Electric Tacho problem

[RogerH]

Hi Folks,

          About four or five years ago I had my TR4A mechanical tacho updated to run off the coil/sparks.

I have always had the feeling that although it reads fairly realistically up to 3000 rpm it stops displaying anything over apprx 3000 rpm. it just sits there at 3000.

I could send it back to the chap that converted it but before I do that is there anything I should be looking at between the coil and the Tacho.

Is it sensitive to the signal amplitude or shape from the coil.?

Any help would be appreciated.

 

Roger

 

[JohnD]

I have a device called a "Picoscope", which allows any laptop to be used as an oscilloscope.     A 2-channel version costs about £100.

It would be ideal for inspecting the signal that your coil is sending to the tacho.    Any dwell problem (for instance) would be immediately obvious!

See: https://www.picotech.com/oscilloscope/2000/picoscope-2000-overview

JOhn

[Nick Jones]

I had similar issues when fiddling about trying to get Smiths RVC tachos to work with Megasquirt and EDIS.  Basically the device is not quite sensitive enough for the signal and as the frequency goes up and the duration of the peaks goes down, it gives up.  This wasn't unexpected for what I was doing as I was trying to trigger it from a 12v square wave rather than the flyback voltage from the coil which can peak at 40v+. 

However, assuming you are connecting to the coil -ve terminal  (I'm assuming it's not one of the series types like the Smiths RVI) you ought to have plenty of signal.  Has it ever worked about 3k?  If not you need to contact whoever did the conversion and check it is connected as he would expect and query why it does this....

[richy_rich]

Sounds daft, but my (homemade conversion) used to do this and it was simply because the needle was catching on the face or somewhere.  The stepper motors in more modern tachos (and the deflection coil thingies in older ones) aren't very powerful and the smallest resistance is enough to stop them moving.

I rebuilt mine recently and changed the face back to the older Spitfire type which was originally fitted to my car ( https://richyrichracing.com/2019/01/replacing-the-dial-face-of-a-modern-tachometer-with-the-original/ ), but it might be worth carefully taking the glass off and checking, you should be able to *carefully* move the needle around using a small blunt tool (or a finger..) see if it's catching anywhere, if it doesn't spring back automatically just move it back to where it started (at zero hopefully).

Alternatively, you can test it with a cheap signal generator - something that puts out a 12v squarewave in the region of 50Hz to 500Hz should do it (would equate to about 750rpm to 7.5krpm)

As Nick says, if it's connected to the coil (as most are) it should get plenty of signal and the fact that it works at low rpm would indicate that it is.

 

[RogerH]

Hi Nick,

 it is happy to get to 3000 but not over it.

I was hopping that somebody would say twiddle this or twiddle that or fit a capacitor. But alas no.

I shall give the man a call.

 

Roger

[SDerbyshire]

If it was JDO then he’s very helpful, and mine flies around to 6000 easily :-)

steve

[RogerH]

Hi Steve,

            it was/is JDO  I shall call John tomorrow.

 

Roger

[oldtuckunder]

10 minutes ago, RogerH said:

I was hopping that somebody would say twiddle this or twiddle that or fit a capacitor. But alas no

my innovate data logger had this problem when taking signal from coil, when configuredfor 6 cyl, lots of suggestions about diodes etc, but never bothered playing, used induction clamp on single lead and worked fine. finally sorted when i went EDIS/MEGAJOLT  and it then got a good lean signal. 

6 minutes ago, SDerbyshire said:

and mine flies around to 6000 easily

if not acurately! i'll bet if you check, converted gauges  are usually either accurate at tickover end or at top end but not both ! if pushed JDO will ask where the important area is for you, surprisingly most people want an accurate tick over, whereas i always regarded that as as irrelevance on the way to somewhere interesting

alan 

[SDerbyshire]

37 minutes ago, oldtuckunder said:

If not acurately! i'll bet if you check, converted gauges  are usually either accurate at tickover end or at top end but not both ! if pushed JDO will ask where the important area is for you, surprisingly most people want an accurate tick over, whereas i always regarded that as as irrelevance on the way to somewhere interesting

alan 

You are right Alan, accuracy above 5000 revs important for me not idle :-)

interestingly the tacho agrees pretty much with the app for the 123tune+ ignition system.

which is more than can be said for the 123’s ‘rev limiter’, which cuts in way too early, like 1000 revs below the set value.

steve

[RedRooster]

1 hour ago, oldtuckunder said:

my innovate data logger had this problem when taking signal from coil, when configuredfor 6 cyl, lots of suggestions about diodes etc, but never bothered playing, used induction clamp on single lead and worked fine. finally sorted when i went EDIS/MEGAJOLT  and it then got a good lean signal. 

 

My MegaJolt used to give a smooth rev counter, now its all over the place which is distracting.

Also the speedo is doing the same but thats the cable or gearbox cog, so overall the dash is a bit crazy with needles going in different directions.

[oldtuckunder]

there are two maybe three possible outputs with mj/edis, i'll take a look at my wiring diagram and see which i found best.

 

alan

[Nick Jones]

Only one from the EDIS-6 IIRC + one from the MJL. The EDIS one isn’t enough to drive a Smiths RVC tachometer without additional amplification in my experience (though I have seen claims of success). No personal experience with MJL.  Erratic behaviour where it previously worked suggests dodgy connection(s) to me....

[RogerH]

Have contacted JDO - not much help.

I shall get my oscilloscope on the job and see what pulses are coming out.

 

Roger

[RogerH]

Just a thought.  Could it be the Lumenition magnatronic module (replacement circuit breaker) damping the pulses etc.

 

Roger

[Nick Jones]

Assuming that the tacho is triggered by a single signal wire connected to the coil negative, the way that the coil is switched (points or whatever) shouldn't matter.  To my mind the problem lies with the tacho....

[RogerH]

Thanks Nick.

 

Roger

[richy_rich]

I'd agree that it sounds like the tacho.  Do you have (or could you borrow for a morning) another old-ish car with a coil?  Perhaps you could try temporarily wiring the tacho up to that and see if it exhibits the same problem - it would at least rule out the signal from the coil.

[SDerbyshire]

If i recall, the tacho needs 12v, signal and 0v, ie earth

perhaps check all are as they should be first

ie, the 12v feed is actually 12v

0v/earth is a good earth

signal connections solid.

so many electrical problems turn out to be ‘dodgy earth’ that it’s always worth checking.

i’d probably try a temporary install with 12/0v from the battery and a short trigger wire to the tacho just to check !

 

steve

[RogerH]

Hi Rich,

that is a good plan that I will try to sort out.

I have found an excellent site that gives all sorts of good info.

Sadly the RVC is so simple that there are very few trouble shooting areas. And non that cover my query.

On another forum a poster suggests that the tacho conversion is suspect.

 

Now if it originally worked on the bench then perhaps an interface board that received the noisy coil signal and transmits a clean square wave may help. But then .........

 

 

Hi Steve,

I'm sure the wires are in the right places as it works - but !!

I may well knock up a signal generator and squirt some clean signals into the tacho and see where they stop. 

Roger

[richy_rich]

Honestly, I reckon it's sticking.  Might have worked on the bench and since a bit of a knock has moved something..  

You should be able use something like this to test it on the bench I'd think https://www.amazon.co.uk/MakerHawk-Frequency-1Hz-150kHz-Adujustable-Rectangular/dp/B07CQG2HSX/ref=sr_1_46?crid=1RZE6CHBOPZDE&keywords=signal+generator&qid=1556210863&s=gateway&sprefix=drive+away+%2Caps%2C248&sr=8-46

As I said, 50Hz would be 750rpm (for a 4cyl) then wind it up from there

[DeTRacted]

In answer to your original query Roger - the tacho should be insensitive to amplitude changes provided the signal is always larger than the sensitivity threshold. Basically the top of the waveform along with any noise should be chopped off by the electronics in the meter and ignored. All it should be doing is responding to the leading edge of each ignition pulse. If for some reason the signal becomes lower than the trigger threshold as the speed increases, the meter should just fail to register and return to zero. As Rich says, if it just sticks at some point, that is more likely to be mechanical than electrical.

Do you have any details of the conversion that was done? Is it a moving-coil movement with a pulse-repetition converter?  If it is, each ignition spike triggers a pulse of fixed width and amplitude. The mark-space ratio of the wave-train is averaged by the meter i.e. the higher the revs, the closer together the 'mark' pulses are and the higher the average voltage which the meter responds to.

 Rob

[spitfire6]

Hi,

 Purchase a one quid squarewave generator from eBay. It will give you an adjustable frequency to check your meter?

Cheers,

Iain.

PS

Buy this for frequency display as well as output. eBay item number:283405071477

Measure tach input frequency with a cheap DVM on Hz?

For a 4 Cylinder;
Frq=RPM x (#Cylinders/2)/60
so, Frq=3000*(4/2)/60. Therefore Frequency = 100Hz. 6K RPM would be 200Hz.

Edited April 26, 2019 by spitfire6
Added