[mmerrill99]

I was prompted to start a separate thread to collate & clarify the information about the modifications to the Hiface that I have done.

M2tech Hiface
Ok, First the stock HiFace is an excellent product built with some particular goals in mind that I'm going to guess at here:
- compactness
- portability; related to above but it also means that it takes it 5V supply from USB & regulates it down to 3.3V & 1.8V not unlike other USB transports
- asynchronous USB - HiFace have written proprietary drivers which do the job
- low-jitter clocks in the right place

Hiface Power Supplies
The Hiface uses 3 different power supplies internally to it’s different on-board chips
5V supply to the SPDIF generating chip
3.3V supply to the USB receiver chip, FPGA chip & clocks
1.8V supply also to the FPGA chip.

Ensuring clean PS is important for every area in the unit but it is crucial in the clock handling areas.

My PS mods address these crucial clock handling functions bypassing the 3.3V regulator & providing a way to connect an independent clean external 3.3V supply to the clocks & a separate 3.3V supply to the FPGA & USB receiver chips. These mods significantly improve the sonics of the unit. This is not a subtle change – it is noticeable from the first note.

Sound Improvements
All the usual clichés apply but they are true:
- The HiFace has better clarity all through the frequency range
- bass is amazingly textured & powerful
- treble has an airiness to it
- finer details are revealed
- soundstage is deep & wide with venues & recording ambience being revealed clearly

as I’ve said before, it sounds just like really good analogue, not a bit of digititis in it.

Other improvements
Other smaller improvements are also achieved through bypassing the USB 5V supply & providing a clean 5V external supply. Not only does this improve the supply to the SPDIF generating chip but it also removes the need for the USB cable to carry power. The cable can now just perform the job of carrying the USB signals and nothing else. This should allow the use of a reasonably cheap USB cable & no more need for expensive “audiophile” USB cables. It should also mean that the variability of the PC/laptop PS is no longer an issue with USB audio.

A unit is being sent to slim.a next week for review & I will be posting pictures of a boxed unit next week also.

 

[slim.a]

Hi jkeny,

This looks very promising. I can't wait to listen to your modded hiface

 

[mmerrill99]

Thanks slim.a - it will be interesting to get another objective opinion of the modded Hiface verses the stock Hiface. Also interesting to hear your impressions of the modded USB cable & what it does for the stock & modded Hiface (as well as other devices) compared against the other USB cables that you have been looking into.

 

[Fujak]

Hi jkeny,

I plan to build in my HiFace into my DAC (inside the box) by tapping the USB 5V feeding from the internal power supply of the DAC. I want to realize this by a selfmade USB-Adapter separating the 5V/ground from the 2 signal wires. Do you think, that could be a good way in the right direction for sonic optimization?

Cheers
Fujak

 

[mmerrill99]

Quote:

Originally Posted by Fujak /img/forum/go_quote.gif .....separating the 5V/ground from the 2 signal wires....

I'm not sure I fully understand what you're doing but the above seems correct to me plus I would provide the 5V from a clean external source rather than the PC/laptop

 

[thoppa]

Hi,

It's great to read about your mods - I look forward to the pictures !

What 3.3v reg do you use/recommend if not using a 3.3v battery ? The best ic regs I have found are the Texas Instruments REG113 smt and these get down to about 7uV noise. I'd like to find something better for the xo in a Teralink X2, but without going crazy with the cost (aka Tentlabs, or Paul Hynes), because it isn't an async. I currently use a 4.4Ah li-ion battery. So I'm really interested to know what reg you are using or can recommend.

Thanks !

 

[Zerotohero]

Hello John,

Does the 3.3v battery and 5v usb cable mods also power the 1.8v fpga chip?

If so, then am I right in thinking that with these two mods the modded hiface is not using any power from the computer at all?

Thanx

 

[rasmushorn]

This is interresting. Looking forward to a review and comparison between the stock model and the modded one.

 

[mmerrill99]

Quote:

Originally Posted by thoppa /img/forum/go_quote.gif Hi, It's great to read about your mods - I look forward to the pictures ! What 3.3v reg do you use/recommend if not using a 3.3v battery ? The best ic regs I have found are the Texas Instruments REG113 smt and these get down to about 7uV noise. I'd like to find something better for the xo in a Teralink X2, but without going crazy with the cost (aka Tentlabs, or Paul Hynes), because it isn't an async. I currently use a 4.4Ah li-ion battery. So I'm really interested to know what reg you are using or can recommend. Thanks !

Here's the thing - I don't use a reg at all. What! Yep, for the absolute best, cleanest power supply I use a 3.3V rechargeable Lithium nanophosphate (LiFePO4) batteries. These are the the ultimate in low impedance (<8mohm), high current delivery(120Amps 7sec bursts, 70Amps continuous - try getting a reg to do this

) & low noise (the noise bit, I'm assuming from research papers about noise levels of other batteries <200dB).

I don't believe there is a reg that can beat them but I recently bought a Paul Hynes to test against them. Now these batteries aren't cheap but also not as expensive as a Paul Hynes reg as you pointed out.

Here's the final bit - they hold their output voltage fairly steady at 3.3V until exhausted.

That's my recommendation!

 

[mmerrill99]

Quote:

Originally Posted by Zerotohero /img/forum/go_quote.gif Hello John, Does the 3.3v battery and 5v usb cable mods also power the 1.8v fpga chip? If so, then am I right in thinking that with these two mods the modded hiface is not using any power from the computer at all? Thanx

I've dealt with the power supplies to the hiface in order of importance, sonically,
1. 3.3V to the clocks
2. 3.3V to the clock handling functions of the FPGA chip
These two mods bring you 95% of the way

3. 5V supply to the USB receiver chip
4 1.8V supply to the core logic functions of the FPGA chip

The really critical one is 3.3V supply as it supplies the clocks & the chips handling the clock & the USB receiver. So I modify the circuit in the Hiface to split it into two independent supplies - on to the clock & the other to the rest. I then supply these two 3.3V from external supplies.

The 5V supply is not as critical, sonically, but some advantage is gained by improving it - so this is an add-on improvement i.e doesn't require the hiface to be opened up. I modify a USB cable where the 5V USB supply is cut & a connector for an 5V external supply provided. Currently, nobody is doing this sort of modified cable, as far as I know. It has another advantage for audio - the USB cable is now solely carrying USB signals and not USB power as standard cables do - so the potential for interference to the USB signal is removed. I believe that this is what makes the "audiophile" USB cables sound different(!) - how well they shield the signal carrying wires from the power carrying wires inside the cable.

I remove the power carrying function of the wire altogether & at 20Euro it's a steal compared to the $100-$300 audiophile cable prices.

So if you get one of these cables you need to supply an external clean 5V supply to the connector provided, otherwise there's no 5V power to the Hiface!

With all the above there is no power supplied from the computer & it's variability & vagaries are removed from sonic considerations.

 

[thoppa]

Thanks for the reply. So are you using a 3.6v (fully charged) battery ?

 

[mmerrill99]

Quote:

Originally Posted by thoppa /img/forum/go_quote.gif Thanks for the reply. So are you using a 3.6v (fully charged) battery ?

So you've looked into these batteries already?
All the chips are tolerant of PS up to 4.0V & the 3.6V voltage of a fully charged battery quickly reduces to 3.3/3.2V level for most of it's output.

 

[hawkhead]

I've copied this post for relevance

I had been following the work done by Jkeny (John) modding the Hiface and decided to contact him to see if he would undertake the work. He agreed which was a relief because my soldering skills will make you wince.

My Hiface winged it's way to Ireland and returned a few days later

Batteries
2 3.3V LiFE batteries were needed to power the Hiface (the standard leads from the connector are a little short so these were perched on top of a box for testing)
This worked fine and the batteries are a great technology but by their nature you have no warning when they will stop, they need charging and won't last forever.

By the time you add up the cost of the batteries (2 sets), a box and connectors for neatness/safety and a suitable charger it begins to stack up

Given all this I decided to switch to a PSU

PSU
I bought a variable regulated PSU
I have extended the leads from the Hiface and will shield and heat tube them

Internal pics of the PSU

AC inlet

Middle

DC out

Bracket
One thing I intend doing is to fabricate a support bracket for the Hiface

Sound

I used three tracks for comparison (from memory)
Eva Cassidy - People Get ready
Deep Purple - Smoke on the Water (made in Japan)
Travis - Turn

The first thing that struck me was the bass - deeper with incredible control, a very dark background, more detail especially noticeable on cymbals, larger soundstage. I could have been in the room with Eva (nature is cruel sometimes :[)
Overall a very positive jump in quality

Software

I am using 1.03 on Windows 7 64bit and have not had any problems with KS or Wasapi with Foobar V1.01 (both sound the same)

thanks John

 

[thoppa]

Hi,

I re-organised the cells from a notebook battery so they gave 8.8V fully charged and 7.5V discharged. I've been using regulators to drop this and hunting around for reasonably priced solutions that improve upon this.

I hadn't considered a separate battery just for the xo. I think that's a very good idea - they use very small amounts of current so it doesn't have to be large and expensive and will last a long long time. Keeping one on a trickle charge while using another seems very doable.

How is the sound affected when the battery is close to exhausted ? I assume the device just shuts down ? With the X2, the sound just suddenly stops. No drama or noise or pops etc and as far as I can tell, no harm done.

 

[mmerrill99]

Quote:

Originally Posted by thoppa /img/forum/go_quote.gif Hi, I re-organised the cells from a notebook battery so they gave 8.8V fully charged and 7.5V discharged. I've been using regulators to drop this and hunting around for reasonably priced solutions that improve upon this. I hadn't considered a separate battery just for the xo. I think that's a very good idea - they use very small amounts of current so it doesn't have to be large and expensive and will last a long long time. Keeping one on a trickle charge while using another seems very doable.

Yes these LiFePO4 batteries are 2300mAH so one lasts for weeks driving the clock(s) at the current draw of these devices. It's also the cleanest, lowest noise PS around which is crucial for clocks.

Just to correct something that Hawkhead said in his review - the batteries will actually last for 10 or 15 years - they have a very high cycle life i.e can be recharged many thousands of times - the manufacturers estimate a lifetime of 10years in use within electric vehicles so for low current audio use double this to 20 years. These are also unlike other batteries where their power output gradually reduces over time & recharging cycles - these don't!!

Quote:

How is the sound affected when the battery is close to exhausted ? I assume the device just shuts down ? With the X2, the sound just suddenly stops. No drama or noise or pops etc and as far as I can tell, no harm done.

Exactly as you say - depending on what is being powered by the battery, the sound just stops.