A look at the Leo Bodnar LBE-1420 , the precision GPSDO, and BG7TBL

[Wim13]

With one clock you know the time, but with two clock's.... a search for accuracy not only stability.

I am trying to find a way to determine the real frequency in real time.
Timelab is more focused on time measurements on the long run.

I have two rubidium clock generators (Efratom), and a leo bodnar LBE-1420 GPSDO (Ublox 10),
and a leo bodnar two ports precision GPSDO (Ublox 8 ).  Edit: and now also a BG7TBL, version 2021-06-10.
Software version V20240822 The BG7TBL has a Ublox 10, and U-center works.

See inside pictures from the LBE-1420 and the precision.

The LBE runs on a ublox M10050KB, an low power ublox M10 and the precision has a ublox M8

The LBE-1420 recieve GPS, Galileo, and Beidou.
I have no info on the precision, only the calculation figures, for different outputs.

The LBE 1420 is very flexible on the output freq. this unit  does have NMEA serial port.
The precision-dual has no NMEA, but there is a topic in which Leo wrote that future update will have.

U-center does not work, only NMEA is on the LBE1420.
On the precision version there is no serial port, just only HID inteface.

Be aware that the Ublox has power protection on the antenne output,
if you use by accident a shorted coax cable the power will be shut down, and poor reception is the case.
Your are lucky if auto recovery is switched on.


So i decided to do some compare test.....

on the scoop, 200Mhz Rigol, channel A the precision and on channel B the LBE-1420. and the Rubidium as trigger.
All on 10 Mhz. The rubidium was not locked to any gps, but was on for hours.

Both units has there own antenna, i have no splitter, not very mutch the same sky view,
i live in a city, and its not easy to put antennes anywhere. Recieves lots and lots of sats

For me there was a surprise view, that every GPSDO has its own jitter drift and offset.
I was expecting more of the same on having more GPSDO's
I can not tell which GPS was the correct one.... there must be a way...

On the scoop i found for de LeoBodnar GPSDO's a drift about 30/40 nSec per hour.
The Rubidiun as a reference.

Jitter, on Timelab, the 1 sec adev is about 4.E-11 the LBE1420 had 5.E-11
Also found this after collection data from a counter, and calculated in excel.

So the stability seams oke, but i am intrested in accuracy, and thats not easy,

The performance of both is looks the same, only the offset frequency is different.

I also have now the TinyPFA to compare the GPSDO's
I will make some Timelab pictures.

The BG7 and my Rubidium have about the same 1 sec adev of 1.E-11

So my experiments goes to the next level..accuracy ??
Problem is , i think i need something better than a rubidium, will see.

[Wim13]

As i want to know the real frequency, and not the stability, in this case.
 
i made some comparisons between the BG7TBL and at first the LeoB dual precision, later i will also do the LBE 1420.

As i can see the closed i can get on this moment is 1.E-11, thats also the conform specs of the Efratom rubidium.
But a rubidium has to be calibrated at times.

I can also see as in my first post, that the LB precision has a drift of about 30/40 nSec, in a hour, which is also close to 1.E-11

The BG7TBL has an OCXO, and has a slow woop, as expected for a OCXO.

For trend analyses i have set the average on the second picture to 100.
The freqs of both are very close. 

[Wim13]

I made some measurments to compare the two Leo Bodnar GPSDO's, The LBE-1420 and the older  Dual precision version.

The reference was a Efratom rubidium.
The older LB dual works better and more stable then the LBE-1420 in my setup.
The LBE-1420 had sometimes trouble , drifted sometime away for about 5.E-11 to 1E-10.

The older LB does more agree with my rubidium and the GB7TBL GPSDO.
The LBE-1420 has SBUS enabled. But for this moment little disapointed about the LBE-1420.

The wobble in the Allen plot of the LBE-1420 at 100 s, occurs because sudden drift at times.
Made several plots but it happens most times, if have orded an antenne splitter,
then i can make a simultaneous plot, of two GPSDO's.
Then i can see if the antenne situation makes the difference, as it is not ideal situation, living in the city.
You never have a whole view at the sky, but i have to deal with that.

On the LB precision dual port, on the same antenna's i have not seen this, this unit is very stable.

I wrote a little program for myself that monitors real time the drift and has a rotating buffer for avarage.
Also i can real time see the Std dev. for 1 sec. in this sample the LBE-1420 is connected and the Rubidium
The LB dual has most of the time a much lower STddev.

[Leo Bodnar]

The reference was a Efratom rubidium.
The older LB dual works better and more stable then the LBE-1420 in my setup.
The LBE-1420 had sometimes trouble , drifted sometime away for about 5.E-11 to 1E-10.

Stability in 10s..1000s region is mostly defined by ambient environment. Airflow and temperature affect the phase shift and as a result — output frequency.
Two port GPS clock has larger internal volume and higher mass which explains why it would look more stable under similar conditions.
Try restricting environmental variations by using styrofoam or cardboard baffle or using a thick towel wrapped around sensitive equipment.

I have run a LBE-1420 against Efratom FRK last night and here is what I got.
Pink trace is LBE-1420 lying on a desk.
Blue trace is LBE-1420 wrapped in a small blanket.

I would suggest enabling error bars (ctrl+e) when you are looking at ADEV/MDEV charts since they are statistical values.
Leo

[Wim13]

Well i made 4 compares, in free air and covered with simple cloth, as Leo suggested.

Compares 1,

I chose a nicer place for the LBE1420, and that works...
LBE-1420 in free air and next with a small cloth, for reference i simultaneously measure the GB7TBL GPSDO
The two antenna's where next to each orher.

Surprise, .. the results compare to the one of Leo are quite good...

Compares 2,

Same picture as picture 1, but now with the error markers on, as requested.

Compares 3,

More interesting for me, now i can get to more accuracy/precision, as i know my FA3 counter has a reciprocal offset of 8 picoSec, 8.E-12.
as also in an other topic measured by others, which is very good compare to even some HP.

I see that my BG7TBL is close to the Rubidium, almost strait line, with indeed an  offset of 8.E-12,
That means that the Rubidium and the GB7TBL are in range of E-12 or better.

Compares 4,

As the green line is original arround zero, i can now calculate the offset of the LBE-1420.
The TinyPFA seems to have an offset error of 2.E-12

Nice i found the accuracy and precision i wanted.
The LBE 1420 need some special treatment for good performance

[dietert1]

If you have two rubidium oscillators you could discipline them to GPS and determine their frequency difference. To get below 1E-11 GPS requires averaging times that the crystal oscillator alone cannot sustain.

Regards, Dieter

[Wim13]

It is not easy to discipline rubidium oscillators, you need a very long feed back loop, because Rubidium is much stable on the short then GPS,
GPSDO's are quite noise in comparison to Rubidiums. IF you don't have a proper feed back system it will cause more damage then good.

For this moment i do the calibration by hand.