I found that it might be the heat problem.
Should I try to buy a cooling fin for the FPGA chip?
@soekris
Please recall I noted an DSD issue with firmware later than 1.06 also but with a dual-mono setup, perhaps it is related. Could it be a filter issue?
Heat could change some timing, but I'm working to redo the whole internal timing, to do the 4K filter I had to double the clock....
Hello all,
After reading this tread for a while, asking some questions sometimes, it’s time for me to present my DAC based on dam1021 rev5.
I decided to put it in a wood case (bamboo) with the Toroidy Transformer(2x14VAC+6x6VAC) outside the box to avoid 230V perturbations.
DIYINHK power supply LT3045 based, LDOVR LT3045 power regulator for iso and HDMI input from audioGD connected on I2S, also for SPDIF input.
Audiozen board and screen (without transformer), OCC wires.
This DAC kicked the audioGD DAC19 big time. It is transparent and sounds very natural. Really enjoy it and would like to thank Soren and the community for the help/advices which helped me a lot in my setup.
Musically,
Faboz
PS: Please stop pollutingthis tread with out of subject and endless discussions, better to listen music
Hello All,
News from my DAC, I recently incorporated into the box a streaming board directly connected to the DAC through I2S. This board eRED-DOCK from Engineered is really really good and matches very nicely with the DAM1021. This board is used in high end equipments (Bel Canto, Audiomat, ...). Designed and manufactured for music.
Compared to my Mano streamer I2S connected but through HDMI, this board goes far beyond and allows to even more appreciate the intrinsic quality of this DAC.
Next step is to incorporate 5V LPS board (better than de zerozone LPS I use for now) and use decent I2S cables since those are the cheapest you can get (test cables).
The board is compatible with native DSD, but I dont know if the DAC is capable of it and I dont know how to connect it to make it work for native. Any idea? For now it is running DSD through DOP.
Musically,
Faboz
Let me introduce the third iteration on the Soekris / MiniSharc DiY combo.
B.t.w what is DiY? You won’t start with a lump of germanium from your mine. Rather you carefully evaluate components somewhere in the value chain as a transistor or a finished board, depending on ambition, skill and resources available.
Here some of the boards with the best technical merits, cost no object where selected. Careful modifications and addons where applied. As this is the final iteration, a modular approach where applied, allowing for later additions.
The goal where to build a unit accepting my sources, processing speaker element correction and crossover with digital output to digital amplifiers and earphone correction. I bet such a unit cannot be procured on the audio market. Hence, a good reason to build beside of the sheer fun.
The modular concept is close to the approach featured by MSB Technology Reference series. Here is a common I2S bus used with multiple Master mode modules using tri-state output ruled by simple priority mechanism. Realized using D-Type Latches With 3-State outputs is used to “mute” the actual unit not in current used. The environmental concerns over long I2S lines were however unfounded, easily proved by empirical evidence and signal integrity is shown by the scope.
The picture shows the fully working one-off prototype of MiniSharc four slots for units, each in plug and play operation. This means that a unit can go active in I2S Master mode on the initiative of the user.
The modules available are these (from right to left):
1. Ethernet/RJ45 interface. Digital Media Renderer for stereo audio UPnP AV 2.0 / DLNA playing and decoding common audio formats from HTTP streams from engineered SA - electronic, engineering and consulting
2. USB audio playback interface. The sampling frequency is defined by the USB streaming itself and depends on the audio source file and computer configuration from engineered SA - electronic, engineering and consulting.
3. Discrete R-2R sign magnitude DAC technology with reclocking FIFO OEM Line Dam 1121 from Soekris Audio ApS, R-2R Sign Magnitude Audio DAC. The Dac is used for the earspeakers using FIR correction of several origins (Stax ED-1 Signature, SRM Monitor or Bob Katz eq’ing).
4. Main DSP unit. Analog Device ADSP21369 always work in Input Slave mode 96kB, 24 bit resolution. The DSP board MiniSharc is from Home | miniDSP. This board is layered above a daughterboard which provide two inputs and four SPDIF outputs, all with BNC contacts from Waveflex - Next Generation DIY Speaker Cabinets. The MiniDSP is fully controlled from a PC-based UI over WiFi using the WI-DG module and from a module VOL-FP using a IR remote-control or with front panel volume and pre-set selection wheel on front panel, all from Home | miniDSP
While Master Clock distribution over coax cable is available from each unit but implementation in multi-master environment is complex and not recommended for the
On the drawing board are now threse modules:
1. An analog conversion unit (ADC) which is based on successive approximation principle. Actually, as far as I can see, there are few if any such implemented modules available for stereo balanced input with I2S output.
2. Speech recognition remote control, using eight Mems microphones in a mic array with DSP complex voice capture function ( beamforming / noise reduction / echo cancellation / de-reverd / Direction of Arrival (DOA) / Voice Activity Dectection (VAD) UMA-8 USB module from Home | miniDSP connected to a raspberry PI accessing Synology server with a Amazon Alexa audio skill software. Amazon Alexa provides a list of supported voice commands, such as “Alexa, ask Audio Station to play the album [album name]”.
A limitation is the compatibility with Amazon Alexa demands any audio files in a format other than MP3 will be converted into 128kbps MP3.
3. A PA213 Bluetooth 5.0 Audio Module BT5.0 CSR8675 Chip with support for APTX-HD from a China-base vendor on eBay.
The two power supplies are a switched medical-grade low noise 5V 15W unit for digital units from www.tracopower.com and a linear +5V / - 5V unit for analog units from www.diyinhk.com.
Each module resides in it’s own module box with an on board master clocking is with low jitter. EMI, Electromagnetic shielding effectiveness using carbon fibre reinforced composites for all units. A simple layer of 2/2 twill 3k dry carbon fibres fabric is able to shield more than 99.999% of the electromagnetic radiation.
https://www.sciencedirect.com/science/article/abs/pii/S1359836818343762
The aluminium chassi, HiFi2000 Slimline is provided by www.hifi2000.it/ with a carbon fibre hardtop.
B.t.w what is DiY? You won’t start with a lump of germanium from your mine. Rather you carefully evaluate components somewhere in the value chain as a transistor or a finished board, depending on ambition, skill and resources available.
Here some of the boards with the best technical merits, cost no object where selected. Careful modifications and addons where applied. As this is the final iteration, a modular approach where applied, allowing for later additions.
The goal where to build a unit accepting my sources, processing speaker element correction and crossover with digital output to digital amplifiers and earphone correction. I bet such a unit cannot be procured on the audio market. Hence, a good reason to build beside of the sheer fun.
The modular concept is close to the approach featured by MSB Technology Reference series. Here is a common I2S bus used with multiple Master mode modules using tri-state output ruled by simple priority mechanism. Realized using D-Type Latches With 3-State outputs is used to “mute” the actual unit not in current used. The environmental concerns over long I2S lines were however unfounded, easily proved by empirical evidence and signal integrity is shown by the scope.
The picture shows the fully working one-off prototype of MiniSharc four slots for units, each in plug and play operation. This means that a unit can go active in I2S Master mode on the initiative of the user.
An externally hosted image should be here but it was not working when we last tested it.
The modules available are these (from right to left):
1. Ethernet/RJ45 interface. Digital Media Renderer for stereo audio UPnP AV 2.0 / DLNA playing and decoding common audio formats from HTTP streams from engineered SA - electronic, engineering and consulting
2. USB audio playback interface. The sampling frequency is defined by the USB streaming itself and depends on the audio source file and computer configuration from engineered SA - electronic, engineering and consulting.
3. Discrete R-2R sign magnitude DAC technology with reclocking FIFO OEM Line Dam 1121 from Soekris Audio ApS, R-2R Sign Magnitude Audio DAC. The Dac is used for the earspeakers using FIR correction of several origins (Stax ED-1 Signature, SRM Monitor or Bob Katz eq’ing).
4. Main DSP unit. Analog Device ADSP21369 always work in Input Slave mode 96kB, 24 bit resolution. The DSP board MiniSharc is from Home | miniDSP. This board is layered above a daughterboard which provide two inputs and four SPDIF outputs, all with BNC contacts from Waveflex - Next Generation DIY Speaker Cabinets. The MiniDSP is fully controlled from a PC-based UI over WiFi using the WI-DG module and from a module VOL-FP using a IR remote-control or with front panel volume and pre-set selection wheel on front panel, all from Home | miniDSP
While Master Clock distribution over coax cable is available from each unit but implementation in multi-master environment is complex and not recommended for the
On the drawing board are now threse modules:
1. An analog conversion unit (ADC) which is based on successive approximation principle. Actually, as far as I can see, there are few if any such implemented modules available for stereo balanced input with I2S output.
- a. Although TI PCM4222EVM has been reported in use for DiY purpose, this board does not fit into the this standard-size module.
- b. LT LTC2380-24 24bits 2Msps ADC that integrate a digital averaging filter to improve SNR. use a serial link (as SPI) to enable conversion and read serial data.
- c. Hopefully, we will see a SAR ADC board from Soekris soon.
- d. Meanwhile an external AK5394A soundcard would apply.
2. Speech recognition remote control, using eight Mems microphones in a mic array with DSP complex voice capture function ( beamforming / noise reduction / echo cancellation / de-reverd / Direction of Arrival (DOA) / Voice Activity Dectection (VAD) UMA-8 USB module from Home | miniDSP connected to a raspberry PI accessing Synology server with a Amazon Alexa audio skill software. Amazon Alexa provides a list of supported voice commands, such as “Alexa, ask Audio Station to play the album [album name]”.
An externally hosted image should be here but it was not working when we last tested it.
A limitation is the compatibility with Amazon Alexa demands any audio files in a format other than MP3 will be converted into 128kbps MP3.
3. A PA213 Bluetooth 5.0 Audio Module BT5.0 CSR8675 Chip with support for APTX-HD from a China-base vendor on eBay.
The two power supplies are a switched medical-grade low noise 5V 15W unit for digital units from www.tracopower.com and a linear +5V / - 5V unit for analog units from www.diyinhk.com.
Each module resides in it’s own module box with an on board master clocking is with low jitter. EMI, Electromagnetic shielding effectiveness using carbon fibre reinforced composites for all units. A simple layer of 2/2 twill 3k dry carbon fibres fabric is able to shield more than 99.999% of the electromagnetic radiation.
https://www.sciencedirect.com/science/article/abs/pii/S1359836818343762
An externally hosted image should be here but it was not working when we last tested it.
The aluminium chassi, HiFi2000 Slimline is provided by www.hifi2000.it/ with a carbon fibre hardtop.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
Faboz - that’s interesting. In theory it is promising in that it cuts out a lot of the interfaces to get files to the i2s-in of the DAC. Do you hear differences you can tell blind?
Looks like a straightforward hookup. Was it? Have you tried playing a DSD file over the network to the Soekris? The DAMs definitely play DSD.
Looks like a straightforward hookup. Was it? Have you tried playing a DSD file over the network to the Soekris? The DAMs definitely play DSD.
Hello,
Compared to my mano streamer, yes the differences are really audible, much better. Details, room ambiances, stage, ...
Yes the PCM hookup is straightforward, but for the DSD L&R from the board, i dont know where to connect them. So it works for DoP but not as native.
Does it even matter? DSD playback is a nice convenience feature but will always remain a pale imitation of a proper dsd dac.
True, as the streamer board is native compatible, i was just wondering if the Dam was, but it is not. I have my answer. Thank you.
th dam1021 plays native DSD on the I2S pins, bclk is bitclk, don't remember which of the lrclk or dat is which channel, did post it here long time ago, try the search function here, otherwise there are only two options to try....
Faboz - thanks, that’s quite an SQ endorsement. Maybe, to avoid the cables, you could consider a simple PCB into which both the DOCK and DAM plug? Along with the power supplies etc.
Ronnhedh - so, if I understand correctly, you’re going into the SHARC with i2s and coming out with SPDIF? I assume to power active speakers?
Are the DACs used on the other end of those BNCs Soekris DAMs? Guess it would get to be quite a sizeable box if you put the DACs (and, horrors, the amps) into one chassis. Would have the potential benefit of keeping all in the i2s domain, though.
I’m just curious what drove the design decision to abandon i2s. Benefit of having each DAC and its amp near the speaker, I guess. But if those carbon boxes (nice find, BTW)are as good as they claim I might have first tried keeping the DACs on the i2s-out of the SHARC and close as possible to the SHARC then go to the amps with the shortest interconnect route possible. Like sit the amps below/above the Big Box Of Tricks. Comes down to long SPDIF versus long speaker cables. Would have been interesting to try both and see if you could hear a difference. Or did you?
Ronnhedh - so, if I understand correctly, you’re going into the SHARC with i2s and coming out with SPDIF? I assume to power active speakers?
Are the DACs used on the other end of those BNCs Soekris DAMs? Guess it would get to be quite a sizeable box if you put the DACs (and, horrors, the amps) into one chassis. Would have the potential benefit of keeping all in the i2s domain, though.
I’m just curious what drove the design decision to abandon i2s. Benefit of having each DAC and its amp near the speaker, I guess. But if those carbon boxes (nice find, BTW)are as good as they claim I might have first tried keeping the DACs on the i2s-out of the SHARC and close as possible to the SHARC then go to the amps with the shortest interconnect route possible. Like sit the amps below/above the Big Box Of Tricks. Comes down to long SPDIF versus long speaker cables. Would have been interesting to try both and see if you could hear a difference. Or did you?
Just to clarify, the I2S bus is bi-directional by separate lclock and data ines.
Directions are out from Master to DSP slave input and In from DSP Master output to Slave units.
The DSP has 2x4 output channels, where 2 are designed to I2S output bus and all I2S-to-SPDIF conversion. The image shows four BNC outputs. These are connected to digital amps using 1m SPDIF and 2-3m to speakers.
The Dam 1121 sits on the DSP I2S output line.
The DSP manage several tasks: Channel management matrix, IIR and FIR filters, amplitude ant time adjustment.
Directions are out from Master to DSP slave input and In from DSP Master output to Slave units.
The DSP has 2x4 output channels, where 2 are designed to I2S output bus and all I2S-to-SPDIF conversion. The image shows four BNC outputs. These are connected to digital amps using 1m SPDIF and 2-3m to speakers.
The Dam 1121 sits on the DSP I2S output line.
The DSP manage several tasks: Channel management matrix, IIR and FIR filters, amplitude ant time adjustment.
Off-topic: Carbon fibrer plates are vey useful for rapid prototyping as well as one-off finished work. Easy to cut with metal tools as saw or model knife.
Easy to glue pieces together with cyano-acrylate based glue.
Carbon Fiber Plates easily available in many different shapes.
I buy from a local vendor specialized in fiberglass, carbon fiber, epoxy, titanium, aluminum etc. They also fix carbon frames in carbon fiber and cut out details to order using CNC cutters.