TESTING THE FE-5650A RUBIDIUM STANDARD AGAINST A LEO BODNAR GPSDO
The simplest useful test is to compare the FE-5650A 10 MHz output directly with the Leo Bodnar GPSDO, also set to 10 MHz.
Feed the rubidium 10 MHz into CH1 of your Rigol DS1102E and the GPSDO 10 MHz into CH2. Display both channels simultaneously and trigger from one of them. Once the FE-5650A has been powered for at least 20–30 minutes, watch whether the two waveforms slowly drift relative to one another.
If both were exactly 10 MHz, their relative phase would remain stationary. A small frequency difference makes one waveform slowly “walk” through the other. The time for one complete 360° phase slip gives us the frequency error very accurately.
===========
CONNECTIONS
===========
FE-5650A RUBIDIUM RIGOL DS1102E
+——————+ +————-+
| | 10 MHz | |
| 10 MHz OUT —–+—————>| CH1 |
| | | |
+——————+ | |
| |
LEO BODNAR GPSDO | |
+——————+ | |
| | 10 MHz | |
| 10 MHz OUT —–+—————>| CH2 |
| | | |
+——————+ +————-+
===========
PROCEDURE
=========
1. Power the FE-5650A.
2. Allow the rubidium standard to warm up for at least
20-30 minutes.
3. Set the Leo Bodnar GPSDO output to 10 MHz.
4. Connect:
FE-5650A 10 MHz –> Rigol CH1
GPSDO 10 MHz –> Rigol CH2
5. Display CH1 and CH2 simultaneously.
6. Trigger the oscilloscope from one channel.
7. Watch the relationship between the two waveforms.
===========
WHAT SHOULD HAPPEN?
===========
If both sources produced EXACTLY 10 MHz, the two waveforms would remain in the same relative position. If there is a tiny frequency difference, one waveform will slowly move or “walk” relative to the other. Eventually it will move through one complete cycle. This is called a PHASE SLIP.
===========
CALCULATING THE FREQUENCY DIFFERENCE
===========
Measure the time required for one complete phase slip.
Let:
T = time for one complete phase slip in seconds
Then:
Frequency difference = 1 / T
or:
Delta f = 1 / T
===========
EXAMPLES
===========
One complete phase slip every:
100 seconds
Delta f = 1 / 100
= 0.01 Hz
At 10 MHz this corresponds to a fractional frequency
difference of:
0.01 / 10,000,000
= 1 x 10^-9
One phase slip every 1,000 seconds:
Delta f = 0.001 Hz
Fractional difference:
= 1 x 10^-10
One phase slip every 10,000 seconds:
Delta f = 0.0001 Hz
Fractional difference:
= 1 x 10^-11
===========
QUICK REFERENCE
===========
Phase slip time Error at 10 MHz
—————————————–
10 seconds 1 x 10^-8
100 seconds 1 x 10^-9
1,000 seconds 1 x 10^-10
10,000 seconds 1 x 10^-11
===========
WHY THIS IS INTERESTING
===========
The Leo Bodnar GPSDO obtains its long-term frequency accuracy from GPS.
The FE-5650A obtains its frequency stability from its internal rubidium atomic reference.
The GPSDO should therefore provide an excellent long-term reference against which to test the FE-5650A. The rubidium oscillator may show particularly good short-term stability.
By measuring the phase drift between them we can estimate the actual frequency error of this particular FE-5650A rather than relying only on its original specifications.
===========