Building a revolutioncounten should be unnecessairy.
Because I think every manufacturer of cars, bikes and mopeds should mount a revolutioncounter as standard.
But for a moped-driving electronics hobbyist should there now is a way to change this.

Setup
The constuction of the electronical revolutioncounter can be done in different ways.
The main difference is in the readout of the instrument.
Rougly there are three ways to show the rpm: with figures using a few sevensegment-displays, with a analoge scale consisting of a LED-bar, or with the old-fashion turningcoilinstrument with a indicator.
The last one of the three is the easiest to realise, but is also a bit vulnerable and shocksensitive.
So for a moped a readout like that isn't very suitable.
A indicator with sevensegment-displays is a real thing for pricisionfreaks, but for this application a digital readout isn't handy.
Because accuracy as high as that isn't necessary and it would also make the circuitboard way more complex.
A LED-readout has the adventage that it's durable and easy to build.
We can use one of the famous steer IC's for this revolutioncounter which can, without using a lot of external compontents, use an anologe tension to display a LED-bar.
If we would take about 20 LED's for this, the accuracy of the indicator should be more than adequete.
The only thing we need now, is a sensor which delivers pulses which correspond to the rpm, and some electronics which converts the number of pulses into a proportional direct current for the LED-bar.
Als you will see, that part is a peice of cake.

From RPM to DC
For starters we have to have a suitable recorder wich produces a puls every engine-revolution.
This can be done in different ways, but the easiest by far is using a coil to pick up the ignition pulses inductive.
Because the ignitionpulses are pretty strong, it's sufficient to wind an ordinary wire 10 or 20 times around the sparkplug cable.
The ignitionpulses are stong, but they differ a lot in the shape of their tensionpeaks.
To get reliable countingoulses out of them, the recordercoil has to be followed by a suitable pulsformer which changes the signal into a series of neat sigle-shaped pulses.
Only this way you can make certuin that acidental variations in the width or amplitude don't influence the countingresult.
Take a look at figure 1, in which the entire circuit board of the revolution-counter is displayed.
The pickupcoil has to be connected to C3.
This condensator functiones in combination with R3/R4 as a differentiator, thus cutting down the ignition pulses to usable triggerpulses - this is done to prevent accidental double triggering of the revolutioncounter.
These modified pulses are then fead to the trigger entrance of on of the as monostable mulivibrator linked 555 (IC3).
This IC then creates pulses with a fixed (with P1 adjustable) width
The only thing we have to do now to get a revolution-dependant-direct curent, is itegrating the pulssignal given by IC3.
Therefor we use a simple lowpassfilter, als done here with R6 and C1.
This filter also makes sure that short vibrations in the pulssignal are "flattend out", because they would make the readout unnecessary unstable.
LED-Readout

To realize the readout we, just as for the pulsformer, used and old acquaintance, namely the LM3914 display driver.
This IC was specialy designed for this purpose and contains a referance-tensionsource and am accurate ten-times tensiondealer.
From the connectionpoints of the tensiondealer ten comparators are driven and in such a way that a next comparator becomes active when the entrancetension of the IC increases.
The comparatorexits are capable of driving a LED directly.
The LED-bar formed this way can be set to "dot"- or "bar"-mode; in the first case pen 9 has to be left "open" and in the second case it has to be connected to the tensionpuls.
Here we have chosen for the second possibility.
A nother nice caraceristic of the LM3914 is that it's very easy to link two together is a LED-bar with a higher resolution is wanted.
Because the maximum of ten LED's is a bit low in our case, we used this possibility gratefully.
As the shceme shows, this resulted in a 20 LED-display, in which, depending on the setup, every LED represents about 500 rpm.
You can use different color LED's and you can make for instance a "safe" (green) area of 500...6000 rpm (D1...D12), an "attention" (yellow) are to 8000 rpm (D13...D16) and a "prohibiten" (red) zone between 8500 and 10.000 rpm (D17...D20).
But ofcourse you can also make your own scale which is completely different.

Practical things
To make the building realatively easy for amateurs, we designed a compact circuit board for the revolutioncounter, which is shown in Figure 2.
When we held a little servey here, we found out that everybody prefered a round scale, and that is why we placed the 20 LED's in a circular shape on the cicuit board.
Since the revolutioncounter has very little parts, the construction of the circuitboard doesn't need a lot of explanation.
Just neatly solder all the compononts in the parts list, and voila!
Jumpercontact JP1 was added so you are able to check the counter after the construction.
If pulses are added to C3, a tiny direct current should be measurable at the first contact of JP1 which can be altered using P1.
If this works, the chances that the pulsformer works are very good.
If you connect an adjustable directcurrent-powersupply of a few volts on the other side of JP1, the LED-bar should come to life; this way you can check the readout part.
Normaly a jumper will be put on JP1.
We've allready talked about the pickupcoil.
It could be possible that this part needs some experimenting, but normaly the inductive pick up of ignition pulses from the sparkplugcable shouldn't be any trouble.
The connection between the coil and the entrancepen of the circuitboard kan be made with a normal piece of wire.
Another thing about the round shape of the circuit board.
Besides our personal preference, there was another reson to choose this form.
Browsing through various catalogs we came uppon a nice housing, which has a shape that fits nicely in the instrumentpanel of most mopeds and motocycles.
That's why we adjusted the dimensions of our circuit board to it.
In Figure 3 you can see this so called "cockpit"-housing, together with a test model op the build circuit board.

Power
We haven't talked about the way to power the revolutioncounter..
Als shown in the scheme, it's designed for a 5 V feed (6V is no problem).
But the power has to be reasonably "clean" and stable, so you shouldn't connect it directly to the moped unless it has a battery.
If you do want to use the available power (which is very handy ofcourse), you will have to place a 5V stabelizer between the moped's power and the revolution counter.
The nessecary extra circuit board for this is can be seen in Figure 1 (it's the one with the dotted line around it.
Because the available power (6 a 7 V) is about 1V above the prefered 5V, you should in this case use a low-drop-power-regulator like the 4805; so a normal 7805 can not be used!
If you want it to be completely self-supporting (so not connected to the moped), you can use a battery.
Four rechargeable penlights are perfect as a powersource and will endure a pretty long time.
A stabelizercircuit isn't necessary then.
And if you want the battery's to last even longer, you should set both LM3814's in to "dot"-mode.
This can be done by not soldering pen 9 of IC1 and IC2.