Replacement tach info request

I've read the other "Electronic Tach( kzrider.com/component/option,com_joomlab...w/id,136171/catid,4/ )" that's active, this is not what I'm looking for.

I'm looking to replace my tachometer in my 1982 GPz 550 (KZ550-H?, I forget). I've removed the gauge cluster, separated the tach, looked for obvious reasons the tach doesn't work (broken wires/connections, obviously fried components, etc) and I can't find any, I suspect the gauge is toast. Also while poking around, I discovered the tach signal wire's voltage varies from ~8.3 volts at idle and increases as the engine RPM increases. I ran it up just below the level where I expect the neighbors to shoot me in the RPM range and got 11 volts. So it looks like the tach is just a voltmeter with different graduations.
Can anyone tell me where I might find the specs as to what voltage corresponds to what RPM? I'm thinking a LED analog bar-meter would be a functional replacement, in addition, I could also have another voltmeter reading volts at the same time, since my alternator doesn't like to charge below 4,000 RPMs.
I realize I could spend almost as much time searching for a stock replacement as I would building an LED voltmeter. Although I'm reasonably sure I can build one that performs both functions simultaneously for under $20 and I would expect the LED tach to last longer than the bike.

My other option is to get (borrow) an external tach and compare the voltage to the RPM.
Any assistance would be appreciated.

Aside from transistors in darlingon pairs, I'm considering using two LM3914(V?) ICs to drive the LED bar.

Check the sender unit first.

do you know what the sending unit is supposed to send? If it's not voltage, is it supposed to be PWM?

Anybody got specs?

There is no sending unit. The signal is from the coil directly.

I'll assume you know the basics of what a coil signal looks like. The signal is approximately a 30% duty cycle (obviously, here the duty cycle is when the voltage is 0v because this is the ground side of the coil). The peak spike is up around 500 volts. The average voltage during discharge is around 20 to 30v, then the rest is at battery voltage.

There is no linear correlation between voltage and RPM. Yes, the voltage varies with RPM, but the reason is because of the variable dwell and because of the variable discharge times of the coil. So a strict voltmeter would be a very poor tachometer.

You need a signal-normalizer. That's a couple resistors and a 12v zener. Then a frequency-to-voltage converter. That's an LM2917. Then send that to the LM3914. All of the circuitry needs it's own regulated supply coming from a 7808 regulator. But there's a lot more to it to make it work reliably.

If you're good you could throw in a switch which will convert the 3914 from tach to battery-voltage.

Post edited by: loudhvx, at: 2007/05/09 10:43

This is the one I've seen. You'll have to do some work, though, it seems his circuit has a few flaws and it's easy to blow up. You'll need to do some good isolation to protect the chips.

www.niksula.cs.hut.fi/~mdobruck/siililan...ien/tacho/tacho.html

If you were to use a cheap automotive tach, like the sunpro supertachII, you could use the diagram in that other tach thread to make it work. Then you could make a single-LED volt detector for your battery voltage. It's pretty simple.

www.geocities.com/loudgpz/GPZ1LEDvoltMonitor.html

I'm working on a 3-LED version which will show yellow, green, and red, depending on the voltage. But that probably won't be done till later this summer.

The signal is direct from the coil? wow, I wouldn't expect that.

I've found out that my original plan isn't quite as simple as I thought, since I built a simple (four LED, eight transistor as darlington pairs) with what I expected I needed as far as resistors and diodes. Got 1 always on (this would be the 500 RPM LED) and dimmed flashing on the others, on revving the engine, the additional LEDS are less dim, after reading your post, this makes sense. Funny how I got no oscilation from my analog meter, I guess I should invest in a low end or pre-owned osciliscope.

I'll give it a shot with www.niksula.cs.hut.fi/~mdobruck/siililan...ien/tacho/tacho.html
On the plus side of things, I made a minor boo-boo during my experimentation, I connected my multi-meter in VDC mode with the probes in ground and AMP<20 slots. When checking the "voltage" it promptly shut off the bike because it blew one of the fuses, fortunately not in my multi-meter. After replacing the fuse, everything was back to normal (tach still dead). Based on the circuit in that link, I'm reasonably sure even frying it to a crisp will still keep the rest of the bike's electrical components intact.

The tacho/volt switch shouldn't be a problem. I'll get back to this thread after I get the chips.

While over simplified (purpose is minimal components) this would probably be a very good start for your battery voltage LED gauge.
www.otherpower.com/images/scimages/236/vm_schematic2.jpg
IIRC, red is < 12 volts, Green 1 is 12.5, Green2 is 13, Green 3 is 14, Yellow (warning) is 15+ volts. I started with this before tweaking it w/ darlington pairs, removing the undervoltage part of the circuit, playing with resistor values and running all the cathodes on LEDs to a common 1K resistor. I also used 1N4001 instead of 1N4004 diodes.

solitary wrote:

While over simplified (purpose is minimal components) this would probably be a very good start for your battery voltage LED gauge.
www.otherpower.com/images/scimages/236/vm_schematic2.jpg
IIRC, red is < 12 volts, Green 1 is 12.5, Green2 is 13, Green 3 is 14, Yellow (warning) is 15+ volts. I started with this before tweaking it w/ darlington pairs, removing the undervoltage part of the circuit, playing with resistor values and running all the cathodes on LEDs to a common 1K resistor. I also used 1N4001 instead of 1N4004 diodes.

Yeah, that's a crude idea of what I'm doing, but that one has no averaging circuit. There is a lot of AC content in the battery voltage due to the nature of alternators, and the AC really throws off the readings. The 1-LED version I made is immune to the AC content, and detect the average-voltage. The 3-LED monitor I'm designing works well (and uses very few parts), in tests, but I have to make the final assembly and write it up.

Post edited by: loudhvx, at: 2007/05/09 12:05

Shouldn't the alternator output be just DC since there's an internal rectifyer?
I understand how an alternator works in concept, but I don't have any experience beyond replacing the unit itself.

Yes, it's AC, the rectifier converts to DC, but doesn't smooth out the pulses. The pulses are considered AC content in a DC signal. There is also a lot of AC induced through the actions of the crude regulator.

If you look at the voltage on the battery with an o-scope, you'll be amazed at the amount of noise on the signal. The battery voltage may swing from 12 to 16v, 100 times a second under certain conditions.

That's why any sophisticated electronic like a tach, or speedo etc. will need it's own dedicated regulator to smooth out the 12~16v.

Post edited by: loudhvx, at: 2007/05/11 10:51

Ah, I see.

I just assumed that after it was run through a full-wave rectifier, it was considered DC, since there's no negative current. But I can also see (now that I've been so enlightened) why the wave counts as AC.
Those o-scope shots are quite amazing. I never and I mean never would have guessed that the alternator output would look like that.

Yeah, I took down the link because it really only shows the AC side of the rectifier.

The DC side looks like 14, with a 2v ripple on it. The DC side of the rectifier IS considered DC, but is considered pulsed-DC until there is a "filter" like a battery or cap in parallel.

If you remove the DC average from the signal, like say through a capacitor in series, the remaining signal on the other side of the cap is AC. That's why it's said to be DC with "AC content", or has an "AC component", but the signal overall is considered DC out of the rectifier.