An Edifying Journey ...

[fas42]

Was thinking of trying one thing; but changed my mind to another - a very bulky, but soft damping arrangement. One alternative to this is to make the whole cable very stiff, which will require a different strategy. Fairly straightforward to add the first, bulky material, in a very rough way ... first signs are good; no obvious negatives to the SQ - some very 'demanding' CDs are working well. The real test will be after the 3 CD mark, with the volume pushed right up - will the integrity be good enough, or not ...

[Confused]

What do you mean exactly by a "damping arrangement"? To be clear, I understand the meaning of the words, but what is not clear to me is what you are specifically referring to in terms of the detail of the actual implementation.

 

A photograph would be great, a picture paints a thousand words and all that. Although I know you do not like posting pictures, so maybe a few more words offering a clearer description might help.

 

And why don't you try the coax S/PDIF connection? If nothing else it would be a useful test to see how it performs subjectively in comparison to Toslink.

[fas42]

9 minutes ago, Confused said:

What do you mean exactly by a "damping arrangement"? To be clear, I understand the meaning of the words, but what is not clear to me is what you are specifically referring to in terms of the detail of the actual implementation.

 

A photograph would be great, a picture paints a thousand words and all that. Although I know you do not like posting pictures, so maybe a few more words offering a clearer description might help.

 

And why don't you try the coax S/PDIF connection? If nothing else it would be a useful test to see how it performs subjectively in comparison to Toslink.

 

Okay, the Toslink cable, being the normal cheap plastic variety is thin, and very flexible. So if it is attached to a socket in a decently stable case at either end, and you give the cable, hanging in free air between, a flick with the edge of your finger, in the middle, it will bounce around quite madly - which is a no-no, I've discovered. The ideal would be to get it to the state that if you tap it, it will move and go back to its rest position in a single oscillation. That is, it's critically damped

 

This would be hard to do. So, I'm experimenting with binding a damping material, like viscoelastic foam, around it; in a sleeve arrangement - and seeing how far I have to go with such, to the point where it makes no difference adding more. I'm also using some foam to act as a bracket half away along; which reduces the potential to wobble, even more.

 

Yes, trying S/PDIF makes sense. But the way I work at things, if something shows high promise then I pursue that single angle as far as I can, trying to squeeze as much out of it as possible ... trying something else midstream just complicates everything - for me, .

 

So far, the optical is a winner - others have also said this; which encourages me to maintain this course, . And to this very second, the latest version of this tweak is hanging in nicely ...

[Confused]

20 minutes ago, fas42 said:

Yes, trying S/PDIF makes sense. But the way I work at things, if something shows high promise then I pursue that single angle as far as I can, trying to squeeze as much out of it as possible ... trying something else midstream just complicates everything - for me, .

 

So far, the optical is a winner - others have also said this; which encourages me to maintain this course, . And to this very second, the latest version of this tweak is hanging in nicely ...

OK, but you will also find plenty of people that state that coax trumps Toslink.

 

I would have thought that after two years of experimenting it would at least be worth trying?

 

Personally I am happy to try all sorts of things of things. (within reason) Sometimes I try things that I suspect might offer poor results (assuming it is easy enough to try), but then the actual results surprise me. 

 

In this case, there is good reason to at least speculate that coax might yield good results, and it is something that would be very easy to try. Who knows what you might learn? Good or bad, it would it least be informative.

[fas42]

Awww ... you're really twisting my arm now, .

 

Ookaaaay ... just for you, I'll whip up a link ... in a couple of days time, when I feel I have a good handle on the latest optical tweak ... but if it all blows up, I know who to blame !!!

 

My god, I haven't even tried one of the other devices in the house that do optical out, yet ...

[Confused]

26 minutes ago, fas42 said:

Awww ... you're really twisting my arm now, .

 

Ookaaaay ... just for you, I'll whip up a link ... in a couple of days time, when I feel I have a good handle on the latest optical tweak ... but if it all blows up, I know who to blame !!!

 

My god, I haven't even tried one of the other devices in the house that do optical out, yet ...

Digital radio? TV maybe? Something else?

[fas42]

3 hours ago, Confused said:

Digital radio? TV maybe? Something else?

 

At least two CD players - one of them could have a better clock, and optical out circuitry than the cheapo Samsung DVD player; which may help things. I've already mentioned in earlier posts here that I'm thinking that at some stage I would try to organise an input switching box, with clean reclocking of the optical out; would be interesting to make that waveform adjustable, to see how the Edifiers react.

[fas42]

Still playing with the optical link - getting it exactly right is critical; and now trying a combination of stiffening, and damping running along the section that is at the speaker end ... too limp and it still wobbles too much; too stiff, and the vibration of the speaker cabinet will be transferred to the internals of the cable. I'm playing this by ear, , and, as would one fairly suspect, the closer to the speaker cabinet you go applying this stabilising procedure, the more it matters ... winding up some pop to bar room levels makes this obvious; get it slightly wrong and the SQ goes sludgy.

 

These speakers have no trouble going to sound levels, cleanly, where it's impossible to hear your own voice - but every anomaly becomes "clear as a bell" when you do this ...

[fas42]

More feedback ... the receiver end of the Toslink plastic is extremely sensitive; the slightest variation in how it's stiffened, or damped quite strongly affects SQ - this won't be so obvious at normal listening levels, with music that is not constantly driving - classical is safe(!) - but bring things up to dance level situations, and there are repercussions if not as good as it can be.

 

So, more exploration here - with a bit of luck, I may be able to constrain the lead just over the first inches or so, to get optimum SQ; and not have to worry further along the cable.

[fas42]

Well, tried constraining at the speaker end - basically, applied a thick sleeve of firm, slightly viscoelastic goo over several inches; which definitely helped ... was it enough? Ummm, no - and some people are going to love this - used a small, stiff square of cardboard to brace the middle of the free hanging section of the Toslink; and yes, this is lifted the quality significantly.

 

So, it's not tiny parts of millimetres that matter; rather, it's how I'm stopping that end of the cable 'wobbling' in a way that degrades the waveform on the digital link, as far as the receiving circuitry is concerned. As I see it at the moment, it's like having a rope securely mounted to the ground at one end, and the other end tied to some object which is vibrating in a fairly random manner - what's required is that the rope under no circumstances flops around unnecessarily; it's an exercise in the art of motion damping.

[fas42]

Still working on stopping that rope flopping ... . Minor variations tried - still not satisfied I've got a good handle on how to deal with this ... last experiment was adding some mass loading on that cable, right at the point where it plugged into the speaker - this most certainly helped. At its best, at the moment, the speakers are extremely close to full invisibility; but then it seems some quality is lost - it's gone backwards.

 

Yesterday emphasised something I was well aware of - and is the trap one can fall into very easily; there are at least two significant factors in play determining the SQ at a specific moment; and focusing on getting one of them optimised may lead to wrong conclusions, because another factor is having too much influence - the DVD player is old, and a cheapie; sometimes it loads the CD poorly, won't be spinning it cleanly, but rather with a wobble. Power supplies are working harder, impacting the optical out circuitry; SQ is down.

 

I thought the mass loading on the Toslink was causing a loss; removed it, but still no better. Cut the power to the DVD player, in case some static buildup was the problem; but no real gain from that. Finally, it dawned on me - was it poor seating, for the CD? Opened the drawer, closed it again - yep, the track was back to where it should be.

 

Which indicates that optimum stabilising of the optical cable is of the same order of importance as the DVD player spinning the CD true - both not at their best effectively adds jitter to the data link. Symptoms? Loss of the "magic", gray, smaller sound; obvious distortion when listening close to a driver.

[fas42]

Yesterday was fiddling with using stiffer damping around the optical cable - it was a swings and roundabouts thing, I was not getting clear feedback as to what was the most effective approach with the damping. Also, noted how much the speaker Toslink socket was vibrating when the music hit the right notes - there was my problem, easily felt. Two approaches, add some loading to the actual speaker back in the critical area to attenuate that vibration as much as possible, which could also help the active circuitry on the other side; or, optimising the linking cable to be as little affected as possible. Or both.

 

Thinking about it, I decided to alter the path of the cable at the speaker end, so that any curves were of maximum radius. This meant that movement would distort the the cable geometry by the least amount; which most likely is the key factor. Trying it now, so far the signs are positive ...

[fas42]

Up and down with this Toslink ...

 

Wanna know how important digital cables can be? Try optimising a cheap plastic link of it - talk about a tone control ! Actually, it's not that; rather, the physical stability of it impacts the ability of the receiver and following circuitry to get the best clocking from it - and what seems to be a smart move one moment then turns out to be a negative ... latest round is focusing on adding maximum volume and weight to that "thick sleeve of firm, slightly viscoelastic goo"; which seems to only drive the SQ in a positive direction - ie. stiff is good!

[fas42]

The process of optimising a system, any system, is locating the weak links of the chain, and experimenting if necessary to find a best value for money and effort method of improving the integrity of that link. First time I've played with optical, and it's very clear that if plugged into a vibrating socket that maximum care has to be taken into getting the entry area, just right. Which means every time an optical cable is plugged into the back of an active speaker. Of course, if the engineering is done well enough then such won't be necessary - what guarantees are made by your manufacturer that their design is, good enough? 

 

So far, if I get the stabilising in the right zone then the SQ is so close to being spot on that nearly every box is ticked. Not 100% invisible drivers, but I have to get extremely close to the front of one side to clearly hear that speaker working. Sweet, sweet strings, huge soundstages, "no bad recordings" , almost zero signature of the setup. But let the last 6 inches or so of the Toslink to less well sorted, and all that goodness fades quite a bit ...

 

So, there will be a range of dealing with the cable entry that will give decent SQ - will there be a critical, 'perfect' organising of it? Very unlikely; something like a smooth hump of peak tweaking is likely to be more the situation - what one can do is push to either side of how things are done, until the fall off is then quite noticeable; and then aim to go for the middle setting of the various adjustments.

[Qhwoeprktiyns]

The issues you are finding with Toslink are well known. A relatively simple alternative is ECDesigns' ElectroTOS connection: https://www.ecdesigns.nl/en/shop/electrotos-cable

 

It connects to the SPDIF RCA out of your source and carries an electrical signal that activates the LED placed at the end of the cable. In that way, you avoid issues with Toslink cables, but you still get the benefit of electrical isolation from the source.

 

To make it work with the SPDIF out of a CD player, you'll have to slightly change the connection of the signal to the SPDIF connector (out), but it is pretty easy to do if you can solder. For details, you should contact ECD.

[fas42]

Thanks for making me aware of that; it's always good to know what people have tried, as a way of circumventing limitations in what is normally used. I don't think I'll be in a rush to acquire one, because at the moment I want to see how far a low cost method can get one, just using the standard protocol.

 

That said, and prior to you posting that, I had a Wha!! moment ... yesterday, I was trying quite a few variations on how the cable was dressed; and it was only having a slight effect on SQ. Then, I decided to improve how a length of lead which was being used to stiffen the run of the cable could be prone to static ... static?!! How on earth could static impact an optical link?! But I don't not try things simply because it doesn't make sense, on first glance - and, you guessed it: big up-jump in quality! What the!! I spent some time considering what might be happening - and then, finally, I realised what had occurred: after adjusting that stiffening lead, and not noticed by me, the run of Toslink had sprung free at one point, and was now continuously curved along that crucial region ... ah, hahh!! Was the fact that at no point was there anything like a relatively straight bit of that cable the key difference??

 

Ran out of time to investigate further - the next step, obviously, is to confirm that the bending makes the difference; and if so, what is the optimum way to organise this ...

[fas42]

Well, so far the bending hypothesis is looking good - did a fairly rough and ready reorganising of the Toslink, so that smooth curving of the cable happened over its full length; what I was focused on previously, trying to stabilise any vibrating, I barely paid any attention to. And the SQ held up, to a very good standard - only one recording, one of the more 'difficult' classic albums, made me pause, and think that more could be done.

 

Is it because the curving of the cable attenuates any vibration in just the right way, or it changes the patterns of the internal reflections of the light beam? Or something else?

 

Anyway, will focus on varying this aspect now, to see how refining of this tweak alters things ...

[fas42]

That 'difficult' recording managed to put into focus what is now, potentially, more of an issue - but is easily dealt with. There are actually 2 CDs in that release, and when I put on the second one, it was fine! So, what was going on? Well, something that has always been there - the DVD player is old, and a cheapy; when the disk is mounted by the load mechanism it doesn't always sit cleanly in alignment, and spin true - and the read laser struggles to track it. This is really obvious when the sound starts glitching and skipping; solution is unload, and reload ... bingo! Everything's back to normal ...

 

A lesser poor spinning just results in lower SQ; the power supplies in the player are working harder, and impact the optical circuitry output waveform. In the case of that CD 1 that didn't sound spot on, that was the cause - playing it the next day, it also was fine ...

 

So, an annoyance. But easily fixed, in the moment. Various ways to deal with this, long term; other items need to be sorted, first.

 

Still very pleased with latest optimisation; the speakers are so very, very close to fully disappearing that it will be only some minor adjustment somewhere that does the trick - not fussed about absolutely nailing it, because the sound now is good enough to make any recording put on work its magic, .

 

 

 

[fas42]

Quality is still going up and down - at first it seemed that the curving of the cable would be enough, without stabilising along its length. But it's riding right on the edge of being, "just right", and the slightest variation, somewhere, for whatever reason, is enough to dethrone the SQ ... I need to go very carefully with what I do, to properly isolate the contributing factors.

 

So, at the moment it appears that I need both curving, and careful stabilising, to optimise enough - other factors, like how recently I ran the setup, and how well the CD mounts, confounds what I hear - in the morning a particular CD will be 'singing'; later on, after some tiny adjustment, it's just not happening, for the same disk ...

 

Going back to max stabilising for next round; and see if I can get a positive result over a full day of trialing ...

[fas42]

And, lookin' good! The difficult classical album mentioned above, some Faure chamber music, sparkled; and everything followed also worked. The warm up cycle is still in play; starts well, then gets a bit too upfront in presentation; and settles after the third CD or so - what happens during that in between period is that more detail is emerging, but too much distortion is coming along for the ride - for  a particular CD, that would mean it was very pleasant to listen to, from cold, but the picture wouldn't be as big as it should be; after stabilising a soundstage as close to as big as it could ever get would be evident.

 

Played an album I normally rarely do,

 

 

And this was another instance where other factors figure - it sounded, not good! Very 2D, small. Which is why I spin it infrequently.

 

But I gave it another go, later today. Tried reloading once - nope! And then again - no! But the third time it came good; the mounting was this time true enough to give decent SQ. Which makes it a good test CD, for the current setup; it needs the digital link to be of high integrity, to overcome the tendency of the type of mastering on this one to cause lacklustre listening.

[fas42]

Okay, a status report.

 

As should be clear from recent posts, dealing with the vagaries of the optical waveform, which varies depending upon the mood of the DVD loading mechanism to get the seating right for the CD, and how optimal is the Toslink cable damping and dressing, shows how much of a bottleneck this now is - so, what's the best way of handling this? Well, with a bit of clear air for reflection, it really seems time to bite the bullet and organise a switching input and optical reclocking unit to be inserted just before the optical disappears into the Edifier box, as I have already mentioned doing in this thread. Best approach: have every possible input; that is, all the types of digital signal coming in, like USB and SPDIF coax as well as multiple opticals, buffer possibly, and reclock out as an optimal optical waveform. If done properly, everything from a multi $10,000 streamer to a barely working CD player with some sort of digital output should be handled with zero impact on SQ - that would be the goal.

 

Being me, this has to be done for cheap as chips money, . And to balance the very low cost of the active speakers. Therefore, I will check out what sort of el cheapo boxes are out there that have the core functionality, try and pick the best one to work on, pull it to bits, and then rejig the internals to get the job of delivering a good enough optical waveform done. Will aim to use the lowest cost method to get there; if the cost of modding gets out of hand then the point of the exercise has been lost.

 

 

[fas42]

So far, looking good. Extremely cheap switching units, with say 4 digital inputs, and sending out Toslink are there, and likewise very low cost USB to S/PDIF boxes. To round it out, will throw in Ethernet, and HDMI converters; aim is to handle everything.

 

To clean up the noise mess that comes with a particular input, probably buffering and regeneration is as good as anything. A technique that caught my eye years ago was the Digital Lens, by Genesis; made excellent sense, and even recent reports by those who have acquired second hand units make all the, ahem, right noises about the effectiveness. Recent products by PS Audio incorporate a modern version - McGowan came up with the idea in the first place - they use a FPGA to do the heavy lifting. Now, the trick is to replicate that functionality as cheaply as possible; some research to be done here. If it turns out that it's going to be too expensive, then another method will need to be considered.

 

As regards tweaking of current arrangement, the portion of the optical cable from where it plugs into the speaker to the point where it's stabilised by a relatively non-moving surface still shows itself to be critical. Wasn't happy, and wondered whether the mass of goo which was damping the cable at the very end was just too much; was working against what the newer curving of the cable gave me. And, yes, removing a major portion of the substance was the answer - the obvious question then is, what is exactly the right amount, and the shape of application to give optimum SQ? Well, I could spend ages trying every permutation ... and this is as good a reason as any as to why I should organise an optical signal cleaning mechanism as just described - to handle every contingency.

 

Running Greatest Hits of Queen at clubbing levels right now; got the house throbbing - this wasn't in the zone at all, prior to that 'minor' adjustment, of the cable damping ...

[fas42]

First look over, the parts needed to do the job of buffering and regenerating are cheap enough to make it viable; that is, available for $10 or less. Overall, I need an assembly of stuff to be an order of cost to match the Edifiers, to make the exercise of doing this meaningful; and this is looking positive, so far.

 

I can't this sort of design fiddling in my head and on paper alone any more; so now looking for a best fit for purpose software tool to do the job - LTSpice is not really designed for digital world exploration.

 

Again, the aim is to create a "digital preamp" - can take any sort of digital in, including ethernet, condition the signal, and spit out a digital waveform that has the right analogue qualities to make the Edifiers work at optimum SQ.

[fas42]

Where's the setup at now, in terms of how well it projects? Well, just this morning, starting from cold, after a couple of days of being switched off, first CD,

 

 

After about 10 minutes or so, no problems presenting the massive space of the church it was recorded in, back in 1967; running it at something like live levels, the power and grandeur of massed instruments and voices is overwhelming ... you know, like being there, .

[fas42]

Have been looking at the hardware, and software to sort out the "waveform cleaner"; getting closer to having the right tools to do this - most likely will use a low cost FPGA board to do the grunt work; but this is the first time I've ever played with this stuff, so working out how to do this in the laziest way possible, .

 

So, overall concept is to have a source switching module at the front, which can take anything in a digital stream coming in - which is always assumed to be 'dirty' - use some memory buffering driven by the FPGA to 'rinse' whatever is being fed in; and then send what is being pointed to in that buffer by a signal from a 'clean room' output section to the latter, which will have an absolute minimum of parts. Here experiments in varying the qualities of the waveform to optimise the 'cleanness' can be tried relatively easily, to make the Edifiers or whatever else follows as happy as possible, . And, the cost of everything has to add up to being sensible with regard to what these speakers cost.

 

With regard to current setup status, it's in that slightly frustrating region where when the integrity is good enough then the SQ is essentially everything you can want. I have been in this situation so, so many times before with all the other rigs in the past - here, it only takes the slightest 'mistake', or accidental upsetting of a tweak to knock the standard off the peak; it just, sounds, "wrong"; and then you have to spend time working out where things have gone awry ... the ultimate aim is to build in a decent "safety margin"; meaning, a couple of things can be less than optimal but subjectively this is not audible. Tidying up the build will be huge part of solving this, but first I want to get enough answers to be comfortable that I have the right combination of ingredients to make that solution robust.