Raspberry Pi Zero 2 W: Nippy stocking filler for the nerd in your life – if you can get one A new Raspberry Pi Zero launches today that comes loaded with speedier silicon, however, the price is roughly a third higher than its predecessor and shipments are limited. The increase in performance is courtesy of the quad-core 64-bit Arm Cortex-A53 CPU running at 1GHz. Memory, at 512MB, is unchanged so perhaps stick with …
I've been running an original Zero W to host my PiHole DNS adblock thingy for a few years now, and while it's never missed a beat, it'd be nice to have some overhead to try other things on it; that single core always made me concerned (probably unjustifiably, with modern schedulers etc) about, say, trying to also run some minor home automation/sensor/network logging/monitoring stuff on it at the same time.
I imagine with a quad core CPU in it, it'll take a lot more to make this new one complain.
Payday tomorrow, hopefully there'll still be stock.
Steven R
Yes it is amazing. Mine too - ~3 years near continuous usage without having to touch a thing. The only time I was forcibly reminded of its presence was when Firefox switched to default DNS over HTTPS - which of course bypassed by LAN Pihole setup so I started seeing ads until turning Firefox DNS over HTTPS off.
(If you want DNS over HTTPS with Pihole on your LAN, you'll have to implement it on the output side of the Pihole - I haven't found the time to try it.)
Why not use a proper PI for that? It's not much more money, or energy consumption, and then you have the added reliability of booting from a USB hard drive, more memory and speed (for anything else you want to add to it).
All it would take is potentially a single power cut to corrupt the SD Card and then no DNS lookups for anyone in your household. I'd also rather have anything critical to the functioning of the network to be hardwired to the LAN rather than Wifi, but that's probably me being oldschool.
A power cut is only likely to corrupt the SD card if it's being written to.
There's no reason for that kind of application to write to any nonvolatile storage at all so the risk is miniscule. Plus it is easily swapped out anyway.
- If you want to keep logs then they need sending elsewhere via syslog so you'll actually get the last few log lines where the problem happened.
Because soldering through the board prevents it from being mounted flush...
Having it there means you can still access all the lines, or even add an IDE connector if you want to use a hat.
If you can't solder, but want to tinker, I suggest you learn! It's not hard I started as an 8 year old.
"mounted flush..."
This is dirty, but a few times I've put plaster spackle into board holes to dodge the "through-hole" feature. You're stuck with drag soldering techniques, but for any type of pin headers this usually isn't a big deal (at least not here on this Pi).
FWIW, you don't HAVE to solder, there's clip in headers. Their probably not the best for serious stuff or non-prototyping (definately not for small footprints), but for a Pi they'd be O.K.
If you're new to soldering just go ahead and buy a CCTV like video microscope. 2 megapixels is fine, just make sure it's 60fps or better. The #1 thing in soldering is visibility and today almost everything is surface mount.
As far as this new Pi, I'm not impressed, especially not with the 50% price increase. While this doesn't fill the gap between itself and the Pico, I'd much prefer the Pico for 50% of the use cases I'd have with this. It'd be great if it had a standard male HDMI header (to just plug it right in to whatever), or even a standard female.
"It'd be great if it had a standard male HDMI header (to just plug it right in to whatever)"
If it did that, they'd have to move it to a different place on the board because otherwise the screen you plugged into would block the power and data ports. They could move it to a short side, but then they're just making a TV stick and not everyone wants that. If you use a short HDMI cable, you can get the same thing and have it flat against the back of the screen (standard thermal problems apply, but the original zero doesn't run very hot).
THIS. SO THIS.
It's NOT THAT HARD. There's loads of YouTube videos teaching you. After about 15 minutes of practice you'll have perhaps not mastered it, but you'll be GOOD ENOUGH. Worst case you fry a $15 board, but really just practice on something else first. Even an old Pi Zero.
Spot on.
Soldering isn't that difficult and soldering the GPIO interface is really easy. My advice is to get yourself some "helping hands" then you just dab the solder onto each pin with the soldering iron.
Personally I started soldering years ago chipping playstation ones. Never looked back and with a few years practice you can get really good. Chipping a Nintendo switch has got to be the most challenging solder I've ever done but it's great to be able to use homebrew apps. Not a pirate by the way, honest guv.
Three other tips:
1) "helping hands" are good, but for something like the Pi Zero a blob of Blu Tac is useful to hold the zero in place when you're soldering it and the header.
2) solder two opposite corner pins to hold everything in place, and easily adjustable by heating them up if things don't sit flat and parallel as you want it.
3) the best advice I was ever given is to remember the soldering iron isn't a paintbrush. The idea is to heat things up to let the solder melt between the component and board, not to melt the solder onto the iron and wipe it on.
It's a fairly easy and rewarding skill to learn, and will hold you in good stead for such work. I did years ago and never looked back, and sat here in my study I can look around and see 5 pi zero's plus three of the larger ones of various flavours. All doing sterling jobs from an internet radio to home assistant server to various displays and info screens.
Or as others have said, either use a hammer header from Pimoroni or others, wait for the headed version to become available or get a friend to do it for you.
"let the solder melt between the component and board,"
You can use a hot bed/plate to mount the board. I very recently solder some attiny402's on the bed of my 3D printer. It's simply about the silicone heater, controller and aluminum plate, so old/broken 3D printers are great to source from.
I thought about getting fancy and dedicating an old 3D printer to this _BUT_t also using the stepper motors to make it spin so I don't have to rotate the board for access/lighting... fancy stuff (at least to me).
+1 for Learn to Solder.
Also, try using the old fashioned 60/40 tin/lead solder - it flows so much better, you don't get 'dry joints'.
While manufacturers are (rightly) forced to use lead-free (we want to avoid too much lead oxide seeping into the water supply from e-waste in landfill sites, and manufacturers don't make boards by hand anyway) But as a home user you are NOT forced to use lead-free (the environment has bigger things to worry about than a few blobs of lead solder on a home-made board - and also the fumes are much less toxic than lead-free solder fumes, due to the acid needed to get lead-free solder to stick to anything..)
It makes a huge difference, as does a decent soldering iron. I recommend the 'pinecil' as a cheap and powerful TS80 clone.
60/40 is much easier to use, but it does not mix at all well with any of the Lead-free solders, so if you are making changes to a manufactured part which uses Lead-free (i.e. pretty much anything made in the last 15 or 20 years) you must also use Lead-free. Obviously the same applies in reverse to older kit.
It's not good practice to use the same soldering iron tip for both, either. I have found that tips are damaged very quickly if you swap between the types. From a hobyist's perspective it's best to stick to Lead-free these days, but if you are working with a variety of kit you will need to keep separate tips, or even two irons to avoid having to swap tips.
Make sure you get a Lead-free solder with a flux core. It seems to be far more common to find Lead-free without a core than it used to be with 60/40. There are also many different formulations, and I have found those with a high Silver content best, but this might just be down to my personal style.
Lead-free solder isn't quite as "eutetic" as 60/40. WIth 60/40, once you take the heat away there is a brief cooling period where it stays liquid, then it all turns solid pretty much at once. Lead-free seems to go through a short "solidyfying" phase, where it goes goey. In other words, make sure to keep the joint still for longer than you'd expect with 60/40.
But the biggest tip of all - and this solves nearly all the problems that "hobbyists" have with Lead-free solder - is to use a powerful iron (at least 40W, but this depends on what you are soldering) which can be set to reach a tip temperature of perhaps 350C - considerably higher than you would use with ordinary Lead solder. If you are soldering correctly you will apply heat for such a short time that it won't damage the component you are soldering, but it does take a bit of practice.
Put it this way, a few years ago I ran a successful "build a kit" workshop with children (mostly 8 - 12 years old) who had never soldered before. After a couple of trial runs, all of them managed to solder together working kits and I only had to "remake" one or two connections for a few of them.
M.
There are two main issues:
Getting the parts to be joined to the same, correct temperature as quickly as possible, and getting enough solder with enough flux to make a good quality joint.
Tinning the tip of the iron is mostly about improving heat transfer, molten solder being good at this, but you only need enough to make good contact with the pad / wire. Contact the wire first in this situation, the pads have very little heat capacity, but don't hang around:
Touch the wire with the tip, then the pad together and then dab a little solder in the hot spot, all within 3 seconds ideally.
If the pad gets too hot, it can come unstuck from the PCB, and you're having a bad day...
Ideally there should be a minimal, shiney solder meniscus, but once the board has cooled don't be afraid of reheating the joint for a second or so.
Practice by tinning wire ends of various length and gauge to get a feel for the heat, solder and flux...
Once your eye is in, any tip that fits is OK, some of the finer tips hold very little heat, and are very fiddly to get right.
Hope this helps...
I suck at soldering, but I still managed to do it a few times!
There are some headers out there for the Pi where the pins spread out and you just hammer the whole thing in with a small acrylic jig. Just Google for hammer headers. I've used those when I felt lazy, and they do work, you just have to be careful with the hammering.
I'm going to use this one for a home made CCTV camera. Small enough to fit into a dummy CCTV enclosure easily, and fast enough not to be laggy.
If it works well, I might even sell the Ring doorbell (because, you know... Amazon), but I doubt it will give it all the things which I've come to like. Convenience. Notifications on the phone app. Able to speak to someone from anywhere, and tell them just to pop the parcel in the box next to them (because they can't read?). And NO you don't need to RIP THE ROOF OFF, just pop it in the hole underneath. It's an old rabbit hutch.
Any hints from knowledgeable techies on a project available that I could copy / steal, which will do ALL the things my Ring doorbell does (apart from the cloud storage, I can do that myself on a NAS or something)??? If this doesn't exist (yet), perhaps it would be a great open source project to help with?
MotionEyeOS is a great pi based cctv solution. I've been running it for over a couple of years - cheap pi w in a dummy cctv camera case (£7 ebay) and pi 4b (again wit motioneye is installed) as the central storage detection and email alerting server which does more CPU intensive processing - you can add more pi w's cameras to the setup it if needed too.
Not on par with ring doorbells in terms of image quality and functionality but a fun project and no cloud subscription fee. Potentially you can tinker with AI and get object detection/recognition feature.
I came across PiKrellCam and intend to base my system on that.
I saw an example of it being used in wildlife trail-cams and thought it had potential. Time will tell, but at least I'll have fun along the way.
I did some motion detect on a zero which then fed into some AI to try and identify non chickens entering the chicken hut but due to networking problems and inquisitive and incontinent chickens I never quite worked out if it was actually powerful enough for this. Might give it another go!
I've ordered mine for precisely the same reason!
Well, almost precisely.
With Ring discontinuing the desktop app and forcing people to use a browser-based interface, too many useful features (such as being able to immediately see who triggered the proximity alert before having to do 2FA to log in) I decided to build my own doorbell system, but as ONVIF so I have full access to the data.
Lazarus is based on Delphi and as such its probably far more ergonomic and less complicated than most RAD/GUI setups. It may be my experience with Access but it just seems a lot easier and less complicated than others and when I play with it I find I can knock up little GUIs with ease which means you can actually SHOW people how things work without extraneous clutter. Pascal is not my language of choice but there's not much wrong with FreePascal for programming up to what I would call degree level.
As I say it could be that it fits my work methods from 35 years ago that really worked for me and I've not managed to replicate with later RADs and that may be me but if it does all you need and is as uncluttered as it can be then it's worth looking at.
I got my Zero2 and its up and running hanging off the router so when I can get off the bloody internet I might start VNCing it to death!
"I thought they were suggesting that the nerds wouldn't fit in the stocking. Given the mental image, that might be a positive spin on it."
All the best heist films have a nerd on the team and they seem to have no issues wearing the same stockings on their heads as the rest of the crew :-)
The Micro:bit is a different beast and fills a niche not really served by any of the Pis. It's more in line with some of the Arduino-alikes such as the Adafruit Feathers, but as a very cheap and very easy introduction to computing systems I can't think of anything else quite as "together" as a Micro:bit, even if I'm not really a fan of the online IDEs, but you can't have everything I suppose.
M.
512MB used to be enough for a large number of tabs in Firefox. I used an Athlon XP machine with that amount of RAM until 2008 and I regularly had a large number of tabs open, as well as the memory-hungry Azureus (written in Java) in the background. I did eventually upgrade that machine to the maximum it could handle (1.5GB) as the swapping was getting bad. Modern websites are probably a lot more memory-hungry these days though...
Depends how much swap space you had. The Pi usually doesn't want to swap at all because a lot of them just have the SD card, and using a section of that for swap is likely to kill it. You can of course use something more reliable for swap or override the configuration and take the risk, but most users will stick with the normal configuration and therefore can only count on the RAM on the board.
I would like to see a Pi mobo for a standard case. Maxxed out Pi paired with an interface board for SATA/IDE/ATX PSU as necessary with any viable mounting arrangement (3 screws or just a backplate strip - like a PCI card but not plugged into a mobo).
Plug everything in, turn it on and you are straight into Linux with a Works package and modern browser, ready to go. A minimum cost system as you are repurposing as much as you can.
Good for poorer countries where they have a lot of old legacy kit (cases, drives, psus). Swop out an old mobo running Windows 90something (with endless security issues and the browser now blocked from most of the net) and away you go.
It may exist. I'm too busy working to monitor the geek stuff so much now. :(
Sometimes the problem is not the tech, so much as making it accessible for consumers.
"[...] while it means all your existing cables will continue to work, [...]"
Have they fixed the Zero W standard case's HDMI plug problem yet? Ever since the first reviews it has been noted that even the apparently official HDMI cable won't lock securely into the board's socket. Every so often someone tells me that the outside screen is showing "no signal" - in spite of the plug being taped to the case. Need a round tuit to take the board out of the case. Then enlarge the case hole to accommodate the HDMI plug shroud for that extra 1mm reach to engage the socket's mechanical locking barbs.
The HDMI socket is set back a bit from the board edge. The cable that came with it from PiHut locked ok on a bare board - but with the case it won't fully engage. The same with other supplier's HDMI cables in my spares.
It took me a couple of hours to get it working on a display the first time. The last thing I suspected was the HDMI socket and case. It was fortuitous that I suddenly discovered it needed a hard continuous push to even make contact with the signal pins. PiHut merely said that I could return the whole order if I wasn't happy - without any indication they knew there was a problem. As it was needed immediately for a Xmas 2019 decoration I decided to keep it - and bodge it.
There was an early blog review which I can't find at the moment that picked up the problem.
Found the review about the HDMI socket/case problem.
"The one minor issue we had was that the case didn’t hold the micro-HDMI adapter dongle in place securely, and ours kept dropping out."
Published in ExpertReviews by David Ludlow 23 May 2017
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
