Tachometer signal multiplier

[ClarkAE]

Hi there, I have just joined this forum to ask for help. I have searched the forums already but unfortunately I have not found anything that is of any help to myself.

I have just carried out an engine conversion on a project car and am trying to convert a 12v 4 pulse per revolution signal to a 12v 64 pulse per revolution signal. Therefore I am looking to make or buy, depending on the difficulty, a 12v 16x signal multiplier so that my tachometer reads correctly.

There are a fair few tachometer signal converters on the market however the most they go to is 10x multiplication and these tend to be quite expensive. I greatly appreciate any help you guys can give me.

Thanks Dan
ClarkAE

 

[4pyros]

Dan; Cant you change the pick up wheel? Andy

 

[MikeMl]

To do 16X electronically, you would have to do a Phase Locked Loop with a 4bit binary counter divider. That is a bit complicated...

 

[Mosaic]

Can't an MCU simply sample the period between pulses and then output a 16 pulses for a matching time period? Thus autocorrecting 4 times per rev. That should be passably accurate as a display device.

However, if the RPM timing signal is DRIVING your injectors you must change the wheel, as any phase shift introduced by electronics will damage your engine.

 

[ClarkAE]

4pyros

Dan; Cant you change the pick up wheel? Andy

Hi Andy, unfortunately I cannot change the pick up wheel as this engine does not have one, nor is there space to mount one on the front pulley. This signal comes direct from one PCM and I need it converting before it goes into the next PCM which subsequently drives the tachometer.

To do 16X electronically, you would have to do a Phase Locked Loop with a 4bit binary counter divider. That is a bit complicated...

Hi, would this be accurate? as at full rpm I believe the pulse rate would be around 448,000 per minute. I don't know the capabilities of the circuit you mentioned nor do I have the ability of creating it however if I know what to ask for I can perhaps shop around and find a company or individual willing to make one for me.

Mosaic

Can't an MCU simply sample the period between pulses and then output a 16 pulses for a matching time period? Thus autocorrecting 4 times per rev. That should be passably accurate as a display device.

However, if the RPM timing signal is DRIVING your injectors you must change the wheel, as any phase shift introduced by electronics will damage your engine.

Hi Mosaic, good point and one I failed to mention. This device is purely to drive the tachometer in the dashboard. I have two ECU/PCM's in the car one that runs the engine and one that runs the dashboard as the engine is OBD1 and the car is OBD2 so I need the original ECU to run all the digital equipment from inputs given to it by the first PCM.

The MCU you talk of certainly sounds something that would be more than good enough for what I am looking for. I am not after something that works one hundred percent accurate, I would just like the tach to give me a rough idea of what the engine is doing and so that I don't have a motionless guage in front of me.

I appreciate your help guys, thanks Dan

 

[Reloadron]

I recently helped a coworker with a tach issue. In this case the tach was as many are, an analog meter. The tach had a small circuit booard driving it that used a F/V (Frequency to Voltage) converter that provided the analog tach with a signal through a resistor network at the F/V output. All we did was modify the tach board resistor network and were able to perfectly scale the thing. Therefore if you have an analog meter as a tach it may be much easier to just modify the analog signal going to the tach.

Ron

 

[DirtyLude]

I had to do this for someone who did a gauge swap on a motorcycle. I didn't have to multiply 16 times, but the code would be very much the same. I used a microcontroller (MSP430) and an optocoupler input.

It's not that hard with a microcontroller and shouldn't be hard to get one of those Arduino's or PICs/AVRs to do it with a simple input/output level converter.

 

[ClarkAE]

Reloadron

I recently helped a coworker with a tach issue. In this case the tach was as many are, an analog meter. The tach had a small circuit booard driving it that used a F/V (Frequency to Voltage) converter that provided the analog tach with a signal through a resistor network at the F/V output. All we did was modify the tach board resistor network and were able to perfectly scale the thing. Therefore if you have an analog meter as a tach it may be much easier to just modify the analog signal going to the tach.

Ron

Hi Ron, unfortunately the only analogue signal I have available to is the one coming from the engine PCM. This then goes into the ECU and goes digital. The instrument cluster is connected through the IBUS, the body network for the car so unfortunately I can't see a way of modifying the signal between the ecu and gauge.

DirtyLude

I had to do this for someone who did a gauge swap on a motorcycle. I didn't have to multiply 16 times, but the code would be very much the same. I used a microcontroller (MSP430) and an optocoupler input.

It's not that hard with a microcontroller and shouldn't be hard to get one of those Arduino's or PICs/AVRs to do it with a simple input/output level converter.

That sounds like just the sort of thing i'm after. As I mentioned before to a previous reply, at full RPM this will be pulsing at around 448,000 ppm so would this circuit be able to cope with that frequency and also is this perhaps a project you would like to undertake? Obviously your time and costs would be paid for in advance.

Thanks Dan

 

[4pyros]

I have two ECU/PCM's in the car

have you looked at the automotive sites to see if this has been done before? Can you use an aftermarket tach? Seems like a lot of work for a tach.Andy

 

[ClarkAE]

have you looked at the automotive sites to see if this has been done before? Can you use an aftermarket tach? Seems like a lot of work for a tach.Andy

Hi Andy, yes there have been many conversions, engine swaps done before but none with exactly this engine model/car model that I am aware of. It does seem like a lot of work for a tach but I like everything to be working exactly the way it should, otherwise it would feel like a bodge job. Cheers Dan

 

[tcmtech]

The simplest way is to use a LM2907 frequency to voltage converter IC and then use its variable voltage output to drive a LM566 voltage to frequency converter IC to convert it back to a new higher frequency signal. I made one a few years ago to convert the low frequency speed sensor signal on a 1985 F150 to a higher frequency so that I could run and aftermarket electronic transmission in it. The electronic transmission needed a much higher input frequency than the stock cruise control sensor supplied.

Its a relatively simple circuit to make and calibrate.

 

[MikeMl]

What is the existing pulse rate at idle and at max rpm? What is the pulse shape and amplitude? What kind of pulse does the tach want?

 

[ClarkAE]

MikeMl

What is the existing pulse rate at idle and at max rpm? What is the pulse shape and amplitude? What kind of pulse does the tach want?

Hi Mike, the pulse rate at idle which is around 800rpm would be equate to around 3200ppm, when to display this 800rpm value on the clocks the ECU would be looking for a signal of around 51200ppm a 16x multiplication of the original signal. As far as I am aware the pulse shape is not of much importance to the signal as the ECU would be expecting a wave signal from a hall sensor but is accepting and using a square signal that works the tachometer, but as mentioned, is 16x below the required frequency. Peak value on the pulsing signal is 12v. Unfortunately this is all the information I have to go on, cheers Dan

 

[DirtyLude]

I'll check to see whether my tach adjuster will work on a 1 to 16 up convert at 1/2Mhz. I made a few nice little PCB's up for it, but the microcontroller is quite limited and only has a single 16bit timer that I use for both capture and output. It may not be able to handle such a wide range. I'll check it this weekend.

If not, I can do a 1 off with a more capable chip, but TCMTech's suggestion sounds like a good one.

 

[ClarkAE]

DirtyLude

I'll check to see whether my tach adjuster will work on a 1 to 16 up convert at 1/2Mhz. I made a few nice little PCB's up for it, but the microcontroller is quite limited and only has a single 16bit timer that I use for both capture and output. It may not be able to handle such a wide range. I'll check it this weekend.

If not, I can do a 1 off with a more capable chip, but TCMTech's suggestion sounds like a good one.

Hi, that sounds great, let me know how it goes. I did pm TCMTech about his suggestion as I do not know enough about these things to go about making one for myself. Unfortunately he does not want to make one for me and I don't have any electronics colleges or centers near me to ask for them to make it. This is the best response I have had regarding this issue in around 3 months as I never thought about the forums until it was suggested to me.

Many thanks, Dan

 

[MikeMl]

The frequency range of the output is about 233Hz to 2.33kHz (800 RPM to 8000 RPM). I would use a Phase Locked Loop (PLL). The Voltage Controlled Oscillator (VCO) must tune the range 233Hz to 2.33kHz. Feed that to a 4bit binary counter (wired as a divide by 16) and then use a phase/frequency comparator to compare the (VCO frequency)/16 to the input pulse rate from the engine.

This could be done with a couple of ICs, like a CD4046 or even the old LM565, plus a CMOS four bit counter.

 

[DirtyLude]

I was just thinking about this. So what is the max RPM? Currently my MCU solution as it's programmed right now needs to handle the range between 400RPM to 16k RPM, but with a significantly decreased range, this should not be hard at all to do the increased spread.

 

[ClarkAE]

Hi, well this engine will never see more than 7500RPM I will be surprised if it ever makes 7000rpm to be honest but we shall say 7500 to be safe. That would mean a maximum output pulse rate of 480,000 p/p/m from a 30,000 p/p/m input at max rpm.

Cheers Dan

 

[DirtyLude]

Okay. I had to setup a new computer and getting CCS running on it turned out to be a pain, but I got it working. I modified my code, because the input range it accepted wouldn't have worked for this application. The code is now simpler and yet more complex, but it will handle just about anything any consumer engine can throw at it.

I also had to do a small modification to the hardware. I made this board with a PNP transistor in mind to drive the output tach, but wasn't until later I realized the MSP430 doesn't have output pins that will sink. I had to make it into an NPN with a pullup. You can see some creative soldering at the top of the picture.

Anyway, send me an address and I'll send it to you.

 

[ClarkAE]

DirtyLude

Okay. I had to setup a new computer and getting CCS running on it turned out to be a pain, but I got it working. I modified my code, because the input range it accepted wouldn't have worked for this application. The code is now simpler and yet more complex, but it will handle just about anything any consumer engine can throw at it.

I also had to do a small modification to the hardware. I made this board with a PNP transistor in mind to drive the output tach, but wasn't until later I realized the MSP430 doesn't have output pins that will sink. I had to make it into an NPN with a pullup. You can see some creative soldering at the top of the picture.

Anyway, send me an address and I'll send it to you.

Hi there, thanks you for doing that it is very much appreciated I look forward to seeing it. My e-mail address is daniel@clarkae.co.uk

I have also just been sent this diagram using a slightly different method to the one you describe, it will be interesting to compare the two methods.

Thanks again, Dan