Toyota tachometer

[hpmaxim]

I'm trying to design a circuit to drive a tachometer in a mid-80s Toyota.  Unfortunately, it doesn't work.  The circuit is just an NPN with the emitter tied to ground and the collector pulled up to 12V through a 10k resistor.  The circuit is functioning as expected, but it doesn't drive the tach.  Looking at the actual tach output, the main difference I'm seeing is it has a big inductive kick when transitioning from ground to 12V (obviously this is just the low side of the primary of the coil).  So the questions are:

1) Is that the problem (lack of inductive kick?)
2) If so, presumably the solution is to put an inductor in series with the resistor.  Anyone have a good set of values for the L and R?  The BJT is being driven from an 74HC chip through a 10k resistor in the base, I presume it shuts off fairly quickly (<1 us).  Albeit I'm not sure how long it takes for it to come out of saturation.

[tonyarkles]

Any idea how much current usually goes through the tach? If it's just a magnetic coil, the 10k resistor might not be enough to drive it.

[Junkers]

Can you post the waveforms you are observing?

[Rerouter]

use a push pull driver rather than just a NPN and pullup, fit a 22uF cap in series, with a 10K to ground pulldown on the tacho side of the signal, most of them have to cross below 0V to register a pulse,

[hpmaxim]

tonyarkles, I have no idea, however the output waveform is exactly what I expected.  It's a 12V square wave, which is normally 12V, but goes low for a few ms on each ignition pulse.  There were actually two 10k resistors, one from 12V to collector, and one between the collector and the tach.  I was looking at the voltage at the tach, and it was still what I expected, so I don't think there's a current issue.

Junkers, I don't have any screen shots of waveforms, but like I said.  It looks like 12V, then goes to nearly ground, for a few ms, then comes back up, and then gives a very damped very high amplitude ringing as the pulse comes back to 12V. 

Rerouter, Is it looking for it to go below ground or above battery voltage, the high voltage looked much more distinctive than the below ground voltage, but maybe you're right.  Could you draw a schematic, I'm really not following what you are describing about a push pull. Are you thinking of something like this (driven by a 5 or 12V CMOS buffer):



Or maybe something more like:



In either case, I'd think R=10k is probably good, but you'd probably want a time constant measured in microseconds, maybe 10us, which would put it at about 1nF, not 22uF.  22uF would could give you a time constant of 220ms, which means the cap wouldn't be fully charged before its discharged, nor would it fully discharge if it were allowed to charge.
         

[Rerouter]

There are a number of different signal types, and have been knee deep into each and every one of them....

i am making the assumption that it is either meant to be a coil pickup, or an inductive pickup for the age, as such your second diagram would be correct, that will AC couple the signal, as its acting as DC blocking, rather than a filter, the resistor pulls the caps bias to ground, so any injected signals will be +- 6V if your original singal was between 0 and 12V, but to do this you need both a push (pnp) and a pull (npn), as a pullup resistor will make it a low pass filter,

You could use something like a high side driver, or an op amp of the equivilent voltage, it just needs to both push and pull current through the cap,

[HighVoltage]

I am pretty sure that this older TACH system was driven by the 400 V kickback of the ignition coil.
Unless you have an ignition module with a TACH output, but those came only in the mid 90s, if I remember correctly.

[Rerouter]

Take a photo of the pcb and tacho, i work with these things every day, there where some weirder ones, and i can probably tell you exactly how its meant to be driven,

[hpmaxim]

I don't have the tachometer out of the car and don't really feel like pulling it.  I will say that the signal that I'm seeing going into the tachometer looks like the low side of the primary of the ignition coil.  Like I said, normally 12V -- during an ignition event its pulled down to nearly ground and then right as its released I get a giant inductive kick and damped oscillation.

My circuit (which doesn't work) looks exactly like your "Electronic_Ignition_Signal.jpg" minus the diode which I don't have (and don't really see the point of).  If anything you'd put it from the collector to the base to get it out of saturation faster.

I was generally assuming I'd want to convert to a circuit like like "Inductive_Pickup" although your R and L values seem unreasonable (L is way too high, and R is way too low).  Presumably that's what I want if I am looking for a high going pulse, and presumably I'd want the CR circuit if I want a low going pulse (although that'd only be a 5V square wave with a spike to -5Vish on the leading edge of the pulse (rather than the trailing edge).  Not sure what its looking for...

HighVoltage, the signal is actually coming from the ignition module, but like I said, it looks like the low side of an ignition coil, its possible its just internally tied together inside the box.  I have a spare, I could open it up and look.  My concern is that I intend to use this on multiple vehicles (probably all Toyota from mid-80s to mid-90s) so I'd like it to be universally compatible.

[G0HZU]

My concern is that I intend to use this on multiple vehicles (probably all Toyota from mid-80s to mid-90s) so I'd like it to be universally compatible.

I'm a bit confused here... are you trying to do a repair or come up with an alternative tach design? It's years since I looked at Toyota TCCS stuff but I thought the rpm meter had a dedicated (pulsed?) drive to some form of integrator circuit that drove the meter dial according to rpm?

[hpmaxim]

I'm trying to design a circuit that will drive the tachometer in the car (or in a variety of cars) with a 5V input square wave (where each pulse represents an ignition event -- let's assume we're dealing with a 4 cylinder car).  This isn't a complicated or trick question, and I'm not trying to deliberately obfuscate details.  As I said, I assume I want either a circuit that look similar to Rerouter's Coil_Pickup or an inverter/buffer driving my RC.

[G0HZU]

Normally I'd expect to see a train of pulses of fixed duration driving an RC integrator that then has a buffer to drive the meter coil for the needle indicator.

So you get a basic frequency to voltage conversion. i.e. the RC integrator will integrate to 'twice' the dc voltage if the train of pulses doubles in frequency. For this to work well you have to have an rpm signal that has a constant pulse width across all engine speeds.

If your driver circuit in the car doesn't do this then you can use your driver pulse to trigger a monostable multivibrator (NE555?) to give you a constant pulse width for every trigger event and use this to drive the integrator.

Or you could use a modern PIC/AVR MCU to do most of the work.

[hpmaxim]

Normally I'd expect to see a train of pulses of fixed duration driving an RC integrator that then has a buffer to drive the meter coil for the needle indicator.

So you get a basic frequency to voltage conversion. i.e. the RC integrator will integrate to 'twice' the dc voltage if the train of pulses doubles in frequency. For this to work well you have to have an rpm signal that has a constant pulse width across all engine speeds.

If your driver circuit in the car doesn't do this then you can use your driver pulse to trigger a monostable multivibrator (NE555?) to give you a constant pulse width for every trigger event and use this to drive the integrator.

Or you could use a modern PIC/AVR MCU to do most of the work.

I'll reiterate.... I am putting a square wave into a transistor whose emitter is grounded, and whose collector is pulled up to 12V through a 10k resistor.  I also have a resistor between the collector and the tach.  The waveform at the tach is exactly what one would expect from this circuit and the tach does not recognize it.  When I look at the tach signal coming from the ignition module, it looks like an identical waveform (both magnitude and timing) except the ignition module's waveform has a large damped inductive kick at the end of the low-going pulse (and that is working). 

So I have to assume I either have to get the waveform above 12V/battery or below ground for it to be detected.  The question is which one.  It might be that I should just throw an inductor in series with the pull-up resistor and be done with it, or as was previously mentioned I could put a resistor and capacitor in series with the output of the 5V signal, which should give me brief bounce below ground on the leading edge of the low-going pulses.

Did I totally misunderstand your post or were you trying to imply that I was not putting in a proper input waveform into my circuit?

[G0HZU]

Did I totally misunderstand your post or were you trying to imply that I was not putting in a proper input waveform into my circuit?

No, I wasn't implying that. I think it's a case of me not understanding you...

I thought you wanted to replace the whole tach drive circuit with your own (maybe because you are running some aftermarket hardware to replace TCCS?)

But now I think you want to emulate the OEM signal. Also, your references to a square wave are confusing me because I would expect to see a pulsed waveform. i.e. rectangular rather than square? Maybe it would help reduce any confusion for any viewers if you drew the relevant waveforms in terms of voltage against time?

Then it would be easier to suggest how to emulate this? However, you may get the best info if you go on a Toyota forum that specialises in relevant models from your era?

[hpmaxim]

Did I totally misunderstand your post or were you trying to imply that I was not putting in a proper input waveform into my circuit?

No, I wasn't implying that. I think it's a case of me not understanding you...

I thought you wanted to replace the whole tach drive circuit with your own (maybe because you are running some aftermarket hardware to replace TCCS?)

But now I think you want to emulate the OEM signal. Also, your references to a square wave are confusing me because I would expect to see a pulsed waveform. i.e. rectangular rather than square? Maybe it would help reduce any confusion for any viewers if you drew the relevant waveforms in terms of voltage against time?

Then it would be easier to suggest how to emulate this? However, you may get the best info if you go on a Toyota forum that specialises in relevant models from your era?

I do want to replace the whole tach drive circuit with my own.  Like I said, it looks like the Toyota tach drive circuit is simply hooking the tachometer up to the low side primary of the coil.  By "square" I simply mean that that it has sharp/step edges, I was not implying the duty cycle is 50%, its not.  The pulsewidths are typically about 3-4 ms, and the period is obviously dependent on engine speed. 

Since the 0-12V "rectangular wave" seems to be unrecognized by the tach circuit, my question is, for a tach that is designed to be hooked up to the low side primary of a coil, what is the best drive circuit, and/or what features is it looking for (going below ground, going above power, etc?)  Does it matter if the spike is on the leading edge or trailing edge?  Does it even matter if there is a rectangular wave present?  Etc..

[kxenos]

I would try a small signal N ch. MOSFET driving the gate of a large (like a TO220) P ch. MOSFET that is used as a switch to drive the tach with 12V pulses and as much current the tach needs, or a push-pull as mentioned. Good luck

[hpmaxim]

It's normally 12V and pulses to ground, not 0 and pulses to 12V....  Also, like I said, there is a resistor between the collector and the tach, and there is no apparent voltage drop across the resistor, so the tach is clearly not drawing much current.  I guess I will just have to try both the inductive and capacitive circuits and see what works.

[hpmaxim]

Okay, I'm dumbfounded...  I tried two different circuits:



I'm feeding a 0-5V rectangular wave signal with a pulse width of between 3 and 5ms into "In".  The output of the RC circuit (Out1) spikes from 0 to 5V and then RC decays on the rising edge and spikes from 0 to -5V and RC decays back to 0.  It has no affect on the tach.

The transistor based circuit (Out2) is pretty much what I expected.  Normally high.  When the transistor turns on the output goes to 0V, when the transistor switches off, there's a nice damped ringing...  It's actually a beautiful waveform, with peaks of at least -7 and >32V, however the ringing is damped out within about 5us.  When loaded by the actual tachometer circuit, the ringing is pretty much gone, it just goes from 0-12V.

The actual tach signal (which I think is directly connected to the low side of the primary of the coil) seems to have a similar (although not identical waveform), but the ringing lasts 2-3 ms.  There is also a single spike that lasts about 2ms immediately prior to the damped ringing.  Loading from the tach circuit, unsurprisingly doesn't seem to affect it.

Other than beefing the thing up drastically (L going to 10mH, R going to 10 ohm, and the 2n3904 being replaced by a Darlington) which I'm not keen on doing (heat, wasted energy, component cost, and size), do I have any options to make this work?  I figure you'd need to be able dissipate 10W at least.