While waiting for a replacement T2295a chip from China, I decided to review the design and come up with an alternative should the Toshiba chips finally run out (which seems possible).
In doing so I have been playing with tacho ECU feed and have noticed that it is very susceptible to interference. Basically any load on the tacho line from the ECU will cause a mis-fire.
There is also a lot of ignition spiking on the tacho feed on top of the normal square wave. These high voltage spikes can damage electronics and obviously the original design deals with this fairly well although it could be a result of my failure.
Initially I tried just changing the capacitors on the board as they are normally the first to fail when old but with new ones in the failure pointed to the main chip being dead, hence the Ebay purchase.
So looking at a new design that doesn't rely on the old chips, I decided to break it into 2 parts;
1. An opto isolator circuit to protect against the ECU spikes.
2. A frequency to voltage converter to provide the 0-4.5v DC needed for the RPM meter.
I selected a simple F to V chip in the LM2917N. An older devise but fairly easy to get hold of. It also has an inbuilt voltage regulator circuit to handle the bikes varying voltages.
The opto isolator is a TLP2962. One minor disadvantage of isolators is that most run at low voltage (5v) so need a separate regulator which I have added. At the end of the day it all aids stability.
So my design so far is;
I've done some basic testing and it seems OK. I am just waiting for some new LM2917 as I blew a couple up testing the ECU spiking issue
If it all works, the PCB will look like;
It fits in the original space and when I unsolder the 3 brass lugs that connect power and tacho feed from the original PCB, I can re-use them.
I will still play with the replacement T2995 when I get them but the new design is probably a better long term solution.