LED burns when oscilloscope is connected to it

Thread Starter

Jean-yves06

Joined Oct 20, 2016
2
Hello,
Working on a project of mine, I have recently bought a USB oscilloscope. It took me only 30 minutes to damage my project using the oscilloscope. I'd like to understand why it happened please .


Setup:

Arduino Mega fitted with a 4xLED driver shield (http://ledsee.com/index.php/en/ardu...-pwm-high-power-led-shield-0-35-0-7-1a-detail) with one LED connected to shield output 1.

Hantel 6022 BE oscilloscope (http://www.hantek.com/en/ProductDetail_2_31.html), powered solely from USB.



Problem : if both Arduino and Oscilloscope are power from the same PC, when I try to measure voltage on the driver output 1 (tip of probe on +, ground of probe on -), I immediately burn the connected LED.

If Arduino is powered not from USB but from an external power supply (and I remove the USB cable between Arduino and PC), the LED won’t burn and I get a realistic read on the oscilloscope.

I am not skilled enough to understand what happens. It is like my probe makes the constant current driver increase its current output, leading to LED burn.

What shocks me is that I thought oscilloscope probes where isolated from the oscilloscope power supply, allowing hot measure on 230 V circuits without risking to inject high voltage back to the PC.
The said probe is 10 MOhms resistive, so it must be something related to the ground of the probe being somewhat connected to the USB ground, but I don't see the big picture here.
 

MrChips

Joined Oct 2, 2009
30,707
This is a common mistake by the uninitiated.

A hand held voltmeter measures the voltage difference between the two nodes where the two meter probes make contact. The voltmeter is "floating" and has no absolute voltage reference.

An oscilloscope probe measures the voltage difference between its probe tip and the GND clip.
NOTE that the GND clip is at absolute 0V (EARTH GND). In other words the oscilloscope probe measures the ABSOLUTE voltage at the tip of the probe. The use of the GND clip is only necessary in noisy high frequency environments or where there might be significant current in the ground loops.

In other words, the oscilloscope is already referenced to GND. It is not "floating". The GND clip is at EARTH GND.

You CANNOT connect the GND clip willy nilly in a circuit if the circuit is already referenced to GND.
Doing so is asking for trouble as you have discovered.

If you want to connect the GND clip to a circuit, it must be connected to a node that is KNOWN to be EARTH GND.

Your oscilloscope probe has an input impedance of 10MΩ. That applies to the impedance seen between the probe tip and GND. The GND clip is NOT isolated from EARTH GND.
 

Thread Starter

Jean-yves06

Joined Oct 20, 2016
2
Thank you a lot for your quick answer, it has enlightened me. I think I understand better the situation. There is one sentence that I don't get though : "The GND clip is at EARTH GND". In my case, the computer is not connected to Earth Ground, the Oscilloscope is powered from the computer, thus not at earth ground neither. I have a tendency to believe that my probe tip is floating, and that my mistake is to have connected probe GND to "-" of the LED rather than to GND of the project board.
Anyway, in the case where I burned my LED, my deduction is that the GND of the board and of the Oscillo where most likely already both connected to the USB 0V. When I touched with my probe GND to the "-" of the LED, I have shorted this "-" to the GND.
My main mistake : beleived all "-" on my board where connected to GND, and generally I thought all "-" where GND. Nooooooo!
 

MrChips

Joined Oct 2, 2009
30,707
Not all computers are grounded.

1) A laptop running off its batteries and with no charger plugged in is obviously "floating". There is no connection to ground.

2) Even if the charger is plugged in between the laptop and the wall outlet, the computer might still not be grounded. If your charger/adapter has a two-prong plug that plugs into the wall outlet it is 'almost' certainly not grounded. If the charger/adapter has a 3-prong plug there is a 'possibility' that it is grounded. A resistance measurement between the ground pin and a known GND point is required in order to verify.

3) A desktop PC with a 3-pin power cord is 'most likely' grounded. Again a resistance measurement will verify this.

In all cases, what is of importance is that the circuitry of your USB scope be referenced to a COMMON node of your external circuits under test.
This is essential otherwise the two devices would not be able to communicate. There are exceptions to this requirement and that is when circuits are intentionally isolated galvanically using optical isolators or other means of achieving galvanic isolation.

Your primary mistake is to assume that you need to connect the GND clip to any point in your test circuit. That is a No-No.
The GND clip is connected to a COMMON 0V reference in your computer, regardless of whether or not the computer is grounded.
You can connect this GND clip only to an equal galvanically connected COMMON 0V on your external circuit and no other place.
You can verify COMMON connection between two nodes by taking a resistance measurement between the nodes with ALL POWER TURNED OFF. The observed resistance reading should be lower that 1Ω.

I should point out that there are different grounds in electronics.
Chassis ground, i.e. computer metal cases, USB metallic housing etc., is not the same as signal GND.
The two ought to be both at 0V but they serve different purposes.
Chassis ground provides electrical shielding and safety. Signal GND is the 0V COMMON reference node.
 
USB- is common throughout all USB ports in a given PC. USB- is the return for the +5V USB supply. USB- is very likely at or near chassis ground. Laptop supplies may or may not have 3 prong power cords.

The scope typically has a 1 M input Z in parallel with about 22 pf. When you add a x10 scope probe the input Z of the system increases to 10 M ohms. The adjustable capacitor on the probe nulls the capacitance of the scope, so the divide is purely restive.

Usually to measure a grounded signal, the scope is placed into A-B mode or A+Invert(B). You still need a ground reference to work.

Line operated Switchmode power supplies create even more hazards. The primary side starts out will 1/2 or full rectified mains voltage depending on 240 or 120 input. That ground (the - of the bridge rectifier) is way off from earth ground.
 

AlbertHall

Joined Jun 4, 2014
12,345
I have a USB oscilloscope, not the same one though so yours may work differently. With the 'scope plugged into the USB, there is about 2.5V between the probe ground clip and the PC 0V.
If a circuit is powered from the same PC USB where USB- is the circuit 0V and I connect the ground clip to that 0V then I am shorting out that 2.5V and bad things happen.

I guess this is down to way the 'scope derives its internal power from the USB.

In this circumstance, I don't connect the ground clip. I connect the probe tip to the circuit 0V and use the Y-shift to put the displayed line where I want 0V to be. Then when connecting the probe to points in the circuit you can read correct voltages.
 

MrChips

Joined Oct 2, 2009
30,707
I have a USB oscilloscope, not the same one though so yours may work differently. With the 'scope plugged into the USB, there is about 2.5V between the probe ground clip and the PC 0V.
If a circuit is powered from the same PC USB where USB- is the circuit 0V and I connect the ground clip to that 0V then I am shorting out that 2.5V and bad things happen.

I guess this is down to way the 'scope derives its internal power from the USB.

In this circumstance, I don't connect the ground clip. I connect the probe tip to the circuit 0V and use the Y-shift to put the displayed line where I want 0V to be. Then when connecting the probe to points in the circuit you can read correct voltages.
That appears to make sense. The USB scope has created a reference at +2.5V in order to be able to measure bipolar signals, i.e. signals going both positive and negative.

Edit: What I just wrote makes no sense. Please ignore this post.
 
Last edited:

AlbertHall

Joined Jun 4, 2014
12,345
That appears to make sense. The USB scope has created a reference at +2.5V in order to be able to measure bipolar signals, i.e. signals going both positive and negative.

Does the User Manual give any warning about this dangerous situation?
I dunno. It long past the lifetime of the user manual.
I just downloaded a copy and it does, though you wouldn't recognise it as such. I paraphrase "Don't connect it to USB while making measurements"

upload_2016-10-21_15-45-29.png
 
Top