did you look at LIVE intro? theres a free demo of it which is probably just time limited, so you can check to see if its useful. It says it supports 6 vst plugins, its a cutdown version of ableton LIVE and costs 99 bucks, limited to 32/192 though.
So now I've had the chance to try the exaU2I with the AD1865. I bought this dac kit off e-bay: DAC kit AD1865N-K NOS 1.0J NOS. This dac is ridiculously cheap, even though the build quality is very high. The PCB is huge, there are discrete circuits all over the board. Separate discrete regulators for the dac and the analog stage, and a very good discrete I/V stage. It runs perfectly at 192 kHz. The sound is really good. Plenty of detail, good timbre, very dynamic and with powerfull bass. And at this sample rate there is no problem with the typical NOS dac early rolloff.
I have encountered two problems with this dac. Perhaps some of you can help? First of all Jriver Media Center stopped playing sound on movie files. I get sound with audio files only. This has never been a problem before.
The next problem is that the exaU2I drives the AD1865 to clipping really early. The digital output is extremely loud, and I have to attenuate the signal by 80 dB in Jriver to get a normal input level. Why could this be? The signal goes through a 74HC2n flip-flop before going to the dac, so perhaps the digital signal from the exa is amplified in some way? The flip-flop runs on 5 V.
I have encountered two problems with this dac. Perhaps some of you can help? First of all Jriver Media Center stopped playing sound on movie files. I get sound with audio files only. This has never been a problem before.
The next problem is that the exaU2I drives the AD1865 to clipping really early. The digital output is extremely loud, and I have to attenuate the signal by 80 dB in Jriver to get a normal input level. Why could this be? The signal goes through a 74HC2n flip-flop before going to the dac, so perhaps the digital signal from the exa is amplified in some way? The flip-flop runs on 5 V.
Yeah, I'm sorry, I see that flip-flop wasn't the correct term. The 74HC02 is connected so that it puts out the left channel data in phase with the LRCK, and the right channel data in phase with the inverted LRCK. No other glue logic than inverting the LRCK for one of the channels.
The AD1865 accepts 2s compliment MSB first, similar to the I2s format. It will only convert the 18 last bits in the stream though, so it's limited to a 16 bit signal, and the two most significant bits will be "empty". So it's definitely not a bit perfect solution. But it works. Only the input of the AD1865 starts clipping unless the input signal is attenuated.
I'll try to find out what kind of output format the CS8414 is configured for in this circuit.
The AD1865 accepts 2s compliment MSB first, similar to the I2s format. It will only convert the 18 last bits in the stream though, so it's limited to a 16 bit signal, and the two most significant bits will be "empty". So it's definitely not a bit perfect solution. But it works. Only the input of the AD1865 starts clipping unless the input signal is attenuated.
I'll try to find out what kind of output format the CS8414 is configured for in this circuit.
I know. I'm working on that. It's gonna be a train of 74HC164's.
The CS8414 is configured for 18 bit left justified output, so it's a bit different. None of this should result in clipping though.
If you want to get best performace bypass CS8414 and feed AD1865 from exa directly. Best solution for glue logic is variable length shift register so you can set delay using dip switches:
http://www.datasheetcatalog.org/datasheet/on_semiconductor/MC14557B-D.PDF
http://www.datasheetcatalog.org/datasheet/on_semiconductor/MC14557B-D.PDF
Have you had any experience with these variable shift registers? It seems like they are quite slow devices. Rise time of 100ns, and a 300ns propagation delay. Could it really be useable for audio? I've never seen anyone use these. The 74HC series are really fast though. Single digit rise time and propagation delay.
Probably the clipping comes about because the word clock (frame clock) is in the wrong phase relationship with the data. Every bit position shift in this clock results in a 6dB gain or attenuation, depending on the direction its shifted. I get great clipping effects on my DACs when the LSB of one sample gets mistaken for the MSB of the subsequent one
<edit> A good way to keep the noise down (and hence sound quality up) is to run the HC logic at low supplies - 3V is plenty fast enough for most audio.
<edit> A good way to keep the noise down (and hence sound quality up) is to run the HC logic at low supplies - 3V is plenty fast enough for most audio.
Last edited:
Probably the clipping comes about because the word clock (frame clock) is in the wrong phase relationship with the data. Every bit position shift in this clock results in a 6dB gain or attenuation, depending on the direction its shifted. I get great clipping effects on my DACs when the LSB of one sample gets mistaken for the MSB of the subsequent one
Yes! That's exactly what's happening. It all makes sense. The timing requirements for the AD1865 are a little different than the I2S spec. I have to delay the LRCK by one bitclock cycle. Any ideas for a simple solution for this?
I2S has the phase of the LRCK advanced by one bit position, so yes if you want a clock which toggles after the LSB you just need a one bit time delay. A 74HC74 flip-flop will do that well enough - its clocked from the BCK and the LRCK goes into the D input. It has both polarities of the delayed LRCK available at its outputs which often comes in handy
I2S has the phase of the LRCK advanced by one bit position, so yes if you want a clock which toggles after the LSB you just need a one bit time delay. A 74HC74 flip-flop will do that well enough - its clocked from the BCK and the LRCK goes into the D input. It has both polarities of the delayed LRCK available at its outputs which often comes in handy
That sounds like a simple solution. My local supplier has the 74hc74 in stock, so I could pick it up tonight and make a quick fix.
I'm planning on making a AD1862 based diy dac from scratch, so this e-bay dac is sort of a test project to see if I can make the signal from the exaU2I work. The AD1862 and the AD1865 have the same timing requirements.
Thanks a lot for the help! I'm new to this digital logic stuff, so all your answers are appreciated.