Hello Everyone,
This is a review and measurements of the LEO BODNAR LBE-1421 Clock Generator.
I previously reviewed the SMSL G1, OCXO-based Clock Generator, and I recommend you read it to familiarize yourself with the two key concepts related to clock generators (Clock Accuracy and Phase Noise). This previous review got positive feedbacks with some tangible measurable benefits, and I was recommended by @MC_RME to consider the above GPSDO (GPS Disciplined Oscillator), as a cheaper alternative, for subsequent testing. So, here we go.
LEO BODNAR LBE-1421 - Presentation
This is a clock generator that use GPS Reference Clock to output an atomic-precision clock at selectable frequencies. Internally, it uses a Temperature Compensated Crystal Oscillator (TCXO) that is steered, or disciplined, with the signals broadcast by GPS (USA) and other GNSS satellites (Global Navigation Satellite System) such as Galileo (Europe), GLONASS (Russia) or BeiDou (China). It can also leverage the Satellite-Based Augmentation System (SBAS), that is a network of satellites and ground stations that improve GNSS accuracy.
Let met quote the vendor's website: "Frequency stability of its output is defined by the accuracy of GPS satellite onboard Caesium references and approaches 1x10-12 or 0.000001 ppm."
This statement is describing the frequency stability of a system that uses GPS signals for timing. Here’s what it means, broken down:
A software is provided for tuning, this is how it looks like:
It shows the number of satellites being tracked and the quality of the signal. The device needs connection to a PC for tuning, but can run on a power bank or USB charger (it requires 5V-250mA). There is a magnetic water-proof active antenna provided together with 5m of cable. As you can see, the Software allows selection of the GNSS providers. By default, the output is set to 10MHz, as it is very standard.
I tested the LEO BODNAR with the antenna outside or close to my window and it worked well:
The atomic precision is not all we need when talking clock generator. We also need low phase noise. The LEO BODNAR uses an internal TCXO with said sub-picosecond RMA jitter that will shape the phase noise. The result is given by the vendor per the below graph:
When comparing the above with the previously reviewed SMSL G1, that uses a higher-end OCXO (Oven Crystal Oscillator), it means the below:
The G1 does indeed better in that perspective. Is it necessary for audio? Haha, I'll let you reply to that question.
On a clock accuracy perspective, the SMSL G1 was given for 3ppb (part per billion) while the LEO BODNAR achieves 3,000 times better!
User Experience
As you can see from the previous photos, we are far from a good looking hifi device, and the added cabling could be a challenge. But, hey, don't forget the asking price... By the way, LEO BODNAR offers a low phase noise 10MHz clock source based on OCXO (LBE-1610) at 1400$...
It takes a little time for the LEO BODNAR to acquire satellites data (roughly 30sec), and the two front LEDs will be flashing during that time.
It does not require to be connected to a PC to acquire data, and it will also continue to run if loosing signal temporarily (digital PLL will maintain best estimated output frequency based on historical data, to prevent frequency or phase noise degradation).
The output frequency is 10MHz square wave, which is nearly standard, so it should minimize the need to tune it before use. Again, I had no issue using it with the antenna close to a window.
One important note, the device is very small, and so the connectors are SMA types, meaning the type of Bluetooth antennas. So I had to use an SMA to BNC adaptor (see photo) to use with the SMSL D200 converter which I used for all tests.
LEO BODNARD LBE-1421 - Measurements
For all measurements, I used the SMSL D200 DAC which offers a clock input.
Again, I strongly recommend that you read the SMSL G1 review as a some sort of reference because:
So, you will only see indirect measurements, that means to check if the clock improves anything tangibly at the analog output of the DAC.
I will follow the same sequence of measurements as I did with the SMSL G1, to ease the comparison.
----
Jitter Test
Of course the first test that comes to mind is the standardized Jitter Test (24bits 48kHz sampling rate).
For the below test, I used the SMSL D200, with its internal clock, and next with the LEO BODNAR. Measurements were done with a punishing lengthy FFT of 512k with 32 averages as to lower the random low level noise so that we can better see potential side bands, and I used the Cosmos E1AD that has a higher resolution for this exercise:
That is an already "nailed" test by the SMSL D200, but let's see if things change with the LEO BORDNAR providing the clock:
To be honest, it's a run to run difference, nothing more. Artifacts are all below -150dBr, only a high resolution FFT can spot that.
Let's try the same from a CD player at 16 bits. The below is a Jitter test comparison when using an old Revox B 226-S that I know to have an aging clock, drifting a little more than all the others I have. The Revox is feeding the SMSL D200 DAC via coax digital output:
No difference really, these are two very good trace as per the signal injected which is a 11.025kHz tone, with the least significant bit (LSB) toggled by a 250Hz square wave (explaining that series of slowly decreasing spikes, that are all harmonics of the initial 250Hz square).
Note that the foot of the fundamental shows very little more random noise (larger foot) when not using the LEO BODNAR. I saw the same effect of the higher precision clock in all other measurements. More on that later.
----
Pitch Error:
I've been talking about pitch error for every CD players that I reviewed. I measure it from a 19'997.00Hz test tone and report the output of the CD player. That way, I can measure down to 0.5ppm precision (if the resulting frequency is 19'9997.01Hz), and with all the test I performed for this review, this is below the limit of the Cosmo that I regularly use. So I relied on my Motu Ultralite mk5 which seems to do better, but in the end, no big difference between the two.
First, let me use the Revox B 226-S that I used when reviewing the SMSL G1 and show you the difference:
Similar to what I saw when I tested the SMSL G1, the LEO BODNAR decreases the pitch error (from +126.5ppm to +3.5ppm) that comes from the CD Player. Indeed, with SPDIF transmission, the clock signal is embedded into the data, so the DAC has to recover it using a PLL and its internal clock. But per my dozens of testing with CD Players, the clock imprecision of the CD player always finds its way into the DACs.
EDIT: as suggested by @Rja4000, I used the Leo Bodnar to calibrate the pitch error of the Motu, which happens to be +3.5ppm. So the improvement is better that shown above, as the real result is therefore below 1ppm.
See the below table that shows internal pitch error measured at the output of 6 CD/SACD players, and how it is passes through several DACs of various technologies, price and age (I updated the table to account for the native clock deviation of the Motu when is is hot):
The internal pitch error of every single CD player can be found at the analog outputs of the 4 different DACs I used for this test.
So let me use the second worse student in the above list, the Sony CDP-337ESD which is more than 35 years old:
And again, when using the LEO BODNAR as the source clock, I see that improvement in frequency stability (from -29ppm to +3.5ppm), which I don't see elsewhere. It was the same positive effect when using the SMSL G1.
EDIT: as suggested by @Rja4000, I used the Leo Bodnar to calibrate the pitch error of the Motu, which happens to be +3.5ppm. So the improvement is better that shown above, as the real result is therefore below 1ppm.
It is important to note, though, that whatever is the initial pitch error of any CD player, I get the same 3.5ppm (EDIT: which is the pitch error of the measurement interface, so the real result is below 1ppm) when using the LEO BODNAR.
Is the above improvement due solely to the better clock precision, the lower phase noise or both? I don't know, but I could measure that difference.
Like I said about the SMSL G1 Clock I was not expecting this type of benefit, and I see it repeats here with a different clock and many different additional devices the I used. The one thing that did not change is the SMSL D200 DAC, so next step would be to confirm all the above with another DAC having a clock input.
----
A bit of SINAD please
Yeah, you can ask for it, and for fun I share, like I did with the G1, let me use a CD player as the source, since it is my motto.
This is my standard measurement of 999.91Hz undithered test tone @0BFS from the XLR analog outputs of the Teac VRDS-25x:
When feeding the SMSL D200 DAC from coax digital output of the Teac, I get:
Yes, much better, at the frontier of CDA max performances: 16bits resolution is what we want and what we get from that full scale signal.
Now, let's feed the SMSL D200 with the LEO BODNAR LBE-1421:
Nearly identical, main difference being two different runs. But look closer at the foot of the fundamental... I already talked about that. Let me help zooming a little:
The foot of the fundamental gets less random and correlated noise. Being located here, this can't make an audible difference, but it again means more precision, ie less (random, deterministic, intrinsic, data, whatever)-jitter, obviously.
Conclusion
The geek of me likes a lot this device
Thanks to @MC_RME for suggesting it!
Having the clock precision of a cesium-based atomic clock, not for the price of a car, is really cool. The possibility to tune the output to any frequency, with the same precision, is also genius and makes it easily reusable.
I love the technical idea behind as well, which comes from the very first atomic clocks that were used to disciplined Xtals in the late 40s. Nearly 100 years later, we get even better than that at home for the price of a good restaurant
The only issue is more cables are required for a very limited use case when talking about feeding only one device with it, as I did here.
At the end of the day, there are many stuff that we buy and make no difference (or have a negative impact). So, you don't need this one, but some of you will still want it
It is always apéro time somewhere in the world, and that time has come for me. A whisky and "aLCHEMY dIRE sTRAITS Live" for me please. Cheers!
I hope you enjoyed this review and I wish you a lovely weekend.
Flo
This is a review and measurements of the LEO BODNAR LBE-1421 Clock Generator.
I previously reviewed the SMSL G1, OCXO-based Clock Generator, and I recommend you read it to familiarize yourself with the two key concepts related to clock generators (Clock Accuracy and Phase Noise). This previous review got positive feedbacks with some tangible measurable benefits, and I was recommended by @MC_RME to consider the above GPSDO (GPS Disciplined Oscillator), as a cheaper alternative, for subsequent testing. So, here we go.
LEO BODNAR LBE-1421 - Presentation
This is a clock generator that use GPS Reference Clock to output an atomic-precision clock at selectable frequencies. Internally, it uses a Temperature Compensated Crystal Oscillator (TCXO) that is steered, or disciplined, with the signals broadcast by GPS (USA) and other GNSS satellites (Global Navigation Satellite System) such as Galileo (Europe), GLONASS (Russia) or BeiDou (China). It can also leverage the Satellite-Based Augmentation System (SBAS), that is a network of satellites and ground stations that improve GNSS accuracy.
Let met quote the vendor's website: "Frequency stability of its output is defined by the accuracy of GPS satellite onboard Caesium references and approaches 1x10-12 or 0.000001 ppm."
This statement is describing the frequency stability of a system that uses GPS signals for timing. Here’s what it means, broken down:
- Frequency stability: How consistently a system maintains its output frequency over time. High stability means the frequency hardly drifts.
- Defined by the accuracy of GPS satellite onboard Caesium references: GPS satellites have atomic clocks (usually Caesium or Rubidium) that are extremely precise. The system locks its timing to these clocks, so its stability depends on their accuracy.
- Approaches 1×10⁻¹² or 0.000001 ppm: 1×10⁻¹² is a fractional frequency error: one part in a trillion. 0.000001 ppm (parts per million) is another way to express the same thing. For example: If the nominal frequency is 10 MHz, an error of 1×10⁻¹² means the frequency could drift by only about 0.00001 Hz. So, at this stability, the clock would gain or lose 1 second in about 31,700 years.
A software is provided for tuning, this is how it looks like:
It shows the number of satellites being tracked and the quality of the signal. The device needs connection to a PC for tuning, but can run on a power bank or USB charger (it requires 5V-250mA). There is a magnetic water-proof active antenna provided together with 5m of cable. As you can see, the Software allows selection of the GNSS providers. By default, the output is set to 10MHz, as it is very standard.
I tested the LEO BODNAR with the antenna outside or close to my window and it worked well:
The atomic precision is not all we need when talking clock generator. We also need low phase noise. The LEO BODNAR uses an internal TCXO with said sub-picosecond RMA jitter that will shape the phase noise. The result is given by the vendor per the below graph:
When comparing the above with the previously reviewed SMSL G1, that uses a higher-end OCXO (Oven Crystal Oscillator), it means the below:
Phase Noise @frequency | LEO BODNAR LBE-1421 | SMSL G1 |
1Hz | -72 dBc/Hz | -105dBc/Hz |
10Hz | -101 dBc/Hz | -126dBc/Hz |
100Hz | -128 dBc/H | -142dBC/H |
1kHz | -147 dBc/H | -151dBc/Hz |
10kHz | -156 dBc/H | -153dBc/Hz |
100kHz | -159 dBc/Hz | -152dBc/Hz |
The G1 does indeed better in that perspective. Is it necessary for audio? Haha, I'll let you reply to that question.
On a clock accuracy perspective, the SMSL G1 was given for 3ppb (part per billion) while the LEO BODNAR achieves 3,000 times better!
User Experience
As you can see from the previous photos, we are far from a good looking hifi device, and the added cabling could be a challenge. But, hey, don't forget the asking price... By the way, LEO BODNAR offers a low phase noise 10MHz clock source based on OCXO (LBE-1610) at 1400$...
It takes a little time for the LEO BODNAR to acquire satellites data (roughly 30sec), and the two front LEDs will be flashing during that time.
It does not require to be connected to a PC to acquire data, and it will also continue to run if loosing signal temporarily (digital PLL will maintain best estimated output frequency based on historical data, to prevent frequency or phase noise degradation).
The output frequency is 10MHz square wave, which is nearly standard, so it should minimize the need to tune it before use. Again, I had no issue using it with the antenna close to a window.
One important note, the device is very small, and so the connectors are SMA types, meaning the type of Bluetooth antennas. So I had to use an SMA to BNC adaptor (see photo) to use with the SMSL D200 converter which I used for all tests.
LEO BODNARD LBE-1421 - Measurements
For all measurements, I used the SMSL D200 DAC which offers a clock input.
Again, I strongly recommend that you read the SMSL G1 review as a some sort of reference because:
- I will run similar measurements, so we can compare
- I will not repeat everything I wrote there
So, you will only see indirect measurements, that means to check if the clock improves anything tangibly at the analog output of the DAC.
I will follow the same sequence of measurements as I did with the SMSL G1, to ease the comparison.
----
Jitter Test
Of course the first test that comes to mind is the standardized Jitter Test (24bits 48kHz sampling rate).
For the below test, I used the SMSL D200, with its internal clock, and next with the LEO BODNAR. Measurements were done with a punishing lengthy FFT of 512k with 32 averages as to lower the random low level noise so that we can better see potential side bands, and I used the Cosmos E1AD that has a higher resolution for this exercise:
That is an already "nailed" test by the SMSL D200, but let's see if things change with the LEO BORDNAR providing the clock:
To be honest, it's a run to run difference, nothing more. Artifacts are all below -150dBr, only a high resolution FFT can spot that.
Let's try the same from a CD player at 16 bits. The below is a Jitter test comparison when using an old Revox B 226-S that I know to have an aging clock, drifting a little more than all the others I have. The Revox is feeding the SMSL D200 DAC via coax digital output:
- On the left side the SMSL D200 uses its internal Clock
- On the right side, the SMSL D200 uses the LBE-1421
No difference really, these are two very good trace as per the signal injected which is a 11.025kHz tone, with the least significant bit (LSB) toggled by a 250Hz square wave (explaining that series of slowly decreasing spikes, that are all harmonics of the initial 250Hz square).
Note that the foot of the fundamental shows very little more random noise (larger foot) when not using the LEO BODNAR. I saw the same effect of the higher precision clock in all other measurements. More on that later.
----
Pitch Error:
I've been talking about pitch error for every CD players that I reviewed. I measure it from a 19'997.00Hz test tone and report the output of the CD player. That way, I can measure down to 0.5ppm precision (if the resulting frequency is 19'9997.01Hz), and with all the test I performed for this review, this is below the limit of the Cosmo that I regularly use. So I relied on my Motu Ultralite mk5 which seems to do better, but in the end, no big difference between the two.
First, let me use the Revox B 226-S that I used when reviewing the SMSL G1 and show you the difference:
Similar to what I saw when I tested the SMSL G1, the LEO BODNAR decreases the pitch error (from +126.5ppm to +3.5ppm) that comes from the CD Player. Indeed, with SPDIF transmission, the clock signal is embedded into the data, so the DAC has to recover it using a PLL and its internal clock. But per my dozens of testing with CD Players, the clock imprecision of the CD player always finds its way into the DACs.
EDIT: as suggested by @Rja4000, I used the Leo Bodnar to calibrate the pitch error of the Motu, which happens to be +3.5ppm. So the improvement is better that shown above, as the real result is therefore below 1ppm.
See the below table that shows internal pitch error measured at the output of 6 CD/SACD players, and how it is passes through several DACs of various technologies, price and age (I updated the table to account for the native clock deviation of the Motu when is is hot):
| CD Player | Native Pitch Error out of the CD Player | Pitch Error out of SMSL D200 | Pitch Error out of SMSL PS200 | Pitch Error out of Topping D50III | Pitch Error out of ASUS ESSENCE STU |
| Accuphase DP-70 | -1.5ppm | -1.5ppm | -1.5ppm | -1.5ppm | -1.5ppm |
| Teac VRDS 25X | +1ppm | +1ppm | +1ppm | +1ppm | +1ppm |
| Pioneer DP-D9 | -19.5ppm | -19.5ppm | -19.5ppm | -19.5ppm | -19.5ppm |
| Yamaha CD-S2000 | -19.5ppm | -19.5ppm | -19.5ppm | -19.5ppm | -19.5ppm |
| Sony CDP-3337ESD | -32.5ppm | -32.5ppm | -32.5ppm | -32.5ppm | -32.5ppm |
| Revox B 226-S | +122.5ppm | +122.5ppm | +122.5ppm | +122.5ppm | +122.5ppm |
The internal pitch error of every single CD player can be found at the analog outputs of the 4 different DACs I used for this test.
So let me use the second worse student in the above list, the Sony CDP-337ESD which is more than 35 years old:
And again, when using the LEO BODNAR as the source clock, I see that improvement in frequency stability (from -29ppm to +3.5ppm), which I don't see elsewhere. It was the same positive effect when using the SMSL G1.
EDIT: as suggested by @Rja4000, I used the Leo Bodnar to calibrate the pitch error of the Motu, which happens to be +3.5ppm. So the improvement is better that shown above, as the real result is therefore below 1ppm.
It is important to note, though, that whatever is the initial pitch error of any CD player, I get the same 3.5ppm (EDIT: which is the pitch error of the measurement interface, so the real result is below 1ppm) when using the LEO BODNAR.
Is the above improvement due solely to the better clock precision, the lower phase noise or both? I don't know, but I could measure that difference.
Like I said about the SMSL G1 Clock I was not expecting this type of benefit, and I see it repeats here with a different clock and many different additional devices the I used. The one thing that did not change is the SMSL D200 DAC, so next step would be to confirm all the above with another DAC having a clock input.
----
A bit of SINAD please
Yeah, you can ask for it, and for fun I share, like I did with the G1, let me use a CD player as the source, since it is my motto.
This is my standard measurement of 999.91Hz undithered test tone @0BFS from the XLR analog outputs of the Teac VRDS-25x:
When feeding the SMSL D200 DAC from coax digital output of the Teac, I get:
Yes, much better, at the frontier of CDA max performances: 16bits resolution is what we want and what we get from that full scale signal.
Now, let's feed the SMSL D200 with the LEO BODNAR LBE-1421:
Nearly identical, main difference being two different runs. But look closer at the foot of the fundamental... I already talked about that. Let me help zooming a little:
The foot of the fundamental gets less random and correlated noise. Being located here, this can't make an audible difference, but it again means more precision, ie less (random, deterministic, intrinsic, data, whatever)-jitter, obviously.
Conclusion
The geek of me likes a lot this device
Thanks to @MC_RME for suggesting it!Having the clock precision of a cesium-based atomic clock, not for the price of a car, is really cool. The possibility to tune the output to any frequency, with the same precision, is also genius and makes it easily reusable.
I love the technical idea behind as well, which comes from the very first atomic clocks that were used to disciplined Xtals in the late 40s. Nearly 100 years later, we get even better than that at home for the price of a good restaurant
The only issue is more cables are required for a very limited use case when talking about feeding only one device with it, as I did here.
At the end of the day, there are many stuff that we buy and make no difference (or have a negative impact). So, you don't need this one, but some of you will still want it
It is always apéro time somewhere in the world, and that time has come for me. A whisky and "aLCHEMY dIRE sTRAITS Live" for me please. Cheers!
I hope you enjoyed this review and I wish you a lovely weekend.
Flo
Last edited:
Cost wise, it’s not too excessive either. Nice review.
