@duball97@i2cjak Home Assistant community is a good source of information for this kind of thing.
Presently, my understanding is that chipsets like the LD2450 (for tracking) or LD2410 (for static presence) are best.
@blind_via@_Blamoo And sum total that gold is probably worth less than $0.25.
If the goal is to reclaim the gold, just scrape the chips off the board and don't waste lots of electricity and time unsoldering it.
@RaisesQuestions Even though a little more than 30% of the voting eligible population voted that way in the last election, those votes don't mean they actually live and/or support that category.
You have to vote for the choices available, and MANY factors go into that decision.
@TheStalwart@ShanuMathew93 It would depend on the model and the silicon.
Speaking generally, custom silicon can sometimes be multiple orders of magnitude faster, to say nothing of more power and space efficient.
But of course, that wouldn't apply to the whole pipeline - hence why it would depend :)
@Rapahelz@paulg As with may attempts to categorize, this falls short in my opinion.
Seems like "compassionate conservatism" would land between the two middle ones.
@blind_via I'd consider my use case(s):
If power goes out, do I want it to power back on automatically, or stay off?
How often do I change the voltage, and how easy is it to change?
How extensive is the remote interface if that's important (program cycles)?
Ramp rate control?
@ptrschmdtnlsn@fuzziphy@blind_via Oh, I misremembered - you're right!!!
It was AGP. We customized the card edge a bit, but the socket would have been off-the-shelf from Molex or Tyco.
those chips that are straddle the heat sink area, are they coming into contact with the sink for thermal releif on their backside? this is the only reason I can see to put them RIGHT THERE along the heat sink line. or room isolation like I previously mentioned. you can even see the dgnd areas that are flooded directly under the ICs which support the single point ground concept
@blind_via TI switchers are there to contact the heatsink (through the TIM). These were old gen, before heat slugs on the bottom of ICs was common.
The heatsink didn't extend into the rear "tall" section, so they couldn't go back there. And inductors wanted to be in the tall area.
@idontwa86202030@blind_via@jlcjak LOL. Two things going on in this case:
1) originally the PCB had a soft TIM (thermal interface material), which leaves an oily residue
2) the board has been in my garage for over 15 years collecting dust
@blind_via Nothing taller than an IC could be in the middle section, so the inductors for the switchers had to go in the rear section. Here's an edge view.
@mzppvhofho16214@D_N_K Having said that, I understand and agree with the original general statement.
Something working doesn't mean it is compliant. You could be "working," yet have no margin.
@mzppvhofho16214@D_N_K Yep, we were using it for multiple lanes of OC-48 (2.488 Gbps) between boards internal to the system.
No PCIe at all on this board.