After a long wait and much discussion exaU2I is now available for purchase! I apologize for the delay in getting the devices out to you. The price is $430 CAD. We are working around the clock to clear the backlog. Thank you for all your valuable input and patience.
The first batch is ready to be shipped out to those of you who received emails containing your Waiting List ID. We are going to send out more Waiting List IDs as soon as devices are produced, and we've chosen to do the shipments in this way in order to minimize the wait.
Meanwhile you can go to the website and look thought the user guide for more technical information. If you like what you see please consider signing up on the waiting list.
The first batch is ready to be shipped out to those of you who received emails containing your Waiting List ID. We are going to send out more Waiting List IDs as soon as devices are produced, and we've chosen to do the shipments in this way in order to minimize the wait.
Meanwhile you can go to the website and look thought the user guide for more technical information. If you like what you see please consider signing up on the waiting list.
I've been notified that some Waiting List emails have been blocked by spam filters. If you are expecting an email with a Waiting List ID, please check your spam folder and add exa065@exadevices.com to your white list.
The concept of master clock is applicable when you have several "boxes". You should think of exaU2I as a part of your DAC. This is true when very short wires are used.
exaU2I should be the master clock for non-reclocking DACs.
Yes. Only one master clock should be used. The situation is different for ES9018 and non-relocking DACs. When ES9018 is used, the master clock is actually the 100 MHz quartz on the DAC board. Ideally if you have 3 two channel DAC boards, they should clocked by only one 100 MHz quartz.
Yes, the board will allow buffering. If you need to you can use buffers. However the GMR isolators should be able to handle 3 DACs. If you decide to use buffers make sure that the cables are short. Only the clock interfaces may need buffering.
This is not exactly correct for ESS DACs. I am on non-disclosure and I can't be more specific. As I said, the master clock is the 100 MHz oscillator on the DAC board. This is the clock that is critical for low-jitter performance. ES9018 is powerful enough to cancel the very insignificant jitter caused by the exaU2I oscillators, the GMR isolators and a short cable.
Jitter is very difficult to measure. We don't have the best equipment, but we did jitter measurements. It makes sense to measure artifacts caused by jitter after the DAC if the DAC deploys jitter cancellation. Our ES9018 DAC board shows total noise due to Master Clock Jitter < -140dB.
In general I2S causes much less jitter than SP/DIF. exaU2I has a very clean I2S implementation.
Mr. exa065 is absolutely correct here. With the exaU2I and a compatible DAC inside a single box, they are effectively operating "as one."
In fact, if you tear into several different high-end DAC units, you will probably find that nearly all of them have two separate circuit boards: One for the digital interface, and one for the analog conversion, with I2S or something similar between. So, a finished exaU2I setup is not any different from any other high-end DAC. It is actually far better to have two separate boards like this, even when the digital board is the master clock and the analog board with the DAC chip is slave. The advantages of two boards in terms of isolating digital noise from your audio far outweighs the slight challenge of having the clock cross between two boards. When the boards share a common power supply, especially with a common ground, and they are in the same enclosure, then there really is no reason to worry about which board is master, provided you don't do anything horribly wrong.
Pepe,
See the Connection Example - Four Channel ES9018 in Quad Mode on the D.I.Y guide page: www.exadevices.com > exaU2I > D.I.Y. Guide. In this scenario every Buffalo kit will use its own master clock. Most likely there will be no problems. If your power supplies provide good decoupling between the Buffalos, there will be no interference between the 100 MHz clocks. If something goes wrong, you may hear a parasite signal in the audible range. You should be able to resolve it by altering the wiring, the distance between the DACs and the power supply filters. The best way to know is to try it.
The ideal solution would be to have one master 100MHz clock for all four Buffalos. May I ask the people that know well the Buffalo boards for help here?
I thought the interest here was seeing the jitter performance without a jitter rejecting circuit in place. Like you'd see with a PCM1792/4 connected to the exaU21. O
We don't use any special cables. You will find that people have very different opinions on this topic.
I would like to further explain the advantages that exaU2I brings when it comes to isolating digital noise. It will be very hard to do something wrong with ground loops. There is no common ground between exaU2I and the DAC board. Look at the picture of the device on the features page - www.exadevices.com > exaU2I > D.I.Y. Guide. You can display the image in a separate browser window to see it in full size. The GMR isolators are the two chips at the top mounted over the air gap. The air gap is the line of galvanic isolation. The only thing on the other side of the gap is the terminal block. The outputs of the GMR isolators are not connected to the exaU2I ground. They must be powered by the DAC power supply. Electrically the portion of the exaU2I board located on top of the air gap is a "part of the DAC".
Don't have or didn't do? I would have thought measuring the performance without a jitter rejecting circuit would be the obvious choice for evaluating your modules performance, otherwise how else would you know how well it performs as an individual unit?
Have you tested 352.8kHz/24bit or 32bit play with ES9018-based DAC of 80MHz master clock? In the case of SDTrans192 which is capable of outputting 352.8kHz/24bit I2S, the use of 80MHz master clock brings certain continuous noises on "Oversampling Mode". We have found the master clock frequency above 92MHz can eliminate this issue actually though the requirement described in ES9018 datasheet is over 67.7376MHz.
An expert who once evaluated SDTrans192 with ES9018/80MHz master clock said that I2S signals of SDTrans192 were not clean enough.
I'd like to know whether exaU2I can play 352.8kHz/24bit DXD audio files without the noise issue on ES9018/80MHz environment or not.
If yours has no noise problem, I think I2S signals of SDTrans192 are not clean enough and it has a room to improve.
Last edited:
I have first hand experience only with 100MHz master clock. I can try to replicate the issue.
I appreciated your immediate reply very much.
For example, though the current revision of TPA Buffalo II DAC uses 100MHz oscillator device, previous version used to equip 80MHz one. Those who own the 80MHz oscillator version may encounter the problem. I really hope your hardware will be able to skip the issue.
Thank you very much again.
Bunpei - in this exaU2I thread you brings up your SDTrans192 and Buffalo DAC and possible limitations regarding interaction between those two products - why ?
I do not believe exa065 have experienced either your SDTrans192 or the Buffalo DAC.
The problems you mentions have not been any issue with my JFET regulated SDTrans192 units or my JFET regulated ES9018 based DAC´s.
As the fact is - I will help you in this regard Bunpei as both I and a friend of me have ordered the exaU2I cards and will be able to compare (A/B) the SDTrans192 to the exaU2I.
We will also test the difference between 352.8k/24bit and 352.8k/32bit audio, but this will only be done on the exaU2I as the SDTrans192 does not support this.
In one setup we will use the SDTrans192 (internal to the DAC) with direct short I2S cable and the exaU2I via I2S over LVDS using a HDMI cable and connectors, and the whole setup run from LIFEPO4 batteries.
In the second setup the exaU2I will be internal to the DAC and the SDTrans192 via I2S over LVDS using a HDMI cable and connectors.
The SDTrans192 units we use are tweaked rev 2.1 with my latest revision JFET regulators and clocks that in my setup outperformed your latest stock SDTrans192 rev 3.0 when compared..
We will post the results both in this thread and in the SDTrans192 thread.
I've just posted screenshots of Foobar performing the Quick test from the DIY Guide. These are the steps that users will go trough to verify that their computer setup and exaU2I device are working properly. An advanced Foobar skin was used to illustrate the great customization capabilities of this player.
www.exadevices.com > exaU2I > D.I.Y. Guide > Quick Start Screenshots
www.exadevices.com > exaU2I > D.I.Y. Guide > Quick Start Screenshots
Bunpei,
The issue is not related to the exaU2I device, it is related to the implementation of the DAC. It will be up to the DAC manufacturer to provide a solution. However for the benefit of the users out there that have DAC boards with 80 MHz clocks I modified one of my DACs and I've managed to replicate the issue. From my experiment it appears that owners of older DACs can simply replace the 80 MHz clock with a 100 MHz clock and the issue will be resolved.
We will know for sure if someone who owns this particular DAC is willing to upgrade the clock on his board to 100MHz and to share his findings with us.