• These commercial threads are for private transactions. diyAudio.com provides these forums for the convenience of our members, but makes no warranty nor assumes any responsibility. We do not vet any members, use of this facility is at your own risk. Customers can post any issues in those threads as long as it is done in a civil manner. All diyAudio rules about conduct apply and will be enforced.

Pulsar Clock - Ultra Low Noise OCXO

In the context of the continuous research that involves many audiophiles and technicians in the improving of the listening experience we concentrated to the central role of the oscillator in the digital to analog conversion process.
The analyses led to the implementation of the Pulsar Clock, an ultra low phase noise oscillator capable to assure significant improvement of the sound quality reconstructed by the Digital to Analogue Converters.
The Pulsar Clock is a low power consumption oscillator compatible with the standard DIL14 clock pin layout, it requires 50 mA only for regular operation and just 150 mA during the brief warm-up phase at 3.3 Vcc.

Pulsar Clock is a so extreme oscillator to be classified as a "Dual Use" device whose distribution is ruled by severe international regulations.
For this reason we are authorized by the European Union Authority to distribute the Pulsar Clock in a limited number of countries:
Australia, Austria, Belgium, Bulgaria, Canada, Croatia, Cyprus, Czech Republic, Denmark, Estonia, Finland, France, Germany, Greece, Hungary, Ireland, Italy, Japan, Latvia, Liechtenstein, Lithuania, Luxembourg, Malta, Netherlands, New Zeland, Norway, Poland, Portugal, Romania, Slovakia, Slovenia, Spain, Sweden, Swiss, United Kingdom, Unites States of America.

For those who are not in the above countries and are anyway interested to the Pulsar Clock a specific quotation request page is available to collect your wish, but a quotation will be provided only after and if a specific export authorization will be released by the relevant authority.

For the above authorized destinations a first batch of Pulsar Clock is now available and it is possible to request a quotation for the following frequencies (specified in MHz):
11.289600, 12.288000, 22.579200, 24.576000, 45.158400, 49.152000, 90.316800, 98.304000, 100.000000

The Datasheet and the Quotation Procedure are available at the following link:

https://drive.google.com/file/d/0B4JU1DLmHzHsTXNKVWN0UzlvQW8
.
 
ADM7150 for clocks and DACs

To work at their best the most performing oscillators need a good power supply; we designed the Pulsar Power regulator based on the ultralow noise ADM7150.
The board is ideal to power the most demanding oscillators and both analog and digital DACs sections.
The gold plated board provides a +3.3 Vdc low noise output regulated voltage accepting in input a voltage in the +5 and +9 Vdc range with up to 800 mA output current, compatibly with thermal environment.
To enhance the filtering capability both C0G, Tantalum and X7R technologies are implemented on the whole regulation line.
During the design tests it appeared clear the great influence of the Cbyp capacitor, the ADM7150 noise spectral density roll-off capacitor, so we allowed the user to customize such capacitor, both SMD and/or TH, with an easy jumper selector interface to calibrate it for best performance.
 

Attachments

  • Pulsar Power Schematic 1.0.jpg
    Pulsar Power Schematic 1.0.jpg
    65.5 KB · Views: 3,696
  • Pulsar Power Datasheet Rev. 1.0.pdf
    464.4 KB · Views: 461
  • Pulsar Power Board.jpg
    Pulsar Power Board.jpg
    87.8 KB · Views: 3,574
  • Items in Packaging.jpg
    Items in Packaging.jpg
    84.8 KB · Views: 3,501
  • Pulsar Power driving DAC and Clock 1.jpg
    Pulsar Power driving DAC and Clock 1.jpg
    81.3 KB · Views: 3,503
  • Pulsar Power driving DAC and Clock 2.jpg
    Pulsar Power driving DAC and Clock 2.jpg
    82.3 KB · Views: 3,765
With the 3-pin configuration Vin-Out,No problem with the mass sharing?.In the Datasheet of the ADM 7150,the best level of noise spectral density is Reached with a value of Cbyp between 500μf -1000μf,why annoy to put a lower value?."A little off topic",The Sanyo Oscon SG serie is a good choice for Cbyp,the SG serie is replaced by the SP,it is also good.
 
Last edited:
With the 3-pin configuration Vin-Out,No problem with the mass sharing?.
In principle a 3-pin regulator doesn't have any problem with the mass-sharing because it operates "in-line" and the ground pin is not used for return of the current flowing through the load, it is instead the reference with respect to the output voltage is regulated.
Through the ground pin flows only the small current that is used by the 3-pin regulator internal electronics, a bias current; I made a quick sketch to show the main and bias currents flow on the PCB where the 3-pin regulator is installed. This current is referred in the ADM7150 datasheet as the Operating Supply Current (Ignd) and is in the range 4 to 9 mA.
A line regulator doesn't galvanically isolates the load by the power source so it must have a unique ground reference. This is also the principle of the star ground, a well defined single reference point for all the ground referenced items on the board. Sometimes this star center could be also the whole ground plane if the application allows it, sometimes the ground plane itself is "star shaped" in accordance to this principle.
In conclusion the 3-pin configuration works on the principle of the mass sharing and EMI/EMC problems could rise only at PCB level, but all the PCBs I have seen in our diyAudio world shows great care to manage this issue.

In the Datasheet of the ADM 7150,the best level of noise spectral density is Reached with a value of Cbyp between 500μf -1000μf,why annoy to put a lower value?
Great question! It is exactly what was in my mind when I saw for the first time the datasheet, I started to immagine plenty of thousend of microfarad installed in the Cbyp location, and so I did as soon an ADM7150 chip was in my hands.
At this point I discovered the first issue (well documented in the datasheet), the output voltage needed "minutes" to reach the working range!
We designed the Pulsar Power mainly to power up the Pulsar Clock so here I'm referring to our experience in this application but the same criteria should apply also to DACs.
The most impressive and unexpected effect was the influence of the value of this capacitor on the sound. The quality of the sound was influenced by the value of this capacitor and the good news was that it was possible to "tune" the system to perfection just seeking for the right value of this capacitor.
We tested the effect of all the capacitors required by the ADM7150 coming to the triple technology solution, but only the Cbyp capacitor showed a so impressive and useful tuning property on the sound system.
On different systems the best fit value for the Cbyp capacitor changed so we decided to install the 1uF and 10uF values that in our tests showed the best results leaving the room on the board to allocate other capacitors for a perfect clock tuning.
Why this unexpected effect? I thought long to this issue, in principle the Cbyp capacitor should just filter the reference buffer but somehow it influences also the output regulation. Maybe it makes the regulator, let me say, "slower" to react to load demands while in audio applications we need a quick and prompt reaction to load changes to drive Clocks or DACs output with great precision in terms of time (so jitter) and level. A too large Cbyp improves the noise shape and this could be a benefit for less demanding loads while our audio applications have larger benefit from smaller Cbyp because the ADM7150 noise shape is in any case excellent.
With a well tuned Cbyp we reached performances very very near to a LiFePo4 battery.

."A little off topic",The Sanyo Oscon SG serie is a good choice for Cbyp,the SG serie is replaced by the SP,it is also good.

The Sanyo Oscon capacitors are for sure a good choice for Cbyp, their reputation in audio applications is great.
Any design requires a verification with physical hardware; our experience on the Cbyp capacitor tells that much more than any other component its analytical choice need to be verified by an "on the field" experience.
 

Attachments

  • 3-pin Regulator Main Current Flow.jpg
    3-pin Regulator Main Current Flow.jpg
    54 KB · Views: 1,205
Pulsar Power board size

clsidxxl, your post makes me realize that I never mentioned the exact dimensions of the board, I post them here with the datasheet accordingly updated.
 

Attachments

  • Pulsar Power Datasheet Rev. 1.0.1.pdf
    458.7 KB · Views: 239
  • Pulsar Power Board Dimensions.jpg
    Pulsar Power Board Dimensions.jpg
    22.7 KB · Views: 943
Pulsar,technology Oven Controlled is it really beneficial for audio XO ?
Both our experience and the preliminary tests described on diyAudio by iancanada and Bunpei says: absolutely yes.
We experienced a sort of Live Effect, the music and the voice was incredibly natural.
We have not yet written about our own experience because we would like that other users could express their own impressions (I could say only any good :)) instead we will post soon how we conducted our test because the description of the system that we used is surely interesting for the discussion.
 
Pulsar Clock Synchronous and Asynchronous Playback Test Setup

A Digital to Analog Converter is like an orchestra that plays music following the exact time stated by the Director, the Clock.
Each hesitation of the Director makes the orchestra less clean, more cloudy. At the same way a jittered clock makes even the best DAC sound more blurred loosing the details that makes music alive.
Due to its exceptionally low phase noise and a consequent negligible jitter the Pulsar Clock pushes at the top the performance of any DAC, to test this performance and compare different clocks we made two different setup for synchronous and asynchronous DAC convertion.
Great care was taken to minimize the hardware induced jitter compatibly with the need to realize a very flexible and easily reconfigurable structure to allocate different clocks and different frequencies.

Synchronous Configuration is described by the block diagram in picture S1, the whole hardware is shown in picture S2:
- Player: Volumio 1.4 from Michelangelo Guarise on Raspberry PI with I2S output, a bitperfect player up to 24/192. Picture S3 shows I2S connections from the Raspberry PI
- FIFO buffer from iancanada to synchronize the I2S data flow
- Isolator from iancanada to ehnance data accuracy providing galvanic isolation
- Pulsar Clock pair on Dual-XO board from iancanada, the location where the ultralow jitter reclock is made. Current DualXO version is able to power the Pulsar Clock pair both during the short warm-up and the operational phases. In our configuration we use a 5Vdc battery to supply the DualXO board with minimum noise. Picture S4 shows the "FIFO - Isolator - DualXO" de-jitter block
- ES9018 Buffalo III from Twisted Pear Audio powered by Pulsar Power boards on 3.3V Analog and Digital lines, the board was modified to accept an external clock on the dedicated three pins with the iancanada adapter. Pictures S5 and S6 show the most critical digital path, the connection between the DualXO and Buffalo III boards, this connection could be slightly improved in a fixed implementation but performs very well also in this flexible configuration
- NTD1 I-V Converter Version 3 from Owen, a stunning Class A converter with a state of the art linearity expressly designed to accomodate the Buffalo III board for maximum quality of the analog signal

Asynchronous Configuration is described by the block diagram in picture A1:
- Player: Volumio 1.4 from Michelangelo Guarise on Raspberry PI with I2S output, a bitperfect player up to 24/192
- ES9018 Buffalo III from Twisted Pear Audio powered by Pulsar Power boards on 3.3V Analog, Digital and Clock lines, the board was modified to accept an external clock on the dedicated three pins with the iancanada adapter. Picture A2 shows the Pulsar Clock on the iancanada clock adapter
- NTD1 I-V Converter Version 3 from Owen, a stunning Class A converter with a state of the art linearity expressly designed to accomodate the Buffalo III board for maximum quality of the analog signal
 

Attachments

  • S1 - Pulsar Clock Test - Synchronous DAC Block Diagram.jpg
    S1 - Pulsar Clock Test - Synchronous DAC Block Diagram.jpg
    97 KB · Views: 1,637
  • S2 - Pulsar Clock Test System Overview.jpg
    S2 - Pulsar Clock Test System Overview.jpg
    300 KB · Views: 1,615
  • S3 - Raspberry PI I2S lines to iancanada FIFO.jpg
    S3 - Raspberry PI I2S lines to iancanada FIFO.jpg
    209.4 KB · Views: 3,153
  • S4 - Pulsar Clock driving iancanada FIFO and Dual XO isolated de-jitter block.jpg
    S4 - Pulsar Clock driving iancanada FIFO and Dual XO isolated de-jitter block.jpg
    180.4 KB · Views: 1,560
  • S5 - I2S and Clock lines from Pulsar Clock on inacanada Dual XO to TPA Buffalo III with Pulsar P.jpg
    S5 - I2S and Clock lines from Pulsar Clock on inacanada Dual XO to TPA Buffalo III with Pulsar P.jpg
    210.2 KB · Views: 1,493
  • S6 - TPA Buffalo III with Pulsar Power boards and iancanada adapters to I2S and Clock input.jpg
    S6 - TPA Buffalo III with Pulsar Power boards and iancanada adapters to I2S and Clock input.jpg
    136.1 KB · Views: 686
  • A1 - Pulsar Clock Test - Asynchronous DAC Block Diagram.jpg
    A1 - Pulsar Clock Test - Asynchronous DAC Block Diagram.jpg
    70.9 KB · Views: 758
  • A2 - TPA Buffalo III Asynchronous configuration with Pulsar Clock on iancanada adapter and Pulsa.jpg
    A2 - TPA Buffalo III Asynchronous configuration with Pulsar Clock on iancanada adapter and Pulsa.jpg
    204 KB · Views: 862
Member
Joined 2003
Paid Member
Also be interesting to see a setup with Acko's S03 or S04 isolator/synchronous reclocker board between the R-Pi and the Buffalo DAC. While the R-Pi can't take a clock input from the S03 or S04, I believe you could use it to provide both isolation and his 'pseudo-sync' mode with or without I2S signal reclocking.

Or even with a BBB & Miero's distro which CAN take a clock input from the S03 or S04?

Greg in Mississippi

P.S. Did you hear improvements in using your regulator module instead of TP's?
 
Last edited:
Greg,
for sure both Acko's "Digital Isolator & Re-clocker" and Ian's "FIFO - Isolator - DualXO" chain are excellent solutions to re-clock. While Ian's solution introduces the FIFO to synchronize data the S03 board delegates this task to the upstream device avoiding virtually any risk of full or empty buffer during long performances. We could implement also an S03 + BBB solution, we already have an S03 + Amanero USB isolator implementation, but the key point is the tricky installation of a DIL14 clock on a SMD native board.
Ian's system is DIL 14 native so it was the natural test solution for Pulsar Clock evaluation. Installing a DIL14 socket on the S03 board we espect the same benefits from Pulsar Clock independently by the upstream player, BBB Rpi or Amanero.
We hope that other users could be able to install DIL14 devices on SMD landing patterns in order to test different configurations.

About the Post Scriptum: Pulsar Power board was originally designed to power the Pulsar Clock at best so in that position it works very fine, I dare to say that in some cases it performs like a LiFePo4 battery, my favourite source for the clock. There are two advantages using the Pulsar Power board in place of a LiFePo4 battery, the first is obvious, it is always "ready to go" without need for battery re-charge.
The second is that LiFePo4 batteries fully charged easily reaches voltages above 4.00 V even using dedicated chargers, such "high" voltage for sure soon will burn a component and if this is an expensive clock or a DAC board it won't be fine.
You asked me if the Pulsar Power performs better than the TP ones. In my opinion we are not in the right position to spend a word on a comparison because obviously we like our products, so sorry for not answering this point.
 
Pulsar,very nice configuration,my set up is very similar,Cubietruck/Volumio-I2S to FIFO/Isolator Digital-Dual XO/ Crysteck CCHD957(for now):),I2StoPCMconvertorBoard to Audio GD Ref 7.1.I'm finished setting up,connexion,case aluminium for Cubietruck,proper implementation is critical.
With what you feed the Raspi?you should use UFL cable between the FIFO/Digital Isolator/XO Board.
You said
'''Pictures S5 and S6 show the most critical digital path, the connection between the DualXO and Buffalo III boards, this connection could be slightly improved in a fixed implementation but performs very well also in this flexible configuration'''' What do you mean by fixed implementation?
 
Last edited:
clsidxxl,
I start from the end, I used with a little bit of fantasy the expression "fixed implementation" to indicate a normal DAC built to perform at home and not to be quickly easily and continuously modified like the configuration that we presented and that we use for test moving among various audio systems sometimes located in cities well far one from another.
Your DAC is what I ment when I wrote "fixed implementation"; I saw in your post #3205 on Ian's FIFO thread the pictures that you published in March and I am really impressed seeing how you modified the Audio GD Ref 7.1 keeping the R-2R converters and changing all the input digital section, a very clean implementation. In that pictures didn't appear yet the Cubietruck, for sure a younger and very valuable substitute of the R-PI.
About your question on the UFL connectors: in the synchronization path from the FIFO to the DualXO you have to consider that the DualXO board doesn't have any UFL connector on the input side making hard the usage of coaxial cables even if you did it on your DualXO where it was motivated by the long lines and the need to preserve the already de-jittered signal.
The key point is to keep under control possible transmission errors induced by the jitter up to the DualXO board where the de-jitter is performed. Consider that the UFL connectors are designed to feed 50ohm lines while digital connections on our boards are not at all impedance matched lines and the cables provided by Ian for sure doesn't induce more jitter than a UFL connection on a such short path.
The most critical data transfer is from DualXO output, where the signal performs its lowest jitter, to the DAC board input and here short coaxial cables on UFL connectors are for sure the best approach.
Your last question: the Raspberry PI power supply is fed via the USB connector by means of a TPS7A4700EVM evaluation board that is partially hidden by the R-PI in the system overview picture.