Adalm-Pluto Rev D und PTT

  • Hallo David,

    I remember a Chinese offer. Look for an OCXO at Aliexpress. It has a 10 Mc/s masterclock and other frequencies, especially 40 Mc/s can be selected by jumpers. It is cheaper than a GPS controlled Oscillator but works as expected. Some friends use it and their signals are stable.
    I am not sure if a portable equipment at different temperatures should have a Leo Bodnar better.
    73

    Andreas

  • Hi,


    look at "Where can I buy a TCXO", I tested the Chinese OCXO


    I've used the OSC101 as external clock. It is completely useless. I transmitted on 2.4 GHz with the internal clock (image left) and with the OSC 101 as external clock (image right).


    Detlev


     

  • Epson TG2520SMN 40.0000M-ECGNNM3


    https://www.dk0tui.de/2021/08/…cht-richtig-funktioniert/

  • Hallo,

    it is not recommended to mount a better tcxo on the board again. It's temperature increases while transmitting. Use some small short pieces of wire stick all trough a nearby hole and let the TCXO flow below the board. The original blue box is not senseful because of thermal reasons also.


    This Pluto is mounted on a copper plated board, the above cover is Acryl with a small, slow running fan. The sensitive connectors are fixed and grounded with thin tin plated metal pieces which gives a better stability.

    73
    Andreas

  • Thank you Andreas Detlev and Frank.


    At the moment my Pluto takes 15 minutes from cold to stabilise on 739MHz (FT8 decode). I can imagine 2.4G takes even longer.

    I will have a try and replace the oscillator with a TCXO. Then go /p. Opening and closing the car door is a good test.


    Thanks for the suggestions. I'll let you know how I get on.


    73 David

  • You may not replace the TCXO because it is not easy to remove it without damaging the PCB. Place the new one on a separate (temperature isolated) small piece of PCB near by and disable original TCXO by EN pin (check schematic).

    Good luck...

  • Hallo,

    the Pluto boards are multilayer and have higher thermal capacity. With two solder irons and some low melting tin it will be possible. With hot air soldering a mask of aluminium foil around the oscillator is recommended to avoid blowing other nearby components away.

    73
    Andreas

  • You can easily remove it with Chipquick if you want to, otherwise disable it. Then you can fit a TCXO on the board and make connections. Note it is on the underside in the Rev D. However, much better to use an external oscillator, you could always put the whole lot in another screened box.

  • Hello all.
    Two months on from my original post and I have learned a lot about oscillators and the Pluto Ver D

    I am currently midway through my mission to make the Pluto stable, on my 3rd oscillator and as yet, it is still not suitable for narrowband modes, but I am making progress. The following info (saga) may be of interest.


    My objective is to make the Pluto stable enough for narrowband digital modes so I can keep it compact and take it /P without having to use my Leo Bodnar external clock. Original issues were the crystal supplied is 25ppm and drifts. Let the saga begin:


    1) Original 40MHz 25ppm crystal oscillator fitted by Analog Devices. These drift both in the short and long term. On receive only in 'the shack' they stabilise after about 15 mins and can be useful for long term monitoring of the digital signals on .540.
    Unfortunately, when you go to transmit, the board heats up and the oscillator drifts making TX signals unreadable and when you go back to RX, decoding (at 739MHz ) is impossible as the oscillator cools down. On the rev D this oscillator has an enable and disable pin.
    This oscillator had to go. Surprisingly, using 2 soldering irons, the oscillators can be removed in 10 seconds. Remember if you are extracting the oscillator, you don't have to worry about killing it. A large blob of solder on each side and 2 irons will melt the solder on all 4 pins at the same time and the oscillator will lift out in seconds. No hot air or Kapton required but a microscope or magnifier is an advantage to see if the solder is molton.


    2) Purchased 2 Epson 40MHz TCXO from Digikey. Part number TG2520SMN 40.0000M-ECGNNM3. These are 500ppb (0.5ppm) 1.8V oscillator modules and were suggested as suitable replacements. They are slightly smaller than the original part, but were fitted fairly easily. As it was late in the evening, the Pluto was switched on and FT8 signals from .540 were quickly seen and began decoding after a minute or so. I watched for about 10 minutes, all seemed well and I went to bed. The next morning I could see a page full of decodes but they had stopped around 7 hours after I started the Pluto / WSJT-X. The waterfall on SDR Console contained no signals from QO-100 and a reboot of the system gave a message about possibly insufficient USB power or a non-functioning oscillator. As the USB power was fine I checked the new oscillator and found that it had failed after 7 hours.


    3) After looking at the circuits for Rev B and Rev D Plutos, I found two possibilities for the oscillator failure. The first is that A.D. have fed DC to the enable pin on the rev D. This was measured as 1.8V. However, the Epson data sheet is clear that there is no enable function on the new oscillators and that Pin 1 should be left floating or be connected to ground. - It should not be at 1.8V. I removed a 1k resistor from the board to make pin 1 'float'. The second possibility for the failure could be the extra circuitry on the rev D board. On the rev D, the oscillator drives the AD9363 but is also connected to U17, the external clock buffer, which adds extra load to the oscillator. (R99 in series with C157). I decided to leave this intact in case I needed it later.

    Eventually I decided not to disconnect the ext osc buffer and left the circuit unmodified. I then fitted the second Epson TG2520 oscillator and hoped this would work.


    4) You get what you pay for !!
    The second Epson oscillator (About 3.5 Euro) has been working for a few days without failure and the Pluto is functioning again.
    However....The performance of the oscillator, although within specification, is completely unsuitable for communication using any sort of narrowband mode. - Let me explain.
    A high performance TCXO for communications should have good long term and good short term stability. Typically you would expect the TCXO frequency to change very slowly over its temeprature range e.g. -40 to +85C. So perhaps 500ppb off at -40 or +85C and something between these limits across the rest of the temperature range.
    With a cheap TCXO, such as this Epson device, the response is very different. When you switch on, the frequency quickly drifts just like a 25ppm or worse crystal. What happens next can be seen in the image from the lower QO-100 beacon shown below. Somewhere in the device, a temperature sensor detects the change and it also senses that the frequency is about to exceed the 500ppb limit. To compensate for this, internal circuitry quickly applies a correction (DAC?) which causes the frequency to rapidly swing in the opposite direction. This process happens again and again with any change in temperature. The result is that the output frequency is always within 500ppb. - So the long term stability is fine. However, the short term stability is terrible, with the frequency moving 400ppb at frequenent intervals. This short term stability variation makes FT8 etc completely undecodable. On the image you can see the oscillator suddenly moving 200Hz lower in frequency. At 739MHz this is just over 300ppb. This is well within specification, but makes the oscillator unsuited for narrowband use.


    4) Next step.....Yet another Oscillator. This time an IQD crystal surrounded by good old temperature compensated capacitors. I'll let you know how it performs. (About 20 Euros)


    73. See you /P sometime, Pluto allowing.