OK, new toy time, one that I've been waiting to find...
So inevitably I have far too many boxes that draw graphs of voltage-vs-time (scopes), too many that draw voltage-vs-frequency (spetrum analysers), but until today nothing that would complete the triumvirate by drawing frequency-vs-time (a modulation domain analyser MDA). I did buy an MDA cheaply last year, but it was effin heavy, junk and ended up in the skip – after I rescued an HP80811-60111 OCXO.
So I've been keeping an eye out for an MDA, but mostly they are £600+P&P (on my limit for a toy), often have dents, and even if a display is shown, the self-check screen shows a fault. Not interested.
So, when I noticed an Agilent 53310A where the owner showed it measuring things and also a clean self-test screen, I looked more closely. Since it had the high stability oscillator and third RF channel and was only £250+P&P, I jumped at the chance, even though that mutated into £350 and would take 4 weeks to get here. In the event, it arrived in 2 weeks
I remembered to change both mains voltage settings, and pulled off the lid, and the tube was still intact, so I turned it on, and bingo: I'm a happy camper.
So, what can you do with it? Obviously look at modulated signals up to 200MHz (ch A) or 2.5GHz (ch C), e.g. FM/PSK, or VCO transients as they change frequency. But, since it is fancifully and accurately described as a
frequency microscope you can do much more interesting things than that.
The first task was to look at the output of my OCXO as it turned on and warmed up. The display isn't wonderful, but the top thin band shows a "panorama" of the entire f-vs-t graph captured over 800s, and the lower main section is a narrow window of that display.
The first thing to note is the number of digits in the frequency display, and that the vertical frequency scale is +-200Hz. The OCXO starts off 200Hz (20ppm) low, but after 400s it looks pretty close to nominal, but how close?
Or the same information presented as a histogram, vertical scale is logarithmic.
Looking at the histogram of the stable oscillator shows that (relative to the MDA's internal oscillator) it is 621
mHz (62pp
b low), with an RMS noise of 216mHz. Frequency microscope indeed
But, arguably more interestingly, even with a 200MHz input, it can measure sub-nanosecond risetimes, roughly equivalent to a 1GHz or faster scope.
The technique is to measure connect both channels A and B to the same input, channel A with a threshold at 10%, channel B at 90%, and measure the time difference between the two inputs with a resolution of 70ps.
I've previously measured my 74lvc1g14 step generator as having a risetime of ~250ps. The MDA indicates 370ps (and 200ps for 20%-80%). That's close since it is pushing the limits of the instrument; I need to sit down and fully understood the specification graphs in the manual.
And I'm sure it can do more strange and wonderful things