Square wave is always relative term; analyzed at sufficiently high bandwidth, square is no longer square. Logic is in terms of settling time required for stable result.
All that matters is recognizing positive or negative going zero crossing. Sine wave and square wave then become same problem.
Why start with sine and turn into square wave with same zero crossings?
All that matters is recognizing positive or negative going zero crossing. Sine wave and square wave then become same problem.
Why start with sine and turn into square wave with same zero crossings?
In which case, it is not the shape of the wave, but where it is at the critical moment and that it is repeatable over time and at every cycle. It's all about stability. If not, then the error actually will take on a phase noise pattern as it moves about. Noise, as always, the most critical aspect of digital playback.
Cheers, Joe
Speaking of digital noise; can we add more dithering noise to add pleasant euphony?
Or can we subtract more, and eliminate jitter to the point of a blacker background?
Is AC noise one of the main culprits? ...RFI, EMI?
Mechanical noise, an attribute?
Secrets of Home Theater and High Fidelity has extensive measurements on the Oppo BDP-105 'Magic-of-all-trade' player.
Or can we subtract more, and eliminate jitter to the point of a blacker background?
Is AC noise one of the main culprits? ...RFI, EMI?
Mechanical noise, an attribute?
Secrets of Home Theater and High Fidelity has extensive measurements on the Oppo BDP-105 'Magic-of-all-trade' player.
Except, for the issue of waveform rise-time. Whatever circuits are being clocked will act as the arbiters of the zero-crossing moment. Or, in the case of digital logic, the state transition thresholds. The longer the rise-time, the wider the time window defining a valid threshold-crossing moment, hence the wider the window for potential jitter introduction, or so it would seem. I agree that a sinewave clock signal would provoke the least supply based LIM and RFI related jitter disturbance. Opinions do seem to vary regarding which approach would result in the lowest net system jitter appearing at the DAC's analog outputs.
Last edited:
Very simple answer to that one - every oscillator is a tuned circuit, whether XTAL or MEMS. Tuning means a single frequency comes out hence there's no option of starting with a square wave.
Thanks all for comments.
Yes, its right that this is an Oppo thread regarding its including Sabre DAC, and other mods. It may looks like a clock discussion it may be out of the thread subject, but I think is not so bead to analyse a little bit in depth some aspects involved in these Oppo player mods, DAC and clock system improvements, and so on.
In fact my digression about square/sine clock is caused by the wish to find the best way to transport this clock signal into the player, best way to interface it to the receiver circuits, etc.
BTW, Crystek oscillators are quite famous for its quality and good figures. These oscillators output square signal. This signal looks quite bad when about overshoots/undershoots, and of course it looks very bead on FFT, comparing with a standard oscillator sine output...
Even more, I tried to filter out a square Crystek 975 output to get rid of the lot of harmonics, terminated as it should, and so on. I got a better clock signal, but the audio result it were not improved, but opposite... I was quite surprised. So I had to let the Crystek as it were designed to output... So, I was asking you here about the reasons of outputting square/sine...
At least a little "action" in this thread is not so bead...
Yes, its right that this is an Oppo thread regarding its including Sabre DAC, and other mods. It may looks like a clock discussion it may be out of the thread subject, but I think is not so bead to analyse a little bit in depth some aspects involved in these Oppo player mods, DAC and clock system improvements, and so on.
In fact my digression about square/sine clock is caused by the wish to find the best way to transport this clock signal into the player, best way to interface it to the receiver circuits, etc.
BTW, Crystek oscillators are quite famous for its quality and good figures. These oscillators output square signal. This signal looks quite bad when about overshoots/undershoots, and of course it looks very bead on FFT, comparing with a standard oscillator sine output...
Even more, I tried to filter out a square Crystek 975 output to get rid of the lot of harmonics, terminated as it should, and so on. I got a better clock signal, but the audio result it were not improved, but opposite... I was quite surprised. So I had to let the Crystek as it were designed to output... So, I was asking you here about the reasons of outputting square/sine...
At least a little "action" in this thread is not so bead...
Last edited:
But the fact that Sabre DACs can be used up to (and some like you have tried even higher) 100MHz clock speeds, then it deserves to be discussed on this thread as it means that we can use clocks on this Oppo that otherwise cannot be used with other players or stand-alone DACs using more conventional Burr-Brown DAC etc. Hence trying non-audio SAW clocks.
Cheers, Joe
PS: Shortly I will be bringing up a discussion re post-DAC filtering for the Sabre DAC, and one that applies to delta-sigma DACs in general (which is 99% plus of modern DAC chips) and the fact that they may need to be damped by a single dominant 1st order pole. I will give an example. But just a bit busy right now. But could be a lively discussion?
In my opinion it will always come something positive out of discussions... But not that kind of "discussions" some few members here are use to... and we have seen before...
BTW, is very right that clock quality is very important for the quality of the audio signals out of a DAC system. But not only this... The same important (I will say huge important) is also the differential processing of the signal out of the DAC chip (post DAC analogue processing).
I just realized in the last time that the simple passive components (resistors) which is to be used in these differential (post DAC) circuits have a huge impact in sound fidelity and for the sound scene definition and resolution. I mean the quality of these passive components... But I will also come back to this subject later on, after I will have some more confirmed facts and informations...
So, let`s discuss! This is at least the meaning of a forum...
BTW, is very right that clock quality is very important for the quality of the audio signals out of a DAC system. But not only this... The same important (I will say huge important) is also the differential processing of the signal out of the DAC chip (post DAC analogue processing).
I just realized in the last time that the simple passive components (resistors) which is to be used in these differential (post DAC) circuits have a huge impact in sound fidelity and for the sound scene definition and resolution. I mean the quality of these passive components... But I will also come back to this subject later on, after I will have some more confirmed facts and informations...
So, let`s discuss! This is at least the meaning of a forum...
Last edited:
Ayre Acoustics; opening up one of their "Universal A/V Engine" players (DX-5?) should give us some useful information.
Lexicon?
And what more can be done to the modified tube 105 from ModWright, with separate power supply (PS 9.1) and top best grade upgrade tubes?
...Roughly $4,000-4,500 US.
Lexicon?
And what more can be done to the modified tube 105 from ModWright, with separate power supply (PS 9.1) and top best grade upgrade tubes?
...Roughly $4,000-4,500 US.
Has anyone with a SABRE DAC experimented with LC filters after the DAC? I ask because I've played a lot with them after my NOS DACs and they're so much better than any other option that I can't go back to RC or not having a filter. From what I gather the 9018 is going to produce lots more RF out than a NOS DAC so would be an even better candidate for an LC filter than say a TDA1545.
Right! Is a god idea about LC filters...
I`m personally in a "not having filter" era now... Apparently Sabre produce enough RF, but comparing with another DAC types it seems to me that this RF is a designed spread spectrum one. At least is few hundred mV after post DAC processing (for many volts of useful audio signal) in a so high spectre at never it will be amplified further, and disturbing of audio spectre is very unsure... At least it can the noise level be reduced at few teens mV with very simple means. Is my now thinking...
LC filters is not a very common solution. So, one may give it a try... You did actually.
Zapfilter 2 as analog output stage
Wouldn't one "easy" solution be to use a Zapfilter 2 as analog output stage? L C Audio Technology / ZAPfilter 2
Pick the signal after the DAC chip and connect it to the Zapfilter 2 and then directly to the output sockets.
I have used the Zapfilter 2 in CD players with very good result.
I'm not sure which kind off output the ESS9018 have, but I think it is balanced voltage... Does anyone know this?
Comments, ideas?
Wouldn't one "easy" solution be to use a Zapfilter 2 as analog output stage? L C Audio Technology / ZAPfilter 2
Pick the signal after the DAC chip and connect it to the Zapfilter 2 and then directly to the output sockets.
I have used the Zapfilter 2 in CD players with very good result.
I'm not sure which kind off output the ESS9018 have, but I think it is balanced voltage... Does anyone know this?
Comments, ideas?
Contact Joe Rasmussen. He has a new trick up his sleeve. I did the "experiment" on a friend's BII and he loves the results.
Cheers.
Here is something somewhat controversial to try with the 9018:
Across the two phases on the output of the ES9018, put a single cap and adjust its value to be around -1.3dB to no more than -1.5dB down at 20KHz relative to 1KHz.
This makes for a dominant single-pole filter, and this filter cannot ring. This trick seems to work with all delta-sigma DACs (which almost all modern DACs are) and not just the Sabre DAC.
I have done quite a few players for clients and they have heard both before and after - and nobody has asked it to be undone.
I believe that this also has to do with side-effects from noise shaping due to single-bit behaviour. As usual, when we find ways to make digital sound better, it comes down to finding better ways to deal with noise. And, let's face it, single-bit delta-sigma DACs are noisy beasts.
Let me give you an interesting example/analogy: Most microphones used in recording studios are Condensor Mics which is to say they are capacitor types and essentially modulated capacitance - but that will also tell you that they must also be a low-pass filter. I am no expert on microphones, but it seems that some tricks are required to get extended frequency response and they show up when comparing them to another classic microphone type:
Ribbon Mics. These are potentially very much more extended and clean up to a much higher frequency.
Yet, when compared to Condensor mics, many will say they sound rolled off, when in fact they are not. Indeed, we listen and get used to Condensor mics and they become the reference we are used to, and then Ribbons sounds a little dull. Until you continue to listen and then, as your ear adjusts, that the Ribbon is telling you the truth.
Condensor mics have a sense of excess energy which is not controlled - certainly sounds like that when the best Ribbons are there to compare with. I would take a Ribbon over Condensors when given the choice. Indeed I know some guys who will NOT use Condensors at all.
I think that delta-sigma DACs and with their attendant noise shaping, not only have a perceived excess energy like Condensors, and a single-pole non-ringing filter can be heard to dampen the DAC.
But the real pay-off is not just the top end, the real gain seems to be in the midrange and some have also reported improvement in the bass - and one of them is a bass player. But to me and others, the best is a more expansive midrange, the sound is more fleshed out, more tonally and dimensional solid - also easier on the ear.
Also, once you get used to it, you may also perceive an increase in the sense of speed (I heard a fast transient on a well-known to me recording, that I had not really heard before and now got my attention) and what I can hear as better damping - and as somebody who has done a lot of work with damping materials and damping techniques (which was what gave me the idea here), that is exactly what I am hearing. Better damping and better control. It is also cleaner sounding.
It also works with voltage DACs as well as current DACs. With voltage DACs, it may be a good technique to add some small value series resistors in series with the two phases and then a cap across - this will give the cap a chance to bite and cause the single-pole filter to work effectively.
I have even used 1:1 transformers on voltage DACs and they require a Zobel network on the secondary, that Zobel can be adjusted to give the required effect as well without adding series resistance (the series resistance is in the transformer itself).
The Sabre DACs and ES918 have a well defined output impedance, so simply add a single cap (one per channel of course) and adjust to be just a bit more than -1.3dB at 20KHz.
You may need sine wave files for 1KHz and 20KHz that can be put on a USB stick or burnt to a CD.
If anybody has a place to park these files, I can supply them.
Using a scope, adjust for full scale and then at 20KHz be down by 0.85 of full scale. That will give you -1.4dB and do the job.
This is fully reversible - so no reason not to try it.
Cheers, Joe R.
.
It could be nice to see few snapshots of how it looks the noise with/without the cap between the phases.
Else I will not comment further Joe`s trick, because I did not try it yet... I anyway think it is worth to try it.
But I feel the need to comment a little bit more in the same direction. I will also repeat my position against using caps in the (audio) signal path.
In my opinion, this RF noises generated in an internal DAC processing is a false/artificial problem. I know well that when one see on a scope that quite high level HF noises, it may think that this is bad for the outputted audio, it is not acceptable, have to be filtered out, and so on. I did actually the same until I have tried the audio ignoring that noise. My surprise was not a little one, I may say...
There is out there big investments/efforts to find best and quite sophisticated solutions to get rid of that noise.
Well, when is about efforts to minimise the RF noises inside a DAC processor (like the ES9018 case), then I find this task as justified. This it have to do with a accurate/best way of functioning for the system. But when is about to filter out that noises in a post DAC processing audio signal, then I strongly disagree. First of all because doing so it will have a negative impact for the audio signal it self. This I have experienced it very well...
We talk here about a range of hundreds Khz to Mhz noises of maybe few, or teens mVpp which is outputted from the DAC chip in the same time with a useful audio differential current signal, which further it may generate few Vpp of an audio signal on I/V or or transformer. Come on, what is wrong with that?
This RF spectre is far away from the most of the highest audio harmonics. This RF noise is very low level. This RF noise it will never be possible to be amplified further by an audio amplifier, even though that amplifier it will have a very large broadband. For not say more about the very simple fact that such noises will never be heard by a human ear...
What about to concentrate us to get the best and the most "natural"/unaltered usefully audio signal out of a DAC, with the largest as possible spectre in the human audible spectre?
No mater where one put that cap/caps in the audio signal path, this it will lead to a alteration of the audio signal (its harmonics).
Any imperfection in the differential signal processing is altering the outputted audio signal.
Any possible noises in the power system it will leads to a alteration of the audio signal. Here I can agree that a HF/RF noises in a power system it have a different impact for the usefully signal than the RF DAC noise does. We can discuss about the reasons with another occasion...
Now about some facts (undocumented for the moment, but my intention is to be done in the near future):
I have eliminated completely any caps in the audio path. Well, I have to use some few pF on the final opamp (recommended by chip`s datasheet), to prevent ringing/oscillations, also in a Rf spectre. I think is just not really necessary to have some more RF noise out on final opamp, than it already come from the DAC chip, and this just on the RCAs connectors...
I use fully differential OPS 1632, which it gave me the best results ever, after lot of experimenting with all kind of I/V designs/chips.
I have used perfectly match for the all resistors involved in the differential circuit, and further processing for the two involved phases. Here there is no way to use 1%, 0,1%, or 0,05% tolerance resistors in this areas. Just perfect match down to mOhm. There is not so much clue to have so precise match for these components, when the thermal stability it may be quite bad... So, using 10 or even lower ppm thermal stability components is a must too. Well, the 1 ppm or lower of such it make the resistors huge expensive. Specially when one need quite many to be possible to match two pair or many for the post analogue processing circuits... But maybe once one have found the right value it may use such expensive components, for the best results ever.
I use also lowest noise and best regulator chips for the power system, with perfect match for the +/- rails on both I/V and final stages.
And at least but very important also, very large capacities for decoupling for all stages involved in audio signal processing, including the DAC chip (the shortest possible connections is a very must too).
Using a perfect match for the components in OPA1632 differential stage it give me the best channels separations and a sound stage with never heard before resolution, even for the high compressed formats, as mp3, or enough bad recordings.
Using large decoupling capacities give me the never heard before low end (bass) deepness and resolution. Even never thinking it were possible to get such low, powerful and good defined bass out of my amplifier/speaker system... I can say that using large capacity decoupling is the most "natural" and reasonable way to get such good definition in the very low end spectre out of an audio system, very high above to the wide spread practice to use caps in the signal path to expand, filter out, or alternate the usefully audio signal...
There is a very simple logic which it explain this fact: to get the best expand in the lowest end of audio spectre, the circuits need more (instantaneous) energy. Where it to be taken from this energy, when the power is very well regulated, and the response of the regulator is enough bad, because it were made like this for the purpose it were designed? The only instantaneous power source is in such case is the decoupling cap. This it will provide the needed energy to the audio circuit to respond precisely to the low end frequencies/impulses. So, one have to use such large decoupling caps, to get the best results, and that broadband which it may go very, very low in audio spectre. There is here not only to get an accurate reproduction of the very low audible spectre, but to get this spectre out of the processing system with quite high energy.
When one face naturally a low spectre sound, one not only hear it, but it feel it with the whole body. This it have to happen too when one reproduce that sound in a HIFI sound system... The human been can not hear well very low frequencies, but can feel very well such low spectre vibrations. When one blow in a trumpet near to a listener, that listener it may feel a kind of pain in the ear... When a big drum is hit it near somebody, that somebody it feel well the vibrations with the whole body. This it have to come out too from a high fidelity audio system!
I have got all these what it were described here out of a (today quite primitive) Burr Brown DAC chip, with that huge RF noises on its output, unfiltered, and undisturbed for the very high fidelity audio outputted signal... I can not wait to reproduce all these using a ES9018...
What for thinking more to that RF noises, when one may have the best out of a DAC chip without filtering, but by other more simple means?
Else I will not comment further Joe`s trick, because I did not try it yet... I anyway think it is worth to try it.
But I feel the need to comment a little bit more in the same direction. I will also repeat my position against using caps in the (audio) signal path.
In my opinion, this RF noises generated in an internal DAC processing is a false/artificial problem. I know well that when one see on a scope that quite high level HF noises, it may think that this is bad for the outputted audio, it is not acceptable, have to be filtered out, and so on. I did actually the same until I have tried the audio ignoring that noise. My surprise was not a little one, I may say...
There is out there big investments/efforts to find best and quite sophisticated solutions to get rid of that noise.
Well, when is about efforts to minimise the RF noises inside a DAC processor (like the ES9018 case), then I find this task as justified. This it have to do with a accurate/best way of functioning for the system. But when is about to filter out that noises in a post DAC processing audio signal, then I strongly disagree. First of all because doing so it will have a negative impact for the audio signal it self. This I have experienced it very well...
We talk here about a range of hundreds Khz to Mhz noises of maybe few, or teens mVpp which is outputted from the DAC chip in the same time with a useful audio differential current signal, which further it may generate few Vpp of an audio signal on I/V or or transformer. Come on, what is wrong with that?
This RF spectre is far away from the most of the highest audio harmonics. This RF noise is very low level. This RF noise it will never be possible to be amplified further by an audio amplifier, even though that amplifier it will have a very large broadband. For not say more about the very simple fact that such noises will never be heard by a human ear...
What about to concentrate us to get the best and the most "natural"/unaltered usefully audio signal out of a DAC, with the largest as possible spectre in the human audible spectre?
No mater where one put that cap/caps in the audio signal path, this it will lead to a alteration of the audio signal (its harmonics).
Any imperfection in the differential signal processing is altering the outputted audio signal.
Any possible noises in the power system it will leads to a alteration of the audio signal. Here I can agree that a HF/RF noises in a power system it have a different impact for the usefully signal than the RF DAC noise does. We can discuss about the reasons with another occasion...
Now about some facts (undocumented for the moment, but my intention is to be done in the near future):
I have eliminated completely any caps in the audio path. Well, I have to use some few pF on the final opamp (recommended by chip`s datasheet), to prevent ringing/oscillations, also in a Rf spectre. I think is just not really necessary to have some more RF noise out on final opamp, than it already come from the DAC chip, and this just on the RCAs connectors...
I use fully differential OPS 1632, which it gave me the best results ever, after lot of experimenting with all kind of I/V designs/chips.
I have used perfectly match for the all resistors involved in the differential circuit, and further processing for the two involved phases. Here there is no way to use 1%, 0,1%, or 0,05% tolerance resistors in this areas. Just perfect match down to mOhm. There is not so much clue to have so precise match for these components, when the thermal stability it may be quite bad... So, using 10 or even lower ppm thermal stability components is a must too. Well, the 1 ppm or lower of such it make the resistors huge expensive. Specially when one need quite many to be possible to match two pair or many for the post analogue processing circuits... But maybe once one have found the right value it may use such expensive components, for the best results ever.
I use also lowest noise and best regulator chips for the power system, with perfect match for the +/- rails on both I/V and final stages.
And at least but very important also, very large capacities for decoupling for all stages involved in audio signal processing, including the DAC chip (the shortest possible connections is a very must too).
Using a perfect match for the components in OPA1632 differential stage it give me the best channels separations and a sound stage with never heard before resolution, even for the high compressed formats, as mp3, or enough bad recordings.
Using large decoupling capacities give me the never heard before low end (bass) deepness and resolution. Even never thinking it were possible to get such low, powerful and good defined bass out of my amplifier/speaker system... I can say that using large capacity decoupling is the most "natural" and reasonable way to get such good definition in the very low end spectre out of an audio system, very high above to the wide spread practice to use caps in the signal path to expand, filter out, or alternate the usefully audio signal...
There is a very simple logic which it explain this fact: to get the best expand in the lowest end of audio spectre, the circuits need more (instantaneous) energy. Where it to be taken from this energy, when the power is very well regulated, and the response of the regulator is enough bad, because it were made like this for the purpose it were designed? The only instantaneous power source is in such case is the decoupling cap. This it will provide the needed energy to the audio circuit to respond precisely to the low end frequencies/impulses. So, one have to use such large decoupling caps, to get the best results, and that broadband which it may go very, very low in audio spectre. There is here not only to get an accurate reproduction of the very low audible spectre, but to get this spectre out of the processing system with quite high energy.
When one face naturally a low spectre sound, one not only hear it, but it feel it with the whole body. This it have to happen too when one reproduce that sound in a HIFI sound system... The human been can not hear well very low frequencies, but can feel very well such low spectre vibrations. When one blow in a trumpet near to a listener, that listener it may feel a kind of pain in the ear... When a big drum is hit it near somebody, that somebody it feel well the vibrations with the whole body. This it have to come out too from a high fidelity audio system!
I have got all these what it were described here out of a (today quite primitive) Burr Brown DAC chip, with that huge RF noises on its output, unfiltered, and undisturbed for the very high fidelity audio outputted signal... I can not wait to reproduce all these using a ES9018...
What for thinking more to that RF noises, when one may have the best out of a DAC chip without filtering, but by other more simple means?
Last edited:
Removing output coupling capacitors and DAC bypass
Well, I have started with some modest modifications. Thank you all for the inspiration.
First I opened up the OPPO, and unhooked the multichannel board. This unit is dedicated for stereo, and I don't need the multichannel board, I left it in place, and taped the cables onto the top of the disc player. This had a small but positive effect on the sound. I would expect this to reduce the amount of noise on the power supplies and the amount of radiated RF in the chassis.
I wanted to hear the balanced outputs of the OPPO, and also benefit from RF and high-frequency isolation, so I installed a high quality Jensen JT-KB10-D 4:1 transformer into my amplifier (a counterpoint solid-one, which had holes for the XLR connectors already). I also included a small resistive divider to reduce the gain. This had the beneficial effect of reducing most of the occasional grain and hash in the upper frequencies, but also left a somewhat distant perspective (too smooth).
Then yesterday, I removed all six output coupling capacitors (RCA and XLR) in the OPPO, and installed four 10 µF ceramic bypass capacitors onto the power supply leads of the DAC.
Joe had suggested these Murata 10uF/25V SMD 1206 capacitors as part of his proposed improvements for OPPO back on page 65. He even included photos to show where they go. Thank you Joe.
These last two changes really had a dramatic and beneficial effect. The details are back, and the bass is tightened considerably. Musical detail and the musical line are more clearly delineated.
There's been some debate on this forum about whether to bypass these capacitors, or replace them with small film capacitors, but I've got to say that it is completely clear to me that one should simply remove them (as advocated by Coris and Ric). With small film capacitors one is still creating a large output impedance at low frequencies. Also, I don't see the need for them as the output has insignificant offset voltage. Previously, using my Sennheiser HD 650 headphones in the headphone jack, it was clear that something was damaging the sound from the RCA outputs, this was indeed the polarized output coupling capacitors (or perhaps also the lack of resistor matching in the unbalanced section).
Eric
Well, I have started with some modest modifications. Thank you all for the inspiration.
First I opened up the OPPO, and unhooked the multichannel board. This unit is dedicated for stereo, and I don't need the multichannel board, I left it in place, and taped the cables onto the top of the disc player. This had a small but positive effect on the sound. I would expect this to reduce the amount of noise on the power supplies and the amount of radiated RF in the chassis.
I wanted to hear the balanced outputs of the OPPO, and also benefit from RF and high-frequency isolation, so I installed a high quality Jensen JT-KB10-D 4:1 transformer into my amplifier (a counterpoint solid-one, which had holes for the XLR connectors already). I also included a small resistive divider to reduce the gain. This had the beneficial effect of reducing most of the occasional grain and hash in the upper frequencies, but also left a somewhat distant perspective (too smooth).
Then yesterday, I removed all six output coupling capacitors (RCA and XLR) in the OPPO, and installed four 10 µF ceramic bypass capacitors onto the power supply leads of the DAC.
Joe had suggested these Murata 10uF/25V SMD 1206 capacitors as part of his proposed improvements for OPPO back on page 65. He even included photos to show where they go. Thank you Joe.
These last two changes really had a dramatic and beneficial effect. The details are back, and the bass is tightened considerably. Musical detail and the musical line are more clearly delineated.
There's been some debate on this forum about whether to bypass these capacitors, or replace them with small film capacitors, but I've got to say that it is completely clear to me that one should simply remove them (as advocated by Coris and Ric). With small film capacitors one is still creating a large output impedance at low frequencies. Also, I don't see the need for them as the output has insignificant offset voltage. Previously, using my Sennheiser HD 650 headphones in the headphone jack, it was clear that something was damaging the sound from the RCA outputs, this was indeed the polarized output coupling capacitors (or perhaps also the lack of resistor matching in the unbalanced section).
Eric
Last edited: