XRef - 10MHz-Locked Reference Oscillator

[VK2GOM]

Absolutely. G4HUP, MiniKits, and VK2XTT already do solutions for precisely that.

73 - Rob VK2GOM / GW0MOH

[VK3HZ]

In fact, no tripling required. Square the 10 MHz up and couple it into the oscillator coil and it will lock.

However, when the 10 MHz is unplugged, you're back to the internal oscillator which wanders around like Brown's cows.

[VK2XAX]

In fact, no tripling required. Square the 10 MHz up and couple it into the oscillator coil and it will lock.

However, when the 10 MHz is unplugged, you're back to the internal oscillator which wanders around like Brown's cows.

That, of course, assumes a coil in the LO to couple to. IC-910's have no such thing so direct connection and/or replacement of the internal LO is required. Injection of an unfiltered signal in to the LO, produces a very dirty output on TX and severely compromises RX .

The IC-7400, on the other hand, has a coil that can be coupled to. I'm about to lock that via injection locking to a Rubidium Standard but i've still gone to lengths to keep the signal clean prior to doing so.

In any case, when I get my hands on a couple of these Xref's I'll be doing A/B comparisons with my test bench to see what wins and then use the best option.

regards

Tim

[VK3HZ]

I was just suggesting coupling 10 MHz into the coil as a simple mod for the 706 (Later models have a can over the coil so it might not be so easy).

The Xref signal is intended to be capacitively coupled into the rig, normally at the output point of the internal reference oscillator, after the rig's own oscillator has been disabled by cutting power to it. Of course, this varies from rig to rig and will require further investigation for some.

[VK4CRO]

hi Our club vk4wie would like to order 1 for our 817nd that we use for 10ghz and 1296
this will be a big improvement ,
many tks on vk4cro

[VK3ES]

G'day David and Graham,

This is a beaut project and I'll be considering one for my FT-897. Have you measured or simulated the phase noise performance of the synthesiser? I would imagine that the phase noise of the original TCXO in the radio would be considerably better than any synthesiser could achieve? Not sure if this will be a problem for any amateur applications though. One possible impact could be reduced performance on HF where there are many strong signals close together?

Just thought I'd raise the issue as it may be a case of giving up some phase noise performance for much better frequency accuracy and stability - something users should be aware of.

73,
Andy VK3ES

[VK2XAX]

G'day David and Graham,
Have you measured or simulated the phase noise performance of the synthesiser?

Hi Andy,

And that's the first test I'll be doing when I get mine.

I have the test bench to test phase noise and I've already measured my IC7400 an IC910, so I'll certainly publish my results when done.

I'll be testing my FT-817 tonight to see what out-of-the-box behavior is too, its only 3 months old so I hadn't gotten around to it just yet.

Besides phase noise, I'll also be testing all my radios for frequency stability, harmonics, SNR, MDS to see exactly what differences occur.

regards

Tim

[VK4RF]

Sounds good too me for my IC 910H as well

[VK3HZ]

Hi Andy,

The phase noise results for the XRefFT board and for an FT-817 when fitted with the board will be put up on the web site once the filters are sorted out.

An important point is that the 10 MHz reference needs to be clean otherwise the old adage applies - garbage in, garbage out.
To this end, a 4-channel, 10 MHz distribution board with inbuilt filtering is being produced.

Regards,
Dave.

[VK3HZ]

After a small delay obtaining some bits to finalise the filters, the XRef-FT board for the FT817/857 and 897 is now complete.

Thanks to Andy VK3ES, we've been able to do some Phase Noise tests of an FT-817. We considered testing just the XRef-FT, but felt that the end result is more important / relevant.
Testing was carried out with the same FT-817 fitted with the original (OEM) reference oscillator and then with the XRef-FT. The XRef-FT was operated from its onboard TCXO and from an external 10 MHz reference from the Spectrum Analyser.

The test setup used an HP70000 series Spectrum Analyser driven by the KE5FX Phase Noise measurement utility.
Unfortunately, some of the test results weren't saved correctly by the software so we'll have to redo some of the plots. Nevertheless, there's enough to see what's happening.

70 cm is likely to be the band most affected, so here are plots of the phase noise operating on the external 10 MHz reference and on the board's TCXO:

432.1 MHz XRefFT with TCXO and External Reference

Unfortunately, the data for the Yaesu OEM reference oscillator didn't save correctly so I can't combine it with the above data. However, here is a screen dump from the day:

432.1 MHz Yaesu OEM Reference

At 1kHz offset, all oscillators give the same result.
At 10 KHz offset, the OEM is best at -102dBc/Hz, 10 MHz about 5dB worse at -97 and TCXO another 5dB worse at -92.
At 100 kHz offset, they are all back to about the same level.

You can make up your own mind, but in my humble opinion, these appear to be very good results and, I have to be honest, exceeded my expectations

[VK3HZ]

We also tested on 40 m to see if the HF side of things was OK.

7.1 MHz OEM and XRef-FT on External Reference

The results seem to mirror those for 70 cm with the XRef-FT showing a rise of about 5dB in phase noise at around 10 kHz, compared to the Yaesu reference, but otherwise quite close.

[VK4GHZ]

Dave, interesting plots.

In using a common 10 MHz reference for both the instrument plotting the results AND being the reference source for the device under test, does this dilute the XRef's Externally locked results at all?
(If only by the 3dB, then I guess that would be insignificant in the bigger scheme of things.)

Unless I have missed it, what is the Phase Detector frequency of the XRef?

[VK3HZ]

We also took some screen dumps from the Spectrum Analyser on 70 cm and 40 m over various spans. Neither results showed anything untoward, so in the interests of saving Adam's storage space, I'll only put up some of the 70 cm plots:

OEM Reference - 2 kHz Span

XRef-FT on External Reference - 2 kHz Span

XRef-FT on TCXO - 2 kHz Span

This backs up the Phase Noise plots showing the OEM and XRef-FT results as almost identical. The TCXO is a little noisier.

BTW, the OEM oscillator was slightly off frequency and they are very tricky to adjust. The XRef-FT onboard TCXO is also off frequency, but it has a voltage adjustment pin (it's really a TCVCXO) and there's a 10-turn trimpot so it's easy to adjust to within a few Hz (which I should have done!)

[VK3HZ]

Hi Adam,

In using a common 10 MHz reference for both the instrument plotting the results AND being the reference source for the device under test, does this dilute the XRef's Externally locked results at all?
(If only by the 3dB, then I guess that would be insignificant in the bigger scheme of things.)

I don't believe so. It just eliminates one source of error - any frequency difference / drift between two 10 MHz sources.

Unless I have missed it, what is the Phase Detector frequency of the XRef?

The PD runs at 1 MHz with the output divided by 32 so any spurs would be at 31.25 kHz. Didn't see any evidence of them.

Regards,
Dave.

[VK3HZ]

Some 200 kHz span plots (Not sure what's going on, but the images have changed format for some reason):

OEM Reference - 200 kHz Span

200kHz spectrum 432.1MHz OEM.gif (11.62 KiB) Viewed 4646 times

XRef-FT on External Reference - 200 kHz Span

200kHz spectrum 432.1MHz.gif (11.63 KiB) Viewed 4646 times

[VK3ES]

I think the test results speak for themselves. Happily, my concerns about synthesiser noise possibly affecting HF performance were unwarranted (that's assuming RX behaves the same as TX, as we expect). Graham and David have done a great job with this design - well done guys! I'll be putting one in my '897 for sure.

73,
Andy VK3ES

[VK3HZ]

We've now set a price of $140 per XRef-FT board, plus $3 postage for up to 3 units within Australia.
Please email me at xref@vk3hz.net if you are interested.

Note that this is for the FT-8x7 version of the XRef. The Versatile version is still a few weeks away.

Regards,
Dave.

P.S. Sorry for the incorrect email - it's now been corrected

P.P.S. I'll be in touch shortly with those who have already contacted me by email.

[VK2XAX]

Hi Dave,
Just attemped to email you for one of these units an I got this error.....

Code: Select all

<contact@vk3hz.net>: host mxd1.myobnet.com[10.194.20.133] said: 550 5.1.1
    <contact@vk3hz.net>: Recipient address rejected: User unknown in virtual
    mailbox table (in reply to RCPT TO command)

regards

Tim

[VK3HZ]

Sorry Tim, try again with the correct address (above).

Regards,
Dave.

[VK2XAX]

Hi Dave,

It arrived today, nice work on such a tiny device.

here's a couple of pics of my 817 with unit the xref attached and SMB added to the back panel.

Bench tests to happen soon

regards

Tim