Am I missing something?

abelk · Post by **abelk** » Wed Mar 15, 2006 12:11 pm

Thanks, Stephen. We will have a look. I have a 0.1uF cap between 5v and ground. Do you think I need another cap on the input (~12v and ground)?

The input impedance of the ignitor/coil pack is about 1.1kOhm (DC). So the circuit only needs to source less than 5mA with the coil connected.

Another possiblity is that when they were testing, perhaps some test leads were too close to the spark plugs/high tension wires. I think that it is probably either something with the power supply or some sort of feedback as this behavior only occurs when the coil pack is connected.

Kent

rshayes · Post by **rshayes** » Wed Mar 15, 2006 2:06 pm

If the input voltages drops below about 8 volts, even momentarily, the output voltage will follow it down. This can also blow a linear regulator if it hass a large capacitance on its output terminals and the input supply is capable of sinking current. I think that the National Semiconductor data sheet on the LM317 discusses this, and it should be covered on some of their other data sheets. It tends to be in the notes of the application section near the end of the data sheet.

A diode to a large capacitor on the input side of the regulator will allow you to bridge short dropouts in the input power. This also gives some reverse polarity protection, but it does add an additional voltage drop.

abelk · Post by **abelk** » Wed Mar 15, 2006 10:15 pm

Thanks, Stephen. That makes sense. The supply voltage shouldn't drop that low as it is supplied from the car battery (~14v running, 12.8v engine off). I will take a look at that too to be sure. Any ideas are welcomed and appreciated. I will let you know what we find out.

Kent

dyarker · Post by **dyarker** » Wed Mar 15, 2006 10:43 pm

During cranking, especially on first couple turns of starter motor, voltage can go that low; lower in cold weather.

Cheers,

positronicle · Post by **positronicle** » Thu Mar 16, 2006 1:24 am

--Edited by Positronicle--

abelk · Post by **abelk** » Thu Mar 16, 2006 2:17 am

Thanks. Which pic are you referring to (one that I posted or one in the link)? There are 2 problems (probably related):

- under certain frequencies, there is a second waveform on top of the input/output signals when the signals are in the high state.

- some conditions with the coil connected causes a retrigger of the circuit.

Normally, the input signal comes from the stock ignitor. Basically an open collector. The voltage stays at battery voltage until the ignitor pull the signal to ground. Now this ground point is on the engine, it may be a bit different than the ground on the chassis (where the circuit/coil gets ground). Maybe this difference is causing a slight negative voltage on the input.

The traces shown were driven by a function generator. All the grounding points should have been at that same potential, but maybe they weren't. The strange behavior only occurs with the coil connected. Maybe we will try to replace the coil pack with an equivalent resistor to see if it is the load that is creating the problem of if it is something else.

Thanks again.

Kent

MrAl · Post by **MrAl** » Thu Mar 16, 2006 2:19 am

Hello there,

I dont know if this has been said already, but,
you really do need power supply bypass caps
right on the board, possibly right next
to the ic's themselves. This is even more
important when the power supply leads going from
the power source (batt) to the board are long.
They could also pick up RF noise (looks like some
in that one scope pic riding on the pulse).

To understand the noise a little better try
looking right across the boards power supply
input lines with the scope, possibly set on
low ac range. See if you can find a correlation
between the noise on the power supply line and
the noise riding on the pulse.
Remember to put the scope lead (tip AND ground)
right across the ps input right on the board
itself, not on the battery.

Take care,
Al

abelk · Post by **abelk** » Thu Mar 16, 2006 2:39 am

Thanks, Al. I have a single 0.1uF cap going between 5v (next to the v-reg) and ground. Do you think I need more than this? The board is quite small (~2"x2"). RF is something I have been thinking. Basically the difference between the coil connected and not is:

- increased load on the circuit (5mA) from triggering the coil pack
- presence of RF due to the dischaging of coil/spark plugs.

Perhaps the addition of a cap between supply voltage and ground would help. I think the data sheet recommends 0.33 uF.

The testing is a bit difficult as I don't have the equipment here. The guys that are testing are in Australia. I tell them what to try and they then report back. I will have them take a look at possible RF on the input.

To give an idea of frequency of the waveform, the main signal is 100-120 Hz in the pics. I believe the scope is set to 2ms/div.

I will let you all know what I find out. Thanks for all the helpful input.

Kent

positronicle · Post by **positronicle** » Thu Mar 16, 2006 4:26 pm

--Edited by Positronicle--

abelk · Post by **abelk** » Fri May 26, 2006 4:29 pm

Hi all,

Following up on this. I finally got an o-scope here to mess around with this circuit. I added a capacitor to the input on the 7805 and that cleaned up the noise on the waveform a lot. I am experimenting with bypass caps. What do you recommend? 0.1uF? Tantalum, disc, .. ? It seems as most of the issues come from noise on the ground line. There is a lot of noise as the plugs discharge and the current returns through the ground. I believe that this noise is interfering with the timing of the multivibrator. Any recommendations to eliminate this? I was thinking of a ferrite choke inline on the ground line on the circuit, but I am unsure of what exact choke to use or if there is a better option.

Here's the scematic again since it wasn't showing up. VCC is from a LM7805 regulator. I have a 0.1uF capacitor on the output and have added 0.33uF on the input. 12v input voltage. "IN" is from the car's ignitor. Basically an open collector. The NAND gates are all 5v. The labels were just from Electronics Workbench. The multivibrator is a HCF4538B from ST.

Thank you for your help.

Kent

philba · Post by **philba** » Sat May 27, 2006 12:43 am

maybe connect to ground via a ferrite bead? Any guess as to the frequency of the noise?

tants or ceramics would be ok for bypass.

abelk · Post by **abelk** » Sat May 27, 2006 1:21 am

I was thinking of using a ferrite bead. Not sure what one to use exactly. The noise is at the frequency of the sparking (~20Hz (idle) to 300Hz or so at high RPM). It may be a bit difficult as the frequency is fairly low. The amplitude I am seeing is about 0.2v or so. Perhaps it would be better using a battery and larger ground wires. I am using a power supply for the power source. The problem is that I don't have my car here in France to test things on the car itself, so I have to bench test. Once I get it working well on the bench, I can have someone else test on the car.

I take take some pics of the waveforms and post them if you would like.

Thank you.

Kent

MrAl · Post by **MrAl** » Sat May 27, 2006 8:17 am

Hi again,

Maybe you could take a pic of the signal on the "IN" terminal
with the scope. This could show lots of switching noise.

abelk · Post by **abelk** » Sat May 27, 2006 8:22 am

Sure. I will post a pic here in a few minutes. The "IN" looks pretty good to me, but I could be wrong.

abelk · Post by **abelk** » Sat May 27, 2006 9:14 am

Here you go. Top trace is input, bottom is output. Top is 5V/div, bottom is 2V/div. Time base is 2ms/div

No coil connected.

Coil connected

Ground, 0.1V/div Time base same as above. Voltage between coil backing and 7805 backing.

Let me know if you need any other pics or information. Thank you for the help.

Kent

T&L Publications

Am I missing something?

Re: Am I missing something?

Re: Am I missing something?

Re: Am I missing something?

Re: Am I missing something?

Re: Am I missing something?

Re: Am I missing something?

Re: Am I missing something?

Re: Am I missing something?

Am I missing something?

Who is online