Cable Modems

[Xymox]

Is all this PS work in your current switch. I have a liner solution switch. Its good. I would like better. And I would like a better server. Again,.what I have is good

Yep... All in my current switch..

 

[Xymox]

Over the weekend I changed some settings on my Technicolor WiFi 6 gateway with Broadcom chip.
The 2.4GHz operates now on WiFi 4, the 5GHz on WiFi 5.
I like the result, greater musicality, more fun to listen to music.
There is a small penalty, the internet speed is reduced by about 20%, that's all.
Thanks @Xymox and @vert

Matt

Yep speed is really not that important for just doing things like a roon GUI. Good to use a dedicated device and then remove all the apps from it except the ones you need. Really dedicate the device just to music..

 

[Xymox]

Xymox
I love reading about your efforts but I have 2 questions (that I'm sort of embarrassed to ask). I obviously don''t know the extent of the noise reduction modifications that you have performed but based on what I see, it appears that there is an obvious addition of a large capacitor on the board that presumably contributes to the improved filtered noise performance you have shown so nicely between 5K and 10K. So my questions are:

1) If the improvements gained are so impressive from what appears to be a minimal additional of hardware, one has to wonder why the manufacturer doesn't employ these modest modifications themselves? Is it engineering competence? Cost? Other?

2) The bigger question is on the issue of presumed SQ improvement. We have seen/read many reports of seasoned users who have made great efforts (often at considerable expense) to use devices that provide "clean" power, but report their disappointment that the resultant sound was either unimproved or in some cases actually worse than using no power conditioning at all. Could the failure of some of these switch manufacturers not to employ what appears to be rather straight forward modifications such as you have performed be due to that fact they don't think that they really improve SQ after all is said and done?

Marty

There are a LOT of caps. The magic is on the bottom of the board mostly. VERY specific caps. Like monolithic ceramic and giant tantalums. These caps are REALLY low ESR at high frequencies. For this type of work, you dont want to use polypropylene ( WIMA ) or electrolytics. Most of these caps are right on the chip input pin or right on the output of the PWM regulator. Or both. There is one really ginormous electrolytic cap that is directly connected to the incoming linear supply. This acts as a power reservoir - not so much noise removal. It stablizes the voltage by providing local reserve right on the board. This helps with overall short term regulation. It smooths out the voltage variations to +- 100 uV under use. Overall the regulation is overkill. its 14.15Vdc +- .003 long term worst case. THEN there are 2 more regulation stages. A internal regulator on the switch then regulates down to 9.5V and then that feeds the regulators for the chips. So its 3 stages of regulation by the time it generates the ethernet signals. So its really precise and very very stable over time. Plus noise is GREATLY reduced, it needs to be very low to attain the precision regulation.

YES... The Voodoo factor. Bill @ GTT convinced me to make a switch. The AppleTV X was a easy win as I could measure huge improvements in HDMI and EVERYONE who has gotten one has seen big improvements. So the ATVX was easy.. When he suggested I do a switch, I spent months saying no. I complained that there was a huge amount of voodoo in the ethernet, server, switch market.. I had spent WAY too much time myself playing with switches of all the brands here and tons of ethernet cables. I was fully baffled as to what was going on. I measured and tested but could not find anything measurable that corresponded with what I heard. I could CLEARLY hear a ethernet cable, but could not explain why it could possibly sound different. Switches, servers, USB, it all sounded different. Things in one system did not react the same way in another system. I heard this myself in a number of systems I have handy. THIS was INSANE to me technically. Me make a switch ? Bahh,, way to unknown..

SO.. Bill finally got me to make a switch. I decided to just do EVERYTHING I could think of to do that would improve ethernet and switch chip, performance. I went at the switch tweaking every signal aspect of performance I could think of. Power supply - all the way to the chips. Better clock chips. RF noise abatement, grounding internally, vibration dampening and a ton of small things.. I sent the prototype off to Bill. Explaing that each person has become accoustomed to a "sound" of the ethernet/switch/server that they like and each system sounds different. Like speakers, some people like horns, some like electrostats, some like traditional drivers.. Its personal taste, same with all the ethernet stuff IMHO.. I told Bill this over and over... SO he gets it and tries it on a bunch of systems in his show rooms. Replacing various high end switch systems that include external clocks and linears. I won out on all of them. I also have a few clients who i did rooms for that I sent switches to, they also liked them. We then sent out 5 to some initial clients, they all liked them.. SO this was encouraging..

Will it be liked by everyone ? WHo knows. Is it better then EVERY switch out there in EVERY system, I doubt that. I still see the voodoo aspect clearly and I see the glass half full. I can produce a switch with crazy mods and extreme ethernet performance. Does it sound good in every system ? WHO KNOWS...

So far tho I am encouraged that my alchemy seems to be besting over the voodoo.

I have some clients with Taiko Extremes and they have the new card and "switch" which I think is more like a buffer. My switch takes SFP and so the plan is for these guys to do a direct AB simply unplugging the Taiko switch SFP modules and plugguing into my switch. I have the advantage of also having a NATed isolated clean output which is where Taiko is going with its "router" that it is working on. IMHO these should be one box.. NOW my unit is not 30lbs of copper. Its also not pretty. But it hopefully has it where it counts..

SO... I don't know how it performs for everyone yet. There is a backlog right now. I am getting linears next week and more will ship to clients. So in the months ahead once I get a bunch more out there, we shall see...

Looking thru all the switches out there, I just don't see anyone doing all the stuff I am doing on the switch. I have cracked open many audio switches and its really boring what they do to them. I never see measurements.

We shall see.. I am going to also dabble with a Intel NUC.

 

[Xymox]

I suspect router manufacturers don't believe routers can affect sound quality.

I find it hard to wrap my brain around it. With the sound bytes being fed into a buffer on the streaming device and the bytes are the same. Its hard to fathom how a ethernet cable, switch or router could make any difference at all. BUT.. I have clearly heard BIG differences. So.. There you have it. Its voodoo..

 

[Xymox]

I tried Aruba AP-11, and it does work without a controller.

But I couldn’t get the WiFi working when the unit was powered by my linear power supply.

Well what do you know.. A Aruba on amazon.. I do not have a lot of experence with that one... The IAP-225-US on ebay tho is a 3x3 and I know that works great. And its $35. I have been doing Aruba for 15 years and I have never seen one fail. So the used ones are pretty safe. I know the IAP-225-US works on a 12V 2A supply. I was playing with one today.. More on that in a minute..

 

[2ndLiner]

Ooooo.... Can you solder ?

No, but I can play a Second Line groove that’ll make you drop your iron and dance! We all have our talents

 

[Xymox]

I made a quick vid.. I show a IAP-225-US powered by a 12V linear. Works great.. Also I give a preview of what makes a Aruba awesome. FPS140 mil specs plus other specs that make it RF hardened.. WHo knew ? So I have now measured and these Arubas dont have any measurable RF noise that they emit. Its JUST 2.4/5 Ghz wifi. Unlike the Ubiquiti, Eero and TPLink which all have TONS of RF noise they emit...

The Aruba's have full mil-spec complaince which also include RF emission specs.. https://www.arubanetworks.com/assets/ds/DS_FEDCERT.pdf

I will do a better vid whowing all the popular access points for audio next week.

 

[Xymox]

No, but I can play a Second Line groove that’ll make you drop your iron and dance! We all have our talents

Oh god.. You would be very distracting to my getting these devices done ! You guys want me soldering, not dancing,, and trust me,, you dont want me dancing, hahaha...

 

[chopchopbin]

Oh god.. You would be very distracting to my getting these devices done ! You guys want me soldering, not dancing,, and trust me,, you dont want me dancing, hahaha...

Is there a link available for pricing and how to order the switch? Thanks

 

[Kingrex]

So, if I were to unplug my ethernet cable and play music off my harddrive, that is the best I should expect from my DAC server combo, correct. If I were to have my ethernet plugged in and compared Qobuz to a stored file and I really could not tell the difference, would that infer that a better switch will do nothing for me?

My system is a server with USB to DAC. The ethernet is in the server.

I guess another way to view it is, if the switch is reclockiing rhe signal and handing it to the server, then the server is not the major influence on noise and performance? No?

 

[marty]

There are a LOT of caps. The magic is on the bottom of the board mostly. VERY specific caps. Like monolithic ceramic and giant tantalums. These caps are REALLY low ESR at high frequencies. For this type of work, you dont want to use polypropylene ( WIMA ) or electrolytics. Most of these caps are right on the chip input pin or right on the output of the PWM regulator. Or both. There is one really ginormous electrolytic cap that is directly connected to the incoming linear supply. This acts as a power reservoir - not so much noise removal. It stablizes the voltage by providing local reserve right on the board. This helps with overall short term regulation. It smooths out the voltage variations to +- 100 uV under use. Overall the regulation is overkill. its 14.15Vdc +- .003 long term worst case. THEN there are 2 more regulation stages. A internal regulator on the switch then regulates down to 9.5V and then that feeds the regulators for the chips. So its 3 stages of regulation by the time it generates the ethernet signals. So its really precise and very very stable over time. Plus noise is GREATLY reduced, it needs to be very low to attain the precision regulation.

YES... The Voodoo factor. Bill @ GTT convinced me to make a switch. The AppleTV X was a easy win as I could measure huge improvements in HDMI and EVERYONE who has gotten one has seen big improvements. So the ATVX was easy.. When he suggested I do a switch, I spent months saying no. I complained that there was a huge amount of voodoo in the ethernet, server, switch market.. I had spent WAY too much time myself playing with switches of all the brands here and tons of ethernet cables. I was fully baffled as to what was going on. I measured and tested but could not find anything measurable that corresponded with what I heard. I could CLEARLY hear a ethernet cable, but could not explain why it could possibly sound different. Switches, servers, USB, it all sounded different. Things in one system did not react the same way in another system. I heard this myself in a number of systems I have handy. THIS was INSANE to me technically. Me make a switch ? Bahh,, way to unknown..

SO.. Bill finally got me to make a switch. I decided to just do EVERYTHING I could think of to do that would improve ethernet and switch chip, performance. I went at the switch tweaking every signal aspect of performance I could think of. Power supply - all the way to the chips. Better clock chips. RF noise abatement, grounding internally, vibration dampening and a ton of small things.. I sent the prototype off to Bill. Explaing that each person has become accoustomed to a "sound" of the ethernet/switch/server that they like and each system sounds different. Like speakers, some people like horns, some like electrostats, some like traditional drivers.. Its personal taste, same with all the ethernet stuff IMHO.. I told Bill this over and over... SO he gets it and tries it on a bunch of systems in his show rooms. Replacing various high end switch systems that include external clocks and linears. I won out on all of them. I also have a few clients who i did rooms for that I sent switches to, they also liked them. We then sent out 5 to some initial clients, they all liked them.. SO this was encouraging..

Will it be liked by everyone ? WHo knows. Is it better then EVERY switch out there in EVERY system, I doubt that. I still see the voodoo aspect clearly and I see the glass half full. I can produce a switch with crazy mods and extreme ethernet performance. Does it sound good in every system ? WHO KNOWS...

So far tho I am encouraged that my alchemy seems to be besting over the voodoo.

I have some clients with Taiko Extremes and they have the new card and "switch" which I think is more like a buffer. My switch takes SFP and so the plan is for these guys to do a direct AB simply unplugging the Taiko switch SFP modules and plugguing into my switch. I have the advantage of also having a NATed isolated clean output which is where Taiko is going with its "router" that it is working on. IMHO these should be one box.. NOW my unit is not 30lbs of copper. Its also not pretty. But it hopefully has it where it counts..

SO... I don't know how it performs for everyone yet. There is a backlog right now. I am getting linears next week and more will ship to clients. So in the months ahead once I get a bunch more out there, we shall see...

Looking thru all the switches out there, I just don't see anyone doing all the stuff I am doing on the switch. I have cracked open many audio switches and its really boring what they do to them. I never see measurements.

We shall see.. I am going to also dabble with a Intel NUC.

Xymox
That was a phenomenally detailed and much appreciated reply. I didn’t mean to understate your considerable efforts but to be fair, I couldn’t see the “action “ side of the boards you worked on. I’m sure your efforts will be productive. Remember that Taiko’s switches will go to Extreme owners first and then possibly to others so your contribution might benefit many listeners.
Marty

 

[Xymox]

Is there a link available for pricing and how to order the switch? Thanks

I need to put the price on the web site. My distributor is Bill @ gtt audio https://gttaudio.com/

There is a little bit of backlog. Right now expect like 3 weeks.

 

[Xymox]

I need to put the price on the web site. My distributor is Bill @ gtt audio https://gttaudio.com/

There is a little bit of backlog. Right now expect like 3 weeks.

There, I fixed the web site.

 

[Xymox]

So, if I were to unplug my ethernet cable and play music off my harddrive, that is the best I should expect from my DAC server combo, correct. If I were to have my ethernet plugged in and compared Qobuz to a stored file and I really could not tell the difference, would that infer that a better switch will do nothing for me?

My system is a server with USB to DAC. The ethernet is in the server.

I guess another way to view it is, if the switch is reclockiing rhe signal and handing it to the server, then the server is not the major influence on noise and performance? No?

"So, if I were to unplug my ethernet cable and play music off my harddrive, that is the best I should expect from my DAC server combo, correct."

But how to do that ? Gotta have a ipad to control it and that usally needs internet for graphics and stuff. Once this is plugged in then RF noise can get into your DAC/Server. And so the battle begins.

"If I were to have my ethernet plugged in and compared Qobuz to a stored file and I really could not tell the difference,"

IMHO.. While the files might be technically the same,,, the streaming format and its long path to you and things like content delievery networks all cause some kind of degradation. While I might be crazy, I think I can tell the difference with streaming files and playing off a local server, like a Taiko. WHY ??? I don't know. I cant answer this technically.

"I guess another way to view it is, if the switch is reclockiing the signal and handing it to the server, then the server is not the major influence on noise and performance? No?"

Its even more crazy then this.. The bytes, the music data, is loaded into a buffer in the DAC and this seriously "reclocks" anything coming in to it. If the bytes are the same, and its fully reclocked, then how can ANYTHING matter like ethernet cables, switches, server, router, modem.. NONE of it should matter AT ALL... BUT YET... I have done a bunch of blind A/Bs with people and everyone can hear things like ethernet cables. WHich makes zero technical sense to me... Once upon a time back at the beginning of digital audio,,, Digital was proclaimed as perfect reproduction because everything we could measure was near perfect.. BUT.. HUGE number of people hated it and said it sounded terrible VS even things like master tapes. Professional recording engineers said it was terrible.. BUT YET.. Everything we knew to measue was perfect... So a huge amount of serious study ensued and JITTER was discovered. Not only jitter but the spectral content of the jitter was all picked up by the brain.

We did not know to measure something.. So we did not catch it.. Now we know.. So I think something like that might be at play. There is something going on we just dont know to measure. Hard to imagine what it is, but, that must be the case. The brain is VERY good at its job and is picking up something science has not figured out about the human perceptual model..

Its a DAMN good question.. I have searched technically for a while to figure this out. Its real. What it is,, I don't know..

 

[Xymox]

Xymox
That was a phenomenally detailed and much appreciated reply. I didn’t mean to understate your considerable efforts but to be fair, I couldn’t see the “action “ side of the boards you worked on. I’m sure your efforts will be productive. Remember that Taiko’s switches will go to Extreme owners first and then possibly to others so your contribution might benefit many listeners.
Marty

Your question was great. It allowed a detailed explanation

Yes the Taiko switches and cards are already in systems now Have been for a while now. I admit, I have already A/B ed. But my view does not count. I need a few Extreme owners to play with them in depth. PLUS the Extreme switch/card needed burn-in and so does my switch. So a proper A/B takes some time.

 

[Pokey77]

Ooooo.... Can you solder ?

Ha!!

 

[msimanyi]

I made a quick vid.. I show a IAP-225-US powered by a 12V linear. Works great.. Also I give a preview of what makes a Aruba awesome. FPS140 mil specs plus other specs that make it RF hardened.. WHo knew ? So I have now measured and these Arubas dont have any measurable RF noise that they emit. Its JUST 2.4/5 Ghz wifi. Unlike the Ubiquiti, Eero and TPLink which all have TONS of RF noise they emit...

The Aruba's have full mil-spec complaince which also include RF emission specs.. https://www.arubanetworks.com/assets/ds/DS_FEDCERT.pdf

I will do a better vid whowing all the popular access points for audio next week.

I know those Aruba IAP-225-US APs are a screaming deal, but have you evaluated any of their current models for similar "non-noisy" characteristics? I know in general to stay away from WiFi-6, but knowing whether or not the IAP-15-US, for example, is an equally good solution - despite the higher price - would be useful info.

 

[Xymox]

I know those Aruba IAP-225-US APs are a screaming deal, but have you evaluated any of their current models for similar "non-noisy" characteristics? I know in general to stay away from WiFi-6, but knowing whether or not the IAP-15-US, for example, is an equally good solution - despite the higher price - would be useful info.

These odd 2 digit Arubas available new on Amazon are odd and new to me.. It appears to be a line created when HP bought them. They are obviously much cheaper. The top of the line Instant access point is $2200 ea and these are close to 10 times cheaper. .. The Aruba access point makes no mention of these "Instant On" products on the aruba access point web site, which is odd. . They appear to be targeted to compete with Ubiquiti. These are some kind of HP pushed product. I am after the shielding and that comes with the mil-spec and these dont have that.

https://www.arubanetworks.com/products/wireless/access-points/indoor-access-points/#products

I suppose they might be good for normal use.. But they are missing the mil specs that require the shielding that make the real Arubas so useful for audio. The specs show they are much lighter.. So less metal and shielding..

They are some separate offshoot that is cheaper. So I think you would spend more because they are new VS used and get less because of the lack of mil-spec rigor. IMHO... This ripping off of the "Instant" to become "Instant On" .. At least they seperated them out so they are easy to spot with a 2 digit model number. They also dont use IAP.. Its just AP-xx...

I made a vid discussing and showing the Ubiquiti and Eero VS the Aruba..

 

[msimanyi]

Good info. Thank you. I dug a little deeper and found the AP 340 is WiFi 5 and what they call a "unified AP", where you can choose to operate it via a controller (expensive) or in an Instant On configuration.

But even there, I don't see any mention of FPS140, though they claim "Advanced Cellular Coexistence" and "Best In Class RF Management" in their marketing document here: https://www.arubanetworks.com/products/wireless/access-points/indoor-access-points/340-series/

 

[Xymox]

Good info. Thank you. I dug a little deeper and found the AP 340 is WiFi 5 and what they call a "unified AP", where you can choose to operate it via a controller (expensive) or in an Instant On configuration.

But even there, I don't see any mention of FPS140, though they claim "Advanced Cellular Coexistence" and "Best In Class RF Management" in their marketing document here: https://www.arubanetworks.com/products/wireless/access-points/indoor-access-points/340-series/

I dont see ANY IAP-345s or any 340's on ebay. I also cant find any reference to a IAP-34x series. I see TONS of AP-345.. You can't convert a AP-xxx to IAP-xxx without one of those expensive controllers as far as i know. In fact, I thought they did not allow a conversion from AP to IAP at all. I know you can convert a IAP to a AP... IAP- is not in the ordering guide for that series.

I don't know enough about this. I Always buy a IAP-XXX-US as I KNOW those will work.

Once HP bought them things got a bit murky.

There is a airheards forum for Aruba. They could answer these questions. Hewlet Packard also screwed this up using a terrible forum format.. https://community.arubanetworks.com/community-home?CommunityKey=ff734391-dcd7-4802-93c6-e0913abba875

There are IAP-325-us.. Not sure if they are all metal inside, but they look like it. https://www.ebay.com/itm/266304988162?hash=item3e0103a402:g:afMAAOSw7HFkihyY