It's unclear... The A+ and Pi3A+ work because they use the native USB 2.0 interface of the SoC. For a Pi4A one would presume a desire for a USB 3.0 port, and that would require some version of the PCIe to USB hub on the Pi4B...and the current location of the Pi4B version of that chip is on the part of the board that the A+ format doesn't have. Of course, something really radical could be done: Just expose the PCIe lane and let people connect whatever they want to it.Moonmarch wrote: ↑Fri Dec 20, 2019 11:21 pmIf the parts needed for the A+ version of the RPI4 computer will be similar or the same as the RPI4B computer then wouldn't a RPI computer with less features be not only lower in cost the computer should cost less to produce, because this RPI computer with less features will need less hardware overall unless the A+ version of the RPI4 computer has a completely new design then you would expect development costs to be about the same as building a brand new RPI computer.
If I had a guess there needs to be a high enough demand for the next release of the RPI computer to be a slim version of the RPI computer or similar most single board computers are the same height, because of the height of the Ethernet port if you do not include the size of the heat sink.
Been discussed before.It's unclear... The A+ and Pi3A+ work because they use the native USB 2.0 interface of the SoC. For a Pi4A one would presume a desire for a USB 3.0 port, and that would require some version of the PCIe to USB hub on the Pi4B...and the current location of the Pi4B version of that chip is on the part of the board that the A+ format doesn't have. Of course, something really radical could be done: Just expose the PCIe lane and let people connect whatever they want to it.
The Pi4 SoC also has a native USB 2.0 interface (it's connected to the USB-C power port on the 4B), so that could be used for the single USB-A on a 4A, and the PICe lane could be available for other use.
I thought about suggesting that, but making custom carrier boards for "a couple of projects" doesn't seem practical.
Indeed. One USB-C for power and USB-A for connectivity, with header pins for PCIe, would create a Pi 4A which is pretty much a direct upgrade from the Pi 3A+.
Given the timing of past A-series Pis, I think that's a bit optimistic. By next anniversary the Pi4B boards will probably still be eating up the full production of BCM2711 SoCs.
You don't even need to do that, use a small flat head screwdriver (the kind sold in electronic toy repair kits works great) and just gently lever it off, working from end to end.PhatFil wrote: ↑Sat Dec 21, 2019 6:09 amFor anyone in doubt.. removing the gpio header is a trivial task even for a total novice with a soldering iron. the trick is to simply use snips/side cutters to cut the plastic base the pins sit in so each pin is separated from the bulk and they can be individually heated up at the solder point and removed.
Confused. The GPIO header is clearly soldered on. Desoldering all the pins cleanly enough to remove easily, is the difficulty. Chopping it up as you go along makes that easy.
Imperf3kt is talking about the black plastic strip the pins are in. It is only retained by friction. There aren't any notches or protrusions.
I would strongly not recommend that technique. I suspect that there is some variation in the level of friction and it will increase if you get the plastic strip "cocked" by levering up on end too much. Plus you pretty much have to lever against the PCB which is a bad idea. If you are going to be modding and soldering you really should have a decent set of side cutters. Cutting the plastic, allows you to remove the pins one at a time via heating the solder until it flows and then gently pulling the pins out, one at a time.
You can pull the pins one at a time while heating the solder WITHOUT removing the plastic strip. The hot pin will pull through it fairly easily.
I have no doubt it will work if done "right". I continue to be amazed what skilled technicians can do, and I have been doing engineering stuff for a long time now. The difficulty comes in trying to explain the "done right" to someone of a vastly different skill level, with perhaps a language barrier thrown in for good measure. Plus there is the issue of posted solutions living on forever...