The Rise of Pyra - A story about the past 6 months

[carlberry]

To be honest communities really changed in the past years, this forum is not very lifely anymore.

It's not anyone's fault but this isn't surprising, not enough users have received units so anyone doing develop struggles to get feedback. It's hard to stay active as a user on a board for a product you don't have.

 

[ClockworkCoder]

It's not anyone's fault but this isn't surprising, not enough users have received units so anyone doing develop struggles to get feedback. It's hard to stay active as a user on a board for a product you don't have.

Indeed. The community is built around the journey we've had from conception right through to production, but unfortunately the last 2 years have stagnated with a notable dearth of updates to keep the conversations going. We need units in people's hands, or at least a little more good news, to grow it back.

There's only so far you can go with bad jokes and what have you bought lately...

Thanks again, BTW, @Askarus for your efforts to stoke the Pyra and warm our spirits.

 

[netcat]

 

[Bosbeetle]

I think the most essential part is a CPU that is maintained/supported for a long period, that already has a working current kernel in place (with the associated memory and extra chips). And try to get that support involved in the project from the beginning on (might be to expensive though)
Keep the keyboard, nubs, big USB ports, Jack (perferably 2 in/out).

 

[netcat]

Exactly. We need a 64 bit cpu or we're excluded from tons of new software. Is this even a possible upgrade?? I wouldn't have thought so.

 

[pyrat]

I don't know. I think the main thing is selling all the Pyras that can reasonably be built with the available stocked components.
Then depending on how many orders remain, and how difficult it is to source components for the same design, decide whether to build more Pyras with the current design.
A new redesign now seems an academic exercise. It's obvious that after 8 or 10 years designs seem outdated, but it seems the available resources now are even less than back then . so any new design now would be as doomed or more than it was.
I don't mean to say we can't all discuss priorities, connectors and wanted features this decade. But someone has to go there, do the designs, build prototypes, write software, test it all, do a small production run, optimize, redesign maybe, do a mass production...
(and put money upfront for all of that).
For what? for 200 orders ? It makes no sense. And 100 of those orders would be from people who already have a Pyra and might forget to contribute to the current Pyra when they have a newer model.

I think there're more handheld devices today, either game-console like or microlaptop like, and it's harder to fit a new one in. And I don't like any current SOC, because they all need blobs, even to boot. I don't like the OMAP in the Pyra so much, but nowadays it's all worse. The last usable SOC was RK3399, and it's now old and I think too hot and power demanding. New Mali GPUs require blobs, TF-A is often released as binary without source, DRAM init is usually a blob... Nothing that inspires any confidence, and if I can't be sure of what the most privileged code does to my device, I could as well buy Android or Apple crap.

I think it's best to sell Pyras with the current design to those who still want them, and hope some more software contributions come this way. If one wants a longer dinasty for the Pyra, one could open source hardware designs, like case or PCBs or something in case someone wants to evolve it somehow. Maybe newer batteries can be smaller in size with the same charge, and win some room to fit some almost standard SOM (Rapsberry Pi CM4, MNT reform, or whatever non Pyra-specific). But even opening it more is a lot of work for no pay, and once done, there's no warranty anyone will take it and evolve it, so... let's just build what's already designed. If there's no time or money to do this, there's still less to do new designs. Unless someone wants to do it as a hobby and pay it from their pockets.

 

[Askarus]

Honestly, I am quite disappointed that I was no able to build and ship way more Pyras. I prepared many housings but because of the lack of PCBs and an updated Debian 12, finalizing units could not happen

 

[elw3]

Askarus proposing the pyra lite here
Do you even know how much voting that will cause?
In anycase at this point we could argue why not just make a case and slop a pi zero in there.
Actually what i would propose is to switch back to the pandora battery, make a bit of room inside, have internal usb solder pads and enough space to let me slab my usb modems/wifi cards/thumbdrives in there.

Post automatically merged: Oct 15, 2024

I think the most essential part is a CPU that is maintained/supported for a long period

Our problem on neither the pandora or pyra is the cpu support however. Its the rest around the cpu like graphics and currently even the soundcard that have too little support.

 

[Askarus]

Askarus proposing the pyra lite here
Do you even know how much voting that will cause?

It would not even be a Pyra Light.
It would be the Pyra as it is now. Most things I would remove are those things that we do not use anyways. Either because they are broken or because we can not use them properly.
The only thing that we really need is a 64bit ARM SoC with proper support.

In anycase at this point we could argue why not just make a case and slop a pi zero in there.

Unfortunately that would be even slower than the Pandora.
If the Pi 4 was open, that would be a great option.
Unfortunately, only the Pi 3 hardware went open source.
The Pi 4 would be about the same as the Pyra but much better supported.

 

[pyrat]

Our problem on neither the pandora or pyra is the cpu support however. Its the rest around the cpu like graphics and currently even the soundcard that have too little support.

But it's a SOC. Sound and video subsystems are in the same chip, not different cards or even different chips. I'm not aware that problems with graphics or sound have to do with the external circuitry?
(there was some issues because of needed screen rotation, and the sound wheel is a bit complex, but the main issues are about things in the SOC, I believe).
I mean if you change the SOC you have a new CPU, a new soundsystem and new graphics (and memory controller, etc. etc.).

 

[Askarus]

So, does this mean the project is officially dead and remaining pre-orders will not receive their units? I would like to confirm the final nails are in the coffin before moving on after over 8 years of patiently waiting.

Why should it be dead? We still have PCBs here and ED sent out new E-Mails just 2 weeks ago.

Wally is still working on the OS and aTc has pushed some updates a few days ago.
We just received new repaired PCBs from GC 2 weeks ago and I finalized 11 this week, that we will ship out to customers next week.

I do not know how the state of the PCB production is but it would not make any sense to build and prepare >100 housings if we would not ship those.

ED even got hold on a PCB with the new LTE modem, so we can test that.

I will visit ED next week and by then we'll probably discuss what the plan is for the future.

 

[Wally]

I'll start working on the OS again shortly, just arrived home last week so still getting back on my feet in terms of jetlag.



It was fun joining in on the fun of building some working Pyras with Askarus while ED was frolicking around his shop doing all the other things (I'm rather impressed at how busy he is!)

 

[oskda]

And with powkiddy and anbernic throwing consoles out ...

But those are gaming consoles, no keyboard. Pyra is/was special because it is a pocket computer with keyboard plus gaming controllers.

Some GPD are compers with keyboard plus gaming controllers but I would say they are not real pocket size.

Of course Pyra is dead.

It is sad to see how other smaller and much later projects, like for example MNT Reform and Pocket Reform, were succesful and done in much less time, while Pyra wasn't, It is sad but Pyra was missguided since begining, not doing CPU/board research until very late, doing molded plastic case with cheap-wrong companies. It woulf be cheaper and much faster milling plastic cases or even aluminium milling, becase you get much more precission and all samples in few time (in theory plastic moulding woul be cheaper, but when you need lots of rectifications/changes because of sinkring and other problems, delays, etc, finally it gets much more expensive). I said it a lot of years ago.

 

[TeDaDeS]

It is sad to see how other smaller and much later projects, like for example MNT Reform and Pocket Reform, were succesful and done in much less time, while Pyra wasn't, It is sad but Pyra was missguided since begining, not doing CPU/board research until very late, doing molded plastic case with cheap-wrong companies. It woulf be cheaper and much faster milling plastic cases or even aluminium milling, becase you get much more precission and all samples in few time (in theory plastic moulding woul be cheaper, but when you need lots of rectifications/changes because of sinkring and other problems, delays, etc, finally it gets much more expensive). I said it a lot of years ago.

The problem I see is that hobby projects have to do everything from scratch. You don't have a relationship with your (European) suppliers yet and as a tiny organization with a tiny budget you'll be last in-line to get your work done. And that work might also be done sloppy, as you're work doesn't get the attention and priority the big budget customers get. Doing stuff in Europe is more expensive to, compared to getting it done in for example China. All of that makes it a slow, more complex and expensive process. Which doesn't help.

 

[oskda]

The problem I see is that hobby projects have to do everything from scratch. You don't have a relationship with your (European) suppliers yet and as a tiny organization with a tiny budget you'll be last in-line to get your work done. And that work might also be done sloppy, as you're work doesn't get the attention and priority the big budget customers get. Doing stuff in Europe is more expensive to, compared to getting it done in for example China. All of that makes it a slow, more complex and expensive process. Which doesn't help.

Look at this:

MNT Research GmbH

It says:
© 2024 MNT Research GmbH / Fehlerstr. 8 / 12161 Berlin
Geschäftsführer*in und inhaltlich verantwortlich: Lukas F. Hartmann.
Amtsgericht Charlottenburg, Aktenzeichen: HRB 136605 B. VAT ID: DE278908891. WEEE: DE 33315564

Can you see it? Berlin, Germany, EU. And they managed to develop and succesfully complete MNT Reform notebook computer and Pocket Reform, both had less bakers/buyers than Pyra:

MNT Pocket Reform had 392 backers​

MNT Pocket Reform

A newer, smaller, lighter, more-affordable, seven-inch mini Reform laptop that remains fully open source

www.crowdsupply.com

MNT Reform had 428 backers​

MNT Reform

The open source DIY laptop for hacking, customization, and privacy

www.crowdsupply.com

 

[TeDaDeS]

Look at this:

MNT Research GmbH

It says:
© 2024 MNT Research GmbH / Fehlerstr. 8 / 12161 Berlin
Geschäftsführer*in und inhaltlich verantwortlich: Lukas F. Hartmann.
Amtsgericht Charlottenburg, Aktenzeichen: HRB 136605 B. VAT ID: DE278908891. WEEE: DE 33315564

Can you see it? Berlin, Germany, EU. And they managed to develop and succesfully complete MNT Reform notebook computer and Pocket Reform, both had less bakers/buyers than Pyra:

MNT Pocket Reform had 392 backers​

MNT Pocket Reform

A newer, smaller, lighter, more-affordable, seven-inch mini Reform laptop that remains fully open source

www.crowdsupply.com

MNT Reform had 428 backers​

MNT Reform

The open source DIY laptop for hacking, customization, and privacy

www.crowdsupply.com

True, this laptop does use an off-the-shelve computing module.
While ED could have done the same thing, to make life easier, a customer in-house board was designed making the development more complex.

And being in Europe doesn't mean the product is produced in Europe, they can still create parts outside of Europe:
Europe / Germany:
* Installation and laser engraving keycaps
* Battery
* Assembly of unit
China:
* PCB's
* Keyboard switches
* Aluminum body
* Hinges
Singapore:
* Keycaps
US:
* CPU module
Unknown:
* Display

 

[oskda]

True, this laptop does use an off-the-shelve computing module.
While ED could have done the same thing, to make life easier, a customer in-house board was designed making the development more complex.

Unfortunately ED committed near all possible erros. For me worst error wasn't using inhouse CPU board but using moulded plastic: in theory it was going to be cheaper because it was a few thousands units, but finally it was a nightmare and CNC milling on plastic block, or even on aluminum, would be cheaper and much faster (you can have samples and modifications in little time, contrary to molded plastic). I followed all those nightmares on moulded plastic during a lot of years.

And being in Europe doesn't mean the product is produced in Europe, they can still create parts outside of Europe:
Europe / Germany:
* Installation and laser engraving keycaps
* Battery
* Assembly of unit
China:
* PCB's
* Keyboard switches
* Aluminum body
* Hinges
Singapore:
* Keycaps
US:
* CPU module
Unknown:
* Display

I have not said parts (some of them or all of them) be created in Europe, I SAID it is an European/German company or initiative. And while being European, they were successful selecting parts from all world, as it is a must in most technology devices today. None expects to get all parts from one country. Pyra get most parts from outside EU.

MNT Reform and Pocket Reform are examples of being European/German and being successfully developing a very small number machines, Pyra is the other side, an example of being unsuccessful. All of them use some parts-works from other countries.

I would prefer facts would be inversed, then Pyra being successful, but facts are facts.

And as you can see, MNT Reform and Pocket Reform are smaller projects in buyers number, so your thesis of being a tiny organization, etc, are wrong: some people with small projects are successful, other aren't.

First worst problem of Pyra was choosing to use moulded plastic. If CNC milling would be chosen (like Nokia did in N9 and some Lumia, milling a block of PC, or being it in aluminum), being it made in EU or un China, it would be much faster and finally cheaper, and more important, probably Pyra would note be so delayed losing its moment. ED thought molded plastic is very cheap for a few thousand units (it was expected to ship around 2k-4k units), but he didn't think about problems with those parts: a lot of time, a lot of delays, shrinking, more imperfections, needing a lot of changes with moulds is very slow and expensive... all those problems would be solved selecting CNC, where you can have modified samples in days even in same day if CNC machine is near.

I appreciate time and effort ED has dedicated, he is a hero (we know it has been a nightmare for him), and I appreciate its tenacity, as a lot of other people would abandoned time ago, but he made same bad choices, while people on some other projects, successful projects, made better choices. Of course there are a lot of failed projects, Pyra is not an exception.

 

[Serenity]

So, does this mean the project is officially dead and remaining pre-orders will not receive their units? I would like to confirm the final nails are in the coffin before moving on after over 8 years of patiently waiting.

Sorry, but what do you expect from a device, where only one guy is building it up and another guy is responsible for the OS? And both of them are not working full time on that?
And where the costs for one device is so high because only a small group of people demands it and buy it even if the price is so high?

 

[ClockworkCoder]

Is it worth doing a survey amongst existing preorderers to get an idea of what they want / are expecting? It's hard to judge on a community that's become less active really how much support there is left in the current backlog, and it sounds inefficient to be sending out small batches of order confirmation requests and waiting however long for any responses...

For the record, I still very much want my Pyra.

 

[TeDaDeS]

Is it worth doing a survey amongst existing preorderers to get an idea of what they want / are expecting? It's hard to judge on a community that's become less active really how much support there is left in the current backlog, and it sounds inefficient to be sending out small batches of order confirmation requests and waiting however long for any responses...

For the record, I still very much want my Pyra.

If somehow magically all Pyra's would be shipped this week, there still would be work to be done to get the software stable and the community active.