Negative Pulse On Fuel Injector

[bostonman]

For fun, I wanted to measure the fuel injector pulses on an old car I'm junking (1997 Buick).

The results were different than my expectations. The pulses were going 0 to -14V with approx. +45V inductive kickback spikes (see attached).

The kick back spikes I expected, but not the negative voltage.

The setup: the injectors only have two wires. After probing, I found common wires between the injectors, and assumed they were ground. Later I used this ground wire to measure battery voltage with my meter. The scope probe was connected to the positive, and ground clip to the common ground wire.

Would a reason exist to use negative pulses? To confirm I didn't have inverse or some weird setting enabled (I was doing this outside with glare on the scope screen), I measured the battery to the probe tip and left the ground clip on the ground wire of the injector. The scope was showing the correct +12V on the battery indicating the negative pulses are real; not inverted due to an incorrect setting.

[langwadt]

there is no negative voltage, one end of the all the injectors are connected to battery+, the other end of each injector is connected to a transistor in the ECU that pulls it to ground to  open the injector

[SteveThackery]

So maybe I've completely misunderstood, but my guess is that the injectors are low-side switched. So the common wire is +12V, and the others are switched to ground to activate the injectors.

So no negative voltages wrt the car ground.


Oops - @langwadt posted at the same time as me.

[bostonman]

Actually, I typed "pulsed" by accident; I would have also expected all the injectors to have +12V on them too and the other wire being pulled to ground through each transistor for that cylinder.

I just remeasured voltages with the car off. Each injector is approx. 12.2ohms (I already knew this) and looking at only two injectors (the others have too much wrapping around the wires to bother accessing), two wires are common among both.

Yesterday I measured from battery positive to each "common" 'wire on the injectors and got +12.3V indicating they are ground. The other wire on each injector didn't measure anything indicating those are connected to the transistors. The only way to have measured (that I can think of) 12.3V is if the wire is ground.

Today I may have got a better connection because both wires on both injectors (i.e. all four wires) measure 12.3V when I connect the meter to battery positive and touch each wire.

I'm confused because now all four wires measure 12.3V. I assume it's because the circuit isn't being loaded due to the car off, however, in order to measure 12.3V, it would mean the wire I assumed is ground is truly ground.

IF this is the case, the  negative pulses I measured are accurate.

[SteveThackery]

I've some moderate experience with this. I think there is a measurement error. I remain of the opinion that the injectors have a 12V feed and low side switching. This is the usual arrangement.

@bostonman: exactly what model and engine size is it? That might help.

[langwadt]

Actually, I typed "pulsed" by accident; I would have also expected all the injectors to have +12V on them too and the other wire being pulled to ground through each transistor for that cylinder.

I just remeasured voltages with the car off. Each injector is approx. 12.2ohms (I already knew this) and looking at only two injectors (the others have too much wrapping around the wires to bother accessing), two wires are common among both.

Yesterday I measured from battery positive to each "common" 'wire on the injectors and got +12.3V indicating they are ground. The other wire on each injector didn't measure anything indicating those are connected to the transistors. The only way to have measured (that I can think of) 12.3V is if the wire is ground.

Today I may have got a better connection because both wires on both injectors (i.e. all four wires) measure 12.3V when I connect the meter to battery positive and touch each wire.

I'm confused because now all four wires measure 12.3V. I assume it's because the circuit isn't being loaded due to the car off, however, in order to measure 12.3V, it would mean the wire I assumed is ground is truly ground.

IF this is the case, the  negative pulses I measured are accurate.

if it is any kind of normal or even semi normal engine one wire of each injectors are connect to battery positive via a relay that is ON when the ignition is ON, the other wire is pulled to ground via the ECU when the injector is opened

measure everything referenced to battery negative (chassis)
 

[bostonman]

I think there is a measurement error.

I agree, I must be doing something wrong. Based on the feedback, no question it can be a negative going pulse, so it's an error for certain. It can't be the probe is reversed because then the on-time would be longer with a negative going inductive pulse; unless I'm thinking about it incorrectly.

The engine is the 3800 series engine - fairly standard engine for GM during this era.

[langwadt]

this?

[Psi]

I think there is a measurement error.

I agree, I must be doing something wrong. Based on the feedback, no question it can be a negative going pulse, so it's an error for certain. It can't be the probe is reversed because then the on-time would be longer with a negative going inductive pulse; unless I'm thinking about it incorrectly.

The engine is the 3800 series engine - fairly standard engine for GM during this era.

This - "0 to -14V with approx +45V inductive kickback spikes"
Is exactly the same as this - "+14V to 0 with approx +59 of inductive kickback"

It's just how you measured it, your scope/meter was assuming +14 was ground.

The on-time is how long it's at 0V for. Which may sound wrong at first, but it's not. 
It's the same as if you had a lamp permanently wired to a battery + and then you touched the other lamp terminal to battery negative for an amount of time. 
The voltage on that wire is 0V for the lamps on-time.

[bdunham7]

I agree, I must be doing something wrong. Based on the feedback, no question it can be a negative going pulse, so it's an error for certain. It can't be the probe is reversed because then the on-time would be longer with a negative going inductive pulse; unless I'm thinking about it incorrectly.

Put the positive lead of your scope on that "common" lead and then put the negative to the actual ground on the car, the negative battery terminal.  You should read about 14V all the time, with a bit of noise from the inductive kick and maybe a small drop during the time the injector is on.  Now put the negative probe scope on the other side of the injector and you should see the pattern everyone expects--zero voltage during the off time, 13-14V during the on time and the spike at the end of the on time.  Note that the scope needs to be completely floating (neither lead grounded) for this to work so if you are using a non-isolated scope this will not work well at all.  So what scope are you using?

[bostonman]

I didn't want to just swap connections until I got the correct results; rather understand the issue.

Although, everyone makes a valid point: connect to the sure thing which is battery ground.

Floating the scope as in no Earth ground? I'm using an Agilent 54831

[bdunham7]

I didn't want to just swap connections until I got the correct results; rather understand the issue.

Although, everyone makes a valid point: connect to the sure thing which is battery ground.

Floating the scope as in no Earth ground? I'm using an Agilent 54831

Yikes!  You're in dangerous territory.  Right now the only reason you haven't blown something up or blown a fuse is that you are floating the car.  If your scope is plugged into a normal 3-prong socket or extension cord, then the BNC shells--and thus the negative lead--are connected to earth ground.  If you connect the negative lead anywhere other than the car's body or negative battery terminal, then if the car became grounded you'd have a short circuit.  And that's a pretty high-end scope to be blowing up working on a junk car!

The only way to do this safely is to use two probes with one of them grounded.  So connect the probes to CH1 and CH2 and set up a MATH display using the subtract  (or add and CH2 invert in the olden days..) function, CH1-CH2.  Then you'll ground the negative lead/ground clip from the CH1 probe and completely omit it from the CH2 probe.  Finally, you'll connect the CH1 probe as your positive lead to the A (common) side of the injector and the CH2 probe to the B (control/negative) side.

And don't let that scope anywhere near the ignition system on that car no matter how you connect it! 

[bostonman]

Yikes!  You're in dangerous territory.  Right now the only reason you haven't blown something up or blown a fuse is that you are floating the car.  If your scope is plugged into a normal 3-prong socket or extension cord, then the BNC shells--and thus the negative lead--are connected to earth ground.  If you connect the negative lead anywhere other than the car's body or negative battery terminal, then if the car became grounded you'd have a short circuit.  And that's a pretty high-end scope to be blowing up working on a junk car!

Oh boy... these grounding issues confuse me; or maybe not. The car is just a 12V DC battery, correct? If I took a wire and connected the positive directly to Earth ground, nothing would happen. The probe is high impedance, so no short should occur. This means I can connect a single scope probe either way without blowing up anything.

If I understand the associated risk and what you mean by blowing up something: I connect the scope probe to the battery with the probe to positive and clip to negative. I take a second scope probe and connect the clip to positive, and the probe to negative. Now Earth ground connects positive and negative through the ground clips?

Since in my configuration I'm connecting the probe ground clip to "common" wires, I so far, haven't created a short?

Also, if i use a single scope probe, much like my single battery example above, I can connect the probe in any configuration without risk damaging anything?

[aeg]

if it is any kind of normal or even semi normal engine one wire of each injectors are connect to battery positive via a relay that is ON when the ignition is ON

Yup and when the relay is off, the injector common is going to find its way to ground, resulting in the bogus measurement observed...

[bdunham7]

Oh boy... these grounding issues confuse me; or maybe not. The car is just a 12V DC battery, correct? If I took a wire and connected the positive directly to Earth ground, nothing would happen. The probe is high impedance, so no short should occur. This means I can connect a single scope probe either way without blowing up anything.

All true as long as the body of the car remains isolated from ground.  If it is just sitting there with nothing around it then the 4 tires should keep it insulated.  Even if there's a spider web or some moisture, that's high enough impedance and you won't burn anything up.  However, if the car body gets connected to a good ground somehow--just use your imagination--then the sparks will fly.  If you don't think that is likely in your particular case then carry on.

[bostonman]

If you don't think that is likely in your particular case then carry on.

I'm second guessing all aspects of the scope ground clip and openly admit I'm a bit confused. Maybe starting with ignoring the four wires will be a good start; possibly the probe tip too since it's high impedance.

Just a quick background about me. For years I worked on DC powered equipment. Most scope measurements at other jobs, at home, etc... was typically with a single channel. The practice of ground clips causing shorts is embarrassingly relatively new to me.

Below are all situations/questions involving scopes that have the third prong connected to Earth ground (I would never attempt to float a scope). In no way am I implying doubt with the advice from anyone; I'm just trying to get a better understanding.

With a single scope probe, I should be able to connect the ground clip to any point in the car without anything happening, correct? Of course, this is ignoring my probe will be referenced to who knows what causing wrong measurements, however, the question is more about blowing up something.

Add a second scope probe, and the ground clip needs to be connected to the same point to avoid a short, correct?

In other words, if I clip the ground onto the positive terminal of the car battery, the clip on probe two needs to be on the same spot.

Assuming the above two are correct. Last night I began questioning just about everything including simple logic circuits.

As an example:

Scope probe one ground clip on ground of a logic board (let's assume isolated from the primary side of the power to make it easier to understand). Scope probe two ground clip accidentally connects to +3.3V. The +3.3V goes through the clip lead, through the scope, through Earth ground, back the other way, down to scope probe two ground clip, and blows up my logic board (assuming the +3.3V doesn't have a series protection resistor/fuse/etc...)?

If this is all true, it seems I've underestimated the risks with the ground clip. Prior I focused on touching the ground clip to the hot of a power supply, but now I'm thinking about the loop(s).

Update: after reading and thinking more, seems for many years I've underestimated the dangers of the ground clip. Most often I connect them to the same point without giving thought. Recent years paid attention to primary side of transformers by using a diff probe, but, after reading more, believe I've learned the answers to the questions above.

[Renate]

I was wondering how the flyback typically gets snubbed since you can't just let it rip and you can't do a straight free-wheeling diode.
The internet seems to indicate a 60-70 V Zener.

[Psi]

I'm second guessing all aspects of the scope ground clip and openly admit I'm a bit confused.

Below are all situations/questions involving scopes that have the third prong connected to Earth ground (I would never attempt to float a scope). In no way am I implying doubt with the advice from anyone; I'm just trying to get a better understanding.

With a single scope probe, I should be able to connect the ground clip to any point in the car without anything happening, correct? Of course, this is ignoring my probe will be referenced to who knows what causing wrong measurements, however, the question is more about blowing up something.

Add a second scope probe, and the ground clip needs to be connected to the same point to avoid a short, correct?

In other words, if I clip the ground onto the positive terminal of the car battery, the clip on probe two needs to be on the same spot.

Yes, you will short things if you connect the ground on one channel to something and then the ground on other channels to something else.
Scope ground clips should go to the same place. For slow DC stuff you can usually just connect one clip and forget the clips on the other channels, but it's bad practice and it come back to bite you at some point.
When you need to look at two signals on the scope and see them referenced to different points you connect a special active isolated probe to the scope, after crying about the price you paid for it.

Scope probe one ground clip on ground of a logic board (let's assume isolated from the primary side of the power to make it easier to understand). Scope probe two ground clip accidentally connects to +3.3V. The +3.3V goes through the clip lead, through the scope, through Earth ground, back the other way, down to scope probe two ground clip, and blows up my logic board (assuming the +3.3V doesn't have a series protection resistor/fuse/etc...)?

Usually your 3.3Vreg just goes into current limit protection, but yes. You're shorting it.
It's a lot more err... entertaining when you do it with 230VAC    Well.. entertaining when it's someone else 
Your more likely to blow up your board shorting higher voltage rails to low voltage rails or GPIO pins than directly shorting the supply rails to ground.

Oh and watch out for circuits that are low side current sensing, each side is at the same voltage and will seem connected with a DMM, but If you ground each side with the scope clip you bypass the current sensing and your device under test will think the current is lower than it is, maybe zero.

[langwadt]

I was wondering how the flyback typically gets snubbed since you can't just let it rip and you can't do a straight free-wheeling diode.
The internet seems to indicate a 60-70 V Zener.

often the driver used is a mosfet with build in voltage clamp, basically a zener diode from drain to gate

something like a VND5N07

[bostonman]

When you need to look at two signals on the scope and see them referenced to different points you connect a special active isolated probe to the scope, after crying about the price you paid for it.

I have one or maybe two from Miscig, and didn't exactly cry over the price, but didn't think I'd have a need for them.

Although I knew long ago ground clips are all connected between channels, and to Earth ground through the scope, for whatever reason the thought of shorting a circuit seemed to be overlooked.

In the case of my car, I was focused on the possibility the wires I measured for continuity were referenced to the same point without giving thought to the bigger picture of that ground clip being a serious risk.

Once it was mentioned, I lost track of the ground clip connections and began thinking too far outside the box of whether the issue was a conflict between 12V car voltage and 120V AC.

I need to put the ground clip(s) as the primary focus in the future.

Now that ground clip issue as been slapped from the back of my head to the front, I'll readdress measuring the fuel injector signals with the ground clips referenced to battery negative; maybe now I'll see +12V on one side of the injector and +12V to 0V on the other. Also, make sure not to connect any other equipment during this process.

I know grounding issues have been discussed numerous times in various messages which has probably been exhausting to repeat, especially those who begin stating they want to or have floated a scope (something I'll never do), so I appreciate the feedback on making me aware of my unsafe setup.