[MWeston]

Thanks again for the details Rob! This sort of feels like it was a "chicken and the egg" situation for you during development. You need audio output to do your testing, but you need to have the DC servo logic in place so you don't ruin any good speakers! Mojo would have been much easier!

 

[Vyyy]

I have a question. Recently got Hugo2. So:

Does USB data signal from battery powered laptop (and power cord unplugged) means i have zero or almost zero RFI (becouse of batteries powering entire laptop) to the DAC?

@Rob Watts what is your source to the Mojo/Hugo2/TT2? Thank you.

 

[Rob Watts]

The problem is that all DACs are sensitive to ground currents flowing into the DAC ground plane. Because of finite impedance, you then get internal voltages, that will be picked up by the analogue electronics, which then cause noise floor modulation if the RF noise currents in the ground-plane is random - and if its correlated noise then you will get some distortion added too.

So if we eliminate any currents, then the ground-plane has no voltage drops, and then no extra noise floor modulation.

By running Hugo 2 off a battery, and feeding it into headphones, there is no ground loop, and hence no currents to flow - so simply running Hugo 2 with a grounded lap-top is fine, as there is no possibility of currents flowing. And indeed, this was verified by listening tests with no change from optical to USB when using a mains powered lap-top and listening via headphones. The problem starts when you ground the output of Hugo 2 - say feeding a power amp - then the source has an impact. But if the source is also battery, then no currents can flow - and again this has been confirmed via listening tests.

So to answer your question, theoretically there is no mechanism for a battery powered source to create noise floor modulation via RF noise, and this has indeed been verified with listening tests comparing optical to USB. Optical of course is the ideal source, as it's impossible to inject currents from the source to the DAC ground plane.

I hope this clarifies!

As to my source this is always via an M scaler, and the M scaler is fed from a Motorola mobile or a lap-top. The dual galvanic isolation on the M scaler means source noise hasn't been an issue with me.

Going back to the DC servo posts - actually the development was very straight forward - lot's of detailed Verilog simulation beforehand to confirm that the DC trim was noise-less, and to confirm the system worked - and although it was some years back - I recall it was straightforward to implement. I also had the option of disabling the servo and putting up with the DC offset and listening tests revealed no change in SQ.

 

[IgeNeLL]

I see there are source on the internet say that they mod the mojo to balance does it possible? When I check the two 3.5 output of mojo or hugo2 still connecting each other. You can input to one 3.5 and get the output on the other while the device is turn off.

I'm really like the sound signature of Hugo2. Do you intend to bring the hugo2 peformance to an DAP?

 

[Mediahound]

I see there are source on the internet say that they mod the mojo to balance does it possible?

It's not a balanced DAC so there would be no point.

 

[Rob Watts]

It's a crazy idea to convert to balanced, it's just adding more complexity - a headphone doesn't care if it's a balanced drive or single ended drive, it's just the same - a voltage across a load.

 

[IgeNeLL]

It's a crazy idea to convert to balanced, it's just adding more complexity - a headphone doesn't care if it's a balanced drive or single ended drive, it's just the same - a voltage across a load.

It is the information from the internet I think it is quite Hilarious ) but they still say it.

 

[odessamarin]

@Rob Watts
Dear Rob, is it save to use 2,4V pp in SPDIF input for Dave?
Seems my player gives this a bit high amplitude digital signal output.
Do we have some isolation transformator or anything else in SPDIF input in Dave to be save?
Thank you.

 

[Currawong]

I see there are source on the internet say that they mod the mojo to balance does it possible? When I check the two 3.5 output of mojo or hugo2 still connecting each other. You can input to one 3.5 and get the output on the other while the device is turn off.

I'm really like the sound signature of Hugo2. Do you intend to bring the hugo2 peformance to an DAP?

Mojo + Poly and use your own smartphone.

 

[IgeNeLL]

Mojo + Poly and use your own smartphone.

I have tried mojo and Poly but the mojo performance is far behind the TOL DAP now.
I expect the Hugo2 performance

 

[Currawong]

I have tried mojo and Poly but the mojo performance is far behind the TOL DAP now.
I expect the Hugo2 performance

Wait for Hugo 2GO? Maybe in a few years when FPGAs become even smaller and use less power, Rob will be able to get Hugo 2 performance into a Mojo.

 

[IgeNeLL]

Wait for Hugo 2GO? Maybe in a few years when FPGAs become even smaller and use less power, Rob will be able to get Hugo 2 performance into a Mojo.

I believe that Hugo2 has been optimized between performance and space using, compacting to mojo size is hardly possible. Why I expect a dap because for kind of hugo2 and mojo, you still need a good source like sp2k to feed through USB line with kind of Nordost Val2 to get the right performance and it is very complicated.

 

[Rob Watts]

@Rob Watts
Dear Rob, is it save to use 2,4V pp in SPDIF input for Dave?
Seems my player gives this a bit high amplitude digital signal output.
Do we have some isolation transformator or anything else in SPDIF input in Dave to be save?
Thank you.

Yes no worries - the M scaler is also at 2.5v pp when terminated, and the interface device will actually tolerate 4.5v PP.

Wait for Hugo 2GO? Maybe in a few years when FPGAs become even smaller and use less power, Rob will be able to get Hugo 2 performance into a Mojo.

I am still hoping for Hugo 2 with integrated M scaler on one FPGA - but we are many years away from that, possibly decades. That said TSMC (they fabricate Xilinx FPGAs) have just announced the starting of construction of a 3nm foundry for 19.5 billion USD - production 2023. But given the massive headaches Intel have had for 10nm, I think 2023 is highly optimistic. Moreover, the economic sweet-spot for FPGAs is still stuck at 28nm and has been for 7 years. In this sense, Moore's law (more gates at lower cost) is well and truly broken.

 

[MWeston]

Hi Rob,

I have another little question that has been in my mind. I've been looking at crystals used by various gear and was curious about your decision to use that 104MHz oscillator. It doesn't divide down to any regular sample rates and I know your approach to sampling isn't so frequency dependent on the industry standard rates, but is there any story behind the choice or was it just in stock that day when you started your first FPGA prototype and it met the timing constraints?

 

[mpickup]

I am surprised to see many recent preamp / integrated amp reviews casually state that all analog inputs are digitized. Given the budget constraints, are these internal ADCs any good?