I'm running an 8-channel active crossover (four 1021 boards via I2S) from the USBstreamer.
There's no difference between 1121 and 1021 when syncing to external bitclock, since multiple boards will lock onto the bitclock independently. The 1121 does have the advantage of perfectly synchronized master/slave mode.
Cool, you're welcome to publish more information and details, so other people could do the same....
My implementation is pretty straightforward:
PC (8-ch crossover/delays/DRC) -> USBstreamer I2S out -> 4 x dam1021 (bitclock and wordclock in parallel).
There's some suspicion of sync issues in this config, I can't hear it, so you're probably right that several boards sync within one wordclock period. Can be verified with impulse response measurements.
Many people would like to implement crossovers with a 2-channel source and multiple DAMs, but it's currently impossible to manage delays/time alignment due to hardware limitations of the dam1021/1121.
Anyone else with active crossover/multichannel setup?
Last edited:
I doubt that it will support a fixed OSC but if it did it needed to scrap the digital PLL and then all jitter from source would play straight out unfiltered or lose data (klicks n' pops) as the internal and external clocks start deviating.
I interpreted the statement about supporting external clocks as in the sync scheme between two (or more?) 1121 described in the manual.
Sören?
//
I interpreted the statement about supporting external clocks as in the sync scheme between two (or more?) 1121 described in the manual.
Sören?
//
Since when "not supported" configurations have stopped DIY audiophiles from doing it anyway? Us DIY folks can fix things even if they are not broken.
The way I understand the master/slave mode, it may be possible to use a fixed XO and run all DAM1121 in slave mode. It would still provide final reclocking while effectively disabling the DPLL and FIFO buffer. The data source should then also be clocked from the same MCLK to avoid buffer under/overruns.
The way I understand the master/slave mode, it may be possible to use a fixed XO and run all DAM1121 in slave mode. It would still provide final reclocking while effectively disabling the DPLL and FIFO buffer. The data source should then also be clocked from the same MCLK to avoid buffer under/overruns.
My implementation is pretty straightforward:
PC (8-ch crossover/delays/DRC) -> USBstreamer I2S out -> 4 x dam1021 (bitclock and wordclock in parallel).
There's some suspicion of sync issues in this config, I can't hear it, so you're probably right that several boards sync within one wordclock period. Can be verified with impulse response measurements.
Many people would like to implement crossovers with a 2-channel source and multiple DAMs, but it's currently impossible to manage delays/time alignment due to hardware limitations of the dam1021/1121.
Anyone else with active crossover/multichannel setup?
Are you using shortened, frequency specific filters in each dam in accordance with your crossover frequencies?
No, all filtering is done on the PC side, each DAM gets already filtered signal with appropriate speaker corrections - baffle step, delay, DRC etc.
Much more convenient for tweaking, and plenty of processing power...
BTW, how is your project with exaU2I and four DAMs going?
I'm not referring to the crossover filters- I am talking about the digital filters in the dam itself. Are you running each board with a full range filter or have you inputted frequency range specific filters into each dam?
My project is on hold as I am contemplating what to do about the mods needed to get the dam to perform as advertised. I wasn't planning on it requiring rework and am not experienced enough with that sort of thing to have the confidence to attempt it on my own.