From Shitty X-DAC V3 To Something

[XTRProf]

Hi all,

Itchy backside weekend was back and continued from where I had left off from the threads (Please click to go to the threads):



DNM T Network Caps

Burson Clock

Shit! Another X-DAC V3 Problem

Collating Favourite OPAs To Replace X-Dac V3 5532

X-PSU V3 Kaput, Dilemma, Pics, Solns & Improvements



Now into beefing up the unregulated PS rails in my X-Dac V3.







Features

1) Elna Audio 12,000uf caps for +12V and -12V of the unregulated analogue rails. Can't afford the BG which will go for hundred+ SGD a pop. Now avoiding regulators for minimal noise to see what is the effect like from KISS. Generally should be great from feedbacks from those hifi freaks locally.

Warning: Please note don't try it when you are using a low VA wallwart or other transformers of this capacity as my toroidal I'm using is 225 VA for comparison purpose after our guru Alex alerted me on this shortfall. Also, don't do it when u have not uprated the bridge diodes for the current surge on switch on with the larger smoothing caps. The diodes I'm using are monster types in the region of at least 60 VDC and 8 A per leg. See my actual diodes used as stated below. Thanks Alex!

2) To improve to almost PS zero impedance, will add in BG supercaps as bypass which will cancel internal caps resonance to achieve that. Have not added yet as no time to go and buy.

3) This is keeping to KISS principle for best sound after eliminating the impedance equation.

4) Using double run of Teflon OCC Cu for cap leads as too narrow to fit in without them.



Overall Sonic Before Run In

1) A bigger volume of the recording space using familiar CDs could be heard.

2) Grains now went down even further as to encourage one to play louder. Usually when we play louder there will be heard some dynamic compression that will give rise to grains and fatigue. Lesser now.

3) The dynamics and PRAT are now even higher with guitarists plucking nylon strings with more snap.

4) Macro and micro details could be heard better with singers breathing before singing the next line heard clearly now. Before, didn't notice that.

5) The overall sound is like changing to the BB 2134 but without the super smoothness of that.

[Deleted]

1) 12,000uf caps for +12V and -12V of the unregulated analogue rails.

XTRProf

The unregulated supply rails are NOT + and -12V.
THey are typically closer to + and -16.5V ! (12VAC x 1.414, ignoring rectifier losses) Some plugged in I.C.s will not like this high a voltage.

SandyK

[XTRProf]

sandyk said:

1) 12,000uf caps for +12V and -12V of the unregulated analogue rails.

XTRProf

The unregulated supply rails are NOT + and -12V.
THey are typically closer to + and -16.5V ! (12VAC x 1.414, ignoring rectifier losses) Some plugged in I.C.s will not like this high a voltage.

SandyK

Ya, Alex u r right provided there are no smoothing caps around. But at 2 X 12,000 uF, it will be closer to my hefty 225VA toroidal of about 13 VAC unloaded. This monster transformer will stay at around that and providing some form of "regulation" at 13 VAC, which will also be 13 VDC RMS loaded with smoothing caps no matter what load. Taking the load of the analog section using 6 OPAs and some transistors, the current required will hardly sag the unregulated PS with 2 X 12,000 uF in it. Can say as good as ripple free until I have an oscilloscope to magnify to prove otherwise.

Don't worry, I will also try regulation when I have a 15 VAC toridal around for other things in future. But as of now, not worthwhile to buy one when there is nothing in sight needing that. I also wanted to hear for myself as I'm Zen also. I will do anything that will improve the sound zenly first follow by Tech when I have the time and doe.

[Deleted]

XTRProf
I was more concerned about other adventurous members jumping in and fitting huge filter capacitors without being aware of the possible consequences. These include a burned out Wallwart , and blown up bridge rectifier. It is one thing to have 2,200uF filter capacitors, but a completely different ball game if you increase this capacity by almost 6 times ! Have YOUR bridge rectifier diodes been upgraded too ? Your technically oriented father would be mortified to see some of the leaps of ZEN faith you take ?
Alex

BTW, Mr. ZEN Master, if your larger toroidal transformer does supply 13VAC under this relatively light load, you may have close to + and - 18.4 V DC , particularly if you use Schottky diodes in the bridge rectifier. I would point out , that this EXCEEDS the maximum + and -18V ratings of many analogue I.C.s like the OPA2134 and LM4562 (+-17V recommended maximum.) Although the original NE5532 are rated up to +-22V, the manufacturers recommended MAXIMUM operating voltage is + and -15VDC !

[XTRProf]

Apr 8, 2008 5:30:50 GMT @sandyk said:

XTRProf
I was more concerned about other adventurous members jumping in and fitting huge filter capacitors without being aware of the possible consequences. These include a burned out Wallwart , and blown up bridge rectifier. It is one thing to have 2,200uF filter capacitors, but a completely different ball game if you increase this capacity by almost 6 times ! Have YOUR bridge rectifier diodes been upgraded too ? Your technically oriented father would be mortified to see some of the leaps of ZEN faith you take ?
Alex

BTW, Mr. ZEN Master, if your larger toroidal transformer does supply 13VAC under this relatively light load, you may have close to + and - 18.4 V DC , particularly if you use Schottky diodes in the bridge rectifier. I would point out , that this EXCEEDS the maximum + and -18V ratings of many analogue I.C.s like the OPA2134 and LM4562 (+-17V recommended maximum.) Although the original NE5532 are rated up to +-22V, the manufacturers recommended MAXIMUM operating voltage is + and -15VDC !

Thanks Alex for the advise. Yes, u are right. I better measure the actual VDC of the caps either tonight or tomorrow.

As for the bridge diodes, I have thought about that before even reaching here. All the diodes r Hexfred for the digital at 600 VDC and 8 amps per leg and for analog all r Schottky rated at 60 VDC and 10 amps each leg. So it wouldn't kill the diodes. To be safe, I will measure the smoothing caps VDC later and will reduce voltages accordingly. So far, no problem yet.

I have again changed my warning above to include the bridge diodes to all seeing. However, the sound is now into another level and better than what I had before in the spirit of Zen. Heh, heh, now I have to look at the tech side as well now to avoid complications later. Much thanks again.

[Deleted]

XTRProf
I am not concerned about the ratings of the Schottky diodes if used. Or the diodes being killed. They are more likely to kill something else, as happened to a member recently. Remember ? I am more concerned that due to their much lower forward voltage drop, the resulting DC supply voltages could be perhaps even half a volt higher again.
Alex

[XTRProf]

Apr 8, 2008 10:49:20 GMT @sandyk said:

XTRProf
I am not concerned about the ratings of the Schottky diodes if used. Or the diodes being killed. They are more likely to kill something else, as happened to a member recently. Remember ? I am more concerned that due to their much lower forward voltage drop, the resulting DC supply voltages could be perhaps even half a volt higher again.
Alex

Don't worry, I don't believe in killing things as rapidly as that member without even the parts being given a chance to survive. That's not true moding but playing casino! True moders make sure their equipments survive when switching on after moding. Same for a great commander during war. Although it may seem that I don't follow the usual but I'm careful even to fusing rating to make sure as many survive as possible during my electric chair episode. So far my 1.6 A slow blow fuse had survived after this monster caps mode on switch on.

[Deleted]

So far my 1.6 A slow blow fuse had survived after this monster caps mode on switch on.

XTRProf
WTF, is THAT fuse supposed to protect, except itself ?
Considering that a 12-0-12 20VA toroidal transformer would quite happily power the X-DAC V3, and they are normally used with a fuse no more than 500mA, it seems that the X-DAC V3 could go up in flames,due to a fault, and perhaps (if lucky) the fuse may blow just as the flames die down !

BTW, I lost a dual secondary 21-0-21VAC 160VA transformer that powered my Class A amplifier recently, when work at a new electricity substation caused a surge, and burned out it's primary.
Not only wasn't my surge protected powerboard inline at the time, but the fuse in use was too high a value, and only blew
after the damage was done.
Alex

[XTRProf]

Apr 8, 2008 11:23:57 GMT @sandyk said:

WTF, is THAT fuse supposed to protect, except itself ?

BTW, I lost a dual secondary 21-0-21VAC 160VA transformer that powered my Class A amplifier recently, when work at a new electricity substation caused a surge, and burned out it's primary.
Not only wasn't my surge protected powerboard inline at the time, but the fuse in use was too high a value, and only blew
after the damage was done.
Alex

No worries here. The rating was arrived at from actual testing with the monster toroidal. Also, in Singapore, the AC voltage is very much well regulated to within +/- 5 % or I can go after the PUB ass.

[Deleted]

XTR prof
That's not much consolation !
Alex

[XTRProf]

Yes, for an alternative ending to PS rails, like in dvd movies, check our resident guru Alex link here:


www.rockgrotto.proboards39.com/v45index.cgi?board=q&action=display&thread=2068


His approach is more Tech than Zen and mine is more Zen than Tech and in keeping to the KISS essence for best sound.

[XTRProf]

Apr 8, 2008 11:53:23 GMT @sandyk said:

XTR prof
That's not much consolation !
Alex

Heh, heh, my consolation will arrive when I have the best sound from KISS.

[Deleted]

Heh, heh, my consolation will arrive when I have the best sound from KISS.

In your dreams my friend ! ;D There is no way your fiddling around the edges is going to produce the same results as a dual regulated supply feeding it, let alone a dual regulated supply with JLH. You haven't even got voltage regulators supplying the analogue I.C.s ! And don't tell me that using 12,000uF filter capacitors is as good as using 2 onboard voltage regulators.
Bollocks !

SandyK

[XTRProf]

Apr 8, 2008 12:30:49 GMT @sandyk said:

Heh, heh, my consolation will arrive when I have the best sound from KISS.

In your dreams my friend ! ;D There is no way your fiddling around the edges is going to produce the same results as a dual regulated supply feeding it, let alone a dual regulated supply with JLH. You haven't even got voltage regulators supplying the analogue I.C.s ! And don't tell me that using 12,000uF filter capacitors is as good as using 2 onboard voltage regulators.
Bollocks !

SandyK

Ha, ha, my dear guru. Then you have not really experience what is for best sound. ;D Even our dear guru Rick has that alternative KISS philosophy which he does not despise. You are more like Stereo Review and Hifidelity mags. Where are they now wheareas Stereophile, TAS and Hifi News still prosper when they acknowledge the Zen? Vinyl is out of date when it is in fact making a big come back. Where is SACD and DVD-A now? The best in sound and not technically is still the tube amps when done properly for those with an open ear and open mind. Don't believe me have a look around for the best amps that people buy. I will stop here and do some calculations for my endeavours here than to have a debate again on this. Bye, bye ........... for here. And now back to "school" to relive some of the engineering mathematics which had long been forgotten for most already!

[XTRProf]

After doing some measurements tonight at the 12,000 uF rails, Alex was right but not in the manner of voltages that will kill the OPAs. The highest DC voltage measured was just 17.90 VDC, short of the 18V that could cause problem to the OPAs. Could cause problem is because there is usually some inherent safety factor above the specs that was not mentioned and those specs were just good guidelines. As most following my threads will know, I had been using the LM 4562 at those voltages for more than 3 months already and until now no problem yet although the specs said max 17 VDC. Anyway, I decided to add a voltage divider in series with the analogue rails. A calculated resistor value for that purpose, based on some assumptions, are in order before I can buy some resistors.

According to specs, the LM 4562 will take about 10 mA nominal. So based on 6 that I will be (6 X 10) = 60 mA. As the rails also supplied some other transistors around the OPAs. I will assume the rest will take about 40 mA. So total will be (60 + 40) = 100 mA for each + and - rails. Based on I want to drop the voltage by 3V thru the divider to get about 15 VDC at the OPAs,


The R = 3/0.1 = approximately 30 ohms.

The wattage P for the resistor = approximately (3 X 0.1) = 0.3 W


I also did a preliminary calculation for the smoothing capacitance at 10% ripple and found = (5 X 0.1) / (1.414 X 13)(50) = 5.44010445 X 10 to the power of -4 = 544.010445 uF. So even at 2,200 uF (the original as used in the MF Dac) the ripple will be more or less be only 1/4 of 10% = 2.5%. At 12,000 uF, the ripple will again be 1/6 of 2.5% = 0.4167% based on 4 significance figures. So the facts are out based on unregulated at 12,000 uF. This 12,000 uF will be supplemented with BG super Ecap of NX Hi-Q at most properly 0.47 uF to give internal resonance cancellation to give almost zero output impedance. Please check the BG webpage for details on this. Anyway, high output impedance according to that page is usually refering to the high frequency impedance which is not so critical at PSU frequency of 50 Hz or 60 Hz unless we are taking of decoupling caps in the signal path. In fact, we want the high frequency high output impedance to reduce the RFI noise.

Hope I had interpreted the datas well and correctly as I try to get up to speed as it had been a long long time since I read anything technical since college. The end for now as I go get or make the resistor value for a more user friendly voltage for the LM 4562. I can also go to the AD8066 which is rated at max 24 VDC to avoid any modification necessary here.

[Deleted]

In fact, we want the high frequency high output impedance to reduce the RFI noise.

Wrong ! We need the flattest,and lowest supply impedance right across the frequency range to >100KHZ. If you read the application notes of some modern higher speed analogue I.C.s you will see that the supply pins are referred to as inputs. Any noise present on the supply lines will be amplified, but if supply impedance is low this will not be a problem, as correct supply bypassing at the I.C. will ensure stability. Using voltage dividers, unless properly bypassed will raise PSU impedance.This is not desireable, and will lead to a non neutral sound.
XTRProf
I will let you go on your merry way. What you are doing is breaking all the rules for good audio, and I highly recommend that members seek out the free application notes from semiconductor manufacturers and take THEIR advice. Hey, what would they know ? They only designed the damn things !
There are plenty of good books on the subject too.

AD8065/AD8066
Rev. E | Page 21 of 28
LAYOUT, GROUNDING, AND BYPASSING CONSIDERATIONS
POWER SUPPLY BYPASSING
Power supply pins are actually inputs and care must be taken so
that a noise-free stable dc voltage is applied. The purpose of
bypass capacitors is to create low impedances from the supply
to ground at all frequencies, thereby shunting or filtering most of the noise.
Decoupling schemes are designed to minimize the bypassing
impedance at all frequencies with a parallel combination of
capacitors. 0.1 ìF (X7R or NPO) chip capacitors are critical
and should be as close as possible to the amplifier package.
The 4.7 ìF tantalum capacitor is less critical for high frequency
bypassing, and, in most cases, only one is needed per board, at
the supply inputs.
GROUNDING
A ground plane layer is important in densely packed PC boards
to spread the current minimizing parasitic inductances.
However, an understanding of where the current flows in a
circuit is critical to implementing effective high speed circuit
design. The length of the current path is directly proportional to
the magnitude of parasitic inductances and therefore the high
frequency impedance of the path. High speed currents in an
inductive ground return will create an unwanted voltage noise.

[XTRProf]

Hi all,

TGIF and I had found 2 18 ohms 2 watter resistor in my "dumping ground". I will fix it to the +/-12 PS rails and report back later on the sound. I expect it to be shitty again from reducing the "biases" in the LM4562s.

So hang around, will u?

[Deleted]

I will fix it to the +/-12 PS rails

XTRProf
You mean the +-17.90V supply rails ?

"The LM4562 has a high slew rate of ±20V/ìs and an output current capability of ±26mA."
The calculations are based around 10mA quiescent current of a typical LM4562. This means that the proposed supply will be inadequate.

I can also go to the AD8066 which is rated at max 24 VDC to avoid any modification necessary here.

Again, you have misinterpreted the data.
The 24V supply quoted is for SINGLE rail operation ! i.e. +24V DC For dual rail operation it is best not to exceed +-12V. This is the reason for the use of the added 78L12 and 79L12 regulators in my modified X-DAC V3
SandyK

XTRProf
I will refrain from posting any further criticisms in this particular thread, if you prefer, but I would hate to see you, or anybody else, damage their X-DAC V3s, or blow up expensive chips like the AD8066 etc.
Alex

[XTRProf]

Apr 11, 2008 13:31:32 GMT @sandyk said:

I will fix it to the +/-12 PS rails

XTRProf
You mean the +-17.90V supply rails ?

"The LM4562 has a high slew rate of ±20V/ìs and an output current capability of ±26mA."
The calculations are based around 10mA quiescent current of a typical LM4562. This means that the proposed supply will be inadequate.

I can also go to the AD8066 which is rated at max 24 VDC to avoid any modification necessary here.

Again, you have misinterpreted the data.
The 24V supply quoted is for SINGLE rail operation ! i.e. +24V DC For dual rail operation it is best not to exceed +-12V. This is the reason for the use of the added 78L12 and 79L12 regulators in my modified X-DAC V3
SandyK

XTRProf
I will refrain from posting any further criticisms in this particular thread, if you prefer, but I would hate to see you, or anybody else, damage their X-DAC V3s, or blow up expensive chips like the AD8066 etc.
Alex

Hi Alex,

Yes, when I said +/- 12 rails, I meant the measured 17.90 VDC.

I'm very open on things and when there is any truth in them I accept them. What is there to loose? Face value? Come on, we have to face the truth, right, and nothing but the truth as in court cases. As you can see, u are absolutely right when u said the LM4562 recommended max is +/- 17 VDC and that's what I'm working on right now with the 2 18 ohms 2 watters which I had found in my "dumping ground" yesternight. However, I don't agree that at 17.9 VDC now the LM4562 will be damaged as usually the specs will not state the safety factor (n) so as ensure better reliability. But it's still safety factor no matter how much. Some may be conservative n and some very thin n. This was shown by my true case doing of running the LM4562 at +/- 17.9 VDC for mre than 3 months now.

Err, I though the AD8066 24 VDC is for +/- and not for single rail as that's what specs will usually state just like the LM4562 17 VDC max? Anyway, I will check the specs again and thanks for bringing that up. So keep your "critisms" coming as I firmly believe in what I'm doing, without blwing things like a terrorist, to stand the test of time. I'm a bonafide moder and not a terrorist from my record so far. May the best sound win!

[Deleted]

XTRProf
Just because your LM4562 continued working with +-17.9V supply rails doesn't mean all specimens will. The SQ is also likely to be better when using the manufacturer's recommended voltages. A while back, I inadvertently supplied an AD8066 with +15 and -12V .It did not survive. When we post, we should be mindful that less technical members expect that our advice is sound. We are in a position where we are able to advise and educate less advanced hobbyists. We should be teaching them the correct way to do things, and that includes not taking liberties with manufacturer's published recommended maximum voltages.
I will offer a word of advice from a member that was posted recently. It is well worth keeping in mind .
"I wouldn't do anything until I'm sure of what to do as that will be costly in the longer term to repair. If not sure just enjoy the stuff until I can get my hands on more info first."
There is a fine line between what you are doing, and being a "Terrorist" who blows things up !


SandyK

[XTRProf]

Apr 12, 2008 3:32:37 GMT @sandyk said:

XTRProf
Just because your LM4562 continued working with +-17.9V supply rails doesn't mean all specimens will. The SQ is also likely to be better when using the manufacturer's recommended voltages. A while back, I inadvertently supplied an AD8066 with +15 and -12V .It did not survive. When we post, we should be mindful that less technical members expect that our advice is sound. We are in a position where we are able to advise and educate less advanced hobbyists. We should be teaching them the correct way to do things, and that includes not taking liberties with manufacturer's published recommended maximum voltages.
I will offer a word of advice from a member that was posted recently. It is well worth keeping in mind .
"I wouldn't do anything until I'm sure of what to do as that will be costly in the longer term to repair. If not sure just enjoy the stuff until I can get my hands on more info first."
There is a fine line between what you are doing, and being a "Terrorist" who blows things up !


SandyK

Ya, sometimes the calculated and the actual defers by quite a lot from what we can deduce from the specs. This because the designer has the liberty of the actual circuit digram whereas we have to backtrack from PCBA tracing like Shellock Holmes. Because of that, disaster can sometimes strikes. But usually with experience, we will be able to tell what is not right and what is right to avoid the blowup, although at times, we come pretty close to the blowup like in my case. But, in my case of the LM4562, it's not a lucky dip as there are 6 LM4562 to jury and all survived at 17.90 VDC for more than 3 months of almost daily dosage of about 2 to 3 hours of using the X-Dac V3. Saturdays and Sundays even longer to say about 5 to 6 hours as I read more to catchup with the news or as usual hifi mags and surf the net for more hifi related news and diys. So 6 juries cannot be wrong. So again, it's the inbuild safety factor that IC designers had in them that was not mentioned in the specs as they wanted avoiding telling us too much and we trying to work to the limits of things. This will give very poor reliability problem and will result in a very poor image of certain brands. I'm sure those designers will be fired as well if this do comes up.

Regarding the AD8066, I have a look at the specs again and although it didn't state +/- 24 VDC, it obviously implies +/- 24 VDC as all OPA operates under +/- VDC. The 5532 from NJM is +/-20 VDC max and the TI 5532 is +/-22 VDC max. As you can see, there is a difference of +/- 2 VDC max between the NJM and TI 5532, althought they are the same products. How come? The unstated safety factor comes into play! Also, the BB range is +/- 18 VD max nd the Linear range is +/-17 VDC max. So the range should be deduced to be from +/- 17 to +/-24 VDC max and a range of +/- 12 VDC is unlikely. So I'm quite surprised your AD8066 kaput at +15 and -12 VDC. Really, this I'm not sure as it deviates quite from norm specs until this can be confirmed with other members with this observation. So in the meantime, I will not touch AD8066 until I'm sure what it is all about and just enjoy whatever I have. I'm no terrorist!

As with regards which supply voltage sounds the best, I can determine that out thru real listening as I have already soldered in the 18 ohms 2 watters. Dun worry, this will sure survive! But from friends' biasing feedback, lower voltagea will thin out the sound and everything seems to have collapsed and higher voltages will make the sound more musical, flesh out, dynamic and more smooth. Well, I can do that now to expereince what they had said.

So be sticky here as I will report back after this trial.

[Deleted]

XTRProf
Haven't you heard of devices that are designed to work with single rail voltages ?
I dare you to plug an AD8066 into your X-DAC V3 with your present voltage !
Seeing you are so gung ho, and disbelieving, cast your eyes over the section from another couple of excellent audio device's specs.
You will probably brush aside their warnings too ?
This does NOT mean + and -27.3V either!
or in the case of the AD8066, working with an effective overall voltage of 52.8V !!!
Still not convinced?
"IC Module AD8066 - Soundlabs Group For this reason we can supply it on an adapter PCB to make this easy! ... Note: The AD8066 has a maximum voltage rating of ±13V, so please ensure your ...
www.soundlabsgroup.com.au/p/IC-Module-AD8066/IC+Module+AD8066 - 12k - Cached - Similar pages - Note this "

Alex

AD8065/AD8066
Rev. E | Page 6 of 28
ABSOLUTE MAXIMUM RATINGS
Table 4.
Parameter Rating
Supply Voltage 26.4 V
Power Dissipation See Figure 3
Common-Mode Input Voltage VEE − 0.5 V to VCC + 0.5 V
Differential Input Voltage 1.8 V
Storage Temperature −65°C to +125°C
Operating Temperature Range −40°C to +85°C
Lead Temperature Range
(Soldering, 10 sec)
300°C
Stresses above those listed under Absolute Maximum Ratings
may cause permanent damage to the device. This is a stress
rating only; functional operation of the device at these or any
other conditions above those indicated in the operational
section of this specification is not implied. Exposure to absolute
maximum rating conditions for extended periods may affect
device reliability.
ESD CAUTION

AD8610/AD8620
ABSOLUTE MAXIMUM RATINGS
Table 3.
Parameter
Supply Voltage
27.3 V
Input Voltage
VS− to VS+
Differential Input Voltage
± Supply Voltage
Output Short-Circuit Duration to GND
Indefinite
Storage Temperature Range
R, RM Packages
–65°C to +150°C
Operating Temperature Range
AD8610/AD8620
–40°C to +125°C
Junction Temperature Range
R, RM Packages
–65°C to +150°C
Lead Temperature Range (Soldering, 10 sec)
300°C
Rating
Stresses above those listed under Absolute Maximum Ratings may cause permanent damage to the device. This is a stress rating only; functional operation of the device at these or any other conditions above those indicated in the operational section of this specification is not implied. Exposure to absolute maximum rating conditions for extended periods may affect device reliability.

Table 4. Thermal Resistance
Package Type
èJA1
èJC
Unit
8-Lead MSOP (RM)
190
44
°C/W
8-Lead SOIC (R)
158
43
°C/W
1 èJA is specified for worst-case conditions; that is, èJA is specified for a device soldered in circuit board for surface-mount packages.
ESD CAUTION
Rev. E

[leo]

XTRProf, you should listen to the guy above, I can assure your going to run into problems my friend if your not careful!

A lot of the newer opamps are now designed to run on lower supplys, running them to over their max rated spec not only reduces reliability but they also go noisy after long term use

Running something like a AD8066 at +/-27v will instantly fry it

ABSOLUTE MAXIMUM RATINGS
Table 3.
Parameter
Supply Voltage
27.3 V
Input Voltage
VS− to VS+

That does not mean to run both its supply pins at both +27.3v to ground and -27.3v to ground , you pick +27.3v to ground or +13v AND -13v to ground

[XTRProf]

BELATED REPORTING PART ONE




Arrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrr ................. Horror! Heh, heh, nothing fried yet but ............ All the predictions about running at lower voltage are true that had even caused me to rewrite the subject title above just for this.

Running at 18 ohms for both /- rails at about 12.6 VDC max, the whole soundstage just collapsed with singer voices seemed to have lost their voices and turn amateurs. The large volume, air, spaciousness, width and depth that were captured earlier were all gone. I can't accept it and had nightmares over it.

Don't worry, I have enough parts today to change all that when I have the time again.

So short and sharp for now. Again, be sticky around here as I do my magic! Part 2 coming and with pictures ............

[Deleted]

Running at 18 ohms for both /- rails at about 12.6 VDC max, the whole soundstage just collapsed with singer voices seemed to have lost their voices and turn amateurs. The large volume, air, spaciousness, width and depth that were captured earlier were all gone.

XTRProf
It has absolutely nothing to do with the lower voltage. The analogue section of my X-DAC V3 is running at +-12V and has a fabulous sound stage. You no longer have a LOW impedance PSU due to your voltage dropping method. You would have been far better off fitting a couple of 78L/79L 12 or 15V regulators and a 100uF capacitor either side. This would have reduced your voltage to a safe limit AND also improved your soundstage on what you had previously. If you don't want to muck around with the PCB, shove the added components on a small piece of veroboard or similar.

SandyK