We provided several free upgrades for the exaU2I at no charge:
- 384 kHz sampling rate was a free software upgrade
- dsd64/dsd128/dsd256/dsd512 was a free software upgrade
- OSX drivers
- We released a second generation ASIO drives and the Dashboard
- We also upgraded the clocks with 1ps crystals without changing the price
- The devices with upgraded clocks have a new firmware with lower jitter
Regards,
exa065
Last edited:
Thanks for the info. I was not aware of the upgrades.
Are there measurements for the exaU2I after the upgrade? Or before and after comparison?
not working means that Audirvana is for example not outputting any sound Amarra is unstable and crashing from time to time under Mountain Lion.
This is not the problem with other USB devices so its a driver / OS problem.
I'm happy to go back to Snow Leopard if that would solve the issues, but so far i have not heard any statement if this would solve the issues, you only promised updated drivers a few posts. I'm also not the only one having issues with Mountain Lion. So either way updated drivers are needed or an older OS.
This is not the problem with other USB devices so its a driver / OS problem.
I'm happy to go back to Snow Leopard if that would solve the issues, but so far i have not heard any statement if this would solve the issues, you only promised updated drivers a few posts. I'm also not the only one having issues with Mountain Lion. So either way updated drivers are needed or an older OS.
The current exaU2I supports DSD in a twisted way. Our e20 DAC and the new to-be-developed USB to I2S interface support DSD natively. I don't like half measures, new driver without hardware support won't make anybody happy. It is time for a new device.
What version of the OS X drivers are you using? Have you reported the issue to customer service on our website? I don't thing we left any request sent to customer service unanswered, and we don't have any open cases on OS X. I use Audirvana Plus, and I recommend it to customers. What is your user name on our website? I would like to check the customer service logs.
Here are the measurements on our website: www.exadevices.com > exaU2I > Measurements
I also did a quick search on this thread for measurements:
http://www.diyaudio.com/forums/exad...ronous-usb-i2s-interface-102.html#post2926393
http://www.diyaudio.com/forums/exad...ronous-usb-i2s-interface-113.html#post3077812
I admit that lately we've been busy with the e20 DAC, and this thread was neglected. I don't know if the world needs another exaU2I, that's why I am asking.
Two years ago there wasn't much choice. We were waiting for the USB to I2S interface from TPA. Now there is a healthy competition. We have the WaveIO, but it has been reported that it has 2 ns jitter because of the internal XMOS re-clocking. My customers have told me that they have bought it, tested it, and sold it because "it is not in the same league with the exaU2I". There is no DSD support either. I know that the use of standard drivers is a big convenience and it opens the door to new platforms, like iPad and Android, but priority must be given to the sound fidelity.
Then there is the new USB to I2S 384Khz - DSD Converter. I am reading on their thread that it cannot pass the bit-perfect test of the Audio Precision APx525 system, and it doesn't sound as good as WaveIO. You get what you pay for.
Asynchronous I2S FIFO device is another alternative. It can provide very good results, but I think it is a partial solution. Perhaps I am wrong, here is my issue with it. The I2S buffer has no way to control the computer. Even if it is attached to an asynchronous USB interface, it has no way to ask the computer to speed-up or slow-down the USB stream of data. Computer timing is horrible. The FIFO board uses its own clocks for re-clocking, so the two clocks will go out of sync. After a while the FIFO buffer will skip, if the computer clock is faster, or it will get empty if the computer clock is slower. It is just a matter of time. For a 100 ppm computer clock the chance is let say 1:10,000 on the bit-clock frequency. Depend on the sampling rate and the FIFO buffer size this can be after 15 min or after an hour. Perhaps it is OK to have one FIFO buffer error per hour. I continue to believe that the top solution is asynchronous USB to I2S, nor asynchronous I2S to I2S.
Two years ago there wasn't much choice. We were waiting for the USB to I2S interface from TPA. Now there is a healthy competition. We have the WaveIO, but it has been reported that it has 2 ns jitter because of the internal XMOS re-clocking. My customers have told me that they have bought it, tested it, and sold it because "it is not in the same league with the exaU2I". There is no DSD support either. I know that the use of standard drivers is a big convenience and it opens the door to new platforms, like iPad and Android, but priority must be given to the sound fidelity.
Then there is the new USB to I2S 384Khz - DSD Converter. I am reading on their thread that it cannot pass the bit-perfect test of the Audio Precision APx525 system, and it doesn't sound as good as WaveIO. You get what you pay for.
Asynchronous I2S FIFO device is another alternative. It can provide very good results, but I think it is a partial solution. Perhaps I am wrong, here is my issue with it. The I2S buffer has no way to control the computer. Even if it is attached to an asynchronous USB interface, it has no way to ask the computer to speed-up or slow-down the USB stream of data. Computer timing is horrible. The FIFO board uses its own clocks for re-clocking, so the two clocks will go out of sync. After a while the FIFO buffer will skip, if the computer clock is faster, or it will get empty if the computer clock is slower. It is just a matter of time. For a 100 ppm computer clock the chance is let say 1:10,000 on the bit-clock frequency. Depend on the sampling rate and the FIFO buffer size this can be after 15 min or after an hour. Perhaps it is OK to have one FIFO buffer error per hour. I continue to believe that the top solution is asynchronous USB to I2S, nor asynchronous I2S to I2S.
Oh I agree, and I'm not sure how much interest there are from others, but I'm extremely interested in the new board.
This is probably the most stupid question you will get any time soon, as I have no idea how fpga stuff works or anything involved in how you design/build the exa.... But how flexible is the output pin configuration, and can it be made to be switched around using software on the fly?
I only ask, as right now obviously a few wires need to be switched around when changing from PCM to DSD and vice versa, when used with an ESS Sabre chip. Obviously isnt that big a deal when using something like TPA's OTTO. But is this the sort of thing that could be taken care of in a future exa design? as in programmed directly into the fpga? Or is it just impossible or too messy?
@exa065
i was not aware that there is support existing. I can also try that one.
Which version of OSX and Audirvana are you running whithout issues ?
I have the issues with 10.8.2 driver 1.6.0a4 and all of the latest Audirvana versions.
I'm happy to downgrade to Snow Leopard if this would fix the problems i do not need all this icrap on a music server ...
i was not aware that there is support existing. I can also try that one.
Which version of OSX and Audirvana are you running whithout issues ?
I have the issues with 10.8.2 driver 1.6.0a4 and all of the latest Audirvana versions.
I'm happy to downgrade to Snow Leopard if this would fix the problems i do not need all this icrap on a music server ...
Pjotr,
The LOTech player uses commercial libraries, and we will be violating the license agreements if we open the source code. Perhaps Flemming can open the source code for his DSD player. He gave it to us, to illustrate his idea, and that's how DSD for exaU2I was born. I don't feel comfortable to release his code.
Here is the way we have it designed for now: On the PCB there will be 13 control pins - one for each sample rate, including DSD sampling rates, one for power and one for the FIFO status. By reading these pins, you will know the mode of operation - PCM or DSD and the sampling rate. You will also know when the FIFO is ready to un-mute the DAC.
Is this OK with you? Do you need any additional signals?
Hi exa065,
How do we identify if the unit we have has the upgraded clocks? When was this introduced?
Thanks
How do we identify if the unit we have has the upgraded clocks? When was this introduced?
Thanks
Hi George
You are welcome to share my code.
But remember is was only a 'proof of concept'.
And the ASIO SDK parts has to be downloaded direct from Steinberger.
Flemming
Here is the clocks upgrade post on this thread:
http://www.diyaudio.com/forums/exad...ronous-usb-i2s-interface-102.html#post2919328
If you have a black plastic clock, it is the old revision. If you have two shiny metal clocks, it is the new one. All devices shipped after 23rd February 2012 have the new 1ps clocks.
Hi Flemming, it is good to hear from you. Thank you for reminding us about the ASIO licensing rules. I will find the zip that you've sent me and I will send it to Pjotr.
Your proof of consent is actually bullet-proof.
Best,
exa065
Volumen control : One or two I2C outputs
The driver supports multiple boards : Do you need to synchronize the clock in the boards.
What I am looking for is a system with two 'black box' dacs connected direct to my power amps, where all the control and display are in the pc, and I only have one usb cable to each dac.
Flemming