I used to browse tunein radio stations like
BBC R3 - BBC Radio 3 91.3 FM London - Listen Online
Does DAM play flawlessly tunein, with MAC OS web browsers?
Can anybody try it actually with a working DAM dac?
I had problems with diyinhk ES9018 board, while ES9018K2M works fine.
BBC R3 - BBC Radio 3 91.3 FM London - Listen Online
Does DAM play flawlessly tunein, with MAC OS web browsers?
Can anybody try it actually with a working DAM dac?
I had problems with diyinhk ES9018 board, while ES9018K2M works fine.
With the 20150826 firmware the problem is still there. The pin 18 glitch has actually become longer in duration.
With v0.99 20150826 firmware:
With previous v0.99 20150817 firmware (the measured output pulse amplitude here is off by 10x due to incorrect probe scale setting) :
Attachments
Did some experimenting with setting up true balanced setup with two dam1021's.
Two dam1021's with input switch and volume pot, no microcontroller.
This setup will allow volume control and input switching only. To make it work, you need to:
Two dam1021's with input switch and volume pot, microcontroller with a single serial port.
This setup will allow volume control, input switching and switching of preloaded filters. To make it work, you need to:
Two dam1021's, controlled only by a microcontroller with a single serial port.
This setup will allow volume control, input switching and switching of preloaded filters. To make it work, you need to:
[*] Do the same as above, except for connecting the volume pot and input switch.
[*]Control the dual-dam setup with V, I and F commands from serial port.
To access umanager functionality, each dam1021 has to be individually connected to a dedicated serial port on the microcontroller. In this case it is the responsibility of microcontroller software to keep both dac settings in sync.
I plan to implement a digital logic circuit that will normally keep both dacs daisy chained as described above to allow automatic synchronization of V, I and F settings. Direct connection to individual dam1021 serial ports will be enabled by GPIO driven control lines. They would only need to be enabled for accessing umanager functionality, thus avoiding the need for synchronizing the basic V, I and F commands in software.
Two dam1021's with input switch and volume pot, no microcontroller.
This setup will allow volume control and input switching only. To make it work, you need to:
- Make sure both DAM boards have the same firmware and filters loaded, and the same presets are configured in umanager.
- Configure one dam1021 with mode=bal-left
- Configure the other dam1021 with mode=bal-right
- Connect the volume pot and input switch to the first DAM board.
- Make sure an isolated 3.3V power supply is connected to ISO +3.3V and ISO GND pins of both DAM boards.
- Connect ISO TXD OUT from the first DAM (with volume pot and switch connected) to ISO RXD IN pin on the second DAM.
- Take balanced output from CH1/CH2 pins of DAM J7 connector (the unbuffered output)
- Connect pin 1 of the XLR connectors directly to the grounded metal enclosure. For prototyping without an enclosure, connect XLR pin 1 to GND at the power supply. Do NOT connect it to the DAM board.
- Parallel the I2S input lines on both dacs.
- Parallel the SPDIF inputs from the TOSLINK connector.
- Parallel the differential SPDIF inputs from the input transformer. Remove the 300R resistor. (Did not actually test it yet, this setup will result in 50 Ohm input impedance instead of the nominal 75 Ohm)
Two dam1021's with input switch and volume pot, microcontroller with a single serial port.
This setup will allow volume control, input switching and switching of preloaded filters. To make it work, you need to:
- Do all the steps above.
- Connect ISO RXD IN pin of the first DAM board to the TX line of the microcontroller serial port.
- Connect the ISO TXD OUT pin of the second DAM board to the RX line of the microcontroller serial port.
Two dam1021's, controlled only by a microcontroller with a single serial port.
This setup will allow volume control, input switching and switching of preloaded filters. To make it work, you need to:
[*] Do the same as above, except for connecting the volume pot and input switch.
[*]Control the dual-dam setup with V, I and F commands from serial port.
To access umanager functionality, each dam1021 has to be individually connected to a dedicated serial port on the microcontroller. In this case it is the responsibility of microcontroller software to keep both dac settings in sync.
I plan to implement a digital logic circuit that will normally keep both dacs daisy chained as described above to allow automatic synchronization of V, I and F settings. Direct connection to individual dam1021 serial ports will be enabled by GPIO driven control lines. They would only need to be enabled for accessing umanager functionality, thus avoiding the need for synchronizing the basic V, I and F commands in software.
My suggestions for a new board revision.
Skip the balanced output opamp section and XLR jacks entirely. This should make the board somewhat smaller and reduce costs slightly. Its far more ideal to use dual boards which provides a true differential pair now that its possible. Smaller boards mean its easier to fit multiples into a chassis.
Expose any unused I/O pins of the uC so that these could be later used for something useful such as multi-board state syncing, easier mute shunting, etc. Tapping off of bare pins isn't so ideal.
I'd go on about removal of AC power input capability but granted this one is likely so Soren doesn't have to deal with people blowing up their board by bunging the polarity on the power inputs, the bridge rectifier prevents many RMA headaches.
Anyone else have any suggestions?
EDIT: Adding
Master & slave configuration for multiple boards? Single master clock from master board? Does it matter?
Skip the balanced output opamp section and XLR jacks entirely. This should make the board somewhat smaller and reduce costs slightly. Its far more ideal to use dual boards which provides a true differential pair now that its possible. Smaller boards mean its easier to fit multiples into a chassis.
Expose any unused I/O pins of the uC so that these could be later used for something useful such as multi-board state syncing, easier mute shunting, etc. Tapping off of bare pins isn't so ideal.
I'd go on about removal of AC power input capability but granted this one is likely so Soren doesn't have to deal with people blowing up their board by bunging the polarity on the power inputs, the bridge rectifier prevents many RMA headaches.
Anyone else have any suggestions?
EDIT: Adding
Master & slave configuration for multiple boards? Single master clock from master board? Does it matter?
Last edited:
Splitting out the analog and digital power input sections, this might simplify the regulation to a bipolar 12VDC supply @60mA and a positive 12VDC supply @250mA
Add onboard an USB input (a suitable STM32 should be able to handle that for a fraction of the cost of an USB interface as it saves lot of here unnecessary components). Provide volume control via USB audio.
Make the DAM modular e.g. input-controll-FPGA-module + 2X ladder module + ladder analog power supply module+ output buffer module.
It would be nice if the input-controll-FPGA-module would have an I2S output at choosable oversampling frequency. I would love to have that as fifo+filter-unit to test with other DACs, bypassing their internal filters as much as possible.
A optional shunt resistor volume control (e.g. 4 bit) after the DAC ladder operated by the ordinary volume control (coarse part the shunts, fine as currently done)
DAC ladder modules mit MSB resistors not fitted (preferable for leaded MSB resistors).
Last edited:
Wow,
Just think of all of the inventory SOEKRIS would have to carry with all of those variations!
Have a feeling we can all wish but the architecture is set in stone, I suspect, and for all of the suggested variations and the reduced production volume for each would require a big price increase. Even with components left out, the cost of components would be swamped by the inventory costs.
Through hole resistors? Cannot imagine that happening. Would have been good if SOEKRIS had used one of the larger SMD formats but he didn't and there are many advantages to the small size of the package. I can see a possibility of a stock board sounding better than an expanded board with provision for through hole bulk foil resistors.
It is all compromise and the longer you think about it SOEKRIS made many good decisions. We have to remember he made and priced these things to sell in quantity. This is not a bespoke product and if it was I would not be able to afford one!
I bet the only changes we will see are firmware changes.
I thought a run of boards was underway so even if there were new options to be added they would be in the future and not the near future.
I can imagine the possibility, if enough could decide on the one variation they wanted, of a group buy based on a minimum number of subscribers willing to put their money upfront for a special production run which would require a willingness to wait between paying and receiving.
Just think of all of the inventory SOEKRIS would have to carry with all of those variations!
Have a feeling we can all wish but the architecture is set in stone, I suspect, and for all of the suggested variations and the reduced production volume for each would require a big price increase. Even with components left out, the cost of components would be swamped by the inventory costs.
Through hole resistors? Cannot imagine that happening. Would have been good if SOEKRIS had used one of the larger SMD formats but he didn't and there are many advantages to the small size of the package. I can see a possibility of a stock board sounding better than an expanded board with provision for through hole bulk foil resistors.
It is all compromise and the longer you think about it SOEKRIS made many good decisions. We have to remember he made and priced these things to sell in quantity. This is not a bespoke product and if it was I would not be able to afford one!
I bet the only changes we will see are firmware changes.
I thought a run of boards was underway so even if there were new options to be added they would be in the future and not the near future.
I can imagine the possibility, if enough could decide on the one variation they wanted, of a group buy based on a minimum number of subscribers willing to put their money upfront for a special production run which would require a willingness to wait between paying and receiving.
That's why I tried to keep my ideas realistic if Soren was wanting to redesign the board a bit. The differential inverter XLR out section is really pointless now that you can gang boards, and if Soren is indeed adding a jfet mute/shunting circuit then he will be redesigning the board at least somewhat.