Installing BNC connectors on digital gear...

[TurtlePaul]

The key to understanding Chat GPT is that its goal is to produce the most pleasing answer for the questioner. If you frame a question differently it will produce conflicting answers. Just like people flock to those that agree with them in real life. ChatGPT is trained on the internet which contains all sorts of conflicting opinions, it then picks the opinion it thinks you will most appreciate.

75 ohm cable terminations were originally designed for composite video signals which are analog ~3-5 mhz signals. Digital audio signals don’t need to worry about analog noise. Analog audio signals have 1/250 the bandwidth and don’t need to worry about this noise unless the cables are miles long (which brings other issues). Audio device designers stole the connector because it was convenient but they didn’t worry about mhz range characteristic impedance matching which only has a signal impact in the transmission of analog video.

 

[G|force]

If you are going to retrofit with a BNC jack, at least use Amphenol brand.

 

[Krusty09]

If you are going to retrofit with a BNC jack, at least use Amphenol brand.

Is kings not a good brand?

 

[klettermann]

Thanks all — this has been useful.

What seems broadly agreed (once you strip away the noise) is that S/PDIF over short coax runs is deliberately very forgiving, which is precisely why consumer gear gets away with RCA connectors in the first place. Designs aren’t so much “optimized around RCA” as they are tolerant enough that it doesn’t matter.

That robustness explains both things at once:
why swapping to a proper 75 Ω BNC won’t produce any audible benefit — and also why it won’t break anything.

One small meta-point, since ChatGPT became a recurring theme: whether a claim is correct depends on the claim, not its source. Several replies here actually addressed the substance, and those were helpful; rejecting an argument purely because of where it came from isn’t analysis.

In any case, the engineering question I was trying to understand is now clear. I’ll still do the mod on one unit because I enjoy hardware work and prefer mechanically robust connectors — not because I expect sonic changes. If nothing else, it’s a practical demonstration of just how tolerant the interface really is.

Appreciate the explanations from those who focused on the “why,” not just the “it doesn’t matter.”

 

[voodooless]

One small meta-point, since ChatGPT became a recurring theme: whether a claim is correct depends on the claim, not its source. Several replies here actually addressed the substance, and those were helpful; rejecting an argument purely because of where it came from isn’t analysis.

It’s a fair point to be honest. But we’ve seen so much nonsense posted with ChatGPT as a source that’s it’s quite easy to reject a lot of what comes out of it outright. We rather argue with actual humans that with a large language model

Have fun with the mods!

 

[Mnyb]

Sadly a lot of buthurt keyboard warriors arrive at frequent rate with cut and paste AI slop to prove Amir wrong or some other senior member that made a real effort writing on a topic ..
I mean literally walls of text in some cases ?

some of the residents may overreact when they sense AI quotes .

So forum policy is a bit stern . It should be obvious that it is an AI quote at least . You can use it as a research tool but think trough the responses and form a real understanding and then write with your own words .

Here is a link to the policy .

Official policy on use of AI

We have had recent issues come about regarding use of AI in member posts. I personally think this technology can be useful to us. And at any rate, hate having to police members left and right on what they post. That said, here are my thoughts: 1. If you are going to use AI, make it...

www.audiosciencereview.com

 

[klettermann]

Sadly a lot of buthurt keyboard warriors arrive at frequent rate with cut and paste AI slop to prove Amir wrong or some other senior member that made a real effort writing on a topic ..
I mean literally walls of text in some cases ?

some of the residents may overreact when they sense AI quotes .

So forum policy is a bit stern . It should be obvious that it is an AI quote at least . You can use it as a research tool but think trough the responses and form a real understanding and then write with your own words .

Here is a link to the policy .

Official policy on use of AI

We have had recent issues come about regarding use of AI in member posts. I personally think this technology can be useful to us. And at any rate, hate having to police members left and right on what they post. That said, here are my thoughts: 1. If you are going to use AI, make it...

www.audiosciencereview.com

You gotta be kidding me. What AI quote? I referenced ChatGBT, didn't quote it. It's even arguabvle that any Sr member made an effort to write on the topic. Mostly the idea was denigrated. And where's the data that ASR holds so dear? Easier to denigrate than educate I guess, especially if it's outside the orthodoxy. And this whole thing wasn't really about audible differences. As I said at the outset, "To be clear, I have little expectation of some sonic breakthrough."

 

[Mnyb]

You gotta be kidding me. What AI quote? I referenced ChatGBT, didn't quote it. It's even arguabvle that any Sr member made an effort to write on the topic. Mostly the idea was denigrated. And where's the data that ASR holds so dear? Easier to denigrate than educate I guess, especially if it's outside the orthodoxy. And this whole thing wasn't really about audible differences. As I said at the outset, "To be clear, I have little expectation of some sonic breakthrough."

Sorry I misunderstood you , apologies.

 

[mcdn]

You gotta be kidding me. What AI quote? I referenced ChatGBT, didn't quote it. It's even arguabvle that any Sr member made an effort to write on the topic. Mostly the idea was denigrated. And where's the data that ASR holds so dear? Easier to denigrate than educate I guess, especially if it's outside the orthodoxy. And this whole thing wasn't really about audible differences. As I said at the outset, "To be clear, I have little expectation of some sonic breakthrough."

You did get a bit of a harsh reception in my opinion, but some including me did give technical responses with references.

For the AI bit what you're meant to do is show the prompt you used, and make clear what (if any) of the rest of your text is AI generated. For example if I prompt ChatGPT 5.2 (pro, with base style set to "professional") with "Is there any downside to replacing RCA connectors with BNC ones for an audio SPDIF interconnect?", it responds with:

"Short answer: electrically, no—BNC is objectively better for S/PDIF. The downsides are practical, not technical." Followed by a quite detailed and correct fuller explanation (see https://chatgpt.com/share/695303a1-31f8-800b-ad9a-c31ee7cbc67d).

The only thing it loses a mark on from me is not being more clear about the lack of audible difference. But an easy A- grade.

This is exactly the opposite of what you said happened in your original post. Maybe "ChatGBT" as you keep calling it is some other, lesser AI. Maybe your prompt threw it off. Without the chat link and prompt we may never know, which is why the policy exists.

 

[EERecordist]

The electrical engineering side is that at 24 bit 192Khz signal is about 25MHz. SPDIF on coax is 75 Ohm characteristic impedance. 25 MHz is not high.

Usually you see 75 Ohm BNC terminated with RG59 cable. It is cheap, easy to terminate DIY, though it benefits from a crimped strain release to the cable. But the cable is very stiff coming off the connector in the rear of your equipment and it is very stiff to route around between equipment. If you are extreme, RG7 is a great cable found in cable TV with even better shielding, my house is wired with it, and it will probably never be used. RG7 is even stiffer, meaning a larger bend radius.

If you have money to spend with no likely return, consider LEMO connectors. They are a Swiss product, very beautiful, and a shame to hide behind the equipment. We used them when we needed to terminate thousands of high energy physics detectors into sense amps with a good signal. You want the B-Series E_G socket and F_G plug wifh RG-179, a small diameter coax cable. Usually you would need a special crimp tool for the F_G. RG179 is also nice on the test bench and test equipment. There are BNC to LEMO adaptors, and companies that build LEMO coax cables to order.

The LEMO connector bodies are beautiful, they could contribute to high-end, even if they are not needed. The connector bodies come in exotic metals. Maybe to take advantage of them, high-end should move the connectors to the front panel, like instrumentation does for practical reasons, mainly that the back bus is very busy.

A less flashy RF connector is SMA which could be used for SPDIF wired connections. It is common to see the connector gold plated.

If you look at your equipment mechanically, what does it take machining-wise to replace a connector?