Probing PECL with scope?

Thread Starter

XenaHagton

Joined Aug 18, 2017
3
I am not certain about correctly connecting scope probes to the output of differential ECL components. The board I'm looking at uses PECL (+5V Vcc, 0V Vee), and I don't have any special probes to use with my Tektronix TDS 3054. Any tips on how to do this would be greatly appreciated! I doubt that it makes any difference, but the specific devices are 100k ECL (e.g., MC100EL34, MC100EP195).
 

RichardO

Joined May 4, 2013
2,271
Rule number one is keep the connection path for the probe very short. This means the shortest ground lead possible and removing the spring loaded clip and using the point directly.

Rule number two. See rule number one. ;)


The edge speed of PECL is a less than a nanosecond. Long leads can cause ringing, overshoot, undershoot and other nasty things in the waveform. If you can use one of the probe tips with a ground pin instead of a ground lead that is even better.

What measurements are you trying to make? If you are trying to make precise timing measurements then you want to use probes of the same model number and length.


p.s. For those not familiar with PECL, it is just ECL operated from +5 volts and ground instead of ground and -5.2 volts.
 

Thread Starter

XenaHagton

Joined Aug 18, 2017
3
Probe length is not an issue, it's the fact that ECL output structures consist of unterminated emitters sourcing small differential currents. The scope acts as a termination, which swamps the output. I know there are some tricks like AC coupling with bias, but I'm not clear on the best way to do it. The circuit itself is essentially just a digital counter and a couple of flip-flops running at around 350MHz - nothing fancy or cutting edge.
 

RichardO

Joined May 4, 2013
2,271
Probe length is not an issue, it's the fact that ECL output structures consist of unterminated emitters sourcing small differential currents. The scope acts as a termination, which swamps the output. I know there are some tricks like AC coupling with bias, but I'm not clear on the best way to do it.
The reason the scope probe cable length can be important is that the signal travels down the cable at about the speed of light. If the cables are a meter different in length, the signals will arrive at the scope about 3 ns apart in time. 3 ns is a huge time difference in ECL/PECL circuits.

I am not familiar with ac coupling tricks either since I have never done it. I believe that some computer buses do what you are thinking about so you might look for application information there.

The circuit itself is essentially just a digital counter and a couple of flip-flops running at around 350MHz - nothing fancy or cutting edge.
I am curious about the circuit. Can you post it here?
 

Thread Starter

XenaHagton

Joined Aug 18, 2017
3
I don't have a way to post the circuit, but my question isn't related to the circuit. I'm just wondering if anyone can tell me how they've used a standard scope to look at PECL.

I had a modular scope with an isolated differential input module Back In The Day, but alas! No more. Oh well, the analog bandwidth was far too low anyway.
 
Top