I am in the lucky position to work on a board (at my day job) that uses a 200MHz LVDS clock signal. What would be better than measuring this clock with the MSO2024B and a DSO7245 (2.5GHz bandwidth and up to 40GS/s woohoo!) to compare the waveform captures. The only two things better than that would being able to report this in the road test review and having more hands for holding probes to be able to show the positive and negative part of the LVDS signal at the same time.
The MSO2024B shows an almost sinusoidal waveform with 360mV amplitude. The measurement clocks it in at exactly 200MHz.
The math function is capable of doing real-time FFT and the signal shows a strong peaks at 0Hz (the DC offset) and 200MHz and still a small signal at the 1st harmonic at 400MHz. This is all expected and good to see confirmed in reality.
FFT:
The DPO7254 with an active probe (2.5GHz bandwidth) capture shows a more LVDS compliant waveform with about 1nsec rise time, a plateau at max and min amplitude and a whole lot of minor reflections and impedance discontinuities.
The same fast scope with a 200MHz probe shows more or less the same signal as the MSO2024B, which confirms the excellent performance up to the specified limit. The amplitude is comparable with all measurements (just forgot to enter the attenuation for the 10x probe to get a correct display).
I wish I had a fast function generator to work out the real limits and the signal strength roll off beyond 200MHz. One thing to try would be using the MSO2024B with a high bandwidth probe, so see if the scope or the probe is the limiting factor.
And here is another trick that might come in handy. Active probes don’t allow AC coupling, which would help to show a good signal amplitude if a small oscillation with a big DC offset needs to be displayed. The trick is to use the offset feature to move the signal down onto the screen to have the baseline outside the visible area and be able to increase the scale and still be able to see the full signal. The offset setting can be found in the vertical channel menu under the 'More' menu.
Discussions
Become a Hackaday.io Member
Create an account to leave a comment. Already have an account? Log In.
Should try doing eye diagram using persistence mode to show what happen.
http://www.edn.com/design/test-and-measurement/4389368/Eye-Diagram-Basics-Reading-and-applying-eye-diagrams
Are you sure? yes | no
Hello KC, I totally agree that an eye diagram is the best way to characterize an differential LVDS clock. Here I was just using the 200MHZ signal to show the limits of the scope, a square wave would have been even more useful in this case. Cheers, W.
Are you sure? yes | no