Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

How to make simple frequency divider?

  1. Nov 14, 2011 #1
    Hi guys, new the forum!

    Long story short, I am doing an engine swap in a 80's 3 series bmw and need to make the tachometer read correctly. Currently the tachometer reads twice as high because the tachometer is made to read a 4 cylinder (now an 8 cylinder). Thus I need to divide the signal in half.

    There is already people selling these dividers for the swap I am doing but they are charging $75-100. I think I can make my own only with a few dollars, and I will learn something along the way!

    I had taken computer engineering in grade 12 where I used breadboards and all types of gates, so I have some very basic understanding of how circuitry works. I am now a 3rd year University Economics student, so feel free to post up equations if need be.

    A simple D-type flip flop (7474) chip should be able to divide the signal for me. However doesn't the chip require a 5v power? The vehicle puts out 12v. Would I have to take the 12v feed and have a resistor (how many Ω?) just before the chip? Then my output will also be 5v right? How do I raise it back up to 12V to continue to the tach?

    Here is one that someone sells if it helps:
    [PLAIN]http://img824.imageshack.us/img824/6421/8a620df5.jpg [Broken]
    Red – Power Supply (12V on ignition position)
    Brown – Ground
    White – Input rpm-signal (from engine control unit)
    Yellow – Output modified rpm-signal (to instument cluster)
     
    Last edited by a moderator: May 5, 2017
  2. jcsd
  3. Nov 14, 2011 #2
    The input signal from the tachometer might need to be cleaned up with a comparator or something to use as the clock signal of a flip flop.

    You can find high voltage versions of logic chips of Mouser or Digikey. Here's a ample search to get you started:

    http://www.mouser.com/Semiconductors/Integrated-Circuits-ICs/Logic-ICs/Flip-Flops/_/N-6j78b?P=1z0w8t2Z1z0vwtqZ1z0wa4cZ1z0wa34 [Broken]

    Max Vcc for these chips is 15V and above.
     
    Last edited by a moderator: May 5, 2017
  4. Nov 14, 2011 #3

    AlephZero

    User Avatar
    Science Advisor
    Homework Helper

    A simple option would be use CMOS logic (4000-series chip numbers) instead of TTL. That will work fine on 12V (in fact on anything between about 5V and 18V) and it has a good tolerance of noise on the power supply lines (you will get noise pulses when turn indicators or brake lights switch on and off, etc).

    You may need a buffer chip (e.g. an inverter that can drive higher power output that standard CMOS) after the divider, depending how much current your tacho needs to drive it.
     
  5. Nov 14, 2011 #4
    Would I be able to use this chip:

    http://www.nxp.com/documents/data_sheet/HEF4013B.pdf

    It is within the range i need (3-15v). Excuse my ignorance but what does is "5 V, 10 V, and 15 V parametric rating" referring to? Is it ok to feed and output 12v?

    Would I get what I need if I connected 12v to VDD, Vss to ground, incoming tach signal to CP, Qbar to D, then Q to outgoing tach signal, and left all other pins alone?
     
  6. Nov 14, 2011 #5
    Look at the tables in the datasheet, and it has values for Vcc = 5V, 10V, 15V. It's a feature of the datasheet, not the chip it seems.

    Clock is a Schmitt Trigger input, which is really nice since the output from the wheel sensor is probably not a nice clean square wave.

    And it specifically mentions automotive in applications. It looks perfect.

    Your wiring is correct, but don't forget to connect SD and CD to ground as well. They are an asynchronous set and clear, so if they aren't connected, any static charge that accumulates on them could screw up the output.
     
  7. Nov 14, 2011 #6
    Will it work even if the incoming frequency is not constant?
     
  8. Nov 15, 2011 #7

    AlephZero

    User Avatar
    Science Advisor
    Homework Helper

    One big difference between CMOS and TTL logic is that TTL was designed to operate at a fixed supply voltage (5V) while CMOS can use a wide range of supply voltages. Many parameters of CMOS, like power dissipation, max operating speed, logic propagation delays etc depend on the supply voltage. That's why the data sheet gives values at a range of different voltages. There is no reason not to use intermediate supply voltages.

    You should never leave any input pins floating with CMOS devices. Tie them to 0V or Vcc as appropriate.

    If you want to be ultra-cautious about static electricity damage, connect unused outputs to ground through 100K or 1M resistors as well.

    Sure.
     
  9. Nov 15, 2011 #8

    sophiecentaur

    User Avatar
    Science Advisor
    Gold Member

    Is the final display analog or digital? If the signal from the chip just goes to an analogue meter then a simple potential 2:1 divider before the meter could do the trick.
     
  10. Nov 15, 2011 #9
    I could be mistaken, but wouldn't it be digital if the car is fuel injected, and the rpm signal is provided by a computer (ECU)? Would the chip still work if the signal was analog?

    AlephZero: Should I ground all the other unused pins also for the second d-type flip flop on the chip, or is it only necessary to ground the unused pins for the d-type flip flop I am using?
     
  11. Nov 15, 2011 #10

    sophiecentaur

    User Avatar
    Science Advisor
    Gold Member

    Ground every unused input. Why not?
     
  12. Nov 15, 2011 #11

    AlephZero

    User Avatar
    Science Advisor
    Homework Helper

    All the input pins should be connected to something (ground or Vcc) for the unused gates as well as the used ones.

    The input resistance of CMOS logic circuits is very high, so small static electricity charges can create very large voltages (i.e. the order of kilovolts) . Modern CMOS chips have better built-in static protection than the early ones, but it costs nothing in components and very little board space to design out the problem completely.

    Also, CMOS logic gates behave as linear amplifiers when the inputs are at about half the supply voltage, and in that state the chip can consume an excessive amount of power, and/or self-oscillate and feed noise into the rest of the circuitry. Floating inputs can allow that to happen, even if static is not an issue.
     
  13. Nov 15, 2011 #12

    sophiecentaur

    User Avatar
    Science Advisor
    Gold Member

    Why did no one pick up on the idea of potting down the feed to the analogue meter? Most tachometers are surely not digital - and that yellow wire in the picture certainly looks like it carries an analogue signal.
    Sometimes less is more.
     
  14. Nov 15, 2011 #13
    It might not be an analog signal, but a frequency, creating a pulse every time a gear passes by a magnetic detector or something.
     
  15. Nov 16, 2011 #14

    sophiecentaur

    User Avatar
    Science Advisor
    Gold Member

    You could find that out with a DC test signal or looking inside the meter. (Or even look at the connections on it - is there a power supply pin, for instance?). Or just llook with an oscilloscope???
     
  16. Nov 16, 2011 #15

    Averagesupernova

    User Avatar
    Science Advisor
    Gold Member

    Chances are the meter movement is a motor. Can't put a series resistor with that and accomplish the goal can we?
     
  17. Nov 16, 2011 #16

    sophiecentaur

    User Avatar
    Science Advisor
    Gold Member

    A 'motor'? How would that work? Something new in the analogue meter display field?
     
  18. Nov 16, 2011 #17

    Averagesupernova

    User Avatar
    Science Advisor
    Gold Member

    Nothing new at all. Some gauges use motors. Have for quite some time.
     
  19. Nov 17, 2011 #18

    sophiecentaur

    User Avatar
    Science Advisor
    Gold Member

    Stepper motors?
     
  20. Nov 17, 2011 #19

    sophiecentaur

    User Avatar
    Science Advisor
    Gold Member

    I'd really like to have a reference to this. I have spent a healthy amount of time, searching, and could only come up with analogue meter displays.

    [Edit: found one or two now. Interesting - I wonder how common they are and when first introduced.]
     
    Last edited: Nov 17, 2011
  21. Nov 18, 2011 #20
    I think I have made a discovery! I know what is called the "coder plug" on the cluster effects how the cluster displays the tach because the car is a bmw e30 (e30 is the generation of 3 series bmw's from 1984-1991). The e30 came in 4 and 6 cylinder models, the tachometer will display correctly in either the 4 or 6 cylinder models IF you swap the correct coder plug (whole cluster replacement not required). Furthmore for this swap, if you start with a 6 cylinder model, and swap in a 4 cylinder coder plug it will show half of what an tach rate the 8 cylinder is running at.

    Thus if we can figure out how, all i really need to do is modify this coder plug. So i decided to dismantle it and took pictures, to my surprise all it is, is one chip, thats it. Question is what chip do i replace this with now or modify this current one to display half? The code on top of the chip reads "HML 087 8701" can anyone identify it? Mind you this is a 6 cylinder code plug so me running this would be off by 3/2 i think? I can very easily and cheaply get a 4 coder plug if it makes this easier.

    Pictures (last two are most important):
    IMG_0716.jpg
    IMG_0717.jpg
    IMG_0719.jpg
    IMG_0722.jpg
    IMG_0724.jpg
    IMG_0725.jpg
     
Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook




Similar Discussions: How to make simple frequency divider?
  1. Frequency divider? (Replies: 10)

Loading...