Topping D50 III optical input bug with 44.1kHz files

[unpluggged]

Khadas TB.

How the TVs were connected? I mean, the KTB has no optical input...

 

[Sokel]

How the TVs were connected? I mean, the KTB has no optical input...

Coaxial of course,SPDIF doesn't care.

 

[VientoB]

Coaxial of course,SPDIF doesn't care.

You do tend to get different jitter profiles with them.

 

[Sokel]

You do tend to get different jitter profiles with them.

..and a little noise too.

 

[VientoB]

Posting an update as I seem to have possibly solved this issue by trying my other Chromecast Audio (I have two) with the D50 III. So far no drop outs with 44.1 kHz music. So it seems to be a problem with my original Chromecast Audio, but only with this DAC. Very strange that it's only with 44.1kHz music though.

 

[MaxwellsEq]

Posting an update as I seem to have possibly solved this issue by trying my other Chromecast Audio (I have two) with the D50 III. So far no drop outs with 44.1 kHz music. So it seems to be a problem with my original Chromecast Audio, but only with this DAC. Very strange that it's only with 44.1kHz music though.

Strange! Perhaps it's a poorly performing clock that's too marginal at 44kHz for the DAC to reliably extract a good clock feed.

 

[antcollinet]

Strange! Perhaps it's a poorly performing clock that's too marginal at 44kHz for the DAC to reliably extract a good clock feed.

I'm gonna bet it has a crystal at a multiple of 48kHz, and is trying to derive the 44.1 imperfectly from that. Whereas all the other most used sample rates (48, 96 etc can be derived perfectly.

 

[VientoB]

I'm gonna bet it has a crystal at a multiple of 48kHz, and is trying to derive the 44.1 imperfectly from that. Whereas all the other most used sample rates (48, 96 etc can be derived perfectly.

88.2 kHz was fine, though

 

[antcollinet]

88.2 kHz was fine, though

Probably rules that theory out then.

 

[MaxwellsEq]

OK at multiplying, poor at dividing? Sounds like a school report

 

[VientoB]

Update.

I got a new optical cable which seems to have solved the problem. The old cable still works with my SMSL DAC but was playing up with the Topping used with both Chromecast Audios.

Very unsure why only 44.1kHz is affected though.

 

[Roland68]

Update.

I got a new optical cable which seems to have solved the problem. The old cable still works with my SMSL DAC but was playing up with the Topping used with both Chromecast Audios.

Very unsure why only 44.1kHz is affected though.

This indicates poor quality or a defect. You would need a suitable analyzer for analysis, but it's not worth the effort.

Often, cheap optical fibers made of cheap plastic fibers (POF) are used instead of real glass fiber optic cables.
Polymer optical fibers, although much cheaper, are more complex and offer more sources of error during production. The transmission performance is generally poorer, and POF cables also have higher attenuation, but this shouldn't matter given the low transmission rates in the audio range.

Real glass fiber optics, on the other hand, are more expensive, stiffer, and mechanically more sensitive.

 

[Rja4000]

This indicates poor quality or a defect. You would need a suitable analyzer for analysis, but it's not worth the effort.

Often, cheap optical fibers made of cheap plastic fibers (POF) are used instead of real glass fiber optic cables.
Polymer optical fibers, although much cheaper, are more complex and offer more sources of error during production. The transmission performance is generally poorer, and POF cables also have higher attenuation, but this shouldn't matter given the low transmission rates in the audio range.

Real glass fiber optics, on the other hand, are more expensive, stiffer, and mechanically more sensitive.

TOSLink uses cheap plastic fiber.

 

[Roland68]

TOSLink uses cheap plastic fiber.

As a blanket statement, that's definitely wrong.
The truth would be that nowadays, almost exclusively cheap plastic fiber optic cables are used, often even cheaper than POF and of very simple processing (without any quality). These plastic fibers are usually just cut and not polished.

I still own several older fiber optic cables with genuine quartz glass fibers. Some of them even came with devices.
But there are still several manufacturers of Toslink cables with genuine quartz glass fibers, with prices starting at around €/$40 per meter.

There's absolutely nothing wrong with a genuine POF Toslink cable with a polished surface.

 

[VientoB]

As a blanket statement, that's definitely wrong.
The truth would be that nowadays, almost exclusively cheap plastic fiber optic cables are used, often even cheaper than POF and of very simple processing (without any quality). These plastic fibers are usually just cut and not polished.

I still own several older fiber optic cables with genuine quartz glass fibers. Some of them even came with devices.
But there are still several manufacturers of Toslink cables with genuine quartz glass fibers, with prices starting at around €/$40 per meter.

There's absolutely nothing wrong with a genuine POF Toslink cable with a polished surface.

Can you suggest a cable of known quality?

 

[Roland68]

Can you suggest a cable of known quality?

Honestly, I can't really recommend much at the moment.
Most manufacturers no longer offer audio cables because people would rather buy rubbish cables with gold-plated Toslink connectors and the cheapest unpolished plastic fibers than high-quality, polished POF or quartz glass fiber cables with plastic connectors.

Tripp Lite is supposed to be well-made, but unfortunately doesn't specify whether POF or something cheaper is used; it costs about €/$5-10 per meter.
Supra Cable is unfortunately very expensive.
You can look for Glass Toslink and Quartz Glass Toslink. The price for these shouldn't exceed €/$40-50 per meter.
High-quality, polished POF cables cost about €/$5-20 per meter.

Personally, I would buy my cables by the meter or pre-assembled from a fiber optic cable shop. Ideally, they also offer polishing.

 

[nanook]

I'm using very simple and cheap POF cables that are sold under the brand name "LogiLink" and I never had problems.
They are a bit stiff, but the connectors sit firmly and the surfaces do look OK.

Fibers:
Attenuation usually is not an issue except for very long fibers.
Multimode fibers like the extremely thick Toslink do add jitter (due to mode dispersion), but this is irrelevant at frequencies used for Toslink. This is why high-speed transmission links use monomode fibers.

The limiting factor for Toslink regarding speed is that it uses LEDs (not lasers) and these have considerable rise/fall times. These will cause so called "Inter-Symbol-Interference" (ISI) and thus jitter.
The performance of LEDs can vary quite a bit depending on technology.

Post #53 mentioned the cause that sounds most likely to me: clock outside specification on either sending or receiving side.

I once borrowed a DAC from a friend and it did only work without stuttering with 44.1 kHz with one specific source (my TV).
When using other SPDIF sources (among them a RME FF UC) it did not manage to achieve a stable lock.

 

[nanook]

Looking at the far-field of the ca. 3 Euro LogiLink fiber tells me that the surface is good enough.
The spot is so round and on the axis.

Edit: A multimode fiber maintains the mode distribution. I.e. the mode distribution at the output depends on the illumination at the source side and thus the spot may look different for different Toslink transmitters.

 

[VientoB]

I'm using very simple and cheap POF cables that are sold under the brand name "LogiLink" and I never had problems.
They are a bit stiff, but the connectors sit firmly and the surfaces do look OK.

Fibers:
Attenuation usually is not an issue except for very long fibers.
Multimode fibers like the extremely thick Toslink do add jitter (due to mode dispersion), but this is irrelevant at frequencies used for Toslink. This is why high-speed transmission links use monomode fibers.

The limiting factor for Toslink regarding speed is that it uses LEDs (not lasers) and these have considerable rise/fall times. These will cause so called "Inter-Symbol-Interference" (ISI) and thus jitter.
The performance of LEDs can vary quite a bit depending on technology.

Post #53 mentioned the cause that sounds most likely to me: clock outside specification on either sending or receiving side.

I once borrowed a DAC from a friend and it did only work without stuttering with 44.1 kHz with one specific source (my TV).
When using other SPDIF sources (among them a RME FF UC) it did not manage to achieve a stable lock.

Given that my setup never had a problem using a Chord Mojo and SMSL DAC, would that not point to the Topping being the slightly dodgy item? Worth noting the problem has been fixed by changing the cable.

 

[nanook]

If using a different optical cable improves the situation, this absolutely sorts out the suspect of "too much deviation of clock frequencies".
This behaviour does rather point to a broken optical cable or dirt on one of the fiber end surfaces.
Of course the Toslink transmitter or receiver can be faulty (or dirty) as well, such that the intensity of the light is marginal and a small difference between cables makes a difference.

Usually you can retract the fiber connector by a couple of mm on the receiver side when you maintain pointing onto the receiver on axis. At some point the protection lid will close and this will shut down the transmission of course.
Retracting the fiber a couple of mm will reduce the amount of light shining onto the receiver photodiode. This way you can check if there's some margin with respect to intensity.

Edit: What I basically want to say with my posts above is, that if you invest money in a better fiber, this may just turn the optical transmission from "marginally not working" into a "marginally working" one.