Protection Zeners
 

The 5.1v zeners across the injector and vss signals are doing nothing, assume they are to limit the input voltage to 5.1V. To work I believe they require at least 5mA before they will dump excess voltage across them.

The current circuit shows 100k resitors in series, so to get 5ma we would need a voltage of v=0.005x100000=500v (worst case in reality of a car is about 100v, but more likely 60-80v)!

Assume that the supply is 12v and the zener is a 5.1v.

We need to drop 6.9V (12-5.1) across the resistor @5mA (+a few uA for the load)

R=6.9/0.0051=1352, but recommend 1K or 1.2K (to be on safe side)

They are doing something, zeners knee to the low side with less current, and I've been using 50k for a long time.

Put it on a scope or something before pulling the fire alarm please.

Sorry, I was not trying to "pull the fire alarm", I was just trying to show that they are not offering the protection that is implied.

Quote dcb "They are doing something"... What are they doing?

Quote dcb "and I've been using 50k for a long time"..Why, when the schematic on wiki shows 100K?

Just to be clear, I have not come on here to bash the project. I think its great, and have built a prototype, only to find I have the dreaded Peak and Hold injector. I am trying to work on a solution, probably involve voltage comparator and diode to hold refernce 0.6-0.7 below supply. This would still give two pulses per injector cycle, but could use the injector delay value to ignore the induced pulse at the peak to hold transition.

Hey, (re TBI) Can you try the circuit mentioned in this thread? It should just need a tweak to injhiUS to compensate for the optoisolator lag

http://ecomodder.com/forum/showthrea...tml#post106855

re: Zener, been there, tested it personally with an oscilloscope, no problem, no cause for alarming thread titles, not that a $3 chip is anything to get too alarmed about :)

http://ecomodder.com/forum/showthrea...html#post73500

The theory was all very well, but here is the real world. The two attachments show that no dangerous spikes are getting through, but the 100k is clipping at 3.8V! The 1k is clipping at 5.04V, which is good enough to backup the theory.

So the fact is that 100k/Zener is protecting against the high spikes, but it appears to be clipping at a lower voltage.

More interestingly is the initial turn on voltage to the injector is negative, and the hold portion is at 0V! I used the negative terminal on the battery as the refernce.

you might be using different Zeners? A bit more current might be in order there, huh?

The only component changed between the 2 traces was the resistor, same zener.

Isn't that where we started?

@DCB could you please change thread title to just "Protection Zeners", as they are clearly doing something. ;)

Here is a bench shot (are you switching 12v on the bench?) and another shot from the car, it is dropping from ~3 volts to 0 at the cpu pin when the injector fires. Using 50k resistors, ironically for protecting the CPU :)

on the bench, using another atmega as a signal generator and a couple transistors to toggle 12v
http://ecomodder.com/forum/attachmen...1&d=1260125024

On the front seat of the saturn
http://ecomodder.com/forum/attachmen...1&d=1260125024

On the front seat of the saturn, giving it a little gas (wider pulse)
http://ecomodder.com/forum/attachmen...1&d=1260125024

All my tests are on the car (Rover 214i (single injection) 95). I will try and get some more data today with the scope, DVM, but with a simple change to the code I am getting some more sensible figures for mpg.

I figure I need my timing to be started on a pulse less than 40uS, and stopped on a pulse longer than 400uS. But I need to see if the shutoff decay is constant at various rpm and loads, not easy by yourself!

Oh yah, you have the peak and hold, I don't have one of those handy, can you try the circuit here:
http://ecomodder.com/forum/showthrea...tml#post106855

I assume the injector is negative switched?

can you scope the injector tap directly so we know what we are dealing with?