There are quite a few of these on eBay again, but for a lot more than the original $20 (looking at $350 minimum) - 3he14833aara01 for sale | eBay
Nice,I did order one from the 'bay that should arrive tomorrow. I'll have some fun with that!
Dutch guy, but I live in Switzerland. I paid around the $270 mark.Nice,
As you are from Switzerland, where did you order? I am located in Germany. And did you pay the 600-1000$ or around 20?
That's a good point. USB polling frame time is 1ms and USB 2+ has roughly 0.125ms jitter. From my personal use case, my avg offset value in chrony is 1.3-16ms (not exact number and it's 5 significant digits long). But a rough check after sync with Time.is - l'hora exacte, a qualsevol zona horària, my clock is +0.001s ahead. Which, personally, is good enough and it goes to your point --- it depends on what is good enough for your usecase or preference.I suspect that this boils down to what "pretty accurate Stratum 1" means to you. The GPS SFP that started this has an OCXO and presumably is good to ~50 ns or so (limited by GPS error, most likely). I don't have a good instinct for how accurate a USB GPS module would be, but I'd guess it'd be in the hundreds of microseconds range. So, good enough for a lot of uses, but ~1000x as much error.
I should probably just order one and test it, because it's an interesting question, but there are lots of interesting questions in this space, and not that much time.
Watch out, I think those are Point-To-Point, not Precision Timing Protocol.I have also recently found these available at a reasonable price, haven't looked for specs though: NEW MEMORYlink ULTRASYNC KIT PTP-600 Series Version GPS-100 | eBay
It's all a matter of what you're trying to accomplish and how accurate you need to be. It's perfectly possible to do better than this with Chrony, though -- here's a random server at home talking to 4 local stratum-1 time servers:That's a good point. USB polling frame time is 1ms and USB 2+ has roughly 0.125ms jitter. From my personal use case, my avg offset value in chrony is 1.3-16ms (not exact number and it's 5 significant digits long). But a rough check after sync with Time.is - l'hora exacte, a qualsevol zona horària, my clock is +0.001s ahead. Which, personally, is good enough and it goes to your point --- it depends on what is good enough for your usecase or preference.
Name/IP Address NP NR Span Frequency Freq Skew Offset Std Dev
==============================================================================
ntp-desktop.internal.sig> 32 7 19 -0.000 0.005 +179ns 44ns
ntp1.internal.sigkill.org 32 15 19 +0.001 0.030 +283ns 254ns
ntp2.internal.sigkill.org 32 10 19 +0.000 0.010 -244ns 88ns
ntp6.internal.sigkill.org 32 5 20 -0.000 0.005 -0ns 44ns
It's all a matter of what you're trying to accomplish and how accurate you need to be. It's perfectly possible to do better than this with Chrony, though -- here's a random server at home talking to 4 local stratum-1 time servers:
Code:Name/IP Address NP NR Span Frequency Freq Skew Offset Std Dev ============================================================================== ntp-desktop.internal.sig> 32 7 19 -0.000 0.005 +179ns 44ns ntp1.internal.sigkill.org 32 15 19 +0.001 0.030 +283ns 254ns ntp2.internal.sigkill.org 32 10 19 +0.000 0.010 -244ns 88ns ntp6.internal.sigkill.org 32 5 20 -0.000 0.005 -0ns 44ns
Name/IP Address NP NR Span Frequency Freq Skew Offset Std Dev
==============================================================================
GPS 8 4 57 -286.039 553.528 +36us 4908us
PPS 9 8 63 +0.329 2.528 +1623ns 28us
ntpi.home 2 0 65 -0.004 2000.000 -156us 4000ms
india.colorado.edu 2 0 65 -0.004 2000.000 -197us 4000ms
time.cloudflare.com 5 3 71 +0.269 4.167 -172us 13us
time1.google.com 5 4 71 -0.344 4.564 -207us 15us
Most SFP modules run at least kind of hot, but I happen to have a fan pointing at my switch to cool the 10GBaseT modules and haven't noticed this one getting too hot (not that I really check.)Are these supposed to run super hot? Do you guys add a fan to blow cool air to it?
I tried this, and although the thing indeed starts responding at 192.168.0.2, and I can even start connect to it on port 22, it refuses to accept 123/123 as credentialsforgot to add here - user @Hroghtar PM'd me some instructions that were found in the manual:
The OSA 5401 unit is always accessible through a Telenet/SSH using ... 192.168.0.2 without VLAN ... for 60 seconds after restarting... Use default username 123 and password 123. An SSH session is open for 5 minutes
![]()
OSA 5401 User Manual 12.1.1 | PDF | Computer Network | Ip Address
Scribd is the world's largest social reading and publishing site.www.scribd.com
I suspect it doesn't have the processing power to do anything more complicated/modern, though it's supposed to be installed on a fairly secure network so it's almost a surprise it doesn't just use telnet.Getting a modern SSH client to connect to it is kind of a pain though; the device uses very old and deprecated key-exchange and other crypto options.
I followed the instructions in the manual to reset the password on mine and it worked fine. Maybe try telnet instead of SSH. Kinda huge security hole but, uh, probably shouldn't put it on an untrusted network anyway.I tried this, and although the thing indeed starts responding at 192.168.0.2, and I can even start connect to it on port 22, it refuses to accept 123/123 as credentials
Getting a modern SSH client to connect to it is kind of a pain though; the device uses very old and deprecated key-exchange and other crypto options.
The OSA 5401 unit is always accessible through a Telnet/SSH client using the default IPv4 IP
address (192.168.0.2) without a VLAN connection for 60 seconds after restarting the system.
You can reset the IP address, username, and password of the OSA 5401 unit using this
connection. Once the system is up and running, the unit is accessible for 60 seconds
provided that the connection is still maintained. You then have five minutes to enter any CLI
command, such as SSH username or password reset, before the connection to the
Telnet/SSH client expires (see CLI Command Definitions).