• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Star clocking and global master timing (iFi Audio)

Veri

Master Contributor
Joined
Feb 6, 2018
Messages
9,598
Likes
12,040
They have decent femto clocks in their DAC, which is used for the xmos controller, 44k and 48k syncing.
They just invented their own names. What they do works, but most competitors use basically the same...
 

mansr

Major Contributor
Joined
Oct 5, 2018
Messages
4,685
Likes
10,705
Location
Hampshire
There's plenty more BS in that PDF too. Let's take a few samples.

"The 3.5 TRRS is not an industry standard. And for not so well-designed amplifiers, plugging in a TRS 3.5 will burn out the 3.5TRRS balanced amp!"
Implying theirs is a poor design. Regardless of the connector, not making the output short-circuit proof is unforgivable.

"When it comes to D/A converter chips, we go all out. Our hip-dac’s heart is the very same as in many of our pricier products – a custom- programmed Texas Instruments Burr-Brown DSD1793."
If going "all out," why not use the better performing PCM1792A chip? And do they really expect us to believe they had a "custom-programmed" version of the chip made? Maybe what they actually mean is that they change (program) some settings from the power-on defaults of the chip.

"In the hip-dac, the clock domain naturally covers USB but also synchronizes the MCU (Master Control Unit) to effectively ensure the hip-dac’s entire digital functionality is slaved to the GMT clock system. It is global by name and by nature."
The clocks are decent, but that "global" nonsense is just that. The USB interface requires a multiple of 12 MHz while the DAC runs at a multiple of 11.2896 MHz or 12.288 MHz. It is of course possible to synthesise the USB clock from the audio clock using a fractional PLL, but there'd be no benefit whatsoever from doing that. The clock domain crossings still have to handle arbitrary phase differences. Oh, and the iFi DACs I've looked inside do have a separate 24 MHz crystal beside the XMOS processor.

"iFi audio’s proprietary high-performance, low-noise, quad J-Fet OV4627A operational amplifiers"
That chip designation doesn't seem to exist outside of iFi marketing blurbs. I'm guessing it means they've renamed a stock product to make it harder for people to look up the datasheet.

"Running alongside the Burr-Brown ‘True Native’ chipset is the 8-Core XMOS based on our own code, which updates it with our own Star Clocking, just as it was done in the micro iDSD."
Wow, they wrote some software. Very impressive. Wonder if they put as much effort into that as they did with the promotional materials.

"The hip-dac implements Version 4 AMR XMOS platform and uses the latest generation 8-Core 500MIPS XMOS1 transputer derived main processor. These processors are quite unique in their architecture and based on a technology once considered to revolutionize computing, the INMOS Transputer."
A slight correction is in order. The transputer designers hoped it would revolutionise computing. It didn't, instead fading into obscurity. Some of the broader design ideas were interesting enough that they show up now and again in specialised applications. As for the XMOS chip, it strikes me as annoying more than anything else. Besides, the architecture of the microcontroller really doesn't matter at all, provided it can shuffle bits from the USB port to the DAC in a timely manner.

"TDK high stability C0G surface mounted capacitors - C0G dielectric close to Teflon in key areas that affect the smoothest sonics."
C0G offers good temperature stability, and the capacitance has little voltage dependency. These properties make it the standard choice for capacitors in the signal path. iFi making a point of this is akin to a car manufacturer boasting about using rubber tyres.
 
OP
S

stalepie2

Major Contributor
Joined
Feb 9, 2020
Messages
1,378
Likes
631
I was thinking it would be interesting if there were such a thing as a hidden seasonal DSP which changed the color of the sound a little by season, or by weather (hot/cold/rain/snow), or by holiday. It could have a jazzier soundstage during certain holidays or on April 1st the DSP goes all wacky :) And then the star clocking, I wondered if that was to suggest alignment with the stars. I did think perhaps it was a real term I wasn't familiar with. I mean, when I looked it up I got a Sky and Telescope article explaining how to make a star clock for kids.

This sort of audiophoolery nonsense may appeal to some, but it actually is instinctively off-putting to many people, so would limit sales.

In regards to why they chose the 1793, it says in the document that the top 6-bits are "true multi-bit" and perhaps that is not the case with the other TI DAC you mention.
 

Veri

Master Contributor
Joined
Feb 6, 2018
Messages
9,598
Likes
12,040
In regards to why they chose the 1793, it says in the document that the top 6-bits are "true multi-bit" and perhaps that is not the case with the other TI DAC you mention.
Conveniently for them, it also costs about half of the PCM1792/PCM1794 (TI's actual flagship), if not less at large quantities. But yeah the multi-bits make for a good marketing blurb ;).
 

mansr

Major Contributor
Joined
Oct 5, 2018
Messages
4,685
Likes
10,705
Location
Hampshire
In regards to why they chose the 1793, it says in the document that the top 6-bits are "true multi-bit" and perhaps that is not the case with the other TI DAC you mention.
Both chips have the same D/A core. The differences are in the digital interpolation filter, power supply, and analogue output stage, the 1793 having integrated I/V conversion.

Conveniently for them, it also costs about half of the PCM1792/PCM1794 (TI's actual flagship), if not less at large quantities. But yeah the multi-bits make for a good marketing blurb ;).
The price difference is $2.50 for the DSD1793 vs $7.20 for the PCM1792A in 1k quantities.
 
OP
S

stalepie2

Major Contributor
Joined
Feb 9, 2020
Messages
1,378
Likes
631
Conveniently for them, it also costs about half of the PCM1792/PCM1794 (TI's actual flagship), if not less at large quantities. But yeah the multi-bits make for a good marketing blurb ;).
Well, they may be stretching the truth with the "all out" bit, but I would guess that the designers at iFi genuinely do believe in these aspects of the 1793 and are being honest in advertising that aspect - why else would they use this chip alone for most of their products? I think they do. but it's an old chip. They could also "go cheap" by using newer chips that measure better if it isn't about something special in the older TI designs that they believe in.

When I've looked up audio chips I'll come across forums, like DIYaudio, where quite a few designers really do believe each DAC imparts a different sound character to the final product.
 

mansr

Major Contributor
Joined
Oct 5, 2018
Messages
4,685
Likes
10,705
Location
Hampshire
Well, they may be stretching the truth with the "all out" bit, but I would guess that the designers at iFi genuinely do believe in these aspects of the 1793 and are being honest in advertising that aspect - why else would they use this chip alone for most of their products? I think they do. but it's an old chip. They could also "go cheap" by using newer chips that measure better if it isn't about something special in the older TI designs that they believe in.
The 1792A is based on the same "advanced segment" architecture as the 1793 but performs better in every way. The only reason one would choose the latter over the former is cost reduction.
 
OP
S

stalepie2

Major Contributor
Joined
Feb 9, 2020
Messages
1,378
Likes
631
Yes, I know, that's what I meant by stretching the truth a bit by saying they went all-out, but if this aspect of the design of the DAC is relatively special to this short line of chips (2 or 3 chip variants, from 20 years ago), and they release product after product using this specific chip, when there are also newer cheaper chips, which measure better, it implies they are sincere in claiming the upper 6-bits being multi-bit vs delta-sigma is important to the perceivable audio quality.

Either that or they bought a stupidly large supply of 1793 chips a long time ago, early in the company's history, and then have been trying ever since to use them up.
 

mansr

Major Contributor
Joined
Oct 5, 2018
Messages
4,685
Likes
10,705
Location
Hampshire
Yes, I know, that's what I meant by stretching the truth a bit by saying they went all-out, but if this aspect of the design of the DAC is relatively special to this short line of chips (2 or 3 chip variants, from 20 years ago), and they release product after product using this specific chip, when there are also newer cheaper chips, which measure better, it implies they are sincere in claiming the upper 6-bits being multi-bit vs delta-sigma is important to the perceivable audio quality.
In whatever way the 1793 is special, the 1792A is more special. It doesn't make sense not to use the best performer of the family unless for cost-cutting reasons. In their cheapest (around $150) models, this may be defensible, much less so as the price goes up.

or they bought a stupidly large supply of 1793 chips a long time ago, early in the company's history, and then have been trying ever since to use them up.
That's certainly a possibility.
 

Veri

Master Contributor
Joined
Feb 6, 2018
Messages
9,598
Likes
12,040
In whatever way the 1793 is special, the 1792A is more special. It doesn't make sense not to use the best performer of the family unless for cost-cutting reasons. In their cheapest (around $150) models, this may be defensible, much less so as the price goes up.
It's actually puzzling to use the 1793 in their 'Pro' iDSD/top of the line statement piece. Maybe they'd think it doesn't matter, when they put tubes in there anyway :D
 
Top Bottom