Terrific
This solution will I preferred to get the 25MHz frequency, so I am in at that, when it will be available, but do You have a SC 6.25MHz crystal to order?
The new Pierce oscillator is designed to implement AT-Cut crystals only, so you need a fundamental AT-Cut at 6.25 MHz.
I can order whatever frequency, as I already said the problem is the MOQ required from the manufacturer, at least 10 pcs for each frequency.
I cannot place an order for 10 pcs if there is not enough interest, I don't stock any crystal since I'm an hobbist not a crystal reseller.
I can order whatever frequency, as I already said the problem is the MOQ required from the manufacturer, at least 10 pcs for each frequency.
I cannot place an order for 10 pcs if there is not enough interest, I don't stock any crystal since I'm an hobbist not a crystal reseller.
You can already join the GB and add yourself to the interest list for 6 and 6.25 MHz crystals, you have only to choice between AT-Cut and SC-Cut.
The prices are the same of the 5.6448 MHz crystals, SC-Cut are more expensive and also you have to add a pair of doublers, while the AT-Cut are a little cheaper and I could consider to develop the All In One board, surely less expensive than 1 oscillator + 2 doublers.
In this way we can understand if there is enough interest to meet the MOQ.
The prices are the same of the 5.6448 MHz crystals, SC-Cut are more expensive and also you have to add a pair of doublers, while the AT-Cut are a little cheaper and I could consider to develop the All In One board, surely less expensive than 1 oscillator + 2 doublers.
In this way we can understand if there is enough interest to meet the MOQ.
Any chance to have SC-Cut version? I think most of us will need 24/25M from 5/6M sc-cut crystal here.
Poting
My info was old. It seems the first generation BVA's were AT cut and later were ST cut. Brief History of the Development of Ultra-Precise Oscillators for Ground and Space Applications | IEEE UFFC
I found the "York" oscillator: http://eprints.whiterose.ac.uk/141563/1/08540461.pdf Its design using a matching transformer is really interesting. I found some other efforts with transformers and balanced circuits also interesting.
Unfortunately no SC-Cut All in One version, the Driscoll and Differential oscillators are too big to fit in a single box togheter with a couple of doublers.
Moreover the best solution for high performance crystals such as the Laptech Sc-Cut 5/6 MHz is using separate boxes for each device.
The Differential oscillator TWTMC-EXO is based on the York design although it has been heavily modified.
It works very good, the phase noise performace are almost the same of the Driscoll one, only a little higher noise floor due to the position of the crystal in the feedback loop, while in the Driscoll the crystal is placed in the emitter circuit where the impedance is much lower.
@andrea_mori: When using a clock distribution chip with differential input, in our case the ADCLK954 ( https://www.mouser.com/datasheet/2/609/ADCLK954-1502655.pdf ). What's better in your opinion, go with the Driscoil and a unbalanced coax cable connection and use the input of the ADCLK in single ended mode or use your differential oscillator?
What do you think about transformers in the clock signal line to achieve ground isolation? Do these have an impact on phase noise performance?
What do you think about transformers in the clock signal line to achieve ground isolation? Do these have an impact on phase noise performance?
Current status and future developments
We will order the proto PCB of the All in One version this weekend, then we have to test and measure its phase noise to publish the plot.
This combo version has a few compromises, it's only suitable for AT-Cut crystals at 5.6448 and 6.144 MHz (maybe 6 and 6.25 MHz if there is enough interest), no other frequencies and no SC-Cut type.
Since it implements a couple of frequency doublers you can configure the output for 11.2896/12.288 MHz or 22.5792/24576 MHz installing one or two doublers.
The board is for a single oscillator, so you need two boards to get both 22.5792 and 24.576 MHz.
Summary of test and measurement:
- New Driscoll oscillator (TWTMC-DRIXO), the board is ready, it has been measured and plots have been published at all crystals frequency (except for the 12.288 MHz since the crystal is not available at now), it's suitable for SC-Cut crystal from 5 up to 24 MHz, the phase noise performance are close to the best oscillators on the market
- New Differential oscillator (TWTMC-EXO), the board is ready, it has been measured and plot has been published at 5.6448 MHz (other frequencies have to be measured, expect something similar to the Driscoll), it's suitable for SC-Cut crystal from 5 up to 24 MHz, the phase noise performance are close to the best oscillators on the market
- New Pierce oscillator (TWTMC-PXO), the board is ready, it has been measured and plots have been published at 5.6448 and 6.144 MHz (other frequencies have to be measured, expect worse performance due to the crystals specs), it's suitable for AT-Cut crystal from 5 up to 98 MHz, the phase noise performance of the 5/6 MHz oscillators are very good, we cannot measure the phase noise from 45 to 98 MHz due to the upper frequency limit of the Timepod
- New Frequency Doubler (TWTMC-DBM), the board is ready, it has been measured and plots have been published at 5 to 11 MHz, 11 to 22 MHz and 5 to 22 MHz (22 to 45 MHz and 45 to 90 MHz phase noise cannot be measured because of the Timepod limit), it's suitable for Driscoll and Differential oscillators from 5 MHz to 24 MHz, the phase noise performance of the 5 MHz oscillator multiplied by 2 and 4 (to 11 and 22 MHz) are superlative, maybe the best oscillators for digital audio ever
- New Pierce All in One (TWTMC-PXO-AIO), the proto board will be tested and measured in a couple of weeks (expect good phase noise performance at 11/12 and 22/24 MHz, but obviously worse than the Driscoll and Differential oscillators followed by one or two dedicated doublers), it's only suitable for AT-Cut crystals at 5.6448 and 6.144 MHz (6 and 6.25 MHz depending on the interest)
In the end how to make your choice:
- if you need 5/6 MHz oscillators, the TWTMC-DRIXO and TWTMC-EXO are state of the art solutions in the region of the best oscillators on the market, the TWTMC-PXO oscillator s a valid alternative at lower cost with good phase noise performance
- if you need 11/22 MHz oscillators, the TWTMC-DRIXO and TWTMC-EXO at 5/6 MHz followed by the TWTMC-DBM frequency doubler are state of the art solutions in the region of the best oscillators on the market, the TWTMC-DRIXO and the TWTMC-EXO at 11/22 MHz are the second option in term of phase noise performance at lower cost, the TWTMC-PXO-AIO combo with AT-Cut crystals at 5/6 MHz oscillator is a valid alternative with good phase noise performance (to be measured and price to be defined), the TWTMC-PXO Pierce oscillator with AT-Cut crystals at 11/22 MHz is the last option, the cheapest solution but obviously the worst performer of the group
- if you need 22/24 MHz oscillators, see above (11/12 MHz), there is also an intermediate solution, the second in term of performance, the TWTMC-DRIXO or the TWTMC-EXO with crystals at 11/12 MHz followed by a frequency doubler
- if you need 45/49 and 90/98 MHz oscillators, best solutions are the TWTMC-DRIXO or the TWTMC-EXO oscillators with crystals at 22/24 MHz follwed by one or two TWTMC-DBM frequency doublers, although we cannot measure the phase noise at these frequencies you can apply the theory adding 6 dB for each multiplication to the phase noise of the 22/24 MHz oscillators (maybe in practice the phase noise will be better than the theory) so again state of the art performance, the other option is the TWTMC-PXO oscillator with crystals at 45/49 or 90/98 MHz, obviously with worse phase noise performance
This GB will include:
- AT-Cut crystals from 5 up to 98 MHz
- SC-Cut crystals from 5 up to 24 MHz (45 to 90 MHz are available for old oscillators only)
- TWTMC-DRIXO New Driscoll XO
- TWTMC-EXO New Differential XO
- TWTMC-PXO New Pierce XO
- TWTMC-DBM Frequency doubler
- TWTMC-PPG New Pierce entry with crystal
- TWTMC-STS-SX Sine to square converter (SX)
- TWTMC-STS-DX Sine to square converter (DX)
- TWRPS-UGL-F Linear regulator 15V for oscillators
- TWRPS-pp Push-pull regulator 12V-18V for oscillators
- TWSAFB-TX I2S Transmitter
- TWSAFB-RX I2S Receiver
- TWSAFB-RPI RPI Power & isolated I2S
- TWRPS-LBS-M LiFePo4 Battery Power Supply Main board and control
- TWRPS-LBS-D LiFePo4 Battery Power Supply Additional board controlled by the main board
- TWRPS-LBS-P LiFePo4 Battery Power Supply Power supply unit and batteries recharge
The next GB will include:
- TWSAFB Asynchronous FIFO buffer Lite version
- TWSAFB-UI FIFO user interface
- TWSAFB-OI FIFO isolated outputs (for DAC switching on LRCK/MCLK such as TDA1541A, AD1865, ESS and so on)
- TWSAFB-OIR FIFO isolated output and reclocker (for DAC switching on BCLK such as PCM1704, PCM1794 and so on)
- TWSDAC-LT 16/24 bit 192 kHz segmented sign-magnitude discrete DAC
All the boards will be offered as:
- Finished board, fully assembled and tested (expect long time since we have to assembly some parts by hand)
- Semi-finished board, users have to assembly a few parts, mostly through hole and a few SMD
- Bare board, users have to assembly all the parts
I will publish soon a spreadsheet with the final price of each device and a comparative table of the prices for each option.
We will order the proto PCB of the All in One version this weekend, then we have to test and measure its phase noise to publish the plot.
This combo version has a few compromises, it's only suitable for AT-Cut crystals at 5.6448 and 6.144 MHz (maybe 6 and 6.25 MHz if there is enough interest), no other frequencies and no SC-Cut type.
Since it implements a couple of frequency doublers you can configure the output for 11.2896/12.288 MHz or 22.5792/24576 MHz installing one or two doublers.
The board is for a single oscillator, so you need two boards to get both 22.5792 and 24.576 MHz.
Summary of test and measurement:
- New Driscoll oscillator (TWTMC-DRIXO), the board is ready, it has been measured and plots have been published at all crystals frequency (except for the 12.288 MHz since the crystal is not available at now), it's suitable for SC-Cut crystal from 5 up to 24 MHz, the phase noise performance are close to the best oscillators on the market
- New Differential oscillator (TWTMC-EXO), the board is ready, it has been measured and plot has been published at 5.6448 MHz (other frequencies have to be measured, expect something similar to the Driscoll), it's suitable for SC-Cut crystal from 5 up to 24 MHz, the phase noise performance are close to the best oscillators on the market
- New Pierce oscillator (TWTMC-PXO), the board is ready, it has been measured and plots have been published at 5.6448 and 6.144 MHz (other frequencies have to be measured, expect worse performance due to the crystals specs), it's suitable for AT-Cut crystal from 5 up to 98 MHz, the phase noise performance of the 5/6 MHz oscillators are very good, we cannot measure the phase noise from 45 to 98 MHz due to the upper frequency limit of the Timepod
- New Frequency Doubler (TWTMC-DBM), the board is ready, it has been measured and plots have been published at 5 to 11 MHz, 11 to 22 MHz and 5 to 22 MHz (22 to 45 MHz and 45 to 90 MHz phase noise cannot be measured because of the Timepod limit), it's suitable for Driscoll and Differential oscillators from 5 MHz to 24 MHz, the phase noise performance of the 5 MHz oscillator multiplied by 2 and 4 (to 11 and 22 MHz) are superlative, maybe the best oscillators for digital audio ever
- New Pierce All in One (TWTMC-PXO-AIO), the proto board will be tested and measured in a couple of weeks (expect good phase noise performance at 11/12 and 22/24 MHz, but obviously worse than the Driscoll and Differential oscillators followed by one or two dedicated doublers), it's only suitable for AT-Cut crystals at 5.6448 and 6.144 MHz (6 and 6.25 MHz depending on the interest)
In the end how to make your choice:
- if you need 5/6 MHz oscillators, the TWTMC-DRIXO and TWTMC-EXO are state of the art solutions in the region of the best oscillators on the market, the TWTMC-PXO oscillator s a valid alternative at lower cost with good phase noise performance
- if you need 11/22 MHz oscillators, the TWTMC-DRIXO and TWTMC-EXO at 5/6 MHz followed by the TWTMC-DBM frequency doubler are state of the art solutions in the region of the best oscillators on the market, the TWTMC-DRIXO and the TWTMC-EXO at 11/22 MHz are the second option in term of phase noise performance at lower cost, the TWTMC-PXO-AIO combo with AT-Cut crystals at 5/6 MHz oscillator is a valid alternative with good phase noise performance (to be measured and price to be defined), the TWTMC-PXO Pierce oscillator with AT-Cut crystals at 11/22 MHz is the last option, the cheapest solution but obviously the worst performer of the group
- if you need 22/24 MHz oscillators, see above (11/12 MHz), there is also an intermediate solution, the second in term of performance, the TWTMC-DRIXO or the TWTMC-EXO with crystals at 11/12 MHz followed by a frequency doubler
- if you need 45/49 and 90/98 MHz oscillators, best solutions are the TWTMC-DRIXO or the TWTMC-EXO oscillators with crystals at 22/24 MHz follwed by one or two TWTMC-DBM frequency doublers, although we cannot measure the phase noise at these frequencies you can apply the theory adding 6 dB for each multiplication to the phase noise of the 22/24 MHz oscillators (maybe in practice the phase noise will be better than the theory) so again state of the art performance, the other option is the TWTMC-PXO oscillator with crystals at 45/49 or 90/98 MHz, obviously with worse phase noise performance
This GB will include:
- AT-Cut crystals from 5 up to 98 MHz
- SC-Cut crystals from 5 up to 24 MHz (45 to 90 MHz are available for old oscillators only)
- TWTMC-DRIXO New Driscoll XO
- TWTMC-EXO New Differential XO
- TWTMC-PXO New Pierce XO
- TWTMC-DBM Frequency doubler
- TWTMC-PPG New Pierce entry with crystal
- TWTMC-STS-SX Sine to square converter (SX)
- TWTMC-STS-DX Sine to square converter (DX)
- TWRPS-UGL-F Linear regulator 15V for oscillators
- TWRPS-pp Push-pull regulator 12V-18V for oscillators
- TWSAFB-TX I2S Transmitter
- TWSAFB-RX I2S Receiver
- TWSAFB-RPI RPI Power & isolated I2S
- TWRPS-LBS-M LiFePo4 Battery Power Supply Main board and control
- TWRPS-LBS-D LiFePo4 Battery Power Supply Additional board controlled by the main board
- TWRPS-LBS-P LiFePo4 Battery Power Supply Power supply unit and batteries recharge
The next GB will include:
- TWSAFB Asynchronous FIFO buffer Lite version
- TWSAFB-UI FIFO user interface
- TWSAFB-OI FIFO isolated outputs (for DAC switching on LRCK/MCLK such as TDA1541A, AD1865, ESS and so on)
- TWSAFB-OIR FIFO isolated output and reclocker (for DAC switching on BCLK such as PCM1704, PCM1794 and so on)
- TWSDAC-LT 16/24 bit 192 kHz segmented sign-magnitude discrete DAC
All the boards will be offered as:
- Finished board, fully assembled and tested (expect long time since we have to assembly some parts by hand)
- Semi-finished board, users have to assembly a few parts, mostly through hole and a few SMD
- Bare board, users have to assembly all the parts
I will publish soon a spreadsheet with the final price of each device and a comparative table of the prices for each option.
Both Driscoll and Differential oscillators are single ended sine wave output so you need a sine to square converter and then you have to use the CMOS input of the fanout buffer.
At now the only available sine to square converter is the TWTMC-STS that was designed to fit Ian's FIFO buffer but you can also use it as a standalone converter.
It's a more than honest device since it adds only a little noise floor to the base sine wave oscillator as you can see in the attached plot, and the close in phase noise is practically superimposable.
BTW, we are developing a state of the art sine to square converter to be used in our top audio system, but will be more complex and expensive and it keep long time to be ready.
Transformers are passive devices so usually they don't compromize the phase noise but obviously a measurement would be needed to confirm the theory.
It's a more than honest device since it adds only a little noise floor to the base sine wave oscillator as you can see in the attached plot, and the close in phase noise is practically superimposable.
BTW, we are developing a state of the art sine to square converter to be used in our top audio system, but will be more complex and expensive and it keep long time to be ready.
Transformers are passive devices so usually they don't compromize the phase noise but obviously a measurement would be needed to confirm the theory.
Attachments
is there any more information available on the TWTMC-STS? And yes, I DID use the forum search and also google, did not turn up anything... Apart from that a very quick mouser search did throw up several possibly useable parts from AD with very good jitter specs, for example the LTC6957.
That's what i thought. Thanks for confirming.
The LTC6957 seems to be a good device but the published phase noise measurement plot starts at 100 Hz that's a little far from the carrier.
In digitale audio the close in phase noise is muchh more important, 10 Hz from the carrier and below.
So the only chance to understand the close in phase noise performance is to measure it in circuit.
The TWTMC-STS is a little board that uses the 74AC04 as the squarer. It has been measured several times, you can find phase noise plots on the web, and it was claimed as one of the best sine to square converter (the plot I have published confirms that), better than using a comparator and almost equal or even better than using a ECL device.
In digitale audio the close in phase noise is muchh more important, 10 Hz from the carrier and below.
So the only chance to understand the close in phase noise performance is to measure it in circuit.
The TWTMC-STS is a little board that uses the 74AC04 as the squarer. It has been measured several times, you can find phase noise plots on the web, and it was claimed as one of the best sine to square converter (the plot I have published confirms that), better than using a comparator and almost equal or even better than using a ECL device.
- Status
- Not open for further replies.