Re: Quick and dirty GPSDO build by F2DC and me.
I have missed that, time to get a proper digital scope and actually see what happens to Si5351 output ...
On Wed, May 12, 2021 at 09:09 AM, Dragan Milivojevic wrote:
I'm not quite sure what? you mean with clock stretching, 24MHz and other
usable frequencies are a direct divide from the internal 48MHz clock?
The phase noise is not that great but not that terrible either (subjective standards I know):
2&3 : stability figures are from this system or when when feeding the Si5351 with Neo directly?
12 MHz and 24 MHz. are not just a direct divide from the 48 MHz. clock.? They also stretch the clock pulse output periodically to get the frequency accurately synced to the GPS timing.
If you quickly look at it on a scope it appears that they don't pulse stretch but if you set the trigger for wider than one pulse you will see the stretched pulses like this ...
Roger
|
Re: Quick and dirty GPSDO build by F2DC and me.
On Wed, May 12, 2021 at 09:09 AM, Dragan Milivojevic wrote:
I'm not quite sure what? you mean with clock stretching, 24MHz and other
usable frequencies are a direct divide from the internal 48MHz clock?
The phase noise is not that great but not that terrible either (subjective standards I know):
2&3 : stability figures are from this system or when when feeding the Si5351 with Neo directly?
12 MHz and 24 MHz. are not just a direct divide from the 48 MHz. clock.? They also stretch the clock pulse output periodically to get the frequency accurately synced to the GPS timing. If you quickly look at it on a scope it appears that they don't pulse stretch but if you set the trigger for wider than one pulse you will see the stretched pulses like this ...  Roger
|
Hi Elia,
That is just too cool. I had to scratch build a V-O-M when I was a teenager (or live without one). I made mine in a cigar box with an old mil surplus meter. The shunt and series resistors all plugged into old tube sockets - I didn't have a range/mode switch. Projects like yours get us intimately familiar with how those measurements are made:)
My XYL and I are both fully vaccinated and now they have removed the ball and chains in our locale. I hope that everybody will get back to real life soon.
73,
Bill KU8H
bark less - wag more
toggle quoted message
Show quoted text
On 5/21/21 6:56 AM, M0ZHN via groups.io wrote: Good day everyone,
Just thought I would share something I started during the lockdown, nothing new or revolutionary but it has kept me entertained :)
My Fluke 77 has served me well since 1984 but is showing sings of age so I said to myself, why not build a bench DMM?
This is a simple DMM, old technology and uses 7 segment display. The heart of the DMM is MAX1496 ADC chip, an arduino controls the functions and range. It is still pretty much a work in progress, attached is a photo of it.
With pubs now open probably this DMM will wait for a while :)
Cheers, Elia, M0ZHN
|
Good day everyone, Just thought I would share something I started during the lockdown, nothing new or revolutionary but it has kept me entertained :) My Fluke 77 has served me well since 1984 but is showing sings of age so I said to myself, why not build a bench DMM? This is a simple DMM, old technology and uses 7 segment display. The heart of the DMM is MAX1496 ADC chip, an arduino controls the functions and range. It is still pretty much a work in progress, attached is a photo of it.  With pubs now open probably this DMM will wait for a while :) Cheers, Elia, M0ZHN
|
Re: Quick and dirty GPSDO build by F2DC and me.
Erik,
I just built this after receiving a NEO-M8N and have the following observation: Occasionally, the 1pps from the GPS skips and the software goes bonkers with a large error and makes a big step which is quickly corrected at the next 1pps update. I am guessing that the GPS signals are changing, like one or more satellites go out of range or others join.? I think the software needs to be modified so that when it gets a large error, it holds the previous correction factor.
Gary
W9TD
|
Re: Convolution - a possible way to provide improved RBW for the tinySA?
I've implemented and tested deconvolution. Step 1: Measure the response to a single signal and use that as the deconvolution filter Step 2: Measure the convoluted response to multiple signals Step 3: deconvolute using fft The outcome was not usable due to the noise in the input. -- HBTE Files section:? /g/HBTE/filesErik, PD0EK
|
Re: Convolution - a possible way to provide improved RBW for the tinySA?
On Tue, May 11, 2021 at 08:45 AM, m0wid wrote:
sampling at smaller frequency step increments and using convolution
Interesting. This sounds like a one dimensional version of what Photoshop does to photos to sharpen up somewhat blurry photos in two dimensions.? Not introducing artifacts might be the biggest challenge. But, you did say the response of the SI4432 is important to measure as part of the process.? Tom, wb6b
|
Please add a picture of the board -- HBTE Files section:? /g/HBTE/filesErik, PD0EK
|
Hello i recently bought a chinese pll adf4351 evaluation board and used the belt with arduino nano my programmed frequency 2.514 GHz with + 5db output works but looking with the spectrum analyzer i see some spurs i wanted to know if there are changes to be made and improvements. Regardless with the green colored card it looks better there are no spurs thanks in advance for the help 73 iw5bsf Roberto
|
Re: Quick and dirty GPSDO build by F2DC and me.
On Wed, May 12, 2021 at 09:09 AM, Dragan Milivojevic wrote:
2&3 : stability figures are from this system or when when feeding the Si5351 with Neo directly?
From this system? -- HBTE Files section:? /g/HBTE/filesErik, PD0EK
|
Re: Quick and dirty GPSDO build by F2DC and me.
I'm not quite sure what? you mean with clock stretching, 24MHz and other usable frequencies are a direct divide from the internal 48MHz clock?
The phase noise is not that great but not that terrible either (subjective standards I know): 2&3 : stability figures are from this system or when when feeding the Si5351 with Neo directly?
On Wed, 12 May 2021 at 16:53, Erik Kaashoek < erik@...> wrote: On Wed, May 12, 2021 at 05:17 AM, Dragan Milivojevic wrote:
Did you by any chance compare this solution to simply feeding the si5351 with
24MHz "pps" output from neo7
The 24MHz pulse has a terrible phase noise making the output of the SI5351 equally unusable as the 24MHz output of the neo-7M itself
?
--
HBTE Files section:?/g/HBTE/files
Erik, PD0EK
|
Re: Quick and dirty GPSDO build by F2DC and me.
Hi
Indeed phase noise is an important issue. So are spurious as Sverre, LA3ZA shows here: Karen, RA3APW has also a page about spectrum performance The Si5351A is a fun and price competitive device. However, for some odd reason 99,99% of all radio amateurs seem to think that the Si5351A with a low pass filter is the greatest invention since sliced bread. But it is not. You really have to put your back into designing a good circuit to achieve a fair spectrum:
Here is an example of the stability of a 2 m signal I.e. more than 14 times worse than it is on 10 MHz.
Bo
www.rudius.net/oz2m :: www.rfzero.net
|
Re: Quick and dirty GPSDO build by F2DC and me.
On Wed, May 12, 2021 at 05:17 AM, Dragan Milivojevic wrote:
Did you by any chance compare this solution to simply feeding the si5351 with
24MHz "pps" output from neo7
The 24MHz pulse has a terrible phase noise making the output of the SI5351 equally unusable as the 24MHz output of the neo-7M itself ? -- HBTE Files section:? /g/HBTE/filesErik, PD0EK
|
Re: Quick and dirty GPSDO build by F2DC and me.
When measuring the output there are for me some important spec points. 1: Phase noise. With the SI5351 this is ok when using a real crystal and not the 24MHz from a neo-7m as there is clock stretching applied to the 24MHz to keep it in sync with the gps time signal. 2: short term stability: When testing with a 70MHz output using a one measurement per second 8 digit full resolution counter running from a verified and calibrated TCXO the short term stability was well within the observable limit of one Hz, to be expected with a good phase noise 3: long term stability: When using the 70MHz output after temperature has stabilized? the small corrections performed are in the order of one Hz so below 0.02ppm But I may be missing something here, please help me better understand -- HBTE Files section:? /g/HBTE/filesErik, PD0EK
|
Re: Quick and dirty GPSDO build by F2DC and me.
Did you by any chance compare this solution to simply feeding the si5351 with 24MHz "pps" output from neo7? I have no means of evaluating the precision, except for comparing to some on the air frequency standard, so I was wondering if there is something to be gained from this solution?
On Tue, 11 May 2021 at 17:06, Erik Kaashoek < erik@...> wrote: Anyone looking for a GPSDO with a frequency accuracy in the order of 1e-8 should have a look at this:? I made a quick and very dirty build of his design using only 4 modules found on eBay - Arduino nano - GPS module with 1PPS output (neo-6M or better, I used neo-7M) - SI5351 module with 3 outputs - 16x2 LCD display and some wires etc,?running from a bench 5V supply.  The Arduino SW was modified (??) to have a adaptive measurement time depending on the frequency error observed. With a stable temperature of the SI5351 crystal and the largest measurement/correction cycle time of 400 seconds the stability was better than 0.01 Hz, or 0.003 PPM? The schematic as published by F2DC  My current SW on github does not need the serial data from the GPS, only the pulse per second. The CLK1 output of the SI5351 is freely configurable to any valid SI5351 output frequency such as 10MHz or 100MHz.? Have fun -- HBTE Files section:? /g/HBTE/filesErik, PD0EK
|
Re: Quick and dirty GPSDO build by F2DC and me.
Hi
Aren't you confusing accuracy with stability? Long averaging does not improve stability over the duration of the measurement period. E.g. the mains frequency is 50 Hz over 24 h but in between it can deviate. Likewise for anything else like a 10 MHz reference. Please see the attached picture. Ask yourself this question: is it good enough if I am "on the frequency over 40 s" but can be "off the frequency in between"? Please see this for MUCH more about accuracy and stability:
The Motorola Oncore M12 GPS modules can be used for some much better GPSDO like this one that uses the Navman Jupiter TU30: Of course a Hydrogen maser or optical clock is in a very different league. But in the end you get what you pay for.
The RFzero that uses a Si5351A and measures every second and is used e.g. by Per, SM0DFP for his 24 GHz portable setup
It is NOT about averaging if the signal is all over the place anyway.
Bo
www.rudius.net/oz2m :: www.rfzero.net
|
Re: Quick and dirty GPSDO build by F2DC and me.
Thanks Erik, I¡¯ll take a look at your software and see if I can adapt it to my i2s display.
Sverre
LA3ZA
|
Re: Quick and dirty GPSDO build by F2DC and me.
Every time the error is measured and the error is above the threshold (which is 1.5 times the minimum observable error) the frequency is corrected. In the original SW this measurement and correction is done once every 40 seconds using a 2.5MHz clock leading to a minimum observable error of 0.01 ppm In my SW the measurement and correction are adaptive between 1 and 400 seconds depending on the stability of the SI5351 crystal. The minimum frequency correction to the SI5351 is 0.01Hz or, if the error is below the threshold, no correction is done. The original SW always updates the SI5351 registers after successful measurement, even when no change is needed, causing the 10MHz output to miss some cycles. Could this be the causing a disturbance? -- HBTE Files section:? /g/HBTE/filesErik, PD0EK
|
Re: Quick and dirty GPSDO build by F2DC and me.
I also built the F2DC GPS-controlled oscillator some time ago and wrote about it on my blog,?
I¡¯ve found however that it makes small jumps in frequency as it updates. As I use it as a 10 MHz input for my Elecraft K3, the K3 will jump in frequency by a fraction of a Hz as the updating of the gps oscillator is performed. I want to decrease the gain in the feedback loop of the update algorithm to see if that helps, but haven¡¯t had time to complete it.
Sverre
LA3ZA
|
Re: Quick and dirty GPSDO build by F2DC and me.
Yes. With an VCOXCO you can do better but the goal was not to use an oxco and still have a factor 100 better performance than any regular free running .2 ppm oxco so I can be used to check and calibrate nanoVNA and tinySA. More accuracy makes no sense for these two -- HBTE Files section:? /g/HBTE/filesErik, PD0EK
|
Roger Need
Gary W9TD
Roberto
Bo, OZ2M
