@mio83
Here's a couple of video clips which might ruin your day (or maybe only my day
).
The jpg is showing the AWGs' settings.
The short 5 second clip shows the 10MHz sine wave output from the SDG1032X on CH2 with CH1 triggering the 'scope from CH2 of the FY6600 which is locked to my 10MHz GPSDO reference. CH3 shows the free running RFS and CH4 shows the FY6600's 10MHz Sinc pulse from its CH1 port.
The SDG1032 is using its internal 50ppm smd XO (25MHz? as in the SDG1025 or is it a 10MHz XO chip?) which is the cause of the flicker noise like jitter (checking the reference output reveals the same jitter). The jitter disappears completely when using the external 10MHz GPSDO reference (but this reveals a +1.43mHz offset which equates to an error of one second per 45 years versus the one second error in
1.8 million years of the FY6600
).
The SDG1032X is a brand new unit that arrived just three days ago. I've posted these movie clips into a couple of other SDG1032X related topic threads to get opinions/evidence as to whether this is just a rogue defect or a more general issue due to Siglent's use of cheap crappy 50ppm smd XO chips as per Feeltech's use of a 50ppm 50MHz one in their FY6600 and FY6800 models (the later FY6900 models use a 10MHz 20ppm smd XO chip).
My hope is that someone will be curious enough to repeat my tests and risk having their day spoilt by the result. I'm curious to know whether, as in the case of my FY6600, this is an inescapable consequence of the use of a cheap smd XO chip or just my bad luck to get a faulty unit. I'd seen similar jitter before upgrading the crap XO chip used in the FY6600 to a 50MHz 0.1ppm TCXO when trying to match its output frequency to some DIP14 XOs in an exercise not unlike the game of "Chase Will o' the Whisp" trying to keep both waveform traces matched in frequency for more than 5 seconds at a time).
After this first oscillator upgrade, this game seemed a little easier to play. Having subsequently upgraded to a CQE branded (as used in some Symmetricom kit) 10MHz OCXO with a 3N502 clock multiplier sat where the crappy XO chip used to reside to multiply this up to the 50MHz used by the FPGA, the only issue with matching the frequencies of DIP14 XOs is their temperature driven deviation which eventually settle down if sheltered from random breezes in my "Lab".
This jitter effect is very disturbing since it makes a total mockery of Siglent's low jitter claims of 300ps (an RMS figure BTW rather than the less misleading Pk-Pk value of 900ns compared to the 200 to 250ns Pk-Pk jitter visible in 'scope traces of the FY6600's Sine and Sinc pulse wave forms).
Quite frankly, as much as the SDG1032X (SDG1062X) is an improvement over not only the FY6600 but the SDG2000X series in the square wave performance department along with other desirable features absent from my much modified FY6600, I'm minded to return it as "Not fit for purpose" by way of this oscillator defect and put the cash back into my account.
The longer 1 minute video clip is demonstrating yet another shortcoming of the SDG1032 over the FY6600 with regard to the stability of the Sinc pulse. In this case, I'm comparing a 6MHz pulse (maximum frequency limit) from the Siglent against a 12MHz pulse (not a maximum limit) from the Feeltech. I won't bother describing this clip (a picture is worth a thousand words and, in this case, this movie clip is worth some 1800 pictures).
I'll post another video clip which may demonstrate the jitter issue a little more effectively - I've hit the 5000KB limit with these attachments, hence the next posting to show another 3.7MiB 20 seconds long clip. You'll have to download them and either remove the ,zip extension or else right click and select your favourite media player to open them in (EEVBlog won't allow you to attach video clip files, hence the .zip filename extension to bypass this file type filter).
Regards, John