Page 4 of 4

Re: any news yet on the PI3A??

Posted: Thu Mar 02, 2017 8:47 pm
by alphanumeric
For me, the reason I bought a couple of A+'s was the small foot print. They are the same size as the sense hat they are plugged into. That makes putting them into a case easier. They are run headless, Astro PI if you will, so I don't need all those extra ports hanging around doing nothing. I only use the HDMI and USB for setup. Having a full sized HDMI and one full sized USB makes things easier. If the 3A ends up being a direct swap, bonus bonus. =)

Re: any news yet on the PI3A??

Posted: Thu Mar 02, 2017 8:58 pm
by fruitoftheloom
alphanumeric wrote:For me, the reason I bought a couple of A+'s was the small foot print. They are the same size as the sense hat they are plugged into. That makes putting them into a case easier. They are run headless, Astro PI if you will, so I don't need all those extra ports hanging around doing nothing. I only use the HDMI and USB for setup. Having a full sized HDMI and one full sized USB makes things easier. If the 3A ends up being a direct swap, bonus bonus. =)
Maybe RPT will release a 2A & 3A so one has a choice of with or without WiFi/BT :D

Re: any news yet on the PI3A??

Posted: Thu Mar 02, 2017 8:59 pm
by DougieLawson
mfa298 wrote:
skspurling wrote:Now that I think about it, are people realy running out of memory on their PI's? You can do some pretty heavy image processing with openCV and not hit any limits other than processor. What are people doing/expecting?
I suspect most of the memory complaints come from those trying to use the Pi3 as a Desktop system with a few other people having issues trying to compile larger/complex applications.

In a few things I've tried to setup I have found that 512MB can start to be a bit limiting but rarely had an issue with 1GB.
We should force everyone who asks for extra memory to run a ZX80 emulator with a massive 3K of memory (that shrank as you painted more stuff on the screen). They'll soon learn memory efficient programming and get away from the bloat.

The first OS/VS2 MVS mainframes I used had 24-bit addressing and we didn't even get that full 16MB to use for our programs (limit was about 6500KB). Everyone learned lean efficient programming back then. If your code went over a 4K base register folks would look at you like you came from Mars (it's a real pain with 12-bit base/displacement addressing when your next sequential instruction bumps on to the next 4K page - you'll have to lose a register to gain the next 4K).

Re: any news yet on the PI3A??

Posted: Thu Mar 02, 2017 9:00 pm
by W. H. Heydt
spock wrote:i think the problem with the a3 is the 25$ price target. :)

it makes sense to include wifi/bt. so that's 5$ (as we see with the pi zero w).
it will be a pi 3 so there will be the quad core cpu and that almost requires 1gb ram?

so the only difference to the b3 would be the usb chip and the additional usb ports and ethernet. does that make a 10$ difference?
*Is* the target $25? The current A+ is $20. I don't see that a Pi3A "requires" 1GB. The early A+ ran fine with 256MB and the original plan for the Model A called for 128MB, though it actually shipped with 256MB. Bear in mind that the smaller PCB probably means that more boards go up to make full boards as they go through manufacturing (take a look at the video that is on the Pi0W launch blog entry...looks like 15 or 18 Pi0Ws are on each physical board). So...less board are, no LAN chip, one single USB connector, no Ethernet RJ-45 jack, more A-series board per manufacturing board...plus the general reduction in chip costs over time due to better yields and other manufacturing savings. All that could add up to a $15 cost savings (at retail) over the Pi3B.

On the other hand...if the Pi3A lists for $25, that might well mean that A+ Pis continue to sell better than they otherwise would if the Pi3A matches the A+ list price.

Re: any news yet on the PI3A??

Posted: Thu Mar 02, 2017 9:01 pm
by W. H. Heydt
alphanumeric wrote:I just assumed, rightly or wrongly, being called a "3" A, that it would be based on the current "3" B. Seems logical, to me anyway.
If your just going modify another current model, why go backwards? Why downgrade? Modify the current top of the line model.
Just my 2 cents and best guess. :D
That appears pretty strongly to be what Eben Upton said in the Reddit Q&A.

Re: any news yet on the PI3A??

Posted: Thu Mar 02, 2017 9:08 pm
by W. H. Heydt
fruitoftheloom wrote:
alphanumeric wrote:For me, the reason I bought a couple of A+'s was the small foot print. They are the same size as the sense hat they are plugged into. That makes putting them into a case easier. They are run headless, Astro PI if you will, so I don't need all those extra ports hanging around doing nothing. I only use the HDMI and USB for setup. Having a full sized HDMI and one full sized USB makes things easier. If the 3A ends up being a direct swap, bonus bonus. =)
Maybe RPT will release a 2A & 3A so one has a choice of with or without WiFi/BT :D
I still think you're barking up the wrong tree. The Pi2B v1.3, by switching to the BCM2837, saved some chip costs by eliminating the use of the '2836 and boosted the performance of the Pi2B by around 20% or so. I don't see any comparable reason to start making two A-series boards, one with WiFi/BT and one without. The WiFi/BT *expands* the peripheral connections available without add on devices. The Pi2B didn't need that expansion since it has 4 USB ports and an Ethernet jack. Like the Pi0W, the Pi3A can use the additional built-in connectivity. In any case, if you want a Pi3 speed A-series board, but without WiFi/BT, you can turn them off.

Re: any news yet on the PI3A??

Posted: Thu Mar 02, 2017 9:12 pm
by W. H. Heydt
DougieLawson wrote: The first OS/VS2 MVS mainframes I used had 24-bit addressing and we didn't even get that full 16MB to use for our programs (limit was about 6500KB). Everyone learned lean efficient programming back then. If your code went over a 4K base register folks would look at you like you came from Mars (it's a real pain with 12-bit base/displacement addressing when your next sequential instruction bumps on to the next 4K page - you'll have to lose a register to gain the next 4K).
Some time I'll tell you about the 11K line main posting program, written in ALC, I had to maintain. It ran on an in-house modified DOS system. We had two systems. A 360/40 and a 360/50.

Re: any news yet on the PI3A??

Posted: Thu Mar 02, 2017 9:15 pm
by fruitoftheloom
If it is a 3A with WiFi/BT the LEDs will be in a different position ?????

Not ideal if it whatever casing the A+ is in requires the LEDs to be visable ?????

Re: any news yet on the PI3A??

Posted: Thu Mar 02, 2017 9:24 pm
by alphanumeric
fruitoftheloom wrote:If it is a 3A with WiFi/BT the LEDs will be in a different position ?????

Not ideal if it whatever casing the A+ is in requires the LEDs to be visable ?????
Same thing happened when we went from 2B to 3B. It's not ideal, but sometimes sacrifices have to be made.
For the most part, the only time I look at the status LED is when I'm powering down, so I know when I can unplug the power pack.

A dremel or drill with a small drill bit will fix it in most circumstances, I think.

Re: any news yet on the PI3A??

Posted: Thu Mar 02, 2017 9:32 pm
by fruitoftheloom
alphanumeric wrote:
fruitoftheloom wrote:If it is a 3A with WiFi/BT the LEDs will be in a different position ?????

Not ideal if it whatever casing the A+ is in requires the LEDs to be visable ?????
Same thing happened when we went from 2B to 3B. It's not ideal, but sometimes sacrifices have to be made.
For the most part, the only time I look at the status LED is when I'm powering down, so I know when I can unplug the power pack.

A dremel or drill with a small drill bit will fix it in most circumstances, I think.
So instead of a free Icelandic Pink Pony we now have to use free Dremel.... :lol:

Re: any news yet on the PI3A??

Posted: Thu Mar 02, 2017 9:38 pm
by alphanumeric
I've used my dremel for several case mods so far. Cutting square cutouts for heat sinks in PIBOW cases. Drilling vet holes in the Pi foundation case. etc. etc. It's come in very handy.

Re: any news yet on the PI3A??

Posted: Thu Mar 02, 2017 9:41 pm
by W. H. Heydt
fruitoftheloom wrote:If it is a 3A with WiFi/BT the LEDs will be in a different position ?????

Not ideal if it whatever casing the A+ is in requires the LEDs to be visable ?????
Only if the case is opaque and a change would be to shift the holes/light pipes. One can, after all put a Pi2B in a Pi3B case...and vice versa.

But that's not an argument for a WiFi-less Pi3A. That's a argument for finding a way to place the antenna so that the LEDs don't have to be moved, such as was done with the Pi0W.

Re: any news yet on the PI3A??

Posted: Thu Mar 02, 2017 9:47 pm
by gregeric
Yes WHH. The obvious looking place to put the ZeroW-style cavity antenna without shifting LEDs or connectors around is between the TRRS & USB sockets. RF knowledgeable people: does a cavity antenna even need to cling to the edge of a board like other PCB antennae?

Re: any news yet on the PI3A??

Posted: Thu Mar 02, 2017 9:51 pm
by alphanumeric
fruitoftheloom wrote:
alphanumeric wrote:
fruitoftheloom wrote:If it is a 3A with WiFi/BT the LEDs will be in a different position ?????

Not ideal if it whatever casing the A+ is in requires the LEDs to be visable ?????
Same thing happened when we went from 2B to 3B. It's not ideal, but sometimes sacrifices have to be made.
For the most part, the only time I look at the status LED is when I'm powering down, so I know when I can unplug the power pack.

A dremel or drill with a small drill bit will fix it in most circumstances, I think.
So instead of a free Icelandic Pink Pony we now have to use free Dremel.... :lol:
The Pi foundation, and others, just made cases with the light PIPES on the other side. There are 3B cases and 2B cases out there. You can put a 3B in a 2B case and vise versa. There are no fitment issues, you just can't see the status LED unless the case is transparent. So an A+ and 3A case isn't the end of the world. ;)

Re: any news yet on the PI3A??

Posted: Thu Mar 02, 2017 10:01 pm
by alphanumeric
gregeric wrote:Yes WHH. The obvious looking place to put the ZeroW-style cavity antenna without shifting LEDs or connectors around is between the TRRS & USB sockets. RF knowledgeable people: does a cavity antenna even need to cling to the edge of a board like other PCB antennae?
I have no experience with "cavity" antenna's like what's on the ZERO W, but I have experience in radio communications and can tell you the less physical obstructions around an antenna the better. Especially things that generate electrical signals or electrical noise.

Re: any news yet on the PI3A??

Posted: Thu Mar 02, 2017 11:31 pm
by mfa298
gregeric wrote:Yes WHH. The obvious looking place to put the ZeroW-style cavity antenna without shifting LEDs or connectors around is between the TRRS & USB sockets. RF knowledgeable people: does a cavity antenna even need to cling to the edge of a board like other PCB antennae?
AIUI the main design issue with chip antennas (as used on the 3B) is that they need a certain amount of clearance around them with no copper on any layers underneath. You might be able to put it in the middle of the board, but that makes the board layout much harder. I wouldn't be surprised if something similar applies with the antenna design on the zero-w

Re: any news yet on the PI3A??

Posted: Fri Mar 03, 2017 5:25 am
by gregeric
Yes I appreciate the requirement for clearance from the ground plane, hence why it's the norm to have antennae on the edge of the board so there's nothing in front, with as much copper cleared around as feasible.

But this cavity antenna seems to be a hole in the ground plane, hence why I wonder whether it could just as well be in the middle of the board.

Re: any news yet on the PI3A??

Posted: Fri Mar 03, 2017 10:53 am
by alphanumeric
For a normal omnidirectional antenna, to get best signal in all directions, you want no obstructions. That's why there are rubber ducky antennas on Routers etc. I've noticed that the ZERO W and 3B have pads for an uFL SMT Antenna Connector. There not as obvious on the 3B but they are there. On the back side I think. I've been tempted to solder a connector on my 3B and see how it does with a rubber ducky antenna. This will almost certainly invalidate the certification though.

Re: any news yet on the PI3A??

Posted: Fri Mar 03, 2017 11:23 am
by fruitoftheloom
IMO there is room for 2A / 2B / 3A / 3B all based on the BCM2837, then the RPi A+ / B+ could be special order only.........the Zeros has filled the low cost space.

Re: any news yet on the PI3A??

Posted: Fri Mar 03, 2017 1:52 pm
by darkbibble
what im expecting from the name 3A is a cut down version of the 3B (no ethernet and only 1 usb port) and maybe under-clocked for lower power requirements.

Re: any news yet on the PI3A??

Posted: Sat Mar 04, 2017 8:40 pm
by gregeric
gregeric wrote:RF knowledgeable people: does a cavity antenna even need to cling to the edge of a board like other PCB antennae?
I had the good fortune today to share a few words with Eben :) He said it works best on the edge of the board. It would work in the middle of the board, likewise the RF would find a way out even if you removed this component or that, but not so optimally. I think that's the gist of what he said.

Re: any news yet on the PI3A??

Posted: Sat Mar 04, 2017 8:45 pm
by W. H. Heydt
gregeric wrote:
gregeric wrote:RF knowledgeable people: does a cavity antenna even need to cling to the edge of a board like other PCB antennae?
I had the good fortune today to share a few words with Eben :) He said it works best on the edge of the board. It would work in the middle of the board, likewise the RF would find a way out even if you removed this component or that, but not so optimally. I think that's the gist of what he said.
Interesting. Of course, now we need threads about the bash, for those of us who aren't there (and who never had a shot at going there).

Re: any news yet on the PI3A??

Posted: Sat Mar 04, 2017 9:02 pm
by mfa298
gregeric wrote:
gregeric wrote:RF knowledgeable people: does a cavity antenna even need to cling to the edge of a board like other PCB antennae?
I had the good fortune today to share a few words with Eben :) He said it works best on the edge of the board. It would work in the middle of the board, likewise the RF would find a way out even if you removed this component or that, but not so optimally. I think that's the gist of what he said.
That sounds reasonable. There's a reason RF often gets described as black magic :twisted:

Re: any news yet on the PI3A??

Posted: Sat Mar 04, 2017 11:08 pm
by W. H. Heydt
mfa298 wrote:
gregeric wrote:
gregeric wrote:RF knowledgeable people: does a cavity antenna even need to cling to the edge of a board like other PCB antennae?
I had the good fortune today to share a few words with Eben :) He said it works best on the edge of the board. It would work in the middle of the board, likewise the RF would find a way out even if you removed this component or that, but not so optimally. I think that's the gist of what he said.
That sounds reasonable. There's a reason RF often gets described as black magic :twisted:
One of the problems is that RF signals pass along the surface of conductors, not through the bulk material. That's part of what was behind Nikola Tesla's trick of demonstrating grabbing two electrodes with a 1 million volt AC potential across them...it was high frequency.