this made me literally lol. my macbook pro is the shittiest bluetooth device i have ever used. just search google for macbook pro bluetooth audio issues, and you will find a slew of forum posts going back years. there's one terminal hack after another suggested just to try and get macbook pros to properly play audio over bluetooth. i don't even bother using my macbook pro these days to play audio through a bluetooth speaker. i just use my iPad or android phone.
Apple has also shipped hundreds of millions of devices over that same time period, so the number of people with failed hardware is non-trivial. More importantly, it is extremely rare for people to perform root cause analysis on the actual failures which combined with the number of vendors and protocol complexity in the Bluetooth world means that a significant number of times the blame is placed on the wrong component.
The main thing I take from this is that vendors need to think about how they can provide a visible way to learn about peer device errors so e.g. your Mac could tell you that sound quality is horrible because the Bluetooth device doesn't support a high-quality codec or sufficient bitrate, is seeing high levels of retransmits or dropped packets, etc.
Like I had mentioned previously, if one side of the paired devices has a garbage antenna setup, it will have poor performance. Things like your soundbar, or a generic bluetooth speaker probably use reference designs for their BT setups and have not tuned them for the enclosure or use case.
What quality is the hardware in something like this:
?
I think that you are remiss if you don't actually name the companies/produucts/devices which are designed well then....
There was a company in Aptos/Santa Cruz who was designing the antennas for a lot of Lockheed stuff back in about 2006/2007 or so who we went to for antenna designs for our active RFID devices - but I cant recall for the life of me the name of the company now...
Your run of the mill bluetooth headset, mouse, or keyboard likely does not spend the time on the RF design beyond "does it work".
1. Do absolutely nothing.
2. Completely lock up the computer so it needs a hard reset.
3. Completely lock up the computer for a few seconds, then work.
4. Start working, but ignore ~40% of the key presses.
5. Swallow the first keypress or two, then work.
6. Actually function like a keyboard.
If Microsoft has a good RF team, I'd hate to see what worse teams do.
I wish there was a USB version of this keyboard, or at least a wireless dongle version (like Microsoft's cheaper keyboards).
The companies that deploy a TI, Nordic, Broadcom chipset are subject to the scrutiny I just mentioned. The chipset manufacturers offer reference designs but generally you need a custom matching circuit and a complex RF design to do it correctly. That's where most manufacturers skimp out because its generally "good enough"
As for WiFi, it does not frequency hop but is capable of "auto" selecting a channel on boot (generally).
W1
Many, many Wi-Fi implementations are capable of frequency hop, including my own at home.
What he is probably referring to isn't frequency hopping, because what it sounds like and that it actually refers to means different things to different people. For someone who is not a RF engineer, hopping might simply mean "an AP that has multiple radios that uses one to scan the 802.11 2.4Ghz space to see what channels have the least utilization, and then restarts the other radios on a that channel, possibly forcing all clients to lose their connection as if out-of-range and reconnecting again on the new channel as they find it".
This, of course, has nothing to do with RF frequency hopping, but if you don't actually know anything about RF, you won't know what hopping means either, and just imagine that what your router with built in AP is doing must therefore be that 'hopping'.
Bingo, except add 5 GHz (and soon, 900 MHz), too.
As far as I know it's not dynamic, though. It just happens when a device connects to a base station and only "steers" the client to connect using one of the 2.4 or 5 GHz bands, and that doesn't interfere with channel selection.
The AP also has a function to scan bands to find the "best" channel to use but that disconnects all devices while the scan is in progress. I think some APs might be doing this automatically (?). My other AP (a Mikrotik) has an "auto" channel selector but I don't know what it does or when.
My neighbor moves around his channels, too. That’s what’s making me comment that it’s likely more common than we think. I’m not at home to prove this and I’m getting downvoted for sharing, but I’ve watched him move between 36, 40, and 44 at a frequency that suggests it is happening automatically, probably when all of his sessions go idle or something. I have observed this with multiple tools.
My 5 GHz was on 44 this morning, because I happened to look for unrelated reasons. I just SSH’d home and it’s now on 36 without interaction from me. Specifications are not implementations. I’m almost positive this is a common feature which I’ve seen in many APs over the years, even though it’s not per-packet like Bluetooth (which I was also aware of, given that I'm experimenting with a Bluetooth 5 mesh system for IIoT).
(it does mention a frequency hopping physical layer but that was part of the obsolete predecessor from 1997, the modern spec uses OFDM so unless all your equipment is from the last century you're not using frequency hopping)
1. IIRC my router is supposed to switch channel (i.e. frequency) automatically unless I explicitly set it to a fixed channel.
I don't know how but my devices figures out and continues working.
2. I think the misunderstanding is related to the word frequency hopping or whatever was used further up in the discussion.
I normally take "frequency hopping" to mean changing frequncy multiple times a minute or even faster while I doubt my wireless access point will switch channels more than a few times in 24 hours.
Disclaimer: I never really sat down and verified if it ever changed.
Also agree with logicalle that I wish for a bit more civility.
jsmthrowaway can you give more information about your setup and what the devices do?
No need for a heated argument. This whole thread is quite instructive!
(No I'm not complaining about being downvoted - this isn't even my comment - just trying to learn.
Most of the issues in this thread are related to poor hardware design more than a crowded spectrum. While the spectrum is in fact crowded in metropolitan areas, most Bluetooth communication doesn't require much bandwidth and can handle error prone areas with ease.
While the frequency hopping helps a ton on BL (and WiFi for that matter), the issues people outlined are due to:
1) Shitty firmware 2) Shitty hardware
Antenna design is black magic and only a few firms in the US do it well. It took us almost 10 months to fully design and test our antenna assembly(s) with a very capable third party firm.
It took dozens of trips to a test chamber, a dozen computer simulations that take a day to run, and PCB samples that take days to verify. They have to be tuned every time copper or mechanical parts move as well.
It's a real pain and most Bluetooth products use garbage chip antennas and baluns or reference designs for antennas. This increases the sensitivity to failure and provides a generally shitty experience.
Most of your product interactions around bluetooth are budget products connected on one side of the equation (e.g. a $50 bluetooth headset). So despite how capable your Mac or iPhone is, if you have a garbage headset on the other side with poor antenna design, it'll be a disaster of an experience.