Continue to Site

Welcome to our site!

Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

  • Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

Differential measuring on an analog scope (again)...

Status
Not open for further replies.
Hello to everybody!

I finally grabbed one used by good analog scope ;)! I'm not total newbie, I worked a lot with oscilloscope on my university, I thought I "know everything", but I missed one thing (fortunately, I RTFM before, just in case:eek:): differential measuring! I read anything I could find on this forum (and others), but still I'm unsure!
I know about methods of such a measuring, like "A-B" method, but for now I would like only to know one "simple" thing: Can I do measurements with normal probe with grounded scope on battery-powered circuits, OR circuits that work with main transformers (the "real" ones, not SMPS), e.g. some CD-player or so, without worries where I attach "ground crocodile" from probe?
In other words, can I measure e.g. voltage across some resistor, wherever it is placed, in devices described above, so I connect probe tip to one end of resistor and ground clip to the other end?
Of course, if I want to work with both channels, I must care about fact that both scope's inputs are tied together, I know that. Also, I do not intend to measure high voltages, or anything tied directly to main-plug! So my GND-clip on the probe would be possibly tied to the LOCAL GND of the circuit (like "minus" end of an battery).

Sorry if I ask stupid question!
TIA!
 
If the circuit is totally floating and the frequency of interest is not too high you can make the measurements single ended with the scope.

Just remember the whole chassis of the scope will be connected to one side of the differential measurement. You might think of this as a large metal plate that is connected on one side of the differential input. The effect of stray capacitance of this large metal object may effect the measurement.
 
Yes. Measure the voltage from the ground of the scope to the point your measuring before using the scope. Make sure it's zero.

Any instrument that you use such as a function generator may also ground the circuit, so again you have to be careful.

To avoid mishaps an isolation transformer on the unit your measuring is preferred, but you still have to be careful. Sometimes that isn't possible. It's also preferred that you don't use the ground clip, but rely on the chassis ground connection.

If you have a 12 V wall wart powered unit that doesn't have a connection to ground (check) an your scoping with a scope, the ground lead can go anywhere. Another instrument that you may be using COULD connect ground to another point. Those instruments should have a value for what voltage the instrument can be floated to.
 
Thank you for fast reply!

Perhaps I didn't explain well what I mean (my English...:eek:). I wish to know can I measure signals from e.g. battery-powered calculator. Or Walkman. I didn't mean on differential measurements right now - that's another story. I'm thinking this way: if I measure anything from e.g. battery powered walkman: the GND clip of the probe can be attached wherever I want - afaik there is no way that ANY signal goes to "safety earth" through the scope - is that OK? Furthermore, the same thing should be true if I power that walkman or whatever from wall wart adaptor (assuming it is galvanically separated, not some SMPS)? I think RCinFLA had that in mind under term "totally floating" (I didn't understand "you can make the measurements single ended" to be frankly!)?

KeepItSimpleStupid said:
If you have a 12 V wall wart powered unit that doesn't have a connection to ground (check) an your scoping with a scope, the ground lead can go anywhere.
Yes, that should be the answer to my question (I saw it later)!
You also mentioned function generator: I'm building FG based on XR8038 chip, and it will be powered by 230 to 2x12V transformer, so I assume there will be no problems, since there is no connection to main's "earth". There will be only local GND, i.e. the central wire from the secondary windings of transformer. It will be isolated even from housing (if it will be made of metal)!
 
Sorry to bother, but if I may to add:
This question IS about differential measurements. I don't understand the following "instruction" from the User manual of my scope:
"Differential measurement techniques allow direct
measurement of the voltage drop across floating
cornponents (both ends above ground). Two identical
probes should be used for both vertical inputs. In order to
avoid ground loops, use a separate ground connection and
do not use the probe ground leads or cable shields
."

So what does it mean (marked red): OK, I will NOT use probe GND leads. But WHAT and WHERE to connect regarding "separate ground connection"???
 
What you do is: You ground your floating circuit at a convienient point. i.e. The ground of the scope gets connected to the (-DC) supply of your circuit. If the circuit is ISOLATED, you can ground the circuit anywhere so you do. You connect the circuit to the ground terminal of the scope or to the outlet ground. The "anywhere" point has to be useful.

If the circuit you were working on had multiple sections that are isolated, you may have to move your ground reference.

The scope will say measure the voltage from the channel A probe to the "outlet strip ground" (lack of a better word) and call it A and the volatge from channel B and the "outlet strip ground" and subtract them.

Using the GND probe clip can cause a GROUND LOOP with excessive currents flowing through the ground lead.

Aside:
(1)This works up until the insulation voltage of the various components. e.g. You would not connect ground to the 10 KV lead of a CRT tube.

(2) In designing things you have to understand grounding. We can classify them in three separate categories in a simplistic model. 1) Protective ground: This gound only carries fault currents and it makes sure that when you tough two metalic objects, they will be at the same potential. It could be the kitchen faucet and an electric skillet.
Ideally this "ground" should be a wire to the outside ground rod that is a reference to the house. This won't happen unless your a hospital or a transmitting station (orange outlets).
2) HI-Current ground: It's where all of the noisy signals go. and 3) Analog ground. This is a very quiet ground and almost no current flows to that point.

Ideally, they all should connect at the ground rod reference for the house.

Now supose the device gets hit by lightning. The protective ground carries the fault current, but the AC neutral would not see any real change in voltage, so nothing likely gets damaged.

Take the string of series connected outlets which is a normal configuration and the furthest away gets hit by lightning. Now the ground potential is raised for each of the other devices connected in the string. Now, we have the potential for damage. Lightning strikes like that are rare.

During a storm, the ground potentials of the earth could be different within a couple of hundred of feet. I knew of a computer installation where devices were destroyed when they were connected cross the room with a cable. They changed those connections to fiber to fix it.

So, the general idea is to connect the types of grounds at one point within the instrument. A LC filter might be used to reduce the effects of noise on the high current ground.

Orange outlets in the US are used for "Isolated Ground" applications. This means that the ground goes all the way back to power distribution and not shared with another outlet.
 
Thank you for your extensive explanation!

Unfortunately, I still don't understand what to do with two probes and "A-B" method, as I quoted from User manual - there is clearly stated that I should not use probe ground crocodile clips/leads or even cable shields! So, I connect the tip of the first probe to one point (e.g. one side of resistor) and tip from the second probe to another point of interest (e.g. other side of that resistor). And then what? Should I use separate wire, connect one side to, say, minus pole of the battery (that powers circuit under test) and other side of that wire somehow connect to scope's GND?

EDIT: Now, when I think better, maybe I don't need that extra wire at all? The circuit will be closed from point A to point B via scope's GND...or...?
 
Last edited:
Yes, it seems it is so! I don't need extra wire at all! I just need to remove ground clips from both probes and just use probe's tips as with multimeter! I just saw video from youtube explaining this!
But, now another question emerges: what about well known "Youscope" demo? I see people are regularly using output from a PC-sound card and connect left and right channel + "common" to the oscilloscope! But, PC's have SMPS power supplies (not galvanically isolated) and GND from the sound card is surely (?) tied to "earth". I guess in this particular case this is not a problem since GND from sound card and GND of oscilloscope are at the same potential (at least those should be)! Is this right? Can I do this, or this is possible (safely) only with laptop?
 
The scope A-B mode of the scope won't work unless there is a connection to ground. This is really called pseudo-differential mode.

The scope only knows how to measure from A to ground and B to ground. It cannot measure from A to B or B to A.

Yes, connect the ground of the scope to the (-DC) of the Device Under Test.
 
In a desktop, the common of the left/right output is connected to power ground.

In a laptop, even if the power cord is 3 wire, and the DC cord to the laptop is 2 wire, the laptop ground is generally not connected to power ground. The incoming power cord ground is used for RF shielding or to reduce RFI interference.

So, with a Desktop, you would not need a ground connection to the scope. With the laptop you do.

There have been stories such as where one side of a lab had the Hot and ground reversed and when instruments were shared across benches sparks flew. One problem, it seems, is that for some reason you don't think that black is the hot wire if you were used to the red/black convention. Initially, I thought white would be the hot lead. Fortunately this erroneous info was short-lived.

When I was in high school and attended a Boy Scouts Explorer's post where we built a clock, they asked which wire goes to the fuse. I had to politely tell the HP employee that his information was wrong. Our group had the first working clock and the second one assembled. Everyone seemed to get a bad component, but I forgot which one and we had a tiny trace (this was the tape method) that could have easily fell off in processing. The effect messed up a segment of the display. I don't remember if it was only 1 digit or not. Inter-digit blanking wasn't implemented. It used the MM53xx chip. HP did the component layout for everybody. Groups of 3 or 4 did the layouts together using the old method of red/black and blue tape. The camera used for the artwork reducing, you could walk inside of it.
 
OK, I will not torture you more!
If I understand well, I can connect ONLY left and right channel and just ignore the "common" wire (I'm using desktop PC). But we are at the beginning again: do I need some extra wire to close "circuit", i.e. first probe (CH1) tip to the left channel, second probe (CH2) tip to the right channel and...
(Did you see youtube video?)
 
No, because the scope ground is connected to Earth and the PC ground is connected to Earth. The wire is somewhat invisable.

If you plug in your scope, no probes connected and plug in your desktop PC to the wall and then take an ohmeter and measure from the PC's chassis to the scope Chassis and you will measure a low resistance.

What's wrong when you can measure about 1/2 the AC voltage between those two points? e.g. ~60 VAC for a 120 VAC line voltage.

Why when that happened did I almost have to have to have 450 outlets replaced and an expensive computer? Only some were blanket replaced because a defective outlet had serious consequences. The color was changed so you could tell. When computers were added to offices, then the outlets were tested. The test happened to be very specific and I developed it.

Testing all of the outlets in the building was also deemed to be too resource intensive.
 
KeepItSimpleStupid said:
No, because the scope ground is connected to Earth and the PC ground is connected to Earth. The wire is somewhat invisable.
OK, I think I see the picture. I suppose, if I use (I don't know English phrase for this:eek:) extension cable with 5 or 6 outlets (you know what I mean) there should be no problem, even if I don't check resistance between two chassis!
Please note that I'm fully aware of grounding problem, but I get different "tips" about differential measurements: from the one side, like in that youtube video, there is no third wire, and man on that video says (if I catched his words good) that with two probes without GND clips I can measure whatewer I want (like with DMM)! On the other hand, like in that "instruction" from User manual there is mentioned some "separate ground connection" so I'm confused!
I don't want to ruin my scope with some stupid action...
 
an example of a differential oscope connection is shown below. the scope ground goes to the chassis ground of the equipment under test, channel b is the differential "reference" and in this example goes to the hot side "ground" (which is floating at a line voltage potential), and channel a connects to the point where the desired waveform to be measured is (in this example, the switching fet drain). before making the measurement, you connect channel a to the same point as channel b, and null out the line voltage component using the gain controls on the scope, then connect channel a to the point to be measured. the scope must be in subtract mode (a-b).
 
Thank you very much unclejed613!!!

I think I understood!

Just to be sure: I suppose your drawing is simplified schematic of some SMPS transformer, isn't it:confused:? Now, if that circuit (whatever circuit) has nothing with mains, e.g. operates from battery, in that case I could use just CH-A and connect probe tip to FET and probe GND clip to "hot side" marked point, without use of CH-B and without that wire that goes to "chassis ground" with no problem - true?
 
You can take a look here: **broken link removed**

A1 & A2 are just like the scope inputs. 2 single ended amplifiers fed into a difference amp. All of the circuits require a reference.
 
Thanks, I will study that!

KeepItSimpleStupid said:
So, with a Desktop, you would not need a ground connection to the scope. With the laptop you do.

Just to be clear about that desktop/laptop thing: if I use laptop only with its own battery, without charger - it should be the same thing as with desktop except I must connect "common GND" from laptop's audio output to the scope's GND?

@everybody: does anyone of you guys saw that video from youtube - is that true?
 
I havn't watched the video yet.

There is a unique property when doing differential measurements. If a signal is common to both inputs of the differential amplifier they get cancelled. The amount that they get cancelled is the CMMR. Common Mode Rejection Ratio.
 
Last edited:
The U-tube video. He doesn't understand what's going on. The first mistake he made was "he know" the voltage expected and he didn't select the lower sensitivity first. Bad. It's a good way to blow up a scope.

He never left the probes on long enough, nor did he show the result being AC coupled.

His measurement is a "special case". Why is it a special case? Because he is measuring a low Z source.

You have to now understand a bit about oscilloscope probes. I'm glad I had the lab in school. Look at figure 3 here: https://www.electro-tech-online.com/custompdfs/2012/07/lecroy_probing_tutorial_appnote016.pdf

The probe is an "impedance divider". It has a series resistor and a capacitor and the scope's input is modeled as a 1 M resistor in parallel with likely 22 pf (Typical). Watch 50 ohm Z scopes. So, compensation means you put the probe in x10 mode and connect it to the scope calibrate pin which spits out a square wave.

You attach your scope probe and it doesn't look anything like a square wave. Why? The Z of the scope's cable + compensation cap and the probe's capacitance are interacting and rounding the corners. What do you do? You adjust the little screwdriver cap until the wave is square.

This is a bit of an aside, but you still need to know it.

So, you x10 probe is now a series resistor in the probe and the 1 meg resistance of the scope. We cancelled the caps by doing the compensation. so, it's behaving as a divide by 10 voltage divider which is called a X10 probe. Who knows where the terminology came from.

What does this mean?

Take a look here: **broken link removed**
and plug in 1,000,000 for the scope R. 100V and 10 V for a divide by 10 and you find R = 900,000 ohms. So, you cable is EFFECTIVELY a 900,000 ohm resistor. There is other stuff going on, but don't concern yourself with that now.

So, what is the measuring circuit then?

He has a nearly ideal voltage source, a voltage source in series with a 0 ohm resistor. This is in series with a 900K resistor and it's in series with a 1 M resistor of the scope. Not much of an error.

Now suppose his voltage source wasn't ideal and was 100k. So there is a 9 V source, a 100K resistor, a 900K resistor and a 1 M resistor. There will be a measurement error.

Remember, that internally, the scope has a 1 M resistor connected to ground. Remember those invisible wires. So, A-B does work without the ground connected. You should get cleaner measurements if you can ground the circuit under test, though.

Your scope probe effectively does have a 10 M input Z in the x10 mode and it does place a 1 meg resistor to ground at the scope's input. The probe tip sees 10 Megs with respect to ground. The scope sees 1M with respect to ground.

So, the A-B mode still has a reference to ground even without the ground connected.

The usefulness of the ground clip comes into play when you are troubleshooting high frequency signals on an isolated circuit.

Now you have something to sleep on.
 
Status
Not open for further replies.

Latest threads

New Articles From Microcontroller Tips

Back
Top