Need help Understanding Oscope vs Multimeter Readings

[skinny]

I recently put together a 555 timer circuit that changes the duty cycle of a signal but keeps the frequency stable. I used this circuit.

https://www.multisim.com/content/tf7Mwpy2CctfFaaVEr4SeN/555-variable-duty-cycle-constant-frequency-astable-multivibrator/

When I measure the output with my multimeter, I get what I expect. When I adjust the pot and cause the duty cycle to approach 100%, the DC voltage reading approaches the supply voltage. Likewise, as the duty cycle approaches 0%, that DC voltage approaches 0 volts. Additionally, I can see the duty cycle change on the oscope as expected with the frequency staying somewhat stable as the duty cycle changes.

The piece that's occurring that I don't understand is that the entire signal shifts up as the duty cycle approaches 0% and down as the duty cycle approaches 100%. Also, even though the entire signal shifts down where the top of the wave form is at 1V @ 84.8% positive duty cycle, the multimeter reads a voltage close to the supply voltage.

My questions are as follows:
1) What causes this shift? I suspect it's something to do with the triggering but I'm not sure.
2) Why would I be reading one measurement on my multimeter and seemingly another on the oscope?

For reference the signal being examined is at 25kHz, 5Vpp, with a duty cyle that can be adjusted from 17% to 83%.
I am DC coupled with the oscope. I am triggered on the rising edge. The circuit is running off of a battery supply.

Thanks for all your help. I'm really looking forward to understanding this.

[JS]

You are saying your circuit's duty cycle can be adjusted from 17% to 83% but you are having trouble over 84.8%, that's an odd one.

What load do you have on your 555?
What voltage peak to peak are you reading when you see 1V as the top of the waveform? can you post a waveform when it's all fine and when it's misbehaving?
What do you see when the duty cycle is low? how low? under 17%?

The first thing that comes to my head is AC coupling, but you said it was DC...

JS

[Brumby]

Agreed.  AC coupling.

Op: You say you are DC coupled to the scope, but do you have the scope set on "AC"?  There is usually a switch that has AC, DC and GND.  If this switch is on AC, then that perfectly explains your observation.

What scope do you have, by the way?

[Shock]

The piece that's occurring that I don't understand is that the entire signal shifts up as the duty cycle approaches 0% and down as the duty cycle approaches 100%.

Shifts up, you mean increases in voltage, offset or frequency? Helps to know what multimeter and oscilloscope you're using.

The multimeter is "sampling" (there's a little more to it) at less intervals than the oscilloscope so therefore is more suited to low frequency applications. So it will not be as accurate as an oscilloscope and in many cases totally pointless to use.

That's not to say multimeters aren't extremely useful they are just not ideal for interpreting signals.

[skinny]

Thanks for the suggestions and help.

Agreed.  AC coupling.
Op: You say you are DC coupled to the scope, but do you have the scope set on "AC"?  There is usually a switch that has AC, DC and GND.  If this switch is on AC, then that perfectly explains your observation.
What scope do you have, by the way?

The vertical shift still occurs regardless whether I put it in AC or DC coupling. I have a Rigol DS1052E.

Shifts up, you mean increases in voltage, offset or frequency? Helps to know what multimeter and oscilloscope you're using.
The multimeter is "sampling" (there's a little more to it) at less intervals than the oscilloscope so therefore is more suited to low frequency applications. So it will not be as accurate as an oscilloscope and in many cases totally pointless to use.
That's not to say multimeters aren't extremely useful they are just not ideal for interpreting signals.

When I say shifts up, I mean the whole waveform shifts up and down on the vertical/voltage axis of the oscope. I agree that the multimeter is sampling. This is a PWM application so the closer the signal gets to 100% duty cycle, the closer the DC voltage reading will approach the supply voltage. However, the same time this is happening on the meter, my scope is telling me that my peak voltage of the waveform is at 1V, not 6.5V as the multimeter tells me.

[Brumby]

Time for some images, I think.

Scope screenshot(s); Schematic; Wired circuit.

[mikerj]

Everything you have described suggests you are AC coupled into the scope.  Are you 100% certain you have the scope input channel set for DC rather than the trigger?

[Richard Crowley]

It could be AC coupled BEFORE the scope probe.  Or the probe could have an AC option.  Or even a breakdown in communication between what @skinny is trying to say and what we are understanding.  No place to go from here without better details. Schematic, photos, screenshots, video perhaps.

[Brumby]

It could be AC coupled BEFORE the scope probe.  Or the probe could have an AC option.  Or even a breakdown in communication between what @skinny is trying to say and what we are understanding.  No place to go from here without better details. Schematic, photos, screenshots, video perhaps.

This is exactly the sort of thing I was concerned about.  More words are not going to have nearly as much success here as a couple of pertinent graphics.

[skinny]

Thanks for the input guys, perhaps this will help.

Video:


Schematic:
http://skinnyrd.com/wp-content/uploads/2018/05/duty_cycle.png

[MK14]

The video seems to mention that you are using a small 9V, almost totally flat, battery to power the thing.
That may be the cause of your problem, because it will tend to have a high internal resistance, and give lots of voltage change, as the current consumption of the circuit changes.

Ideally, you want to use some kind of very steady power supply, to power it. Such as a bench power supply.

[skinny]

The video seems to mention that you are using a small 9V, almost totally flat, battery to power the thing.
That may be the cause of your problem, because it will tend to have a high internal resistance, and give lots of voltage change, as the current consumption of the circuit changes.

Thanks for the reply. Switched it out for a DC power supply. Same phenomenon occurs.

[MK14]

Thanks for the reply. Switched it out for a DC power supply. Same phenomenon occurs.

Ok.
Also, the ground lead of the (visible on video) scope probe (I can't easily see how the other probe is connected, at least at, that point in the video), seems to be NOT connected up. This might explain why the oscilloscope image is so shaky.
Ideally, the ground would be connected, where the 555 gets its ground (as close as reasonably possible).

[Brumby]

The signal is definitely AC coupled somewhere along the line.  The proof is when you switched over to AC coupling on the scope and the behaviour did not change.

For DC coupling of a signal - EVERY point along the path from source to scope must be DC coupled. ^**
For AC coupling of a signal - you only need ONE point along the path to be AC coupled. ^**

** This can be affected by active stages

Thanks for the reply. Switched it out for a DC power supply. Same phenomenon occurs.

Ok.
Also, the ground lead of the (visible on video) scope probe (I can't easily see how the other probe is connected, at least at, that point in the video), seems to be NOT connected up. This might explain why the oscilloscope image is so shaky.
Ideally, the ground would be connected, where the 555 gets its ground (as close as reasonably possible).

Ah, I hadn't looked that closely yet - but I did notice the shaky trace.  This would certainly explain the level shift since you would be relying on stray capacitance to complete the ground connection.  This is automatic AC coupling by default.

The really jittery trace is another indicator, since such a ground connection would be rather high impedance, allowing all sorts of electromagnetic hash to get into the signal ground.

If you don't have a ground connection from the scope to the circuit, connect one to the negative of the battery - and watch the magic.  (My punt)

[Brumby]

Here it is ... flapping in the breeze and no other connection visible on the scope.

[skinny]

Here it is ... flapping in the breeze and no other connection visible on the scope.

Corrected the issue with the hanging lead and got the same response. Did this for both a battery and the DC power supply voltage source.

[alsetalokin4017]

I think that the vertical jitter is because the probe isn't grounded and you are picking up the usual 60Hz from the powerlines. Ground the probe to the negative power rail of the circuit and this should stabilize.

The other thing: Please change the timebase to display only one cycle, so we can see just where on the upward slope of the pulse the scope is triggering. This may reveal more information for us to mull over.

[Brumby]

While we're on the subject, ground connections for battery operated projects are usually quite straightforward.  Same thing for isolated bench supplies ... but it is always good to have seen this video of Dave's - especially on gear that has its own power supply...

[Brumby]

Here it is ... flapping in the breeze and no other connection visible on the scope.

Corrected the issue with the hanging lead and got the same response. Did this for both a battery and the DC power supply voltage source.

* scratches head *

... and the coupling has been set back to DC on the scope?

[Brumby]

What about the power rails on the left?  Are they connected?

I briefly saw something red in the bottom left corner of the video when you first turned towards the breadboard, but I have no idea if the second set of power rails are actually connected...

[skinny]

To all of you, thank you immensely for your help. After several of your comments I started getting real suspicious of my equipment after checking for the 20th time. Everything everyone was saying sounded logical but reality wasn't adding up. 

Turns out I've got a bad probe. After taking about a 15 minute break to mull over what everyone was saying, I rechecked everything again. You guys said the ground was the issue but nothing was changing on the display. Finally I decided to change out the probe. After switching it out, the problem went away. Glad it worked because I was really starting to question my sanity. Looks like the ground on that probe is bad.

Thank you all once again for your patience and getting me to a solution!

[Brumby]

I was going to suggest changing out the probe if the last couple of things I mentioned weren't a problem.  I'd traced out the circuit from your video and that was fine - so the probe was about the only thing left!

Glad you got there - and you can pat yourself on the back for working it out!


It's also a good lesson in questioning everything if something isn't making sense.

[exe]

Why did you delete video? I'm curious to see the problem. Also it may have educational value for future viewers if you described the problem in the video description.

[JS]

Nice to know, is usually quite a pain to detect a problem in what we think it's a "good known" that just started to fail.

JS

[skinny]

Why did you delete video? I'm curious to see the problem. Also it may have educational value for future viewers if you described the problem in the video description.

As requested, the video is back up.