LM2917 Tachometer for V12

[alexb1406]

I see what you mean - and also, with diode no3 connected between input line and ground, the oscillations should not drop below -0.7v, but they go to almost -2v.

Let me check it again and come back to you.

 

[alexb1406]

Scope image taken again as the original ones were wrong - not sure why but I suspect something wasn't connected properly. At least this is more credible. As before blue is before diodes and red is after.

 

[eTech]

Scope image taken again as the original ones were wrong - not sure why but I suspect something wasn't connected properly. At least this is more credible. As before blue is before diodes and red is after.

Yes. That's looks better.
I think what may be happening is the tach input signal (before the diodes) at high RPM is falling below the "on" threshold of the input transistor, which is about +0.65 volts. Each diode drops the voltage about 0.6v, so with two diode that's 1.2v plus 0.65 volts to turn on the transistor. This equals about 1.85v total. So the input signal needs to stay above this minimum level at all RPM's. The LM2917 is a little better in that the threshold is much lower at about 250mV.

I've attached a revised schematic using the LM2917. Please note for R1 I showed a 39.2k resistor but its value can be 39k.

 

[alexb1406]

Yes I see that - but look at the attached scope traces on post #34. If you look at the red trace which is the BC547 collector output you can see that without the diodes in place, the transistor switches on much more slowly than when the diodes are in place. This has the effect of not allowing the input signal (blue trace) to pin 2 of the 555 to drop low enough, (because the edge trigger starts to raise the signal before it has been able to drop low enough to trigger the monostable). When the diodes are added, the transistor switches off sharply before the edge trigger has time to do its thing, and consequently the signal drops low enough to trigger the monostable. This is the bit I don't understand, i.e. why does adding the diodes cause the transistor to switch off more sharply? Is it because of the 1.2v lower input to the 1K resistor and its base? And, why does increasing the engine RPM gradually make it switch off more sharply, so that by the time it gets to around 2500RPM, it is triggering the monostable?

And thanks for your schematic. I will try that at some point and see how it works. I think I'll stick with the 555 for this rev counter, but I do have another that needs the same so I'll have a play with that one when I get some time.

 

[eTech]

Yes I see that - but look at the attached scope traces on post #34. If you look at the red trace which is the BC547 collector output you can see that without the diodes in place, the transistor switches on much more slowly than when the diodes are in place. This has the effect of not allowing the input signal (blue trace) to pin 2 of the 555 to drop low enough, (because the edge trigger starts to raise the signal before it has been able to drop low enough to trigger the monostable). When the diodes are added, the transistor switches off sharply before the edge trigger has time to do its thing, and consequently the signal drops low enough to trigger the monostable. This is the bit I don't understand, i.e. why does adding the diodes cause the transistor to switch off more sharply? Is it because of the 1.2v lower input to the 1K resistor and its base? And, why does increasing the engine RPM gradually make it switch off more sharply, so that by the time it gets to around 2500RPM, it is triggering the monostable?

And thanks for your schematic. I will try that at some point and see how it works. I think I'll stick with the 555 for this rev counter, but I do have another that needs the same so I'll have a play with that one when I get some time.

Hi

Can you get an image of both waveforms at high rpm?
It might help answer those questions.

eT

 

[alexb1406]

Ok I'll do that when I can

 

[eTech]

HI

So I went back to the bench to test a 555 version.
I made a few changes...I changed the base resistor from 47k to 22k to ensure the transistor switches to a low enough voltage to trigger the 555. I also didn't have a BC547 so used a 2N2222A. I also removed the input diodes and ac coupled the input and ac coupled the output of the transistor to the 555. This actually worked pretty well. Tested the circuit by generating a square wave input signal of +0.8, -0.2 from 50-700 Hz. The tested circuit is attached.

eT

 

[eTech]

Yes I see that - but look at the attached scope traces on post #34. If you look at the red trace which is the BC547 collector output you can see that without the diodes in place, the transistor switches on much more slowly than when the diodes are in place. This has the effect of not allowing the input signal (blue trace) to pin 2 of the 555 to drop low enough, (because the edge trigger starts to raise the signal before it has been able to drop low enough to trigger the monostable).

After thinking about this, you might try lowering the value of the transistors' base pull-down resistor from 47k to maybe 22k

When the diodes are added, the transistor switches off sharply before the edge trigger has time to do its thing, and consequently the signal drops low enough to trigger the monostable. This is the bit I don't understand, i.e. why does adding the diodes cause the transistor to switch off more sharply? Is it because of the 1.2v lower input to the 1K resistor and its base? And, why does increasing the engine RPM gradually make it switch off more sharply, so that by the time it gets to around 2500RPM, it is triggering the monostable?

The diodes might be changing the transistors bias, try removing the diodes and install a 0.1uF cap.

eT

 

[alexb1406]

After thinking about this, you might try lowering the value of the transistors' base pull-down resistor from 47k to maybe 22k

Ok I'll try this when I get back to the car, but...won't this have the opposite effect? The problem is the transistor is slow switching on and going into saturation, i.e. the base voltage is slow rising.

But, I'll try it. And many thanks for taking the time to do this for me.

I took some scope images of what is happening yesterday. RPM is approximate, especially when the tacho was not working and I had to guess the RPM. Blue is transistor collector and red is edge trigger output/pin 2 of 555. It is clear that without diodes the transistor switches on much slower at low RPM and the edge trigger does not allow the 555 input to go low enough to trigger the monostable.

 

[eTech]

Hi
When you get a chance can you post a photo of the top and bottom of the original tach circuit board? I'd like to look at the components connecting the chip. Especially the tach input circuit.

 

[eTech]

Ok I'll try this when I get back to the car, but...won't this have the opposite effect? The problem is the transistor is slow switching on and going into saturation, i.e. the base voltage is slow rising.

Yes...place the cap at the front and the diode next, remove the second diode.
So the signal input path should look like this: Signal input -> cap -> diode -> 1k res -> transistor (base).

BTW- your schematic shows the transistor emitter disconnected. I hope its really connected to ground.

eT

 

[alexb1406]

BTW- your schematic shows the transistor emitter disconnected. I hope its really connected to ground.

Oops! Yes it is definitely connected to ground - I can't see it working if it wasn't. I've amended the schematic.

Hi
When you get a chance can you post a photo of the top and bottom of the original tach circuit board? I'd like to look at the components connecting the chip. Especially the tach input circuit.

I can do better than that - I've laid down a schematic on paper. I did that first when I was trying to get a useable signal for the 555 and the LM2917 but wasn't successful. Let me see if I can find it later today and come back with it.

 

[alexb1406]

Photos of board and schematic.

Green = +12v
White/Blue = Tach signal
Black = Meter -
Red = Meter +

I know there is a broken track on the board but I'm sure that must have happened later as I would have spotted it. It has been lying aroung and knocked about since being removed.

 

[alexb1406]

I tried what you suggested in post #47. Firstly I have to admit to another mistake - the base resistor was not 47K but 4.7K. I realised this when I went back to change it to a 22K as you said. I also removed the diodes from the input line and inserted a 0.1μf cap instead. What I found now was that under 1500 RPM the needle was very erratic, and I took some scope pics to show the problem. The first image is the tach signal that comes from the ignition system. This was taken with a 20:1 attenuator so the initial spike is around 160V. I took this to show you the kind of signal we have to deal with. When you look at the next two at idle RPM you can see that some of the coil oscillations are also causing the transistor to switch on, and we get additional triggers of the monostable when we don't want them, hence the erratic needle. At higher RPM (3000 in this case) this does not seem to happen - the transistor is switching on and off cleanly.

 

[eTech]

I tried what you suggested in post #47. Firstly I have to admit to another mistake - the base resistor was not 47K but 4.7K. I realised this when I went back to change it to a 22K as you said. I also removed the diodes from the input line and inserted a 0.1μf cap instead. What I found now was that under 1500 RPM the needle was very erratic, and I took some scope pics to show the problem. The first image is the tach signal that comes from the ignition system. This was taken with a 20:1 attenuator so the initial spike is around 160V. I took this to show you the kind of signal we have to deal with. When you look at the next two at idle RPM you can see that some of the coil oscillations are also causing the transistor to switch on, and we get additional triggers of the monostable when we don't want them, hence the erratic needle. At higher RPM (3000 in this case) this does not seem to happen - the transistor is switching on and off cleanly.

Hi
A couple of questions.

1. The input needs at least one diode. Did you test as I described in post 51?
2. Where is the source of the tach signal connected?
Is it connected to the (-) side of the ignition coil?
Or is it connected to the ballast block? if so, how many terminals does the ballast block have, and to which terminal is it connected?
Or is it connected to the ignition amplifier? Which terminal?

eT

 

[alexb1406]

Hi,

1. I tried it both with and without a diode on the input...but having just read your post again, I can't remember if I put the cap and diode in the right order. I'll have to go back and try it again. Thanks for pointing that out.
2. The car has electronic ignition, currently running on the later Lucas AB14 amplifier, and the tach signal comes from the amplifier. At some point I plan to go back to the original Lucas OPUS system if I can find one that actually works!?! But my signal looks like it is coming from the coil (-). When I go back to the OPUS system, that does have a ballast block, one of which is for the tach.

 

[alexb1406]

1. The input needs at least one diode. Did you test as I described in post 51?

Actually I just realised why I got that wrong. Your schematic in post 47 doesn't show a diode - I was following that when I tested it. But it's no problem, I'll go back to it later this week.

 

[alexb1406]

Hi eT,

I went back and tried it again, this time making sure I got it in the right order, but I could only get it to work erratically under 1500 RPM. I tried it both with and without the diode and 0.022μ cap to ground between the signal diode and 1K resistor (as you don't show these in your schematic) but none of the combinations worked. I think the problem is that a signal generator on the bench cannot replicate the exact signal from the coil/ignition amplifier, so while it works on the bench, making it work on the car is much more tricky.

Looks like I will have to go with the two diodes in the signal line. I've already designed the pcb to be the exact same size and shape as the original - I will make it when I get some time.

And next I will try out your current controlled circuit attached to post #43 - it may not be for a while though as I have other jobs to do. If that works I will make if for the other tach I have lying around - in fact I think I have two more...and a Rover V8 tach all requiring repair. I'll be an expert by the time I've finished.

But thanks for your input, and of course everyone else in the early part of this thread. I think I've taken a little bit from everyone and eventually got something that works.

 

[eTech]

hi

Yes...it would have helped to have a Jaguar V12 to experiment with..
I'm not clear though on how your schematic is working. Looks like the tach signal peaks at 160 volts...are there any zener diodes being used?

I'm still curious about where the tach signal source is connected. In post #56:

2. The car has electronic ignition, currently running on the later Lucas AB14 amplifier, and the tach signal comes from the amplifier. At some point I plan to go back to the original Lucas OPUS system if I can find one that actually works!?! But my signal looks like it is coming from the coil (-). When I go back to the OPUS system, that does have a ballast block, one of which is for the tach.

Usually there is a resistor between the coil(-) and the coil driver.
Is the tach signal connected to the (-) coil or is it connected to the amplifier? Also, It sounds like there is no ballast block?

Anyway...hope your circuit works out ok.

eT

 

[alexb1406]

I'm not clear though on how your schematic is working. Looks like the tach signal peaks at 160 volts...are there any zener diodes being used?

I'm still curious about where the tach signal source is connected.

I got the 160v peak with the signal wire not connected to anything - Once connected that dropped considerably, I can't remember exactly what, but I think it varied depending on what I did with it. The wire itself on the current system comes from the ignition amplifier, so my guess is there must be some resistor in line with the signal before it leaves the amp. There is no ballast block with this system.

However, the original Lucas OPUS system that I intend to go back to does have a ballast block with several resistors. The signal does come from the amplifier but there is also an internal connection in the ballast block between the ignition positive and the tach signal. Also, the diagram shows the tach signal being a straight through connection through the ballast block, but the accompanying text says there should be a 100Ω resistance, so I don't know. There were several versions of the ballast block and ignition system and some of the later ones had an external resistor on its own and in line for the tach signal.