Now, if only we could get those chips back. All my projects are on hold for now.
Did I miss something, is there another shortage about to make progress difficult again?
I am not used to design with what's available, from whatever company. And the effort to adapt the requirements to what's available is something I got tired of.
Try to find some CS2100 before mid 2022, or some XMOS XE216-512-TQ128. I don't want to give up the XMOS Ethernet support for some internal flash, and the FBGA version of the XE-216-512 is not something I feel like dealing with now. I solder my stuff manually, to the finest pitches, but have no tooling nor experience with handling BGAs.
Try to find some CS2100 before mid 2022, or some XMOS XE216-512-TQ128. I don't want to give up the XMOS Ethernet support for some internal flash, and the FBGA version of the XE-216-512 is not something I feel like dealing with now. I solder my stuff manually, to the finest pitches, but have no tooling nor experience with handling BGAs.
I've seen a few really cheap actively crossed over systems using analogue solutions. They sounded way better than they had any right to. You are right, but I still think people deserve much better than they are getting.
Systems sold today are scraping the bottom of the barrel compared to available technology. We were looking at technology exceeding our limits of hearing for quality, now you are describing junk.
Systems sold today are scraping the bottom of the barrel compared to available technology. We were looking at technology exceeding our limits of hearing for quality, now you are describing junk.
I talked to the XMOS guys about that Ethernet interface. It seems you would be on your own getting anything useful from it.
What I'm hearing now is that projects don't start until production parts are in hand. Its too risky to bet on getting delivery, especially when a $.50 part availability can stall shipping a $500 project.
What I'm hearing now is that projects don't start until production parts are in hand. Its too risky to bet on getting delivery, especially when a $.50 part availability can stall shipping a $500 project.
In a successful product the design effort should be a small part of the economics. And doing it right costs little more than doing it cheap. I have seen vendors strain to get it right when its much easier and quicker to ship what everyone else gets, but that model has no future.
Current high end audio has little real connection to the best and latest audio technology (unless buzzwords count). I would say a speaker without an amp integrated is as current as a 3 wheel Morgan. Fun and interesting but clearly obsolete.
Current high end audio has little real connection to the best and latest audio technology (unless buzzwords count). I would say a speaker without an amp integrated is as current as a 3 wheel Morgan. Fun and interesting but clearly obsolete.
I have a project at work being partially held up by Intel (Altera) parts fabbed at TSMC. Fairly expensive parts (~2k each), near 7 figure order and still couldn’t get a commitment under 52 weeks. Already had to design around Cypress PSoC unavailability. Have trays of Ethernet PHYs with no home.
Last edited:
Yes, the XMOS is yet another link in the chain like a deck of cards its been, but i've been waiting on that for 6 months or more, with it just being another extension .... I did manage to get some proto quantities, just 3 pieces iirc, so I have that to work with in the meantime. I' actually finally in a position where I have everything I need to proceed. this project has been like pulling teeth!! I can't imagine what its been like for those that this has become just the way you have to do business.
I am using some HDI technique on the boards anyway for ethernet PHY and PCIE, so BGA will be a necessary evil moving forward and a good learning experience for me. I'm just using toaster oven/arduino for reflow and its achievable, plus initially at least the xmos resides on a daughter card, so wouldnt blow the whole board if it doesnt work out. i'll be doing a layout for both formats. i've already been soldering various other leadless packages and choose them these days, so fingers crossed there arent any problems I cant work through, even though the eyes are starting to fade a touch.
We are using it in conjunct with an RPI compute module and will have access to memory and PCIE, so wouldnt expect to be using xmos for that (ethernet)
It is a PITA i'm sure. I feel you. How many pieces do you need?
*ahh, my apologies, i'm using/have the XU216-256-TQ128 on hand, not the XE
Last edited:
In my case it is one notch worse: even when they have the parts, TI refuse to sell me some of them.
They sold me OPA862 once, but since approx. 6 months it's a no go.
And in order to get INA849, I'd have to reduce the quantity to less than 0 (see att.).
Regards,
Braca
They sold me OPA862 once, but since approx. 6 months it's a no go.
And in order to get INA849, I'd have to reduce the quantity to less than 0 (see att.).
Regards,
Braca
Attachments
Last edited:
It is not about the number, it is about the challenge. An intellectual stimulus!
The human race has always been racing forward to more and more and better and beter. We cannot stand still, can we?
The human race has always been racing forward to more and more and better and beter. We cannot stand still, can we?
I believe you've perhaps misinterpreted what it was I was trying to say. I don't have any problem with passive crossovers but only when they are done right. The main issue with them is they are expensive. In cheap speakers the crossover is sometimes there just to stop the tweeter exploding and then you might get a resistor for it too and then a coil on the woofer. You probably have perfectly good budget drivers and an adequate cabinet but because a well designed passive crossover would likely cost as much as the drivers did the speaker ends up not getting it. You end up with a poorly designed speaker that sounds like crap.
Even in more expensive speakers crossovers are oversimplified, to save a penny, at the expense of performance. No one wants to buy more passive crossover components than they have to. I certainly don't. Designing a good passive crossover isn't hard. What's hard is coming up with ways to simplify it but without sacrificing (too much) performance. This seldom happens in inexpensive speakers and sound quality suffers tremendously as a result.
Introduce a DSP to the mix and all of those constraints evaporate. You can just add as much crossover complexity as is required to get good sound.
@5E :I still think burning for being a heretic would be in order.
@syn08: Customers are reluctant for extra complexity in a lot of areas, but smart active home speakers controlled by your phone are selling by the million and I wouldn't be suprised if active monitor sales outstrip passive 'hifi' sales at least in volume if not in $$$. So it's just plain old fashioned domestic stereo that is stuck in the dark ages. Genelec has active speakers for every budget and apparantly some of them are rather good.
@syn08: Customers are reluctant for extra complexity in a lot of areas, but smart active home speakers controlled by your phone are selling by the million and I wouldn't be suprised if active monitor sales outstrip passive 'hifi' sales at least in volume if not in $$$. So it's just plain old fashioned domestic stereo that is stuck in the dark ages. Genelec has active speakers for every budget and apparantly some of them are rather good.
Sure, DSP+multi channel amplifier in each speaker, that would not add any complexity from a customer perspective.
For whatever reason, don't know exactly why, I don't like the idea. Perhaps I'm a control freak enjoying being in control of every little detail in my system, and relying on a smart phone is not my idea of "control". Or I am a dinosaur on it's way out.
For whatever reason, don't know exactly why, I don't like the idea. Perhaps I'm a control freak enjoying being in control of every little detail in my system, and relying on a smart phone is not my idea of "control". Or I am a dinosaur on it's way out.
Have for some time. The largest selling speakers are the Alexa/Google Home things. Everything else is trivial by comparison. Who would want a speaker you CAN'T talk to??? (I won't have those in my house, it was bad enough to work on the design of them.)
Today I would describe the amp speaker being separated as being on the order of attaching a horse to pull your car. Quaint.
And from my experience designing speakers- if you need a complex crossover send the driver back and have the vendor fix it. The complex crossover to deal with a driver is a bandaid. If you are small time and only buying a handful of drivers your leverage is limited, but if you are volume (1K /yr or more) you can invest the time to get the drivers right. Then a passive crossover is pretty simple. And active/dsp makes it easier to optimize everything.