After receiving some opamps and a fairly decent phase noise DOCXO the next step of the 1Hz to 20kHz offset Phase Noise Analyzer was implemented. The simplest method to get the DOCXO locked to the signal from the DUT was to also send the DOCXO output as reference to a frequency counter and to measure the frequency from the DUT. Using some course/fine potmeters the DOCXO was tuned to be within 0.01Hz of the DUT and this allowed the PLL to lock and after some seconds  the frequency difference between the DOCXO and the DUT disappeared. The opamps provided 35dB gain to hopefully reduce the noise floor. First measurement was of a -90dBm/Hz noise source to confirm the -35dB gain and the corner frequency of the PLL. The Phase Noise plots uses 1Hz resolution so both noise and narrow band signals use the same power level scale. As can be seen in attached PN_Noise_-90dBm_Hz.JPG the response is fairly flat down to 3Hz and the 35dBm gain was confirmed The second measurement used a Rigol SG at 10MHz modulated with a PM signal at 220Hz at -78dBm which was measured at -42dBm and this also roughly confirms the 35dBm gain. A third measurement uses a AR60 Rb as DUT. The DOCXO used in the Phase Noise Analyzer should have a PN at 10Hz offset of -100dBm/Hz and this is what is measured so the AR60 Rb and the noise is not the limiting factor here. At 10kHz offset the Phase Noise of the DOCXO should be -150dBm but unfortunately either the noise of the ultra low noise opamps or the Phase Noise of the AR60 is almost 20dB higher. As last measurement a really bad GPSDO where the measurement is identical to a previous, just shifted up 35dB Next step is to find way's to reduce the noise level.

Are you sure the PN floor of the measurement instrument isn't the limiting factor? David -----Original Message----- From: Erik Kaashoek via time-nuts <time-nuts@lists.febo.com> Sent: 04 July 2022 10:14 To: time nuts <time-nuts@lists.febo.com>; Erik Kaashoek <erik@kaashoek.com> Cc: Erik Kaashoek <erik@kaashoek.com> Subject: [time-nuts] DIY Low offset Phase Noise Analyzer : At 10kHz offset the Phase Noise of the DOCXO should be -150dBm but unfortunately either the noise of the ultra low noise opamps or the Phase Noise of the AR60 is almost 20dB higher.

Hi David, Let me explain the DIY "measurement instrument" DOCXO into LO port of a mixer, DUT into RF, IF port with low pass filter to both steer the Vtune of the DOCXO and into an opamp to amplify 1-100kHz into PC audio input. On PC audio spectrum analyzer. So yes, the limiting factor is the measurement instrument, probably the opamp Further improvements of the setup has reduced the noise level at 10kHz offset to -145dBc/Hz so the phase noise of the DOCXO plus the AR60 is at least -145dBc/Hz at 10kHz offset. If the DOXCO is, as the spec states, -150dBc/Hz at 10kHz offset that the AR60 can not be above -145dBc/Hz at 10kHz. On 4-7-2022 12:01, David C. Partridge via time-nuts wrote: > Are you sure the PN floor of the measurement instrument isn't the limiting factor? > > David > > -----Original Message----- > From: Erik Kaashoek via time-nuts <time-nuts@lists.febo.com> > Sent: 04 July 2022 10:14 > To: time nuts <time-nuts@lists.febo.com>; Erik Kaashoek <erik@kaashoek.com> > Cc: Erik Kaashoek <erik@kaashoek.com> > Subject: [time-nuts] DIY Low offset Phase Noise Analyzer > > : > > At 10kHz offset the Phase Noise of the DOCXO should be -150dBm but > unfortunately either the noise of the ultra low noise opamps or the > Phase Noise of the AR60 is almost 20dB higher. > _______________________________________________ > time-nuts mailing list -- time-nuts@lists.febo.com > To unsubscribe send an email to time-nuts-leave@lists.febo.com

Hi If you are running a high gain op-amp to buffer things into a sound card *and* using the same op-amp output to drive the EFC, then you will have problems. Simple answer is to use a couple of op amps. Buffer the mixer with something low noise. Get the output of the mixer up to the point it almost saturates the op amp. Just how much gain that is depends a lot on your parts and power supplies. Running +/- 18V supplies into this op amp is no at all unusual. Since this output is linear, you have the full range of the beat note present. Nothing has been lost (yet). One path off this device goes to the high gain stage to the sound card. If the beat note is present, you will have clipping there. The other path goes to whatever you do to run the EFC. There are *many* approaches that could be used. One of many is a variable gain / variable roll off amp to “set” the PLL corner. That is followed by a simple summing amp to tune out the DC offset on the EFC. One thing to note is that once you get past the fancy amp in the first stage, the noise properties of the other amps are much less important ( assuming the supplies are fairly high ). This makes the rest of the circuit pretty cheap to build. Bob > On Jul 4, 2022, at 6:02 AM, Erik Kaashoek via time-nuts <time-nuts@lists.febo.com> wrote: > > Hi David, > Let me explain the DIY "measurement instrument" > DOCXO into LO port of a mixer, DUT into RF, IF port with low pass filter to both steer the Vtune of the DOCXO and into an opamp to amplify 1-100kHz into PC audio input. On PC audio spectrum analyzer. > So yes, the limiting factor is the measurement instrument, probably the opamp > Further improvements of the setup has reduced the noise level at 10kHz offset to -145dBc/Hz so the phase noise of the DOCXO plus the AR60 is at least -145dBc/Hz at 10kHz offset. If the DOXCO is, as the spec states, -150dBc/Hz at 10kHz offset that the AR60 can not be above -145dBc/Hz at 10kHz. > > On 4-7-2022 12:01, David C. Partridge via time-nuts wrote: >> Are you sure the PN floor of the measurement instrument isn't the limiting factor? >> >> David >> >> -----Original Message----- >> From: Erik Kaashoek via time-nuts <time-nuts@lists.febo.com> >> Sent: 04 July 2022 10:14 >> To: time nuts <time-nuts@lists.febo.com>; Erik Kaashoek <erik@kaashoek.com> >> Cc: Erik Kaashoek <erik@kaashoek.com> >> Subject: [time-nuts] DIY Low offset Phase Noise Analyzer >> >> : >> >> At 10kHz offset the Phase Noise of the DOCXO should be -150dBm but >> unfortunately either the noise of the ultra low noise opamps or the >> Phase Noise of the AR60 is almost 20dB higher. >> _______________________________________________ >> time-nuts mailing list -- time-nuts@lists.febo.com >> To unsubscribe send an email to time-nuts-leave@lists.febo.com > _______________________________________________ > time-nuts mailing list -- time-nuts@lists.febo.com > To unsubscribe send an email to time-nuts-leave@lists.febo.com

Another great post from Bob (as usual). Bob's advice is exactly correct. Unfortunately, the HP 11848 phase noise test set (part of the HP 3048 PN measurement system) is poorly designed, and does not break out the low gain non clipped signal. I modified mine to bring out this signal and it was MUCH easier to use. Follow Bob's advice if you are DIYing. Rick N6RK On 7/4/2022 9:12 AM, Bob kb8tq via time-nuts wrote: > Hi > > If you are running a high gain op-amp to buffer things into a > sound card *and* using the same op-amp output to drive the > EFC, then you will have problems. > > Simple answer is to use a couple of op amps. > > Buffer the mixer with something low noise. Get the output of the > mixer up to the point it almost saturates the op amp. Just how > much gain that is depends a lot on your parts and power supplies. > Running +/- 18V supplies into this op amp is no at all unusual. > > Since this output is linear, you have the full range of the beat note > present. Nothing has been lost (yet). > > One path off this device goes to the high gain stage to the sound > card. If the beat note is present, you will have clipping there. > > The other path goes to whatever you do to run the EFC. There are > *many* approaches that could be used. One of many is a variable > gain / variable roll off amp to “set” the PLL corner. That is followed > by a simple summing amp to tune out the DC offset on the EFC. > > Bob >