SDRPlay TCXO?

[jazz57]

Hi there,

I couldn't find any information on the oscillator used on the SDRPlay? What is its temperature stability and maximum offset? Is it a TCXO?

Thanks!

[Sideteacher]

We do not use a TCXO in the design. We use as crystal the main reference. The tolerance over it's working temperature range is +/-10ppm. The crystal itself is fixed in frequency. Any frequency offset applied in the plugin is applied using the fine frequency trim within the tuner. Hope that helps.

[jazz57]

Yes, that definitely helps understand how the SDRplay works. I'm curious, is there a technical reason why a TCXO can't be used in your design or would that cause a significant increase in price?

[usagi]

is there a way to retrofit it to accept external clock? i have a nice gpsdo with stable output frequency. +/-0.1ppm

[Tech_Support]

Hello Usagi,
It is possible, but it requires modification of the PCB. To do this, you would need to remove C65, C66 and X1 from the PCB and apply the new clock signal to Pin X0 of the MSi2500. Care would be needed with shielding as any noise pick up on this clock line could result in a serious degradation in performance. The required clock frequency is 24 MHz.

For schematics, please see: http://www.sdrplay.com/docs/SDRplay_RSP ... s_r2p3.pdf

I should stress that this is NOT something that we would recommend unless you are both experienced and very handy with soldering equipment and that anyone who does this, is doing so at their own risk.

Sincerely

SDRplay Tech_Support

[g1hbe]

One thing that puzzles me is that there seems to be a discrepancy in the exact tuned frequency depending on 'where' you do it. If I trim the oscillator (via the EXTIO panel) with the test frequency at one end of the displayed spectrum, it is a few tens/hundreds of Hz out when I check it at the other. I know this doesn't really matter as the error can be tuned out anyway, but can you tell me why this is happening? Thanks.

[Tech_Support]

It is not really clear what might be going on here. If you have not already done so, try ensuring that the reference signal that you are using to trim the frequency error is centred on the receiver bandwidth and that you are not selecting it with a VFO. That way, you know that the displayed frequency is what the synthesizer is set to.

Sincerely

SDRplay Tech_Support

[g1hbe]

Thanks for the quick reply. Here's what I do:
1. Set sig gen (Marconi 2024) to 109 MHz.
2. Start SDR#. Set the tuning cursor to somewhere on the left of display.
3. Dial up 109.000000 MHz on SDR#.
4. Either - select SSB and trim oscillator for zero beat OR select AM and trim osc for peak in centre of IF passband.

Now I do exactly the same again, but with 109 MHz positioned to the RH side of the display. When I do this, there is an audio beat of a few tens or hundreds of Hz/the peak in AM mode is off centre.

Am I doing something wrong? I'm not sure what you mean by 'not selecting with a VFO'. I'm an analogue type who was brung up on knob-tuned radios. Like I said it doesn't make a jot of difference really as the error can be tuned out, but I'm the curious sort....

[Tech_Support]

I think we will have to investigate this further. It will take us a little while as our apps engineer is away on vacation until the end of this month.

Please bear with us whilst we look into this a bit deeper.

Sincerely

SDRplay Tech_Support

[g1hbe]

Ah - I've been having a prod around and discovered that the xtal osc is running at 24.006 MHz. As this osc provides the reference for the synth(?), I'm assuming this could be the problem. Most likely c65 or c66 has died, probably as a result of my rather rough handling when re-housing the PCB in a box with some additional front-end filters. I'll do more later on today.