RVI to RVC Tacho Conversion

[codekiddie]

I have been building a Tacho conversion circuit for my RVI current sensing Tacho, to an RVC contact breaker sensing one, so it will work with my electronic ignition. I will add more details here when I can about the project, but here is a little taster

The circuit board fixed on the Tacho mechanism.

Inside the case:- a little bit tight :?

Not much headroom either :shock:

Perfectly aligned calibration pot over the original calibration hole I don't just throw these things together you know 8)

Calibration in progress

I created Windows .WAV files with square waves to generate pulses, and then fed that signal through my amplifier and then on to the tacho so that the signal was at a level it could see.

The required signal frequencies required for calibration is worked out as follows:

For example, at 2000RPM, (revs per minute as we all know), we get 2000/60 revs per second, which is 33.333 revolutions. For each revolution of the engine only half of the cylinders are firing, so for my V8 I multiply by 4 giving me a calibration frequency of 133.333Hz.

The easy math is to divide the RPM by 15

For simple calibration I injected a 133.333Hz signal for 2000RPM, a 266.666Hz signal for 4000RPM, and a 400Hz signal for 6000RPM.

I managed to calibrate it fairly OK, but the meter was reluctant to reach a reading of 6000RPM. To obtain a reading of 6000RPM, I had to inject a 410Hz signal, which was the equivalent of 6150RPM, so only an error margin of 2.5%

I have added a link to the calibration video below. The screen in the background is the .WAV file generation program I used, and you can see the frequency increasing when the tacho RPM indicator moves 8)

Calibration Video

All good so far

 

[ghce]

Sticky required on this!

wot no birds nest, far too tidy

Graeme

 

[vin-kohler]

But where is the circuit diagram?
Vin

 

[SydneyRoverP6B]

Very clever 8)

Ron.

 

[codekiddie]

wot no birds nest, far too tidy Graeme

Sorry, must be old age creeping in and wanting to do things properly :shock:

Is this more of what you wanted to see?

But where is the circuit diagram?Vin

Sorry, just a bit rushed this weekend

The data sheet for the LM2917N has a couple of circuits for tachometers, and after looking into what others had done I found this circuit, which adds a couple of transistors to allow for meters which may draw more than the rated current for the device, and since my meter only read 10ohms I opted for this design :shock:

Very clever 8)
Ron.

Thanks Ron.

I promise to add a full write up when I have more time

 

[vin-kohler]

Before diving into the details of the circuit operation two queries come to my mind.

1) Is there any kind of overvoltage resp. reverse voltage protection upstream of the power connection? Those semiconductors are so unforgiving against voltage spikes.

2) Concerning the unlinearity in the upper range: My first suspect would be the 8.2V zener close to the T2 voltage follower whose stability is questionable under changing load. A voltmeter across the zener - or even from 'Meter +' to GND during the calibration might confirm/refute this.

Vin

 

[codekiddie]

Before diving into the details of the circuit operation two queries come to my mind.

1) Is there any kind of overvoltage resp. reverse voltage protection upstream of the power connection? Those semiconductors are so unforgiving against voltage spikes.

2) Concerning the unlinearity in the upper range: My first suspect would be the 8.2V zener close to the T2 voltage follower whose stability is questionable under changing load. A voltmeter across the zener - or even from 'Meter +' to GND during the calibration might confirm/refute this.

Vin

1) The device input voltage range is up to 28V, and a simple series diode in the power connection should cure any reverse voltage possibilities, and will be easy to add. The device itself has an internal Zener Diode regulator, so possibly nothing to gain with an external one.

2) I have just measured the voltage across the meter connections and at 2000, 4000 and 6000RPM, and they are 0.25V, 0.50V, and 0.745V, so may account for some of the error (0.6% ish)

As with most projects I am sure there might be a version 2 if version 1 isn't up to it:- time will tell :?

 

[vin-kohler]

I have just measured the voltage across the meter connections and at 2000, 4000 and 6000RPM, and they are 0.25V, 0.50V, and 0.745V, so may account for some of the error (0.6% ish)

What I meant is: Measure the voltage from 'Meter +' to GND. This one should be constant at ca. 7.5V if the zener is not to blame.

Vin

 

[codekiddie]

I have just measured the voltage across the meter connections and at 2000, 4000 and 6000RPM, and they are 0.25V, 0.50V, and 0.745V, so may account for some of the error (0.6% ish)

What I meant is: Measure the voltage from 'Meter +' to GND. This one should be constant at ca. 7.5V if the zener is not to blame.

Vin

Sorry Vin.

Across the Zener the voltage does seem to fluctuate a little, around 8.36V quiescent, 8.34V @ 2000RPM, 8.33V @ 4000RPM and 8.32V @ 6000RPM. I am always open to suggestions for circuit mods and enhancements

 

[vin-kohler]

Ok, Phil, and mind you: I just wanted to niggle. I fully agree that one should be satisfied with the observed tolerances.

Although for your Mark-2 model you might want to select a three terminal voltage reference instead. That is provided that the instrument mechanics itself responds linearly.

Concerning the voltage spikes (i.e. transients) I found this one http://www.industrologic.com/autotransients.pdf.

And mind that: (copied from another forum)
"The word "spike" is copyrighted by Mr. Spike Lee. You should expect a call from Johnny Cochrane anytime now for your flagrant misappropriation of the word "spike", and it's use without prior written consent. The authorities have been notified."

Vin

 

[codekiddie]

Here is the full write up of what I did to convert my tacho from an RVI type to an RVC type. Sorry if I get a bit technical, but I like to know how things work, and thought others may be interested :shock: :shock: :shock:

If your head starts to hurt simply scroll to the first picture

As mentioned above I used an LM2917N frequency to voltage converter IC for the tacho circuit. Although the original data sheet had circuits for tachometers, I found a circuit, which adds a couple of transistors to allow for meters which may draw more than the rated current for the device, and since my meter only read 10ohms I opted for this design.

Data Sheet

Also mentioned before, for my V8 the relationship between RPM and frequency, (lets call it Fin), is:

Fin = RPM/15

so at 6000RPM

Fin = 6000/15 = 400Hz

The IC uses a simple formula to determine certain values to give the desired results.

Vout = Fin x Vcc x VR2 (R1 in the data sheet, but VR2 in my circuit) x C2 (C1 in the data sheet, but C2 in my circuit)

Vout - voltage output for a specific frequency input
Fin - Frequency input
Vcc - The value of R4 being 470R, fixes this to 7.56V
VR2 - calibration pot (0 - 100K)
C2 - 22nF (expressed as 0.000000022F for the formula)

I decided to use an output voltage of 6V for the maximum reading of 6000RPM (Fin = 400Hz), so...

6(V) = 400(Hz) x 7.56(V) x VR2 x 0.000000022

or

VR2 = 6 / (400 x 7.56 x 0.000000022)

so VR2 should be around 90.187K. Well within the limits of the 100K calibration pot 8)

Now the head hurting stuff is done lets move on to the easier bit

Firstly the instrument panel needs to be extracted from the car, the front of the panel removed, and then the three screws holding the tacho to the panel unscrewed thereby releasing the tacho to be worked on.

Markings on the face show the tacho to originally be an RVI type.

Release the clip on the metal band and remove

Place some protective material on the face and under the needle and pry off with a fork. Mine took quite a bit of force so be patient :shock: :shock:

Unscrew the fixings holding the face on to the body and remove it.

Now undo the screws and remove what I believe is a light diffusion panel. I imagine this works in the same way as the lighting for the centre controls for lights, wipers etc...

This will now show you the inner workings of the tacho.

Remove the two screws shown (or not shown since I have already removed them )...

...and withdraw the mechanism and circuit board.

Remove the two screws either side of the mechanism...

...unsolder the wires to the meter (noting which side was positive:- RED wire in my case)...

...to separate the circuit board and connector plate. The male and female bullet connections simply slide off of the plate, and unsolder the yellow power wire from the plate.

A test piece of board was used to ascertain the best mounting position that would give me the most board space in the available room. Having measured the distance between the screw holes, I made 90 degree bends on two lugs and soldered them to the board...

...and screwed it in place with two screws for the original circuit board. The lugs are also the ground connection for the circuit.

I checked for height clearance as well...

...and with some components to check the gap.

The new board I used allows you to make connections in either direction by simply soldering adjacent pads together.

Now to get this little lot onto the board!!!

A short while later, all present and correct and another test fitting...

...including location of the calibration pot using the original calibration hole!!!

At this point I checked the board out to make sure all was as it should be, and temporary long wires were attached to the it. The meter was then re-assembled with the face and needle so that I could calibrate it. It is far easier to calibrate with the board out on the table

The +ve side of the meter was temporarily connected up to a 6V supply of batteries, (remember from above I wanted 6V at 6000RPM), and a short placed between pin 8 and 10 of the IC socket to connect the -ve side of the meter to the 1K current limiting pot (not the 100K calibration pot).

The pot was then adjusted to give me the full 6000RPM of the meter, i.e. 6V = 6000RPM. According to the formula above, when I inject a 400Hz signal I should get 6V out and see 6000RPM on the meter

The calibration signals I used were simple .WAV files using square waves. I created a single .WAV file starting at 0.0Hz and increasing to 133.333Hz (2000RPM), pausing at 2000RPM for a while then increasing to 266.666Hz (4000RPM), pausing a while longer then increasing to 400Hz (6000RPM) pausing for a final short while then decreasing back to 0.0Hz. This ensures that the needle returns in an 'orderly' fashion, and not suddenly plummeting from FSD to zero :shock: :shock:

Here is the link to the calibration video again:

Calibration Video

Once initial calibration was complete, new shorter wires were connected to the board...

...which was then fixed in place, and the wires to the meter soldered on...

...and a bullet connection added for the -ve coil connection (this was soldered on, not crimped). The 12V connection was soldered in place and sleeved with heat shrink sleeving.

The bullet connection in place.

Fitted back into the case...

...and fully re-assembled.

The unit was then run through the calibration process again, and all still seemed well

Now just to find time to fit it in the car for a real test:- watch this space :shock:

And a little side note...

One thing to be wary of, if you are simply stripping the tacho for a clean or some fault finding, is that there is an insulated washer on one of the screws, which separates the metal body from one of the PCB tracks.

The metal body is an earth connection, but the track is the battery connection, so don't forget to replace the washer, especially since, (according to my circuit diagram), there is actually no fuse on the tacho circuit :shock: :shock: :shock:

 

[codekiddie]

Ok, Phil, and mind you: I just wanted to niggle. I fully agree that one should be satisfied with the observed tolerances.

Although for your Mark-2 model you might want to select a three terminal voltage reference instead. That is provided that the instrument mechanics itself responds linearly.Vin

I'm actually sure that the result I am seeing is the lack of use of the meter in the higher RPM range, compared to the frequently used 0-4000RPM range. After 40 odd years of only using the lower range I expect the top end may be a bit lazy comparatively :shock:

Concerning the voltage spikes (i.e. transients) I found this one http://www.industrologic.com/autotransients.pdf.Vin

Oh God:- more complicated math :shock: :shock: :shock:

And mind that: (copied from another forum)
"The word "spike" is copyrighted by Mr. Spike Lee. You should expect a call from Johnny Cochrane anytime now for your flagrant misappropriation of the word "spike", and it's use without prior written consent. The authorities have been notified."

Vin

:shock: :shock: :shock:

I need a lie down....

 

[vin-kohler]

A very good presentation. Congratulations!

Vin

 

[codekiddie]

A very good presentation. Congratulations!

Vin

Thanks Vin

 

[SydneyRoverP6B]

That's fabulous Phil.

Ron.

 

[WarrenL]

Tremendous how-to, Phil. I'd better bookmark it - if spindle lubrication doesn't stop my wobbling tacho needle, I might end up going down this path.

 

[ghce]

Tremendous how-to, Phil. I'd better bookmark it - if spindle lubrication doesn't stop my wobbling tacho needle, I might end up going down this path.

Warren the fact that it is indicating at all means yours will be just a repair not replacement, from what I understand the other type will not work at all with electronic ignition of the type you have. The spot of oil should do the trick, would not use wd40 or 3 in 1 oil, a light lubricant such as a sewing machine oil or a 10 weight bearing oil are needed and even then I think they may be a little viscous and should be very sparingly applied or applied via a carrier fluid to reduce the amount of oil. In the past I have used lubricating contact cleaners on these types of bearings to good effect.

Graeme

 

[WarrenL]

G'dam, I'm getting all confused now. Might just be the cold still hanging around and misting my brain. I'll lubricate the spindle before anything else - I even have genuine Singer sewing machine oil in the garage. If that doesn't work I'll reread the threads on this topic to try and get my head around it all. Brown Rover facts:

- RVI tachometer.
- Pertronix Ignitor electronic ignition.
- I think (although my memory is horribly unreliable) that the tacho was wobbling pre-Pertronix. Suggests lubrication is the answer.
- THAT instrument company "overhauled" my instrument panel some years back. Suggests lubrication should have happened, but potentially didn't.

 

[codekiddie]

G'dam, I'm getting all confused now. Might just be the cold still hanging around and misting my brain. I'll lubricate the spindle before anything else - I even have genuine Singer sewing machine oil in the garage. If that doesn't work I'll reread the threads on this topic to try and get my head around it all. Brown Rover facts:

- RVI tachometer.
- Pertronix Ignitor electronic ignition.
- I think (although my memory is horribly unreliable) that the tacho was wobbling pre-Pertronix. Suggests lubrication is the answer.
- THAT instrument company "overhauled" my instrument panel some years back. Suggests lubrication should have happened, but potentially didn't.

You could try lubrication, (very light I read somewhere), but there are still others that have had similar experiences to yourself.

Bouncing Tacho

I had a quick re-read of your 'Wobbly Tachometer Needle' thread, and you don't mention how you wired the +ve side of the Pertronix.

I am lead to believe that most instructions for electronic ignitions, (mine included), simply say 'Red to coil +ve, and Black to coil -ve', but then the current drawn by the unit will also be going through the (RVI) tacho, and since it is a current sensing device will register that current as well as the current through the coil.

This was also bought to my attention by Darth Sidious in my thread:

That Ballast Resistor Again

Have you tried connecting the Pertronix directly to the battery, (via a suitable fuse of course), or is that the way it already is connected?

 

[codekiddie]

Have you tried connecting the Pertronix directly to the battery, (via a suitable fuse of course), or is that the way it already is connected?

Obviously what I meant was an ignition switched battery feed, then through a fuse to the ignition module, but basically anywhere but directly to the +ve side of the coil :?