List of test equipment with external 10MHz reference

[picitup]

@cncjerry thanks for the update.  Hopefully this list will be helpful to others.

My SDG1025 arrived today, despite the rather long timescales for delivery suggested by the seller which was between 26th Jan and some time in Feb so I'm pleased.

I like the SDG1025 which I find very intuitive to use.  I found most of the functions just by playing around.  I was hoping for the frequency counter to have a higher resolution, but it seems to mostly have 6 digits after the decimal point, offering a resolution of 1Hz at 10Mhz.  So I shan't be ditching my TF830 counter/timer which is 10 times better at 0.1Hz.  Could be wrong as I'm still figuring the unit out.  To access the frequency counter, you press Utility, 1/3, 2/3 and press counter.

The 10MHz ref is at the back and to enable it, you press Utility, 1/3, System, 1/2, and CLKSource.  It would be nice if the display showed you which clock source it was using.  It's a bit of a faff to go through the menus to check.  My TF830 displays "Ext" on the LCD which I find helpful.  I think I'll pop this into Siglent as a feature request.

Anyway, this isn't supposed to be a review of the SDG830 so the interesting bits as far as the 10MHz ref go are as follows:

The output of the BG7TBL works fine driving both my TF830 and the SDG1025 directly with no amplification, just a 50 ohm terminator on the end of the chain to stop those reflections 

Just as a sanity check, I connected up the 10MHz to the TF830 and SDG1025 and generated a 10MHz sine wave which was fed back into the TF830.  With a 10 second gate time and corresponding 0.1Hz resolution, the display was bang on the money (carry + 0.0000000Mhz) and rock solid so that's a good start.

Well after all this messing around with standards, for some reason I really feel the urge to drive my £1.00 quartz clock from the pound shop from the GPSDO.  But that's another thread.

Thanks for reading and all those that have contributed

Steve

[B1nary]

Keysight InfiniiVision 4000 X-Series Oscilloscopes
Keysight InfiniiVision 6000 X-Series Oscilloscopes
Agilent Keysight InfiniiVision 7000A & 7000B Oscilloscopes
All current Keysight Infiniium Oscilloscopes

Having read through these posts and recognising that my 3104T does not have a external 10mHz reference I wondered if it was even possible to add an external frequency reference to the 3104 or other scopes. More curious than serious. 

[usagi]

has 10mhz reference input:
R&S CMU 200
Agilent E7495 series
Agilent N9342 series

[edpalmer42]

Keysight InfiniiVision 4000 X-Series Oscilloscopes
Keysight InfiniiVision 6000 X-Series Oscilloscopes
Agilent Keysight InfiniiVision 7000A & 7000B Oscilloscopes
All current Keysight Infiniium Oscilloscopes

Having read through these posts and recognising that my 3104T does not have a external 10mHz reference I wondered if it was even possible to add an external frequency reference to the 3104 or other scopes. More curious than serious. 

If you're willing to work hard enough, it's probably possible to add an external reference input to almost anything.  You'd have to find the appropriate clock inside the unit and then figure out how to get from 10 MHz to the frequency of that clock.  Typically, the solution will be a PLL and/or synthesizer.  Whether the result is worth the effort will depend on many variables.

[TheSteve]

Keysight InfiniiVision 4000 X-Series Oscilloscopes
Keysight InfiniiVision 6000 X-Series Oscilloscopes
Agilent Keysight InfiniiVision 7000A & 7000B Oscilloscopes
All current Keysight Infiniium Oscilloscopes

Having read through these posts and recognising that my 3104T does not have a external 10mHz reference I wondered if it was even possible to add an external frequency reference to the 3104 or other scopes. More curious than serious. 

The reference in the 3000T series is pretty decent, so there likely isn't too much value in an external reference(if it had it I'd still use it with my 3000T scope though).

[tautech]

Keysight InfiniiVision 4000 X-Series Oscilloscopes
Keysight InfiniiVision 6000 X-Series Oscilloscopes
Agilent Keysight InfiniiVision 7000A & 7000B Oscilloscopes
All current Keysight Infiniium Oscilloscopes

Having read through these posts and recognising that my 3104T does not have a external 10mHz reference I wondered if it was even possible to add an external frequency reference to the 3104 or other scopes. More curious than serious. 

has 10mhz reference input:
R&S CMU 200
Agilent E7495 series
Agilent N9342 series

Please.....yes we know you mean Mega not milli but please set an example on an engineering forum by using the correct abbreviations ! !
Laziness is no excuse. 

[macboy]

HP 8657B 2GHz Signal Generator has a 10 MHz reference input (also accepts 1 Mhz or 5 MHz). It also has a reference output. The high stability option adds a few inches to the depth of the unit, with an ovenized reference having a separate 10 MHz output which must be externally connected to the main reference input. The main reference output tracks that input and can be used to supply an additional instrument.

LeCroy WavePro 960 oscilloscope (2 GHz, 16 GS/s) has 10 MHz reference input.

Would be very nice if the OP would maintain a list in the first post. Otherwise, the thread misses it's point.

[Tony_G]

I love my 8657B - It came with the high stability timebase but I use it on an external 10MHz clock anyway.

Personally it seems a bit silly to list all the gear that takes an external 10MHz reference because just about every professional signal generator (now available on the used market) has a reference input - Just in my lab alone I have a 3325B, 3335A, 5351A, 8657B, 8340B, 8673B, 8902A, 8903B, 8563E, 8714C, E4406A, E8285A, 8672A, DG1032A, 53131A and I'm probably missing mentioning some other pieces.

That said, the one thing I haven't seen people mention, and it may be of use to the OP, is a distribution amp for the reference signal - A lot of people use old Extron RGB distribution amps (some even modify them down from 75ohm to 50ohm), there are several kits (TAPR TADD-1, etc) - I use a Symmetricom 4036B which I picked up from a Go-Dove auction.

TonyG

[Leo Bodnar]

Some LeCroy LC series LC3**/LC5**/LC6** and DDA-120/DDA-125 scopes have optional 10MHz reference input.
I have DDA-125 with that option.

[G0HZU]

It's nearly xmas and I'm considering buying a 10MHz GPSDO but am struggling to find kit with an external 10MHz reference input.  It would be great to have this accuracy and sync my kit.

When wishing to lock high performance RF test gear to a common reference I think it's best to choose a 'master' piece of test gear that has the best internal 10MHz OCXO in terms of noise and stability and use that as a reference to connect all the other gear to.

I'm not sure it is always wise to use a GPSDO for this. The extra accuracy will usually be irrelevant and the risk of adding excess noise or spurious terms could be significant. Even if somebody gave me your BG7TBL for free and it was working perfectly I doubt I would ever want to use it as a master reference here for all of my RF test gear because I'd expect it to be too noisy.

However, it would be fine for use as a reference for a decent frequency counter if you wanted to check/adjust the reference OCXO in a decent spectrum analyser or sig gen.

[SoundTech-LG]

Yes, and even the Dinosaurs of yesteryear TEK 2465B 2467B had options for it...

https://www.eevblog.com/forum/testgear/tektronix-2465b-oscilloscope-teardown/msg600790/#msg600790
 

[SoundTech-LG]

and what the 10M Ref will do for you...

https://www.eevblog.com/forum/testgear/tektronix-2465b-oscilloscope-teardown/msg620497/#msg620497

[Tony_G]

I'm not sure it is always wise to use a GPSDO for this. The extra accuracy will usually be irrelevant and the risk of adding excess noise or spurious terms could be significant. Even if somebody gave me your BG7TBL for free and it was working perfectly I doubt I would ever want to use it as a master reference here for all of my RF test gear because I'd expect it to be too noisy.

I measured mine and the STDEV of the period is 7.6e-18s and the 20/30MHz harmonics are -70dB down and rest are close to -90dB down. General spurs out to 2 GHz were at least -50dBc.

Your mileage may vary.

TonyG

[G0HZU]

I'm not sure it is always wise to use a GPSDO for this. The extra accuracy will usually be irrelevant and the risk of adding excess noise or spurious terms could be significant. Even if somebody gave me your BG7TBL for free and it was working perfectly I doubt I would ever want to use it as a master reference here for all of my RF test gear because I'd expect it to be too noisy.

I measured mine and the STDEV of the period is 7.6e-18s and the 20/30MHz harmonics are -70dB down and rest are close to -90dB down. General spurs out to 2 GHz were at least -50dBc.

Your mileage may vary.

TonyG

It's the impact on close in spurious and phase noise on a (synthesised) high frequency local oscillator that I would be most concerned about.

For example, the oldest bit of RF kit on my bench right now is an HP8566B analyser and the first LO typically runs at over 4GHz. So in terms of 20logN that's maybe 53dB wrt 10MHz.

So any close in phase noise and spurious on the reference will rise by 52dB at 4GHz. I think the 10MHz OCXO inside the analyser has a phase noise of close to -158dBc/Hz at 1kHz offset. So it's contribution to overall phase noise on a 4.4GHz LO will be 158-52 = -106dBc/Hz at 1kHz. It's probably better than this in reality.

But a typical mid range GPDSO OCXO might only achieve -140dBc/Hz at 1kHz meaning the noise contribution will be -88dBc/Hz. The synthesiser loop in my 8566B typically manages to produce a combined divider + reference noise level of about -97dBc/Hz at 1kHz offset (at a 300MHz test freq) so you can see why it's important to have a reference that is maybe 10dB cleaner than this. At a 1kHz offset a 10MHz reference in a mid range GPSDO would be maybe 15dB worse than the OCXO in the 8566B in this respect.

The same applies to close in spurious. If there are spurs close to carrier, eg at -100dBc on the reference they can rise by 52dB to look huge on the analyser display at -48dBc on a 4.4GHz local oscillator in the analyser. But a typical 5kHz span on an HP8566B when run from its internal reference will show no close in spurs down to the -80dBc limit on the display where the loop noise dominates at a 30Hz RBW.

There seem to be lots of spurs at 1kHz offset in the plot of the GPSDO here:
http://www.ke5fx.com/gpscomp.htm

I don't know if these spurs are AM or FM or if they are really there or not and maybe this will depend on the quality of the OCXO fitted to the GPSDO and the rest of the PCB layout/design but there's enough there in terms of noise and spurs to put me off wanting to swap the clean OCXO in the 8566B for that GPSDO.

[Tony_G]

Not disagreeing, just adding some measurements from my unit in case it's helpful for people to understand what you're saying (though it is good enough for my purposes).

Here is a phase noise plot - my environment is a bit noisy right now but hopefully it will help:



TonyG

[G0HZU]

Thanks.
Just to clarify, was that taken of a clean 10MHz test signal with the analyser running its reference externally from the GPSDO?

It would be nice to see the GPSDO when fed to something like an E5052A SSA. I have access to these analysers at work but I don't have anything here that can measure the phase noise of a fairly reasonable OCXO. The nearest I have is a Tek analyser. It can just about measure a low cost, low performance crystal oscillator close to carrier but nothing as exotic as an OCXO.

[Tony_G]

No that was taken with internal OCXO on the unit - Apart from an E4406A which I'm still waiting for a spare part, that is the best I can do here at home.

TonyG