Rigol DSA-815-TG External 10Mhz Reference problem / Question

[Boatanchor]

I've been trying to inject a OCXO 10Mhz reference signal into the External reference input on the rear of the DSA-815.

Why? Because I can..

As documented, when I inject the 10Mhz signal, the 'Ext Ref' display comes up on the screen.

The problem is that the internal reference is not being disabled. The internal and external 10Mhz references are mixing, which causes distortion on displayed carriers. Not only this, but the displayed carrier's amplitude drops (depending on the displayed span) by 15dB or so.

Has anyone else had these issues, or does anyone know of a fix?
 
I have emailed RIGOLNA, but have not yet had a reply.

Will try and post a screen capture of the distortion soon.

Cheers,
Rick

[pa3bca]

I use a GPSDO as an external 10 MHz reference for my 815-TG (because I can  ). Actually the GPDSO is also used for a siggen and counter.
But I do not see the problems you describe. The ext reference works as expected.
What input level for the external reference do you use? The 815 is not very critical – my 10 MHz is approx 2Vpk-pk.
Minto

[Boatanchor]

I've tried adjusting the ext ref input level from about -20dbm right up to +10dBm and it appears to makes no difference.
The 'Ext Ref' indicator comes on when the external level exceeds about -20dBm.

The problem is most noticeable at narrow spans of <50Khz.

I can definitely see the heterodyning effect of the internal and external frequency references when looking at an unmodulated carrier at these low spans.

It's strange. You would think the internal reference would be disabled completely once an external ref is applied.

Rick

 

[Rory]

I use an external 10 MHz reference on my DSA815-TG as well. Sounds like you need to send it back to Rigol.

[Boatanchor]

OK- Sorted 

Turns out I had some noise on the output of the external reference that was superimposing itself on the displayed signals on the DSA.

Phew!

Thanks to the above posters for posting their observations on their DSA's and their successful use of external references.

Rick

[fqahmad66]

For Rigol SA, what is the main advantage of using external reference?.

[Boatanchor]

Probably only relevant if looking at very, very narrow spans 

[casinada]

To have a more precise and stable frequency source (OCXO, Rubidium, cesium, GPS..) 

[babysitter]

Bet it was not a OCXO but a "great" source like Rubidium, Cesium or GPS - which usually have included high quality noise and spur generators, at least without any care taken to turn a good timing/exact long term frequency source into a spectral clean source.

[Rufus]

For Rigol SA, what is the main advantage of using external reference?.

The DSA815 phase noise performance is poor.

Somewhere here in one of the early DSA815 threads someone did some testing with an external reference which indicated a lot of the phase noise was down to the internal reference oscillator.

As well as improved frequency accuracy an external reference may offer reduced phase noise.

[pa3bca]

For Rigol SA, what is the main advantage of using external reference?.

The DSA815 phase noise performance is poor.

Somewhere here in one of the early DSA815 threads someone did some testing with an external reference which indicated a lot of the phase noise was down to the internal reference oscillator.

As well as improved frequency accuracy an external reference may offer reduced phase noise.

That is very unlikely and also not what I see.
(Phase) noise performance of the 815 is indeed not very good (-88 dBc/Hz at 10 KHz – so next to useless in measuring phase noise fo oscillators and such) but the 10 MHz reference is not a factor. Even a simple xtal oscillator easily does better than -140 dBc/Hz @ 10 KHz. The lousy noise performance comes from the VCO of the 815. I see no difference between the internal reference and my 10 MHz GPSDO (with a 10 MHz OCXO in it).

[Mr Simpleton]

Ho-hum... given a XO with -140 dB/Hz at 10 kHz... and this is the reference to which a VCO running at say 3-4.5 GHz (assuming high IF at 3 GHz).
Multiplying the 10 MHz up to 3 GHz equals 300x or 20*log(300) ~50dB degradaton i.e. -140+50=-90 dBc/Hz.
Now this is very simplistic and several things would likely change but I just want to show how important the reference is!

Key to low noise is low multiplication and high performance reference plus VCO. But you can do miracles with the new fractional N chips from Analog devices and Hittite, especially for frequencies below 1 GHz