41 replies to this topic

[Leykor]

Model: Austin Mini Leykor
Year: 1971
Description of problem: I would like to build a electronic tachometer

Hi all

In a burst of inspiration I decided that I would like to build a tachometer. . Not just build, but design and build from the bottom up.

I've got the bulk of it figured out however there is one point I need clearance on. I’m hoping that someone might know.

My circuit design needs to count something, I need to know where the signal for the tachometer comes from. I have heard it originates from the coil? -ve? If so, does anyone know what type of signal is obtained? Like is it -12V, 0V, -12V, 0V ... ect. as the distributor fires to each plug?

Thanks

P.S. Will post my designs for anyone else interested in a project of this nature soon.

[Kerrin]

Keep us posted, very interested in what you end up building.

AFAIK. and may be wrong, the feed would be a 12v pulse each time a spark is generated to a plug, It is definately taken off the coil -ve lead on the OE tacho's.

Some ecu's ie MJ have a tacho output which creates pulses to run a tacho. but that is probably complicating things for you.

There was some american web site with a DIY digi tacho on it, I will have a look and see if its in my favourites somewhere.

[Kerrin]

No it was a digi speedo

Digi speedo

But maybe some help

Edited by Kerrin, 25 August 2009 - 10:17 PM.

[dklawson]

Google for information on the LM2917 series of chips from National Semiconductor. You're likely to find a couple of articles on the web generated by Stuart King. I'm not sure if he's living in the U.K. or NZ these days.

The LM2917 chip is basically a single chip supporting tachometer development. With a handful of caps and resistors you will have a tach with an analog output. This signal can be sent to a high impedance (low current draw) meter movement if you want an analog tach. If you want a digital tach you need to use the chip's output to feed basically a digital voltage panel meter and adjust the chip's gain to calibrate the output voltage.

The only limit to the LM2917 as a starting point is that chip itself cannot drive low impedance loads that require a lot of current. The work Stuart King started was trying to use the chip to power an old Smiths RVI tachometer movement. He got it to work but found the current draw was so high that he saw the accuracy of the circuit drift as the components heated up. Eventually I think he gave up.

I worked on a similar circuit but I tried a different approach than Stuart. I took the output of the chip and used it to feed a power op-amp which in turn fed the Smiths movement. Unfortunately the op-amp required both positive and negative power supplies. I found a cheap way to do that, but the circuit was suddenly much more complicated (many more components) than what Stuart was doing.

Regardless, you'll find lots of topics on the web discussing the use of the LM2917 chips for tachs, shift lights, rev limiters and other RPM specific functions.

[Leykor]

Hi good people of TMF

This is just an update on the digital tachometer design progress. I will describe my design ideas and would like some constructive critism. Thanks to all for their responses.

Doug I have considered using the mentioned LM2917 IC however I have decided against using it for my main design reasons that follow.

The LM2917 is a frequency to voltage converter; in a nutshell it outputs a DC voltage proportional to the input frequency. This output voltage is analogue and is thus extremely hard to convert to a format displayable on 7-Segment LEDs without employing the use of a microprocessor chip (PIC or the likes)

SO...

I have decided to follow 2 routes:

1 - A simple design based using the LM2917 and a PIC microcontroller for conversion to Binary Coded Decimal (BCD) format that can be displayed on four 7-seg LEDs. I will make the code available and final circuit schematics.

Now I understand that some of us have no experience using/programming microcontrollers, for this reason my second design will use basic IC`s available at any electronics store. All that will need to be done is a bit of soldering.

2 - Let me see if I can explain this correctly.

Theory of operation is to count how many pulses of a certain set frequency that can fit into the period of one engine revolution.

Four BCD counters will be cascaded ... this will enable the display to the 7 seg LEDs via a latch.

I will use a 555 timer (or the likes) to generate a high frequency continuous pulse train, this pulse train will increment the counters mentioned above.

A second counter will be used to count the amount of sparks occurring by using the (-)ve terminal of the coil as a signal. Once four sparks have occurred (all cylinders have fired) I will use this counter to do the following:

- Reset itself
- Latch the counted value to the 7 Seg display
- reset the 4 BCD counters


Its a bit vague but will get something working soon.

[dklawson]

I am not an electrical engineer (I'm mechanical) so take my comments with a grain of salt.

You can certainly use the 555 for your time base. However, there are probably nicer, more stable oscillators you can buy these days. However, 555 circuits are well documented out there.

When you are counting your pulses from the coil (-) terminal, your math has to be adjusted a bit. Four (4) voltage pulses is not a single (1) engine revolution. Since this is a 4-cycle engine, each cylinder fires once for every two revolutions (4 pulses = 2 revolutions) therefore, each time you count two pulses would represent one (1) revolution.

I haven't worked with PICs. I'm too old for that. I can see that you will come up with a nice digital gauge that way though and hope you keep us updated with your work.

For additional ideas on tachometer circuits that do NOT use the LM2917 chip, visit the web site of Theo Smit in Canada. It can be cumbersome navigating his site, but he has developed an upgrade board for older Smiths tachometers. His board can be used in current pulse or voltage pulse sensing mode. He also made his circuit with discretes so there is a possibility you might find part of his circuit useful.
Theo's page:
http://members.shaw....d/tachmod0.html

[Nightrain]

Interesting link doug could this be the answer to making RVI revcounters work with the Megajolt ?

[Ethel]

LM3914

[Leykor]

LM3914

This is my problem I realise you get IC's that give you a bar display. Im looking at taking analog to digital, such that it can be displayed of those number LED's(7-seg displays).

Thanks anyhow

[Ethel]

I'm not sure it doesn't also do 7 segment arrays, I'm pretty sure there are similar chip that do even if it doesn't

http://electronics-d..._volt_meter.php

[dklawson]

Interesting link doug could this be the answer to making RVI revcounters work with the Megajolt ?

Theo's board may work with MegaJolt... I simply don't know. Theo is very approachable and responds to most emails in a few days. I encourage you to send him an email and ask. Contact info should be on his web site.

[Leykor]

I'm not sure it doesn't also do 7 segment arrays, I'm pretty sure there are similar chip that do even if it doesn't

http://electronics-d..._volt_meter.php

Thanks for the link. I did come across the ICL7107, the problem with it is that its literally a 3.5 digit driver. The thousands digit only swings between -1 and 1. In terms of counting revoloutions this posses a problem as we need to go to 9999.

I am still searching for a four digit, 7 segment driver however if I cannot find one I do have a workaround that can be implemented.

[Leykor]

Hi All ... Just an update

Okay so i`ve been making really slow progress, trying to balance collage thesis and tacho project.

Non the less so far i've got a simple dispaly going as shown below . YEY

 Counter.jpg   43.39K   22 downloads
 Display.jpg   44.97K   13 downloads

Ive got the design for the rest on paper just need to build, test, calibrate and should be good to go.

Will keep you updated

[Ethel]

Impressive progress, I'd have just opted for the 3 digits n stuck a nought on the end so all credit to your perseverance.

[stormintrooper]

was this done with PIC chips or simple logic chips etc?