I want an RPi...
With 32Gb of DDR5, a 10Tb NVMe4 SSD, and an iron clad alibi for what I get up to next.
I probably shouldn't have said that last bit.
Raspberry Pi fans have another way of powering their mini 'puters and attached gizmos thanks to an impending update to the Power over Ethernet (PoE) Hardware Attached on Top (HAT), as the silicon required for its predecessor gets increasingly difficult to come by. The update for the original PoE HAT is imaginatively dubbed PoE …
"I probably shouldn't have said that last bit."
As someone prone to delayed scolding by the other half, after nearly every case of sudden hardware manifestations let me tell you that no alibi will suffice.
Yet another bunch of pi with different specs would be nice, maybe the keyboard version with actual RAM and peripheral slots.
I know that amazon is somewhat of a no-no in these parts, but as far as i could see, there are several offers in your desired price range for a pi 3B+.
As you quoted Dollars, i assume you want the pi delivered to the USA so i checked amazon.com (and discovered that sending a pi back to europe would cost additional import fees and $22 shipping).
Now i need to ask my crystal ball why the 3B+ has a price range for the single board from 36 to 50 euros...
This post has been deleted by a moderator
I'm getting more of a curmudgeon as time goes on...so:-
Is this FULL compliance with the at/af standards, including the power control negotiation/signalling, or just the 'non standard passive POE only' stuff. Also, does anyone know what pins the power is on? Mode A or Mode B?
The official Raspberry Pi page for this claims it's compliant to IEEE 802.3at-2003 PoE. Last time I checked, 802.3at wasn't a thing until 2009...
....or have I just made a fool of myself?
Thanks for the reply, however...
I don't disagree...they do indeed claim 'at' and 'af'. I am bitching about them claiming compliance to something that doesn't exist ('at' arrived 2009 IIRC) and they don't say if they are completely compliant to either i.e. do they do the signalling/negotiation or just 'passive'. Mode info i.e. power pins expected would be handy too....
I've asked the same thing of them, so I am interested to see what they say.
Of course, if I am being foolish, tell me:-)
Lol, I'm not complaining about El Reg's reporting...
I'm moaning about the inaccuracy at the Pi Foundation site, and their claim to support a standard that does not exist. Typo or not, that's their problem and an org with the worldwide reach that it has shouldn't be making them.
My point is really that the errors (The 2003 vs 2009 part) , and poor detail (Signalling/negotiation, power pins Mode A/Mode B) make it hard for people starting out to make informed decisions.
Most of us here know POE, have used it for years and understand it's subtleties . If I put myself in the shoes of a noob desperate to learn from an affordable Pi system (We were all there once, with me it was the ZX81), I'd be disappointed if things didn't work due to either wrong info or misleading info on the manufacturer's site.
If the Pi Foundation really is trying to help educate the next generation, then they should pass on the full story, not just part of it. If you're going to reference specifications, then quote the right ones and tell the whole story. Teach, not confuse!
Here endeth this rant, as this evenings ration of Merlot is taking hold :-)
If you read the blog post on the pi website you will see both a statement from Eben that
we use an Ethernet jack with the appropriate taps on its windings, and connect those taps to an additional 4-pin header, located just beneath the top-right mounting hole. A HAT can pick up these signals, request power from the switch, and regulate the resulting 37-57V DC down to 5V to power the Raspberry Pi.
The key word seems to be "request". There is also a table which claims the PoE+ HAT has "current sense". As I understand it (and I'm quite prepared to be wrong) a switch compliant with "at" will not attempt to deliver high power unless it is specifically requested by the client, thus the client must "negotiate", as you put it, if it wants high power.
I note your question about Mode A / B, but there is also an actively monitored comment section below the article where you can see that Eben and Liz directly respond to queries. Ask away.
Personally, I'm fuming that two bits of TP-Link wireless kit I've bought appear to have such a proprietary form of "passive" PoE that it would actually be dangerous to plug them into a normal PoE switch. The injectors supplied with the kit weren't a problem for a couple of years, but I've just reorganised everything and it would be great to tidy things up a bit...
A lot of industrial usages are allergic to untested/unproven (by them) change, even if it's a simple change and for the better. "Copy exact" and such like.
So if they have existing kit and want more, often they will insist on the older and obsolete version rather than going to the effort of doing extensive tests on the new one. Even if the new one passes the test, as soon as it's installed then that would be the trigger for all the older ones to be removed and retrofitted with the new version too.
That can be considerable time, effort and cost, which in some cases isn't worth the investment (depending on what is new/improved on the new version compared to the old). It can also occasionally lead to odd scenarios of having new tools operating "too well", and having requests to detune them to match the operation of older kit.
""Even the Raspberry Pi team are not immune to the shortages affecting the industry it seems, although Upton said "Raspberry Pi computers, Zeros and our other products have not been affected," before adding modestly: "We're very good at pipelining."
I'm reading that as "there will be a delay in when you start to see shortages, but they're coming..." Pipelines are all well and good, but they're not immune to stalling.
Biting the hand that feeds IT © 1998–2021