Stock Electronic Image FAILS

[McBryce]

OK this one is right from the cover of the last Electronics Design magazine. No one needs stinking ground reference when doing differential "A-B" on a conventional scope.

I've spent way more time on this than I should have but I just don't see the issue here. I've held 2 probes on a board before. Please fill me in so I can sleep tonight.

Did you also just connect the probe tips without any connection to ground?

McBryce.

[TerraHertz]

Here's one from the digikey "catalogue":
https://www.digikey.com/product-detail/en/panasonic-electronic-components/ECW-U1102JX5/PCF1348CT-ND/353833

Look closely at the picture of the capacitor.

Click on the small image for the full size mystery.
What IS that? A capacitor made of masking tape and foam rubber? Or perhaps someone made a cap-cake as a pun? (Which actually, come to think of it, would be pretty good.)

OK this one is right from the cover of the last Electronics Design magazine. No one needs stinking ground reference when doing differential "A-B" on a conventional scope.

Never mind the completely missing ground clips, are they seriously pretending anyone can hold round-point probes on the tiny leads of those packages, and never slip off, shorting adjacent pins together?

Why does that cover photo look so 1970s? Something about the colors...

[Cyberdragon]

OK this one is right from the cover of the last Electronics Design magazine. No one needs stinking ground reference when doing differential "A-B" on a conventional scope.

I've spent way more time on this than I should have but I just don't see the issue here. I've held 2 probes on a board before. Please fill me in so I can sleep tonight.

Did you also just connect the probe tips without any connection to ground?

McBryce.

Guys, there are situations which you don't need a ground connection. There are certain machines out there that have a true ground rail connected to the plug ground. In that case, the scope is already commoned to the circuit through the ground in the plugs.

[spudboy488]

OK this one is right from the cover of the last Electronics Design magazine. No one needs stinking ground reference when doing differential "A-B" on a conventional scope.

I've spent way more time on this than I should have but I just don't see the issue here. I've held 2 probes on a board before. Please fill me in so I can sleep tonight.

Did you also just connect the probe tips without any connection to ground?

McBryce.

Guys, there are situations which you don't need a ground connection. There are certain machines out there that have a true ground rail connected to the plug ground. In that case, the scope is already commoned to the circuit through the ground in the plugs.

I normally don't use the ground clip that attaches to the probe, especially if I might be probing around on a board that size. The leads are too short. I'll run a separate lead from the scope front panel ground lug to a common connection on the board or the power supply. I'm sorry, but if this were Mythbusters, I'd call it plausible.

[McBryce]

OK this one is right from the cover of the last Electronics Design magazine. No one needs stinking ground reference when doing differential "A-B" on a conventional scope.

I've spent way more time on this than I should have but I just don't see the issue here. I've held 2 probes on a board before. Please fill me in so I can sleep tonight.

Did you also just connect the probe tips without any connection to ground?

McBryce.

Guys, there are situations which you don't need a ground connection. There are certain machines out there that have a true ground rail connected to the plug ground. In that case, the scope is already commoned to the circuit through the ground in the plugs.

And all the noise across 5 or 6 metres of ground cable lying beside AC power lines, connected through oxidated ground pins in the socket and passing through an SMPSU isn't having an effect on the signa you're measuringl?? There's a reason why those ground clips are so short. Or are you measuring very high voltage signals where even that would be irrelevant?

McBryce.

[spudboy488]

And all the noise across 5 or 6 metres of ground cable lying beside AC power lines, connected through oxidated ground pins in the socket and passing through an SMPSU isn't having an effect on the signa you're measuringl?? There's a reason why those ground clips are so short. Or are you measuring very high voltage signals where even that would be irrelevant?

McBryce.

In my case, I work in a relatively clean electrical environment. The ground lead from the board to the scope is 12 inches. I mostly measure TTL level digital and analog from about 100mV to 5V. Nothing earth shattering. Two probes on a board when I use one for the trigger and the other for what I'm measuring. As I said, I've spent way too much time on this. I'll go back to grabbing my soldering iron like a pencil and probe the light bulb that's in my PC power supply. 

[tggzzz]

And all the noise across 5 or 6 metres of ground cable lying beside AC power lines, connected through oxidated ground pins in the socket and passing through an SMPSU isn't having an effect on the signa you're measuringl?? There's a reason why those ground clips are so short. Or are you measuring very high voltage signals where even that would be irrelevant?

McBryce.

In my case, I work in a relatively clean electrical environment. The ground lead from the board to the scope is 12 inches. I mostly measure TTL level digital and analog from about 100mV to 5V. Nothing earth shattering. Two probes on a board when I use one for the trigger and the other for what I'm measuring. As I said, I've spent way too much time on this. I'll go back to grabbing my soldering iron like a pencil and probe the light bulb that's in my PC power supply. 

You may measure TTL levels like that, but you don't measure TTL transitions like that.

[james_s]

It will "work" in that you'll measure something, but with a long ground it may not resemble the signal that is already there. For signals up to a few kHz without sharp transitions it may be ok and may be suitable for a quick go/no-go but without a properly grounded probe take any measurements with a grain of salt. Up in the tens of MHz even a 6 inch ground lead can cause substantial ringing. Risetimes in the nanoseconds will give you more ringing than signal.

[PointyOintment]

You have gone too far if your probes are smoking.

I just realized that that appears to be a dark-trace oscilloscope.

[Bud]

It will "work" in that you'll measure something, but with a long ground it may not resemble the signal that is already there. For signals up to a few kHz without sharp transitions it may be ok and may be suitable for a quick go/no-go but without a properly grounded probe take any measurements with a grain of salt. Up in the tens of MHz even a 6 inch ground lead can cause substantial ringing. Risetimes in the nanoseconds will give you more ringing than signal.

I'd say I agree, though Tektronix accepts no compromise, see attached.

edit: link to the Tektronix article
http://info.tek.com/rs/tektronix/images/3AW_19134_2_MR_Letter.pdf

[tggzzz]

Up in the tens of MHz even a 6 inch ground lead can cause substantial ringing.

As I'm sure you are aware, the fundamental frequency (or baud rate) is irrelevant; only the transition is important.

Rule of thumb: a 6" lead on a 15pF *10 probe resonates at ~100MHz.

[james_s]

As I'm sure you are aware, the fundamental frequency (or baud rate) is irrelevant; only the transition is important.

Rule of thumb: a 6" lead on a 15pF *10 probe resonates at ~100MHz.

Sure, but there is a relation between the frequency and transition rate. Typically low frequency signals are not fussy about the transition rate, and the lower the frequency, the smaller the percentage of each state that consists of ringing.

[tggzzz]

As I'm sure you are aware, the fundamental frequency (or baud rate) is irrelevant; only the transition is important.

Rule of thumb: a 6" lead on a 15pF *10 probe resonates at ~100MHz.

Sure, but there is a relation between the frequency and transition rate. Typically low frequency signals are not fussy about the transition rate, and the lower the frequency, the smaller the percentage of each state that consists of ringing.

No, in this context there is no such relationship. The "percentage of each state" is, in most cases, irrelevant.

If a receiver is "confused" or damaged by ringing (etc) on a transition, it doesn't matter whether the subsequent transition occurs in 1us, 1s or 1hour.

You would benefit from understanding
http://www.edn.com/electronics-blogs/bogatin-s-rules-of-thumb/4424573/Rule-of-Thumb--1--The-bandwidth-of-a-signal-from-its-rise-time
and I recommend all his other "rules of thumb".

[james_s]

Yes I understand the theory, but in practice I have not encountered many situations at all where a low frequency signal had a particularly fast rise time or the rise time was all that critical. I can look at the square wave from my function generator using no probe ground at all, just both instruments sharing the same earth ground and it looks fine at low frequency, maybe a bit of ringing if you zoom in on a transition but the period is long enough that the ringing is nothing more than a little glitch on the edge. Dial it up and it starts to get ugly, at some point there is practically nothing but ringing and the waveform looks nothing like it should. This is all academic anyway though because there's no sense in not properly grounding the probe.

[tggzzz]

Yes I understand the theory, but in practice I have not encountered many situations at all where a low frequency signal had a particularly fast rise time or the rise time was all that critical.

I have, many times. The issue isn't whether or not the signal needs a fast transition time, but whether it has one or not.

How about a system reset signal or a powerfail signal? Low frequency (one hopes), and if you get it wrong you can have all sorts of "interesting" behaviour.

I can look at the square wave from my function generator using no probe ground at all, just both instruments sharing the same earth ground and it looks fine at low frequency, maybe a bit of ringing if you zoom in on a transition but the period is long enough that the ringing is nothing more than a little glitch on the edge. Dial it up and it starts to get ugly, at some point there is practically nothing but ringing and the waveform looks nothing like it should.

Function generators usually have analogue outputs, so the frequency content is limited (and well-controlled) by the output amplifier. Ordinary logic outputs have no such limitation.

Modern jellybean logic has sub-nanosecond transition times, e.g. 74LVC devices. And then you can consider fast logic outputs with transition times of <50ps. We're not in TTL territory any more!

I've built an edge generator from 74LVC devices which have, I believe but have not proven to my satisfaction, risetimes of around 300ps into 50 ohms. For the uninitiated, 300ps => 60mm on a PCB, so a track on a PCB can be long enough to "contain" several such transitions simultaneously.

This is all academic anyway though because there's no sense in not properly grounding the probe.

It isn't academic, it is extremely real and practical.

Others have suggested in this thread that grounds aren't really needed, and that needs quashing.

[CatalinaWOW]

Instead of quashing the lack of ground probes, let's work on quashing the lack of understanding that forces rote rules like that.  As pointed out previously, short ground leads can be problematic, for a variety of reasons, including masking problems with ground integrity.  Whichever way you hook up the probes you are measuring something.  The key is to understand what you are measuring, and whether what you are measuring has any relationship to what you want to measure, and what the relation is between what you want to measure and how the circuit works.  There can be failures at all three layers.

[TerraHertz]

http://www.smbc-comics.com/comic/how-to-photograph-science

[lmester]

 This came from an advertising link on ZDNet for an investment company.
By 2020 everyone will have one    I'm ahead of the crowd. I have some 8 pin DIP's in my parts bins. I'm sure a few of them have bent pins. 

[Lord of nothing]

ok not electronic releadet but a translate fail:


from a Ebay Seller.
Run for my life?  Does the sell so much crap....   

[Cyberdragon]

This came from an advertising link on ZDNet for an investment company.
By 2020 everyone will have one    I'm ahead of the crowd. I have some 8 pin DIP's in my parts bins. I'm sure a few of them have bent pins.

I think it's on purpose, I think it's supposed to resemble a spider crawling on the finger.

[Urs42]

WHY?!

[McBryce]

To catch the electrons that fall out?

McBryce.

[WorldPowerLabs]

Adding GLUE logic??

[Cubdriver]

Retain the CPU in a high vibration environment?

-Pat

[Lord of nothing]

What you dont Glue the CPU the the Board?!
How did you prevent it to fall it out.