ESS Saber DAC

[jkeny]

Yes I have used 2 DPDT switches as I was worried about the order of powering up the various PS of the DAC but I since learned that the Sabre DAC is tolerant of the power on sequencing so reverted to the JST connectors. These are not low impedance but they are OK. What I'm really experimenting with is what PS need to be individual & which ones can be combined.

I'll be moving all of the connections to Mosfet switches when I've finalised the PS configuration

[grottojones]

I use a single 4PDT one of these:

www.nkkswitches.com/pdf/MtogglesBushing.pdf

Contact resistance is spec'ed at 10 milliohm max. Don't know what the impedance may be, but as it is similar to the one Charles Altmann uses in his DACs for switching clock frequencies and SPDIF, its probably pretty benign.

You can get them at Digikey or Mouser. I got mine at work from HP labstock.

I've thought about relays, but its seems to be working well with the single manual swiitch.

[Deleted]

Oct 2, 2010 1:11:42 GMT grottojones said:

I use a single 4PDT one of these:

www.nkkswitches.com/pdf/MtogglesBushing.pdf

Contact resistance is spec'ed at 10 milliohm max. Don't know what the impedance may be, but as it is similar to the one Charles Altmann uses in his DACs for switching clock frequencies and SPDIF, its probably pretty benign.

You can get them at Digikey or Mouser. I got mine at work from HP labstock.

I've thought about relays, but its seems to be working well with the single manual swiitch.

grottojones
I also suggested a 4PDT to John.
Alex

[jkeny]

The DPDT switches that I use on the boxed Hiface are 10mOhm contact resistance too - silver plated.

I mentioned this thread to the person who compared them to the PH regs & he said "I found the PH reg to have good transparency and was nicely dynamic, but was a little splashy in the treble, perhaps due to the Oscon that Paul Hynes fits on the output. The battery was just purer sounding, with real hear-through transparency."

[leo]

Mine is also PDT silver contacts, contact resistance 0.01R max, they was from Farnell, make is C&K I think

There is no Oscon on the output of the mini shunts , not the ones I have anyway, the Oscon is used for something else .
Certainly no splashyness here and I have hearing like a bat

[Deleted]

Leo from Transylvania.

[leo]

[jkeny]

Leo,
I think there are Oscons on mine too (from memory). Are you using his Z1703.3 or Z7803.3 shunt regs?

[leo]

John,

There is Oscons on mine but they are not used on the output of the regulator , they are used for something else.
I use the 3v3 with the 78** type pin outs

[jkeny]

Oct 2, 2010 2:00:29 GMT leo said:

John,

There is Oscons on mine but they are not used on the output of the regulator , they are used for something else.
I use the 3v3 with the 78** type pin outs

Yes, I think he made a slip here as he said in the past that the Oscons were filtering the voltage reference & he was going to try replacing them.

The ones i have are Z1703.3 which PH recommends for the Buffalo. Is there any difference between these & the 7805 replacements other than the pinouts?

[leo]

Oct 2, 2010 2:06:55 GMT jkeny said:

Oct 2, 2010 2:00:29 GMT leo said:

John,

There is Oscons on mine but they are not used on the output of the regulator , they are used for something else.
I use the 3v3 with the 78** type pin outs

Yes, I think he made a slip here as he said in the past that the Oscons were filtering the voltage reference & he was going to try replacing them.

The ones i have are Z1703.3 which he recommends for the Buffalo. Is there any difference between these 7805 replacements other than the pinouts?

Apart from the pin out I don't think there is any difference between the Z17 or 78 (I could be wrong) , I used the Z17's in the Buffalo32 and 78's in the BII

[squeezethatbuffalo]

Hi guys,

I'd like to share my Buffalo II build. Big thanks to Leo for his posts and advice (mostly on another forum )!

The DAC module is powered by Paul Hynes Z1P5v45, Volumite by a separate LT1086 5v supply, and each channel of the Legato gets its own "Golden Reference" bipolar supply and transformer.

Legato is mostly stock, I switched the FW caps to BC037s, and used MKP2000 3.3uf over Muse ES. The Lundahl transformers are hidden under the Legato board. I tried the compact Wima 6.8uf caps, and they actually sounded fine. However, the bass was a bit bloated and muddy in my system, so they had to go. The MKP2000s are still running in, but they already sound better to my ears. More transparent sound, better separation, and "normal" bass (probably due to lower capacitance though).

Although I like my old Buffalo alot, the BufII is obviously in a different class. I'd like to compare the BufII+Legato to much pricier Sabre 9018 DACs, such as the DAC202 and Moon 750D. Now I just need find someone who owns one of those DACs in Taiwan

Sorry about the iPhone quality picture. I'll update a better one soon®.



Cheers,
Alex

[paulhynes]

jkeny,

Regarding your statement :

I mentioned this thread to the person who compared them to the PH regs & he said "I found the PH reg to have good transparency and was nicely dynamic, but was a little splashy in the treble, perhaps due to the Oscon that Paul Hynes fits on the output. The battery was just purer sounding, with real hear-through transparency."

Ask your source how he powered the PH reg. If it was a mains powered DC source I am not surprised there were some mains power related artefacts present. This low level interference can mask audio transparency and give an edgy HF performance. Some regulators deal with mains issues reasonably well, but no regulator that I know of (including mine) can eradicate it completely. This is why there are so many mains treatment products on the market.

I am currently using spiral wound gel batteries in place of my mains power supply for my Altmann Attraction DAC and this battery pack feeds all of the regulators I have applied to the DAC. The relatively low HF edge that was still apparent with the mains supply has gone leaving a nice open mid and high frequency presentation. I wasn’t too keen on any type of batteries that I tried directly connected to the standard Attraction DAC, as, whilst the mid and HF was fine, the bass was loose and sounded out of time with the rest of the bandwidth, but the regulators cured this with tight but open bass without loosing the transparency benefit of being off grid. I have found this battery bass anomaly happens with other items of equipment too so it is not just an artifact of the Attraction DAC. The regulator’s very low output impedance (lower than any of the batteries I have tried) also benefits the dynamic scale and the image stability under large signal dynamic swings.

I would suggest that if you or your source have not tried your PH regs with a battery supply it would be worth setting this up to get a fair comparison with direct battery connection and a regulated battery supply. You may be surprised. Leo's findings would suggest this is very likely.

Incidentally, the Oscon on the shunt reg board is used to provide HF decoupling for the error amplifier. This error amp has a gain bandwidth product of 1 GHz. I am sure I do not need to tell you how critical careful decoupling is with a device like this. The Oscon has the right HF characteristics for this job. Don’t even think about changing this cap unless you can verify stability with a wideband oscilloscope with at least 250MHz bandwidth.

Also the Z17 and Z78 shunt regs are the same circuit just with a different pinout to aid orientation.

Regards
Paul

[jkeny]

Paul,
I appreciate you posting here & helping out. I would prefer the person answering this himself but he is gone away for the weekend. I'll try to see if he is interested in joining up here, I don't think he's a member.

From our previous correspondence here's what he said - "I've been using it to feed the analog supply of a Sabre DAC. The PH reg is in turn supplied by a very heavily filtered supply. The whole DAC is feed from a 1KVA isolation transformer, so as you can see, as mains based supplies go, this is pretty damn clean."

[paulhynes]

John,

I am interested in how LiFePo4 batteries behave in case they will be of benefit to my system.

When comparing different devices in audio equipment, with a view to improvement, it is important to ensure a level playing field to reach any valid conclusions. I would be surprised if the mains power supply you are describing is able to stop all common mode and differential interference, or all rectification and ground return current artefacts from getting through to the load. The mains power supply would have to have infinite supply rejection to do that and I have not found one of these yet. The only way to be sure whether the mains supply or the regulator is the cause of the veiling and HF edginess is to set up a battery driven regulator verses direct battery connection trial.

You have a shunt reg and some LiFePo4 batteries so it should be easy enough to do such a trial. The shunt reg dropout voltage is around 1v5 so two LiFePo4 batteries in series will happily run a 3v3 shunt reg for long enough to get some idea of which is the better method. You will need to get the shunt reg properly run in before doing this trial. Because of the relatively high current draw of the shunt reg I would suggest you burn in the reg on a mains driven 6 volt DC supply so you are not constantly recharging the batteries. A couple of weeks of burn in should be enough for a valid trial.

Regards
Paul

[jkeny]

Paul,
Thanks for your suggestions. I don't wish to upset your commercial activity in any way so I'll refrain from further public posting on this. I've PMed you!

BTW, I've no commercial interest in these batteries (I don't sell them) I just use them to power some clocks inside my modified Hiface) so I'm agnostic as far as their performance in this role of powering the AVCC

BTW here is the original post www.head-fi.org/forum/thread/446375/usb-24-192khz-m2tech-hiface/930#post_6581482

Regards
John

[paulhynes]

For the record this is not about my commercial activity. I have no wish to stand in the way of real performance improvement and would happily incorporate such improvement into my own system as well as into my design work if the cost/return ratio proves worthwhile and there are no intellectual property right infringements involved.

I am concerned that assessment procedures for upgrades can leave something to be desired. I see this with all sorts of equipment, across many projects on many forums.

At the risk of repeating myself, as an example, I do not see how you can assess the performance of any voltage regulator against a battery, if you are feeding the voltage regulator from a mains powered DC power supply. You can only assess the combined performance of the mains driven power supply and the voltage regulator against the battery. You have to get the regulator power feed off grid too, to be able to eradicate any mains related artefacts from the assessment equation.

If anyone has anything further to add to this feel free to comment.

Regards
Paul

[Deleted]

Hi Paul
Wecome aboard. Your fame precedes you.
A good heavy duty mains filter using several varistors never goes astray, even with the best of regulators, as well as helping to protect that expensive iron. Quite often a selected value capacitor to match the VA of the transformer can help matters further when connected across the outers of the secondary windings. Leo will also verify that using fast diodes in the BR section such as the SF12 and SF14 also makes a worthwhile improvement.
Regards
Alex

[leo]

Hi Paul,

Great to see you on here !
As mentioned further up I've been trying a few different options including running the shunts from the batteries, I wanted to spend a little more time listening before commenting too much , so far IMHO running the AVCC from the battery direct wasn't as good as the batteries-shunt-AVCC.
Adding the shunt improved the timing, depth and gave more weight and realsim to the performers in my system which I felt was missing from running the AVCC direct from the batteries.

[paulhynes]

Hi Alex and Leo,

Thanks for the welcome.

There is a lot that can be done to mains driven DC power supplies to reduce their annoying artefacts and I have tried and used most of these techniques over the years. Using all the tricks available can take a product into the high end of performance and it will also take it into the high end of cost too especially commercially. A perfect power supply would have infinite transient response, zero settling time, zero noise, zero impedance at all operating frequencies of the load and infinite supply line rejection at all frequencies of any upstream interference. Would I like one of these? Yes, but we are stuck with the real world and anything we can do to progress things without high-end expense is laudable.

Using a battery as the master supply neatly sidesteps the mains issue and eradicates a whole chunk of expensive conditioning electronics which at best only reduce any problems on the mains layer by layer and by finite amounts. Also there is more to this than just supply line rejection. We are still dealing with a power supply whichever method is chosen so it is important to consider how the power supply reacts with the load and visa versa. Here the ideal zero impedance, at all load operating frequencies, would be very nice. You cannot generate any voltage change across zero impedance no matter how much current you pass through it. I haven’t been able to achieve this and as far as I am aware neither has anyone else. However the closer you get to this ideal the less reaction you will get on the power line from load current changes. There are a number of batteries available that have reasonably low impedance, although I suspect that the impedance is rather frequency dependant with capacitive and inductive effects within the battery. It would be interesting to run some impedance tests on some of the typical batteries used in audio. I have put this on my list of things to do but it may be a month or two as I am still pushing hard to catch up with work.

Thanks for presenting your results Leo. Your assessment techniques should be an example to all. I hope I haven’t thrown your normal assessment procedures by starting this discussion.

Out of interest I am now running my (Altmann) DAC using a spiral wound gel battery stack with the usual voltage regulation applied to the key supply sections of the DAC. I wasn’t happy with direct battery connection with any of the batteries I tried, for the same reasons Leo has presented. What I do like with the battery master supply plus careful regulation is the more open mid to high frequency range and the lack of edge in the top end compared to any of the mains driven solutions I have at my disposal. This has now become my preferred choice of power for the Altmann.

There are a few operational issues with batteries that must be addressed for safety. The major one, of course is the correct charging procedure to maintain the batteries in tip top condition, and fully charged ready for listening. There are plenty of chargers around for this. Another one is connecting a charged battery to an electronic system. Two important issues here are battery polarity (get this wrong and you will let the smoke out of your electronic components) and surge current arching across the on/off switch, where there is any capacitance on the power line. This arching can damage switch contacts very quickly and is worth eliminating this problem to prolong switch life and ensure the impedance remains low over the life of the battery, which is not the case with a badly burned switch contact.

I have prepared simple low loss solutions for these two issues and if any experimenters want to look at this I will be pleased to prepare and post a diagram showing how to do this safely with very little voltage drop, minimal impedance penalty, and, if required, a slow run up of the supply voltage.

I have to go do some work so will sign off for now.

Regards
Paul

[grottojones]

FYI: the spec sheet of the A123 LiFePO4 battery that I'm using:
www.a123systems.com/cms/product/pdf/1/_ANR26650M1A.pdf

"10A DC resistance 10milliohm, 1kHz AC impedance 8milliohm, 120A pulse discharge for 10 seconds, 70A continuious discharge."

Pretty impressive, or so the marketeers write.

I wouldn't mind seeing actual measurements of the DUT. I'll see if I can get it done here at work (Hewlett Packard).

I first tried these batteries as a replacement for the Optima Red Top 12V automotive battery feeding the 5V48 ALTREG feeding my Buffalo II with PH regs on the AVCC supplies. They were IMO such an improvement over the Red Top, that I started replacing the active supplies. Now I'm running the four 3V3 supplies directly off of LiFEPO4 batteries and am quite impressed with the sonic improvements that running the DAC off of LiFePO4s brings to my system.

Of course, YMMV.

[dangiovanni]

Hi All,

New here - I'm the chap Jkeny was referring to who has done a number of tests including the PH shunt regs and the A123 batteries. There's been a few pages of discussion, so not sure where to start. If anyone has any questions then please ask.

I have noticed that already my hearing, the quality of my system relative to someone else's and my testing methodology have already been questioned. This is strange, as no-one here has any knowledge of these things at all except perhaps Jkeny. Kinda disappointing - no-one is subjecting Leo's experiences to a similar type of scrutiny.

First, I found the PH shunt to sound very good, roughly equivalent to the modified PFM Flea I was using previously, and easily better than a number of other reg's I'd tried. I did find it slightly splashy, as noted, in comparison to the Flea i.e. all fed from the same heavily filtered/conditioned/pre-regulated mains supply.

The battery was easily better in terms of hear-through transparency, which really surprised me. As noted earlier, I speculated leakage noise that is eliminated with a battery system may be responsible, but I'm not sure. Either way, the batteries can give really excellent results.

I'm not convinced how important ultra low output impedance is - as it gets lower it's going to get dominated by other sources of impedance anyway - tracks, contacts. It's also going to depend heavily on the particular application - some really won't need 1mOhm Z at 200KHz.

To finish for now, enjoyed reading others' experiences, even if they differed from my own.

[leo]

I've tried various versions of the Flea including alterations to make it 3v3 output, very surprised at your findings in all honesty.
Anyway as long as people are happy , I've gone to the trouble of faffing about adding and removing parts with the batteries, at the end of the day I still prefer the regs , I don't see what the problem is.

Because I liked the batteries as a main source supplying the regs I'm going to try the series regs instead of the shunts. These will be a much lighter load for the batteries than the shunts

[Deleted]

Kinda disappointing - no-one is subjecting Leo's experiences to a similar type of scrutiny.

Perhaps that is because on many occasions, what Leo has reported hearing has been confirmed by several other RG members ?
I wouldn't be surprised if the reason that designers such as Paul Hynes and Russ White have become members here, is because the findings and reports by Leo have also been widely accepted in DIYAudio and other forums. Still, when you get down to the incredibly high performance of DACs like the Buffalo series after modifications , in comparison with earlier popular DACs, it shouldn't seem surprising that other areas of the chain will be shown to have a marked influence on the listening experience. Even the types of diodes used in the PSU area has an influence, with several RG members now reporting (me included) that the faster SF12 and SF14 series, lifts a veil in comparison with the much touted UF400x series.
Things like that may well have an influence on the amount of HF rubbish that gets through, to sometimes make the Paul Hynes regulators seem a little "splashy " ?
SandyK

P.S.
I will not use the LiPo batteries with my SC DAC , because the DIR9001 maximum working voltage is given as 3.6V.
A freshly charged LiPo will be above that voltage.
I will however, in the next few days be trialling 2 Energiser Ultimate Lithium batteries in series for the DIR 9001, and another 2 for the 3.3V supply to the DAC

[jkeny]

Alex,
I've said it to you privately but maybe worth repeating it publicly - that is not the max working V 4V is the absolute max; 3.6V is the recommended max. The LiFePO4 batteries (not LiPO) when fully charged are 3.7V but you can discharge them back to 3.3V within seconds before operation if it worries or just charge them up to 3.3V. They really don't work any worse at 3.3V than at 3.7V.

Not all batteries are the same & your trial with Ultimate Li batteries will reveal what Li Ultimate batteries sound like in your system - nothing more - results good or bad cannot be generalised to other batteries. It may well help ground loop or RFI issues but that is perhaps all that can be generalised.