Tacho Repair

[Chrispy]

Yes, I did note that. For some reason the article didn't say to replace it and I didn't purchase the part to do it...

[mbell]

It calibrated ok and seems ok up to 4000 rpm but doesn't move any further.

Need to spend some time charactising the issue thou.

Had to get the car tested today, so had quick drive. Not much chance to fully test as raining but mine seems ok upto indicated 4.5k now and then starts flapping around a bit. Pretty sure it is slightly over reading at idle and under reading quite a bit at high revs. So much that I am wondering if I am hitting the rev limiter and this is just a calibration issue.

Could really use another method of reading the rpm. Maybe I'll adjust the rev limiter down very low and use that.

This is directly connected to the coil which is a pertronix one, with pertronix electronic ignition and rev limiter box.

[MarkDa]

My engine definitely revs up to the 6k limit but the rev counter stops (a bit wobbly) at 4k.

[billwill]

Basically these tachos square up the pulse from the ignition and then make them a fixed width. Then it averages the result to feed to the moving coil meter.

If the width of the generated pulse reaches the time interval between pulses then the average reaches 100% and it won't go any higher.

So the fix is presumably to make the pulses taller (more volts) but narrower in time.

[MarkDa]

Spiyda offer several solder points -apparently they just vary the resistance at input.
So on the basis of Billwills analysis we need low resistance to keep voltage up which can be effected by user.
But frequency is beyond us.
The input is determined by engine and as designer knows revs will be up to (say) 7k surely he'd allow for that.
Anyway when I get mine back I'll have a word with him to see if I can get it working at the upper end.
To be honest I don't really use the tacho as I have ears, a working Speedo and a rev limiter.
But having spent the cash it would be nice for it to somewhere near!

[billwill]

Spiyda offer several solder points -apparently they just vary the resistance at input.
So on the basis of Billwills analysis we need low resistance to keep voltage up which can be effected by user.
But frequency is beyond us.
The input is determined by engine and as designer knows revs will be up to (say) 7k surely he'd allow for that.
Anyway when I get mine back I'll have a word with him to see if I can get it working at the upper end.
To be honest I don't really use the tacho as I have ears, a working Speedo and a rev limiter.
But having spent the cash it would be nice for it to somewhere near!

>So on the basis of Billwills analysis we need low resistance to keep voltage up which can be effected by user.

No.. the voltage of the input signal is not what I am referring to. The different resistors for input are to bring the somewhat ragged signal from the coil down to a level where only the most significant part of the signal triggers the internal fixed width pulse. I.e ONCE per spark, not several per spark.

[MarkDa]

Electronics aren't my strong point so I'll await further information from Spiyda.

[rjaxe]

My experience of using the LM2917 IC circuit showed me that you need a well defined signal to avoid the sort of issues reported here. I built my circuit some years ago basing it on a design from a mini web site. My converted tacho worked fine on the bench and in the car when the lights were on and only noticed subsequently that the tacho stopped working when the lights were switched off. Eventually after weeks of fiddling with the power supply with no success I started to investigate the quality of the signal. First test was to leave the tacho powered from the car but feeding the signal from a generator and it worked fine lights on or off. So concluding that the problem was with the signal I put my scope on it and it was awful, noisy as hell with spikes all over the place, which part of this mess actually signalled the rpm was open to doubt. Also I could not determine what difference the lights on or off made to the signal. Speculating that the multitude of spikes could be seen as a high frequency signal, with the tacho on the bench I fed it with increasing frequency signal and at around 3-4kHz the tacho reading dropped to zero, as had happed in the car. So I concluded from this that the circuit needs a reasonably clean signal avoiding too high a frequency.
I stumbled on an article on the web which showed a pulse clean up circuit applied to a different tacho but ever optimistic I tried this on my set up and it worked. I modified the circuit board to incorporate the resistors, caps and diode needed and installed it in my car and been pleased with myself ever since. Anyone interested in this pulse clean up circuit let me know and I will point you in the right direction.
No doubt some learned fellows on this forum can make a much better analysis of this than I have but having spent many fruitless hours chasing phantoms once I got it working satisfactorily I was only too pleased to get back to driving.
Out of interest my car is a 1971 Plus 2 S130 with an Aldon ignition.

[billwill]

See https://www.princeton.edu/ssp/tiger_cub ... eforms.pdf for an article on ignition waveforms.

[Chrispy]

An update on my tacho.

It was running really well for the first 2 drives but when my and the girlfriend went up the local mountain for coffee on Sunday morning we hit a bump in the road and the tacho needle jumped. Didn't think much of it, but a few minutes later it was jumping around all over the place and occasionally locked at full sweep

Pulled it apart and found out it was the variable resistor which is used for calibration was loose. Tightened up the little nut and it was all good! Must have been jiggled a bit when I was soldering the new resister in.

Here's a quick vid showing what was going on and how much the rheostat was moving. https://www.youtube.com/watch?v=LHzjSY3n0GQ

Originally thought it might have been something to do with electronic ignition, but turns out I'm still running normal points.

[nmauduit]

Here's a quick vid showing what was going on and how much the rheostat was moving. https://www.youtube.com/watch?v=LHzjSY3n0GQ

It could be that the solder is broken at the component being touched on the video, so that the conductivity or contact resistance varies (with bumps, or when touched by a finger) : you may want to try resoldering anew that component and check if the tacho still behaves erratically when touched.

[Chrispy]

Nope, I checked that. It relies on straight contact rather than a soldier joint from the body of the rheostat to the circuit board. Bit different...

Went for w good drive yesterday and the needle was rock solid and behaving beautifully.

[spiyda]

As the designer of the Spiyda Kits, If anyone has an issue with a tachometer after fitting one of our kits,
just send it to us with a description of the issue and we will sort it for you.

All we ask is that you have made a good attempt to fit it, and you cover the postage costs.

It isn't always necessary to send it to us and many of the possible issues are covered in the online FAQ.
https://spiyda.com/faq

One of the issues mentioned is only working up to a certain rpm.
This is most likely due to the input signal getting weaker as the rpms rise.
It can be fixed by moving the input to a more sensitive solder pad on the board (one nearer the chip)

In extreme cases ( where the tachometer is being driven by an alternator for example.. and yes it will do this !) there are instructions in the kit on how to make the tachometer more responsive to high frequencies.

The output of the board is very linear but sometimes this does not match the response of the movement,
Non linearity can be fixed by

a) adjusting the bottom stop tension ( the little fibre plate behind the dial) which can move the whole scale downward, effectively reducing tick-over rpm

b) adjusting the needle counterweight ( the big brass disc that is friction mounted to the movement) moving it outward makes the first half of the movement easier and the second half harder and vice versa.

c) adjusting the small spring like balance weights ( mounted at 90 degrees to the large counterweight) These can adjust the mid position a little.

with a standard coil and a movement in good order, none of these adjustments should be necessary, but in practice, different methods of triggering a spark or indicating a spark event and the many different coil profiles mean that some adjustment is sometimes required. So far, in the many thousands of kits made, of the tachometers that have been sent to us, all have eventually performed well.

To give you a better idea of why the input sensitivity is critical, below is an image of what you would see if you connected an oscilloscope to the input to the tachometer, The input sensitivity has to be such that it picks up a wide enough part of the wave to trigger, but be insensitive to the "bounces" after the main spark event, these bouces are called "ringing" of the coil. If the tacho is too sensitive, it will read too high, too insensitive and it will not work at higher rpms.

Chris

[mbell]

I just wanted to follow up on this thread, firstly to tank Chris for jumping in to respond on this and update it for the record.

I finally got time to revisit my tacho yesterday (Easter Monday is possibly best day of the year with wife at work and Kids at school!). I pulled the tacho, opened it and soldered (very badly) external connections for all four inputs on the board and re adjusted the needle position which was slightly high.

I refitted it in the car, this time connecting the "12V connection" to the "TTL" output of my digital rev limiter. Fired the engine up and got sensible behavior straight away! I then used the adjustable rev limiter to test accuracy which seem to be ok, maybe reading ~300 over fairly consistently and probably improvable with some adjustment to the calibration.

I'd previously tried the rev limiter on the 2V-5V/calibration input and a direct wire to the High Energy coil connection. So my issue was really "user error", as 12V connection works and likely normal coil connection would work.

One thing I would comment on is that there is a lot of information around on how to do the conversion but it is a little sporadic in location/presentation and I feel lacking information/advice on the inputs. I think a little time restructuring the doc and providing more information in places would make life easier for users and lower the support burden on the product.

[spiyda]

One thing I would comment on is that there is a lot of information around on how to do the conversion but it is a little sporadic in location/presentation and I feel lacking information/advice on the inputs. I think a little time restructuring the doc and providing more information in places would make life easier for users and lower the support burden on the product.

You are absolutely correct, the documentation and videos need tidying up and pulling together,

perhaps like the wizard page I did a few weeks ago http://www.spiyda.com/Link/WIZARD.html

Chris