A few short measurements
Contents
Now we are already quite impressed with the Mutec MC3+ as a reclocker. The jitter that we measure within the digital domain is around 1 to 2 Ps. That is incredibly low. We did some quick measurements with and without the REF10 and it’s not so much that the jitter gets lower when we do standard measurements in the digital domain (Digital I/O measurements).
What you see in the pictures is first of all the clock signal from the REF10 (10 Mhz) clock. The second is a clock signal (wordclk) from the Mutec MC3 +. You can see that these look different. The Mutec overshoots a bit without a 75 ohm termination; that is fixable. The second measurement is an eye-pattern measurement. Tighter is better. However, we see very little difference. We also did a jitter spectrum measurement. Here we see subtle differences, but very much within range. We will of course continue to look to determine how we can make measurements that will give more insight here.
What you see above is two “Jtest” jitter measurements on two different dacs. The first is an oversampling dac. The second is a NOS dac where a Jtest is not actually possible at all, so take those with a large grain of salt when it comes to jitter performance. NOS dacs should be measured differently.
What I found interesting, however, is the reduction in spikes when the REF10 is used as the master. With just the Mutec we see spikes everywhere, they are all gone when the REF10 is coupled to the MC3+. Interesting… On the measurement of the oversampling dac we see nothing. No lower jitter… nothing at all. Probably because it simply locks to the incoming signal with its own PLL.