Raspberry Pi is now manufacturing 70,000 Pi 5s per week, will surge to 90,000 in February

Lee Duna@lemmy.nz to Technology@lemmy.world – 317 points –
Raspberry Pi is now manufacturing 70,000 Pi 5s per week, will surge to 90,000 in February
tomshardware.com
55

While it's good that they have been ramping up production, their attitude towards consumers during the shortage is something that some users won't forget, as well as them seemingly ignoring that they are an education charity.

At least the Pi CEO acknowledges this in the CES interview with Jeff Geerling, where he mentions that the company has been "burnt" from a customer perspective. While they do contribute a lot to mobile linux development (indirectly), I think most people here would probably prefer the company just focus on their original mission of getting an affordable, credit card sized computer into users' hands... not scalpers and hardware developers' warehouses.

Also, I personally don't really want to support Broadcom seeing the horrible decisions they've been making recently - why would they buy VMWare, then proceed to drop ALL of their partners, and put a ton of their staff out of work??

3 more...

Stop reposting this corporate press release. Fuck the Pi foundation, and frankly, fuck the tech "journalists" and YouTubers who shill and cover for their anti consumer backstabbing.

Not familiar with their anti consumer backstabbing, could you share share some links please?

There were 40,000 PI4s a week produced during Covid, the shortage on the consumer websites was because the entire production was sent to industry users, and there was the barest dribble left over for the hobbyists that made them popular.

Every time there was an increase in production, it all went to shore up backlogs in industrial orders. Why an industry player would use an rPi instead of purpose-built PLCs is beyond me, but that's what was happening.

The rPi foundation will drop hobbyists like a hot potato when the 5s start being specced for industry and we'll be back to the same shit. Pretty sure that's why they didn't bother with H265 hardware licensing, because no industry player will need that.

TL;dr - They're going to fuck you, find another source.

I deal with both PIs and PLCs for a living. I don't have much faith in the future of PLCs to be honest. They just don't seem to be willing to move forward in any sense of the word. The price for the same hardware tracks inflation, the lead times are getting worse, no version control, no higher level code development, still struggling over basic driver stuff, almost no interoperability, basic things that are wrong aren't getting fixed, almost no code sharing, everything locked down....

Basically they fit 1994 and decided to just stay there. The only good stuff they offer is greater reliability and more I/O. Right now I can buy an HMI-PI-PLC that can do everything my old systems can do and more for lower cost.

As far as guys I know that do industrial MMI, it's PLCs all the way down for reliability. They'd desperately like to avoid the Siemens and ABs of the industry, but nobody ever got fired for buying those and they cratered. Which they do, no doubt, but I wouldn't be trusting much to a hobbyist SBC.

Basically your argument is of the form: IBM is the best and there is no future in the desktop. No one is going to want to have to buy software from multiple vendors. They want the same guys who made the hardware, to make the OS, and to make the software.

You have no idea how frustrating it is to deal with companies like Rockwell and Siemens. From endless tech support licenses, to special cables, to refusing to support any new features, to lockdown down protocols. I can share my python code with anyone, not my compactlogix code. Every single sin of the tech industry you can name these guy implement.

Eventually they will lose. Eventually the PLC will either become like every other embedded dev platform or get replaced by ones that are.

Also, fuck having to deal with stl and ladder logic, if the industry minded towards more common languages and frameworks, you wouldn't need to have mechanical engineers learning plc programming, you could have actual developers working on it.

Every second that I had to spend on software like tia portal drove me further and further away from industrial automation

I had to get a CRC done in ladder once and this is my goto (hehe) war story.

Be easier to teach the mech-es I work with python. "Ok I setup some free software on your computer. Go find some tutorials online and follow it. Let me know if you get stuck" vs "alright I offered one of my children to our Rockwell Sales rep who is cough...working...cough from home. He has agree to give us a 30 day license as part of our Faustian bargain. I talked to IT and we think we can get it on the license manager sometime this week. Did you setup an account on the Rockwell site so you can read tutorials? No? Ok get on that. It has surprisingly complex password rules and requires the account manager to setup your account.

When that is done I will build you a 2 or 3 thousand dollar test system so you can test code. No there is no simulator you should trust you need physical hardware. Did you install the Studio 5000 software? Shit you installed the old version. Rslinx is broken now. Fuck fuck. Ok I can fix this, we just need to delete the old Rslinx and make sure the registry is clear. After a day or so...

Ok so download Studio 5000 from Rockwell, make sure you get firmware release 33. Yes they have updated firmwslare 33 times. That should make you very confident. Now I want you to follow good coding practices as you learn. Unfortunately since no one is legally allowed to share code and even if it was legal it would be near impossible good coding practices are defined by whomever old timer teaches you. Which is me, hi. What you don't want to learn from one flawed imperfect person you want to have an entire community sharing, growing, creating together? Hehe stop you are killing me.

Oh you are getting a weird error code? Go tell Rockwell. You can't email them. You have to call them. Use the direct dial numbers and you might reach someone within the hour who might know how to fix it. Or you can bother me and I will check my to see if my dead tree notebook has the answer."

You people think I am exaggerating, I really am not. I am understating how truly messed up it is. I do not know any controls engineer who are fully sane and not addicted rageaholics.

It's actually making me mad how close this was to my onboarding experience when I first started working with PLCs.

The test setup especially. Having to create this complex pseudo machine in order to make a preliminary proof of concept, only to find my IO block poached off my desk the next morning because someone else needed one.

That's rough. For my interns I try to shield them from most of it, start with some real basics like uploading and downloading, changing constants, cross-references etc.

On the bright side I will be employed for life.

Dang, that makes Cisco look competent and approachable by comparison

And that's saying something. Another company that went to shit when they became a quasi-monopoly.

Did they go to shit or were they always shit? They've never been able to code software worth a damned and I don't know if there's any combination of model + IOS version that's fully compatible with any other, they changed the syntax of commands pretty much every chance they got back in '05 and everything I've seen suggests they've not changed at all in that

W. T. FUCK.

And here I bitch about the bullshit UI/UX devs are doing to every piece of software I use (the usual office and file management stuff, on Windows, Linux and Android).

That's just insane. The stuff that specialized industries do to protect their profits sometimes is really effing galling. I've seen somewhat similar illogic in security and surveillance hardware. Wait, you designed it to work on a network, but you left the ethernet port off, it has to be physically added to each device? And if a device ever fails over to another path (another network port, serial, telephone, cell card) it will never retry the primary connection until we signal it too, and maybe have to go put hands on it? You know it's the 21st century, right?

It’s not about how many they can manufacture, it’s about how many they actually sell to consumers. I have given up trying to buy them. It’s just not worth the hassle.

Just checked for my country, and 4 out of 5 places had them in stock. Might be a local problem?

The issue isn’t the stock, it’s the price gouging.

The registered retailers are selling them at normal cost I think? Or I got ripped off and didn't notice πŸ€”

So what is the price supposed to be? I'm seeing ~90€ for the 8gb variant

Β£79 is official RPi5 8GB price in the UK. So €90 sounds correct.

MSRP is $80 I believe

That sounds okay with taxes and stuff

It's "okay" but you can do much better for the price.

I bought an old thinkcentre for $50 on ebay that trounces the Pi's performance.

Reduce, REUSE, recycle.

I agree that's a better device but I don't think it's fair to compare new to used. You could probably buy a used Pi4 for $20, and if that is sufficient for your needs, would be a much better deal.

I suppose looking at it from a gaming/emulation perspective. While the Pi4 was fantastic, there are definitely out there foe that usecase

Can you attach any RPi HATS to it? No? Not an alternative then.

That’s decent. They sell for over $130 on Amazon. And that’s the issue. You can check decent stock and prices here: https://raspistock.com/

Yeah okay, that is way out of line, are there really scalpers, that buy raspis? 50 bucks for shipping one of those sounds like a decent business model ... But scalpers suck nonetheless

Yes plenty of scalpers were mass buying boards to increase the shortage. Now adafruit requires an authentified account to buy them with a quantity limit.

There's plenty available for sale at non-gouged prices since production ramped back up again. Last 6 months have been fine.

Supply and demand? If you flood the market with stock, everyone can sell them and out bid each other until it's as cheap as it can get while still turning a profit. That's competition.

The fact that there isn't enough stock is why it's so easy to price gouge...

I always open these threads to find out from the experts what they recommend to replace RPi as established, novice-level mini computers, but sadly I don't see any here yet

From what I've gathered from various sources:

  • Orange Pi: Good documentation, but prices of newer models are not as affordable as previously
  • Radxa/Rock: Poor hardware support apparently
  • Pine64: Amazing hardware variety (phone, smartwatch, IP camera, soldering iron), but documentation can be hit or miss. Check the Pine64 wiki and search around for other documentation by community members
  • Khadas: Good documentation, and support directly from the hardware developers, but this comes at a cost
  • MilkV: Poor documentation - Ideal if you want to tinker
  • Libre Le Potato: Generally hear positive things about their hardware. Hundreds of these were used on a recent YT project in lieu of a Pi with great success, so may be worth a look.

Another thing to check would be Armbian's site - if something is supported by that distro, it might be worth taking a closer look at

A lot of the companies producing these "Pi killers" made them to survive the shortage, because their Pi accessories weren't selling. This means that generally they'll work great with the accessory, but support may be hit or miss outside of that.

I would lean towards Orange Pi personally, mainly due to cost and how long they've been around. Avoid the very early models as there were some overheating issues on a minority of the Allwinner chips - iirc their recent boards are using Rockchip instead.

Edit: add Libre Le Potato

Orangepi, rockchip, Arduino.

There's nothing novice about wanting to learn.

I'm really liking my orange pi 5 Plus. Wasn't able to get the 32GB version, but 16GB is realistically more than I need anyway.

Main bonus for me over RPi is the RAM and storage --- SD, eMMC, and NVME. The dual NICs and extra efficiency cores are a nice perk, too.

My top recommendation is going to be an old desktop PC. Something with an Intel processor that ends with t.

X86 just opens up so many more options over ARM.

However if you want something new, the Zima Blade seems like a good alternative at a similar price point. And even includes 2x SATA ports and a PCIe slot.

If you want to learn embedded systems, the RPi is vastly superior to an old PC because of the variety of hats and the community support around it.

It's not a perfect replacement because of increased cost, but there are plenty of sub $200 mini pc options these days. It's all included unlike the Pi which is still going to need a case, storage, and power supply. I bought one recently that blows away the Pi5 ,and it should because it cost more.

At publishing time, Raspberry Pi 4 boards were widely in stock at all the U.S. and UK outlets we checked. However, given that the Pi 5 models with 4GB and 8GB of RAM cost only $5 more than their Pi 4 equivalents, most individual makers would be right to prefer the new model.

However, companies that are using Pi 4 either within products or for enterprise use cases may want to buy more of the older board, because the Pi 5 isn't a drop-in replacement. It requires new chassis, a higher-wattage power supply and (in most use cases) an active cooler.

Adafruit had pi 5's in stock a couple weeks ago and they didn't sell out instantly. I could have ordered but decided I didn't have an immediate use for it, so it could wait.

Pi Zero 2's as of the same time were fairly easy to find. I don't know about now. Those had been extremely scarce for a while.

Pi 4's are now plentiful. But, Pi 400's (4 with a keyboard more or less) have been fairly easy to get all along.

Unless they drop the price significantly I'll stick with used x86 minis until risc-v is more viable.