sure if they're right.

 

[Theatronics]

Vi Wh Bl Rd - Gr   
7  9  6  2    5% 796 * 100 =79600

Rd Yl Or    - Gld
2  4  3      10% 24 * 1000 =24000

Vi Bl Vi Br - Gr
7  6  7  1    5% 767 * 10                 =7670

Vi Wh Bk Rd - Gr
7  9  0  2    5% 790 * 100 =79000



Those switches can be tricky on the inside.  Some have been set up to short multiple contacts and create unexpected outputs.  Others are just straight forward 3 position switches.

In any case it appears that the resistance is used to control the measurement.

That trim resistor on the side looks very promising.

The shadow and the parts placement make it hard to follow the traces. 

It appears like there is a chip under the gauge and the gauge has 2 large pins that straddle the chip. 

I would try and work out the circuit diagram of at least the switch the chip and the gauge.
The other parts may be just power regulation.

Once we understand how the switch works to provide resistance to the chip I think we will find that the number of calendars Vs. Revolutions per a second is a simple formula and by replacing some resistors we can do 2, 4, 16 or 24 calendars just by putting the correct resistor on the board.

-Mike

 

[ante1]

The IC seems to have 16 pins which exclude the in this kind of devices often used LM2917 which have only 14 pins if I remember correctly.  :-\

 

[Theatronics]

I would guess that is is a rare custom chip.  It could be a common part, but my experance has been that devices like that often are built around a custom part or one where they scratched the part number off the top.

One way to be sure would be to use some solder wick and just desolder the gauge
and take a look at the chip.  If there is a part number we might just be able to find a Data sheet.

I am guessing it is something like this:


http://www.national.com/pf/LM/LM2917.html

But it could easily be a small microcontroller that has an onboard signal counter and a PWM output driving the gauge.

-Mike

 

[ante1]

Yes Mike!

That’s the one I mentioned above but also discarded because of the 16 pins (at least I think there are 16 on the pictures) since that’s too many for the LM2917!!

 

[65ShelbyClone]

Yeah, the IC has 16 pins.

However, you might be surprised to see that it still has the part number. Yes, I desoldered the gage part.

It's an ON Semiconductor CS8190ENF16 presision Air Core tachometer/speedometer driver!

ON_Semi_air_core_tach_driver_IC.zip

 

[ante1]

Great, it’s hacked! You got the formula right there!

 

[65ShelbyClone]

Now I just have to sort out what all these other extra resistors and caps are. 8)

 

[ante1]

Please let us know! 

 

[65ShelbyClone]

Alright, it appears to be just a variation of the example circuit in the data sheet. The cylinder count selector switch just changes the resistance between the negative charge pump input and the frequency-to-voltage output pins (15 & 16). On the 8cyl setting, the resistance is 196k, on 6cyl it is 26k, and 4cyl is 195k....the same as the three blue five-band resistors.

There is a 68k difference between 4 & 6 (33%) and a 62k diff between 6 & 8 (25%). Based on my calculations, I need 449k for a 2 cylinder setting and 573k for a single cylinder. Sometimes I get mixed up in the math, so maybe way out in left field.

EDIT: my calculations were off as I expected, but I got it right after some trial-and-error. It took a 1M resistor and a 470k in series. I tested it on a truck that I know idles at 750rpm. I figured that on the 1cyl setting, 750 x 4 = 3000rpm on the tach. Thats what I got with 1.47M ohms.

 

[ante1]

I think you have a voltage divider on the board, since that would make sense with the values you mentioned. Never mind that though, it works now I suppose so why worry!
Is this taco going to be fitted on one of yours Briggs&Stratton twin turbo nitro engines? ;D

 

[65ShelbyClone]

Is this taco going to be fitted on one of yours Briggs&Stratton twin turbo nitro engines? ;D

I can't hide anything from you guys!  Now that I know how to modify the tach, I might get a couple more and use one on my Honda XR600. This one I took apart was supposed to go into my car, so I need at least one more tach, but at $29.99 each, who can complain?

Actually the engine will be single turbo, MegaSquirt injected, and probably running methanol.  8).

 

[ante1]

Yeah, that’s about what we would expect from you! ;D

 

[adalex]

Not to hijack but I'm having the same issue. I have a 99 camy 4cyl and swapped a instrument cluster from a 98 es300. Everything works except for the tach and I bought a dakota digital sgi-100bt. The problem I'm having is the higher the rpms get, the less accurate the tach gets. I took apart the tachs from both the 4cyl and v6 to see if I could try swapping resistors over but I'm having trouble understanding what each resistor does etc. I have close up pictures but I might get 2 unmarred ones from the junkyard to start over. Is this something that someone could possibly help me with? I'd really appreciate it.