Another lm2917 thread

Thread Starter

propertyco

Joined Nov 8, 2022
14
Working on recreating a tachometer board for a rarish car.

Car is a 4 cylinder and tach receives signal from negative side of the ignition coil.

The EMA for the mag needs 1.75v at 7,000 rpm (117hz)

Using circuit attached for values for capacitor.

6k resistor is actually two 12ks in parallel and the 100k resistor on vout is a trimmer pot

This circuit will receive input from the negative side of the coil, and of course I built the circuit then realized (after testing and frying my lm2917) that I must condition the input to be a cleaner square wave without any huge spikes.

Anyone have suggestions on the best method of approaching this?
 

Attachments

MisterBill2

Joined Jan 23, 2018
18,167
For the circuit between the coil terminal and the LM2917 input, start with a series capacitor, o.o1MFD, followed by a series resistor, probably 4700 ohms, and then a small capacitor, 0.001Mfd to common. FRom the junction of the capacitor and resistor have a diode clamp circuit, two diodes in series, cathodes toward the V+, bottom anode to common, and the junction of the two to both the RC connection and the LM2917 input. The series R and C will keep the DC out and limit the power to the clamping diodes, which are there to hold the input voltage withing the IC limits specified. The circuit is possibly a bit of overkill but the total parts cost is less than a new LM2917.
Hopefully my description is adequate, if not others can produce the circuit drawing.
 

Ian0

Joined Aug 7, 2020
9,667
Have you tried connecting the input to a piece of insulated wire and wrapping that wire around the HT lead between distributor and coil? (No physical connection to the points)
 

MisterBill2

Joined Jan 23, 2018
18,167
I suggest not doing that because it is not at all a known coupling arrangement. And even just 100 volts will destroy another IC. Use the input protection circuit I described and avoid causing damage.
 

Thread Starter

propertyco

Joined Nov 8, 2022
14
I suggest not doing that because it is not at all a known coupling arrangement. And even just 100 volts will destroy another IC. Use the input protection circuit I described and avoid causing damage.
Think that works perfect. Both 600v and 12v provide less than 12v wave.

Do you have a suggestion on what specification diodes I should use?

Appreciate the help!
 

Attachments

Ian0

Joined Aug 7, 2020
9,667
I suggest not doing that because it is not at all a known coupling arrangement. And even just 100 volts will destroy another IC. Use the input protection circuit I described and avoid causing damage.
I have used both this and a diode clamp circuit similar to yours, and they both worked.
The pickup wire can be used with the 8-pin LM2917/LM2907 and all that is needed to protect the input is a back-to-back diode clamp.
Another possibility is the one favoured by Smiths - a current transformer in the supply wire to the coil.
Both that and the pickup wire can be used equally well with negative-earth and positive-earth vehicles.
The photo shows one half of the current transformer. There is another u-shaped piece of steel that fits external to the case held in place by the brass screw, and one and a half turns of wire goes around it, so there is no break in the supply wire to the coil. This also introduces a gap into the transformer which would be necessary to prevent it from saturating due to the net DC current in the circuit.450D33C2-F381-41FC-AB28-C63EB724EB43.jpeg
 

MisterBill2

Joined Jan 23, 2018
18,167
for the circuit that I suggested a 1N914 diode works well, and is fast enough to clamp any reasonable pulse getting thru that low-pass filter. Any signal diode with at least a 20 mA current capability and at least a20 volt reverse voltage withstanding capability should be satisfactory. Tis circuit was used to protect CMOS devices in assorted testers used for automotive engine electronic fuel controls. It protected from all sorts of transients.
 
Last edited:

Thread Starter

propertyco

Joined Nov 8, 2022
14
I'm not sure what i'm doing wrong, still no luck with this circuit.
When bench testing it, the needle will move initially when receiving power, and when start cranking the car the needle pegs then the whole unit dies.

Lost my spare lm2917s, maybe i'll reroute the tach- to ground as my mind tells me i should, and rebuild the whole circuit this week.

No oscilloscope so i'm going in rather blind.
 

sghioto

Joined Dec 31, 2017
5,378
I'm not sure what i'm doing wrong, still no luck with this circuit.
The LM2917 has an internal zener diode that requires a 470 ohm resistor to Vcc.
Most of your schematic and values do not resemble the circuit from the data sheet.
Try this arrangement and component values.
1670037915325.png
 

sghioto

Joined Dec 31, 2017
5,378
Running that in LTSpice i get an odd output, at 117hz its starting at 1.1v and dropping down.
Not sure why that is. I have a 2917 I can breadboard to verify. Basically you are converting frequency to voltage, correct?
At 117 hz what voltage are you expecting?
 

Thread Starter

propertyco

Joined Nov 8, 2022
14
Not sure why that is. I have a 2917 I can breadboard to verify. Basically you are converting frequency to voltage, correct?
At 117 hz what voltage are you expecting?
Frequency to voltage is my goal

At 117hz I am expecting 1.75v

I am drawing signal from the negative side of the ignition coil.
 

MisterBill2

Joined Jan 23, 2018
18,167
In that circuit posted I would recommend that thr 10K resistor marked "R7" be between the capacitor and the diodes. That will limit the diodes current. Consiider that the ignition coil effective source impedance is close to zero, the high voltage spike can deliver a fair amount of power.
 

Dodgydave

Joined Jun 22, 2012
11,284
An easy way to calibrate it is to use a 50Hz signal (mains frequency) and set it to 3000rpm, i used to use a 6V AC transformer and inject the signal via resistors..
 

sghioto

Joined Dec 31, 2017
5,378
Frequency to voltage is my goal
At 117hz I am expecting 1.75v
I don't have the 14 pin LM2917 but the 8 pin version so I can't duplicate the circuit as posted.
This version from the spec sheet was tunable to your requirement using a signal generator for the input.
1670106909018.png
 

Thread Starter

propertyco

Joined Nov 8, 2022
14
I don't have the 14 pin LM2917 but the 8 pin version so I can't duplicate the circuit as posted.
This version from the spec sheet was tunable to your requirement using a signal generator for the input.
View attachment 282094

I'm not sure if I can use 8 pin version with the provided clamping circuit as the 8 pin version requires a dip below 0v.

The r1, c1 and c2 values there don't work in splice but I'll verify in practice this week.

Using the clamping circuit provided doesnt seem to work either, when starting up the car it'll pull the entire 12v circuit down partially like a high resistance short to ground and the car won't idle.

Such a simple circuit, not sure why I am struggling so much
 

Attachments

Ian0

Joined Aug 7, 2020
9,667
This is the clamp circuit I used:
Screenshot at 2022-12-10 15-07-15.pngL1:L2 is the coil
C1 is the "condensor" across the points
D1 was 1N4007 or UF4007
D2 was a schottky
R1 was quite small, 47Ω or so, and just serves to limit the current through the diodes if the points waveform rings below 0V.
R2 was a pull-up, probably 4.7k. It might have gone to the Zener-regulated Vcc of the LM2917.
Quite a while since I made one!

If you've got the 8-pin LM2917 then it's worth trying a current transformer circuit, or my capacitive pickup off the HT cable. Current transformers are a lot easier to get than they used to be, but you do need one that will withstand DC.
Do you still have all the parts to the current transformer that was on the original tacho? If so, use that.
 

sghioto

Joined Dec 31, 2017
5,378
This is the circuit from the spec sheet that matches your setup. Additional signal conditioning as Ian0 suggested maybe required. The component values should be right on for 1.75 volts at 117hz.
1670691152681.png
 
Top