Post B64XygjhMbD4FAiIr2 by azonenberg@ioc.exchange
(DIR) More posts by azonenberg@ioc.exchange
(DIR) Post #B64W87Y4qQR46DeE64 by azonenberg@ioc.exchange
0 likes, 0 repeats
@petrillic wait till you learn about the cheap symmetricom rubidium oscillators
(DIR) Post #B64WQlfqOlWVl4AkhE by azonenberg@ioc.exchange
0 likes, 0 repeats
@petrillic although it looks like the eBay supply is drying up they're $250 now I think I got mine for less than half thathttps://www.ebay.com/itm/256841284276
(DIR) Post #B64WY1ysAU7uMOXIpc by azonenberg@ioc.exchange
0 likes, 0 repeats
@petrillic https://www.thinksrs.com/products/fs752.html you know you want it
(DIR) Post #B64WfCvNsUkOKRKIfw by azonenberg@ioc.exchange
0 likes, 0 repeats
@petrillic (if you want me to use mine to calibrate your alarm clock to better than e-11 accuracy let me know lol)
(DIR) Post #B64XPyRXXM0J1ffPmK by azonenberg@ioc.exchange
0 likes, 0 repeats
@petrillic Yeah it's not a bad price. And it's well made.I cracked mine open and did a TSP-style teardown video https://www.youtube.com/watch?v=mbS7AiLWte8
(DIR) Post #B64XVb6hDHhUO5bCF6 by azonenberg@ioc.exchange
0 likes, 0 repeats
@petrillic I got a symmetricom rubidium since doing the video and did a similar accuracy comparison between the two. It was much more stable and after trimming I got their 10 MHz outputs to agree within about 120 degrees of phase over ~4 hours
(DIR) Post #B64XnQ5UvCJVQMXLHs by azonenberg@ioc.exchange
0 likes, 0 repeats
@petrillic I had a HDO9204 as my flagship lab scope for quite a while, only selling it when I got the 4 GHz WaveRunner (or maybe when I got the 16 GHz SDA, I forget which purchase it helped fund but one of them).I quite liked it, I didn't use the HD mode feature very often but it's a nice scope. Big display, good performance, and works great with ngscopeclient.
(DIR) Post #B64XygjhMbD4FAiIr2 by azonenberg@ioc.exchange
0 likes, 0 repeats
@petrillic It's a full Windows platform, probably W7 or W10 depending on how old it is and how recently it was serviced by the manufacturer. You should still be able to get service and calibration for it although you'll have to ship it out to New York to get a manufacturer cal or repair which will be a bit pricey.Keep it away from the internet, obviously, but it should be fine on a closed lab LAN.
(DIR) Post #B64YE7Kyk74L5dVZgm by azonenberg@ioc.exchange
0 likes, 0 repeats
@petrillic The LeCroy UI is my favorite on-scope interface I've used, although it seems to be pretty polarizing and some people very much don't like it. I love it.You'll definitely see some inspiration from it in ngscopeclient, although there are plenty of things I did differently because I wanted to put my own spin on things rather than copying, and was optimizing for mouse-and-keyboard operation rather than touch.I know the platform inside out and backwards though, so if you have any questions about features I can definitely help. Or if you want to pop over my lab at some point and test-drive mine to get a feel for the UI before making the purchase, we can probably work that out. It's not the exact same model but the firmware is unified so it'll have the same feel
(DIR) Post #B64YktgWWVgKItDTNo by azonenberg@ioc.exchange
0 likes, 0 repeats
@petrillic yep same here, all my test equipment is on a lab SCADA network that's got no outbound at all, and blocks all inbound except for SCPI and RDP from my workstation vlan
(DIR) Post #B65UyuMAcKvFma57Im by Quantensalat@scicomm.xyz
0 likes, 0 repeats
@azonenberg @petrillic So that's be like what, 2 x 10^-12 ?Sounds pretty optimalWe'll do the same game (GPSDO vs. Rb) at the club, looking forward to it
(DIR) Post #B65UyuZHpZPkRGDakK by azonenberg@ioc.exchange
0 likes, 0 repeats
@Quantensalat @petrillic yeah I recall it being somewhere in the e-12 range but not the exact number.Several more sig figs than i need for practical work, that's for sure
(DIR) Post #B65WnSGQvMXTNV449I by azonenberg@ioc.exchange
0 likes, 0 repeats
@Quantensalat @petrillic I had to tweak the trim pot on the carrier board for the Rb to get it that close, at the default midrange setting it was off by a fair bit.Getting it dialed in was a slooow process, took a day or two, because once you're close you have to make a tiny adjustment then wait like 30 minutes to see if there you need to go further or you overshot lol.Then once you're close, give it a couple hours to measure stability
(DIR) Post #B65WwdwcrjNYD8BIPY by azonenberg@ioc.exchange
0 likes, 0 repeats
@Quantensalat @petrillic I use the FS752 as my primary lab timebase with all of my test equipment synced to it, the rubidium mostly sits on a shelf unused.The long term plan is to send it out to a third party calibration lab to get trimmed as close to perfect as possible then have them issue a traceable calibration cert, then take it back home as a transfer standard and compare to the GPSDO so I can verify it's working properly and in tolerance.
(DIR) Post #B65WzuNuCNE04EUvya by azonenberg@ioc.exchange
0 likes, 0 repeats
@Quantensalat @petrillic like GPS is realistically good enough but there's always the chance of something breaking so if i want the best possible accuracy I need either a second GPS or a traceable transfer standard to check it against
(DIR) Post #B65Y8434FhFaJICiQa by Quantensalat@scicomm.xyz
0 likes, 0 repeats
@azonenberg @petrillic Very interesting I'll report how we fare and what we find
(DIR) Post #B65Y84HxML9z3TAbdQ by azonenberg@ioc.exchange
0 likes, 0 repeats
@Quantensalat @petrillic One thing to keep in mind is that once you get down to this many sig figs in time, you cannot see frequency drift within a single scope capture using the normal "frequency" measurement (plus the scope timebase is not nearly good enough).What you need to do is set the scope to trigger on one of the clocks, the one you're calling a reference, then measure the phase of the test signal compared to the trigger point (averaging over the entire waveform to get better accuracy).Then trend *that* over time, maybe with some averaging, so you can see the beats between the two frequencies. When they're way off you'll clearly see the test clock sliding left or right in phase in the live scope view, but as you get closer in you'll be looking at the slope of the phase trend line integrating over minutes to hours and trying to get it as close to flat as possible, with beat frequencies in the sub millihertz range.
(DIR) Post #B65YcxwF1GFAKxkJnM by Quantensalat@scicomm.xyz
0 likes, 0 repeats
@azonenberg @petrillic Sure, we plan to do just that! Also, we have TinyFPAs which should resolve it
(DIR) Post #B65YcyA4BrIp1qDMLQ by azonenberg@ioc.exchange
0 likes, 0 repeats
@Quantensalat @petrillic Interesting, haven't seen those before.I'm generally a fan of doing as much as I can with as few instruments as possible, so if I can use an oscilloscope to capture the data then do the analysis in DSP (rather than buying a dedicated instrument) I'll prefer that approach unless there are unacceptable tradeoffs in performance/accuracy etc