Tacho Signal Converter

[petert 586]

This is where it may fall down with a lot of people.... I used a known signal input (from a signal generator) to make sure that the tacho was reading right. Adjusting the pot changing how the tacho reacted.

 

I'm familiar with wave forms, voltages, frequencies, duty cycles etc, I was hoping you could be more specific about where you injected the test signal.

[madspikes 41 1 Cars]

People always want more!

 

My inital testing was done using a square wave from a signal generator with a 50% duty cycle (0-6v). I cant remember how it reacted to a sine wave inject but I'll try and do some more work over the weekend.

 

If you give a rough idea of the duty cycle and voltages required I'll give it a bash

 

This work was no by means complete because I just got to work for myself, I wasnt expecting soooo many people to be intrested!

 

Mad.

[madspikes 41 1 Cars]

Ok, next update.

 

I couldnt borrow a fully programmable signal generator for the weekend, so I'm stuck with old clunky!

 

Voltage - As long as the low is 0v and the peek is above +4.5v tacho works fine.

 

Signals - Square / Triangle / Sine wave forms all give the same output on the tacho. The input wave forms are generated -x volts to +x volts then run through a diode to leave just the +ve side.

 

Duty cycle - the MS gives out a wave form where its off (0v) for 1ms and on (+5v) for the rest time period, and the tacho readings are as before.

 

So, duty cycle, signal input voltage and wave form seem to have no effect on the tacho output. The only thing that does change the tacho read is the input frequency!

I hope this helps.

 

Mad.

[niklas 1]

Ok, next update.

 

I couldnt borrow a fully programmable signal generator for the weekend, so I'm stuck with old clunky!

 

Voltage - As long as the low is 0v and the peek is above +4.5v tacho works fine.

 

Signals - Square / Triangle / Sine wave forms all give the same output on the tacho. The input wave forms are generated -x volts to +x volts then run through a diode to leave just the +ve side.

 

Duty cycle - the MS gives out a wave form where its off (0v) for 1ms and on (+5v) for the rest time period, and the tacho readings are as before.

 

So, duty cycle, signal input voltage and wave form seem to have no effect on the tacho output. The only thing that does change the tacho read is the input frequency!

I hope this helps.

 

Mad.

 

Excellent information!

 

My only question is regarding calibration. Did the calibration alter the rpm returned for a specific duty cycle?

If so, do you have any reference values where duty cycle n% = m rpm etc?

[madspikes 41 1 Cars]

Err.... much beer had... christmas doooo.

 

 

I cant really comment on that at the moment as I said that the tacho had been played with before I got it. If I've got time this weekend I'll take the virginal from the pug and pop is its after market cheery!

 

 

 

[madspikes 41 1 Cars]

Did the calibration alter the rpm returned for a specific duty cycle

 

As far as I can tell the input duty cycle has nothing to do tacho output, its the frequency of the pulses that makes the difference. This is show by the fact the MS and signal gen produce pulses at vastly different duty cycles but at the same frequency and the tacho reading is the same.

 

image from jochem

 

This shows the gti6 tacho pulse, which is very simular to the MS pulse... So in theory it should work with the GTi6.

 

gti6 tacho pulse

[Super Josh 4]

All I did was cut the track near the first pin of the IC (chip) isolate it from everything else. Then I applied my rpm signal to that.

 

So this should end up being a very simple mod, looking at the schematic posted by Mad Prof. Rather than cutting any tracks, we should be able to replace the in line resistor to pin one of the I.C. with a wire link, best to leave the DC blocking cap in place. Remove the resistor and capacitor from the signal conditioning RC network to ground.

 

The Schematic shows a variable resistor in circuit already, Madspikes, was this not on your particular Rev. counter?

 

 

 

 

I have a decent programmable Sig. Gen. at home and an untouched Rev. counter, if you want me to have a play. I am tempted to connect up my generator to my in car Tacho anyway, just so that I can calibrate it Because they all seem to be a little inaccurate

We could offer a calibration service to other Pug. owners, it would be the most accurate thing in the car

 

 

 

Josh

[jackherer 543]

I am tempted to connect up my generator to my in car Tacho anyway, just so that I can calibrate it Because they all seem to be a little inaccurate

We could offer a calibration service to other Pug. owners, it would be the most accurate thing in the car

 

I've seen up to +/-700rpm variation (at the rev limiter) between tachos when swapping engines/ecus from one 205 to another!

[niklas 1]

I've seen up to +/-700rpm variation (at the rev limiter) between tachos when swapping engines/ecus from one 205 to another!

 

The obvious question is what causes these variations!

[jackherer 543]

The obvious question is what causes these variations!

 

they're cheap mass produced cars so Peugeot just made the tachos 'good enough'...

[niklas 1]

they're cheap mass produced cars so Peugeot just made the tachos 'good enough'...

 

Definitely!

If I rephrase it a little, which part of the revcounter is usually the inaccurate part?

Is it the dash? Or is it the ECU?

What I'm after is that if the tachometer is used with megasquirt (or other), will the revcounter still be inaccurate?

[jackherer 543]

Its the tacho itself that is inaccurate, like I said above I have had one engine/ecu in several 205s and seen several different RPM figures at the rev limiter.

 

Based on this thread it looks like you will be able to calibrate it when you adapt it to work with a megasquirt anyway.

[madspikes 41 1 Cars]

So this should end up being a very simple mod, looking at the schematic posted by Mad Prof. Rather than cutting any tracks, we should be able to replace the in line resistor to pin one of the I.C. with a wire link, best to leave the DC blocking cap in place. Remove the resistor and capacitor from the signal conditioning RC network to ground.

 

 

Ok, this basically what I've done. I though it was just simpler to cut the couple of tracks! Leaving the DC blocking is prob a good idea.

 

 

The Schematic shows a variable resistor in circuit already, Madspikes, was this not on your particular Rev. counter?

 

Nope it had a perment resistor.

 

We could offer a calibration service to other Pug. owners, it would be the most accurate thing in the car

 

It is a very sensitive little device... It has little balance screws in side which you can use to adjust it, 1/4 turn can make a large difference. Its a kinda electrial mechanical device!

Bare in mind, its french, its old, its electrial its never going to work exactly right! its for 'indication only'

 

We could offer a calibration service to other Pug

We could indeed. Just remember who did it first!

 

Ps.

 

How do you do the quote thing properly?

 

Mad.

[petert 586]

I finally got around to modifying my tacho in my track car as per some of the instructions posted above. However, rather than breaking tracks etc., I simply "jumped" a wire around the first resistor (unknown value) in the diagram below:

 

Once working, I then turned my attention to calibration. Rather than using a 100 kohm trim pot like in the diagram, I replaced the original 16.4K resistor with an 18K resistor. It was impossible to get it accurate everywhere, so I settled on making 7000 the accurate point. I used a signal generator for the calibration signal. It was important to make sure the tacho was at the same angle as per the dash. Laying it horizontally gives different results. This was the final result:

 

RPM - Hz - tacho readout

1500 50 1600

2000 66.7 2100

2500 83.3 2600

3000 100.0 3150

3500 116.7 3700

4000 133.3 4250

4500 150.0 4800

5000 166.7 5250

5500 183.3 5700

6000 200.0 6200

6500 216.7 6550

7000 233.3 7000

7500 250.0 7500

 

It's a bit academic anyway, because you rarely have time to look at the tacho anyway in a track car, relying on the shift light and rev. limiter.

Edited March 26, 2007 by petert

[madspikes 41 1 Cars]

I also used the little 'weight' screws to help get it more accurate across the rev range.

 

mad.