Measuring the voltage drop across a VCSEL- or I am too stupid to use an oscillos

[Bigman]

Hi,

this already feels like a stupid beginners problem.

Background
I have build a standard TIA-circuit with a fotodiode. To play around with that TIA-circuit I am using a VCSEL, which is powered by a constant current circuit as shown in the picture. Perhaps not the best way, but a beginner's start. When there is no light on the photodiode, the TIA's OpAmp-Output noise is really low. But when the VCSEL irradiates the photodiode, the noise is very, very high (more than expected). Well, there maybe a million reasons for that, but also out of curiosity, I wanted to check if perhaps the voltage at the VCSEL is distorted.

How I connected my oscilloscop
To get an idea about the voltage curve/drop across the VCSEL did it like this:

• Probe's tip --> VCSEL anode
• Probe's/oscilloscpe's GND --> VCSEL cathode
• The oscilloscope's input is labeled with "1MOhm || 15pf, 300V RMS".

Well, also I can not say it for sure, but I killed two VCSEL's with that setup (not during normal operation, but when I tried to do the voltage measurement).

My assumption
By connecting the VCSEL's cathode with the oscillosopce's ground I created a short-circuit.So far, I assumed the oscilloscope's ground is not really connected to any ground (e.g. main's ground). Instead it would be somehow "only floating." ... but honestly speaking, I have no glue.

My question
In case my assumption above is right, what is then the explanation why I fried the VCSEL's?

[Marco]

By connecting the VCSEL's cathode with the oscillosopce's ground I created a short-circuit.So far, I assumed the oscilloscope's ground is not really connected to any ground (e.g. main's ground). Instead it would be somehow "only floating." ... but honestly speaking, I have no glue.

Scope ground is earth if you're using an earthed plug ... generally you should earth the scope and float the DUT.

[Bigman]

By connecting the VCSEL's cathode with the oscillosopce's ground I created a short-circuit.So far, I assumed the oscilloscope's ground is not really connected to any ground (e.g. main's ground). Instead it would be somehow "only floating." ... but honestly speaking, I have no glue.

Scope ground is earth if you're using an earthed plug ... generally you should earth the scope and float the DUT.

thanks a lot for your kind reply. Since I don't want to destroy a 3rd VCSEL, pls. allow me to clarify this.
When you say "generally you should earth the scope" do you mean "I should make sure, that the main's power supply is connected with ground? Sure, this should be the case (also for safety reasons - or?)

"float the DUT" ... hmmm ... the whole circuit was always only connected with the battery's + and - terminal. Perhaps the GND-triangle symbol is misleading in my picture, which I updated.

In the picture, I also added the probe .. and I still don't understand, where I should connect the black wire. With point (1) ... well .. this I did and fried the VCSEL. With (2) or (3)? With nothing (I tested this with a simple two-resistor voltage divider, and the scope showed always 0V).

[w2aew]

Connect the scope ground lead to the circuit ground, and the probe to point (1).  It won't measure the voltage "across" the VCSEL, but since the Anode of the VCSEL is at a constant DC value, the voltage seen at (1) will give you a good representation of the voltage variation across the VCSEL.

When you connected the scope probe ground lead to (1), you effectively shorted that point to ground, putting the VCSEL directly between the circuit's power rail and ground.  This certainly cooked the VCSEL.

[dadler]

Hmm-maybe I misunderstand. He seems to indicate the DUT is floating, so unless he has a second probe connected with its ground lead to circuit ground, how could he fry the laser diode in that fashion?

[w2aew]

Hmm-maybe I misunderstand. He seems to indicate the DUT is floating, so unless he has a second probe connected with its ground lead to circuit ground, how could he fry the laser diode in that fashion?

I'm making the assumption that the 3.6V supply for the VCSEL circuit is from a power supply whose ground is common to the scope (and is not floating). 

[dadler]

Hmm-maybe I misunderstand. He seems to indicate the DUT is floating, so unless he has a second probe connected with its ground lead to circuit ground, how could he fry the laser diode in that fashion?

I'm making the assumption that the 3.6V supply for the VCSEL circuit is from a power supply whose ground is common to the scope (and is not floating).

You are probably right. He did add this though:

"float the DUT" ... hmmm ... the whole circuit was always only connected with the battery's + and - terminal. Perhaps the GND-triangle symbol is misleading in my picture, which I updated.

I guess we'll need to hear more from him.

[w2aew]

OK.  VCSELs can be pretty fragile devices - so it could be that it was floating, and simply connecting the scope probe ground may have induced a fatal spike in current due to the voltage difference (capacitance to ground, etc.) between ground and the circuit.

[Bigman]

Sorry, that I haven't been fast enough to give further comments.
1) the circuit was powered by a Lithium-Ionen battery (state of charge around 3.6V),
2) there was no other probe

However, I have a further ASSUMPTION about a possible mistake. The probe's ground was also connected to a metal-box housing (for shielding). Althought this housing was connected to nothing else, than with the probe's gnd, I could image I touched it while testing and greated through my body a short-circuit path. Again, this is just an assumption and I am willing to risk a further VCSEL to find out.

Suggestion: I will make a picture of the real set-up before I run a further test. Give me a second, to set-up the equipment again and to prepare the picture.

[dadler]

Alan is probably onto something with the floating vs ground issue.

One other approach is this: clip the ground leads of two separate probes together (don't connect them to anything else, no shielding boxes etc!). Then probe the anode of the diode with one probe, and the cathode with the other. Invert one channel and use the math function to show A+B.

Or just the A-B math function.

[Bigman]

Okay, here a picture of my setup. (Pls. don't laugh at it ... I am doing this as hobbist and I am on a beginner's level). One assumption is that I created a short-circuit through my body by accidentently touching the metal-box ... (and it could be that at that time I was also wearing an anti-static band ... but I cannot remember exactly)

At my side, it's now already after 11:00pm, but the last suggestion with the two probes and using the scope's math function sounds interesting. For sure, tomorrow I will consider this and post the results.