(joshua appreciate you're busy with teardown, let's make this one short)
Mike got back to me, his engineer completed the review of the BOM, and we managed to work out all of the information that had been lost by Mike's long standing manager quitting with no notice.
The BOM still had components in it from the RTC, Ethernet and ONFI NAND, and we had had to change 10uF 0805 to a pair of 0603 4.7uF - all 12 or so of them - when there was that shortage caused by Apple, last year.
All of this was *supposed* to be documented as part of the pre-production runs that cost USD 2500 a shot, so that the longer runs are far less risky because a short trial run is supposed to prepare the engineers for the longer one. We cannot keep doing that, it takes 2 to three months each time, so we have to take the risk and go straight to QTY 100.
This is just how it is.
Patience is required. Success happens by solving each unknown and unknowable issue that comes up. However, it would be appreciated if people would accept that this is not something that cannot be predicted, neither what might happen next nor how long it will take. Continuously asking "when is shipping going to happen" the answer is, always and will always be: until we have actual finished product ready to put into the hands of the shipping agents, we don't know and we can't know. It really is as simple as that.
L.
Thanks for the update. your detailed updates are far better than most companies! Golden standard :)
Shame about the lack of meant to be done doc but i guess all those mailing list posts contain reminders of things to check :)
For me, Posts like these are far better than a date :)
On Friday, June 21, 2019, Alexander Ross maillist_arm-netbook@aross.me wrote:
Thanks for the update. your detailed updates are far better than most companies! Golden standard :)
Shame about the lack of meant to be done doc but i guess all those mailing list posts contain reminders of things to check :)
Preetty much yeh
For me, Posts like these are far better than a date :)
The irony is that the updates become a canonical list of Everything That Could Go Wrong on an ooen hardware project...
L.
For me, Posts like these are far better than a date :)
The irony is that the updates become a canonical list of Everything That Could Go Wrong on an ooen hardware project...
L.
Seems like a lot went wrong, Still, I am certain you will succeed. :)
And yes, better to be very late than have a product that sucks major dong. ;)
Tell me, when I showed you that A64 processor, with the sunxi link, does it look like it could be run with only free software? If so that could be your next processor to use with the eoma68 standard.
That aside, I wish you the best and hope you will succeed! I am very hopeful you will. Same with the libre risc-v processor. We need one of those for the future.
Quad core with 2.4 ghz/5glops + under 2.5 watts would be sick man. I hope as you learn more about how to do that stuff it will be achieveable. :)
On Saturday, June 22, 2019, zap calmstorm@posteo.de wrote:
For me, Posts like these are far better than a date :)
The irony is that the updates become a canonical list of Everything That
Could Go Wrong on an ooen hardware project...
L.
Seems like a lot went wrong, Still, I am certain you will succeed. :)
And yes, better to be very late than have a product that sucks major dong. ;)
Tell me, when I showed you that A64 processor, with the sunxi link, does it look like it could be run with only free software? If so that could be your next processor to use with the eoma68 standard.
I was annoyed with Allwinner and deleted the design I was working on. Also, I think they restricted the RGBTTL output to 1280x800 and EOMA68 requires 1366x768 minimum.
After using a superb LCD driver under Mr Ding's design direction the other Business Units started to license some shoddy LCD driver that restricted the framebuffer to 1MByte in size.
Any product that went beyond that *also licensed PowerVR* and there's no way in hell I'm doing a Card with a PowerVR GPU.
All that having been said, it turns out yet again that it is Allwinner Marketing Dept lying and incompetence that shoots themselves in the foot.
They started lying about the A20 a few years back, saying it is only capable of 1024x768 because otherwise it made other offerings, more "advanced", on the same marketing page, look bad.
http://linux-sunxi.org/A64 says the A64 can do up to 1920x1080p60 on the RGBTTL interface.
Sigh.
That aside, I wish you the best and hope you will succeed! I am very hopeful you will. Same with the libre risc-v processor. We need one of those for the future.
Quad core with 2.4 ghz/5glops + under 2.5 watts would be sick man. I hope as you learn more about how to do that stuff it will be achieveable. :)
:)
arm-netbook mailing list arm-netbook@lists.phcomp.co.uk http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook Send large attachments to arm-netbook@files.phcomp.co.uk
I was annoyed with Allwinner and deleted the design I was working on. Also, I think they restricted the RGBTTL output to 1280x800 and EOMA68 requires 1366x768 minimum.
After using a superb LCD driver under Mr Ding's design direction the other Business Units started to license some shoddy LCD driver that restricted the framebuffer to 1MByte in size.
Any product that went beyond that *also licensed PowerVR* and there's no way in hell I'm doing a Card with a PowerVR GPU.
All that having been said, it turns out yet again that it is Allwinner Marketing Dept lying and incompetence that shoots themselves in the foot.
They started lying about the A20 a few years back, saying it is only capable of 1024x768 because otherwise it made other offerings, more "advanced", on the same marketing page, look bad.
http://linux-sunxi.org/A64 says the A64 can do up to 1920x1080p60 on the RGBTTL interface.
Sigh.
So then, is there any 64 bit arm processor you think you could achieve this with? I mean one that is not allwinner
That aside, I wish you the best and hope you will succeed! I am very hopeful you will. Same with the libre risc-v processor. We need one of those for the future.
Quad core with 2.4 ghz/5glops + under 2.5 watts would be sick man. I hope as you learn more about how to do that stuff it will be achieveable. :)
:)
Glad to be able to lift your spirits. :)
arm-netbook mailing list arm-netbook@lists.phcomp.co.uk http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook Send large attachments to arm-netbook@files.phcomp.co.uk
On 21/06/2019 4:13 pm, Luke Kenneth Casson Leighton wrote:> On Friday, June 21, 2019, Alexander Ross maillist_arm-netbook@aross.me
wrote:
For me, Posts like these are far better than a date :)
The irony is that the updates become a canonical list of Everything That Could Go Wrong on an ooen hardware project...
also shows how your one of the few projects to get a lot of things right:
*Is possible to devolop, make stuff in china. Without years of experience. By being in china.
*You made contacts with locals like mike that have been very helpful.
*Is possible to use china tricks to save money, lower bom. Cheaper, easy to get components in china markets that do the job. *-Getting the balance right with how cheap ya go. (the li batt charger, usb power IC pain). With success with components like cap choices, headphone amp,...
Most techys avoid making stuff in china like the wind, buy everything from digikey at X markup. You sodded that and documented and helped shown it can be done. Lots of lessons like with anything but it is possible. Where’s most techys can’t face it. I do wish they adapt a can do attitude too.
*Demonstrated tons of perseverance. if someone whats someone that gets the job done and doesn’t give up easily (If at all?!), your the person to hire :)
* Example of careful considerate engineering. You thought about environmental impact of may parts of your designs.
* You didn’t keep oppies behind closed doors.
* you made sure you didn’t make stupid common proprietary mistakes, like accepting proprietary drivers. Unlike fairphone that disregarded that everything needed to use phone should be libre software. Then they had problem of there anti obsolescences phones becoming obsolescent like most phones which they marketed them selfs as not being! All because of the drivers! Then they do shiny marketing post to try and undo the damage. oow look we listening, a bit more now... urrgh well if only there where considerate in the first place!
* librem had to do shiny oow look of course intel wont release there cpu firmware like we can encouraged you to think would happen but lets do a little campaign against mutli billion company, maybe if more of us ask nicely they will give us the code with all its usa backdoors :)? huh yea right.
though they did seam to get something out of it after the campaign?? or was that someone elses work that helped save the day? i forget.
* nokia made linux phone but key drivers where secret... well that phone didn’t get so much adaptation... put me off.
* Ind.ie Sold the dream of a libre software phone, in the end delivered just the social network but for mac only.
* Ubuntu nearly got enough millions of funding for phone. yet if they had gone for a not so hard to make design, hired you :) and got the same funding they maybe they could have made a phone after all.
* intel thought eoma68 was a good idea too but just had to make there own version so they thought they could turn it into a intel only monopoly. Snag is that’s what killed it. If they had been more considerate and less fool. They would have then let go of sabotaging it by thy own monopoly effort in order for it to have a better chance of adaptation.
Refreshingly your one of the few projects/people to not make those awful mainstream mistakes. No over selling, over promising, monopolization, Making practical choices for products more in reach with small production.
The fact you have made such refreshing decisions means i am very willing to put up with lots of woes, lessons because you trying to do what others have not managed to do. I’m also quite forgiving anyway :)
On Saturday, June 22, 2019, Alexander Ross maillist_arm-netbook@aross.me wrote:
On 21/06/2019 4:13 pm, Luke Kenneth Casson Leighton wrote:> On Friday, June 21, 2019, Alexander Ross maillist_arm-netbook@aross.me
wrote:
For me, Posts like these are far better than a date :)
The irony is that the updates become a canonical list of Everything That Could Go Wrong on an ooen hardware project...
also shows how your one of the few projects to get a lot of things right:
*Is possible to devolop, make stuff in china. Without years of experience. By being in china.
This is a really funny summary, to be fair to purism they are now a Benefit Corp, and they had some influence with Intel, their requests for switching off the ME by default actually resulted in internal buzz around a "ME-less BIOS".
Then the "NSA off switch" was discovered and they were able to do FW BIOS patches for their products, without Intel's help, permission or involvement.
Even as far back as 2011 I could have compromised, released some SBC design, made some cash, established a business, and funded EOMA68 with it.
Or made some other compromise, done something a la fairphone or etc etc and got up and running that way.
The risk was, getting sidetracked by running such a business, instead of focussing on EOMA68.
Or, being criticised for compromising, undermining the whole purpose of the exercise.
What can you do, eh?
L.
This is a really funny summary, to be fair to purism they are now a Benefit Corp, and they had some influence with Intel, their requests for switching off the ME by default actually resulted in internal buzz around a "ME-less BIOS".
Then the "NSA off switch" was discovered and they were able to do FW BIOS patches for their products, without Intel's help, permission or involvement.
Even as far back as 2011 I could have compromised, released some SBC design, made some cash, established a business, and funded EOMA68 with it.
Or made some other compromise, done something a la fairphone or etc etc and got up and running that way.
The risk was, getting sidetracked by running such a business, instead of focussing on EOMA68.
Or, being criticised for compromising, undermining the whole purpose of the exercise.
What can you do, eh?
L.
Frankly we need more people who will not compromise and will do what is right for the world, the future and for the love of God, what is good for freedom and justice alike!
My point being, I thank you for going full freedom, integrity is a lost art in our world it seems like.
On Saturday, June 22, 2019, zap calmstorm@posteo.de wrote:
.
Frankly we need more people who will not compromise and will do what is right for the world, the future and for the love of God, what is good for freedom and justice alike!
It is not a lot of fun, to be honest. I am really nit surprised that more people do not put their foot down.
Then they go "err that really didn't work out, er what do we do now", and many of them regret how they shouted and blamed me, but cannot bring themselves to admit that. Several people in the samba team went through that painful lesson, 15+ years ago.
My point being, I thank you for going full freedom, integrity is a lost art in our world it seems like.
Appreciated.
arm-netbook mailing list arm-netbook@lists.phcomp.co.uk http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook Send large attachments to arm-netbook@files.phcomp.co.uk
On 21/06/2019 7:47 pm, Luke Kenneth Casson Leighton wrote:> On Saturday, June 22, 2019, Alexander Ross maillist_arm-netbook@aross.me
wrote:
On 21/06/2019 4:13 pm, Luke Kenneth Casson Leighton wrote:> On Friday, June 21, 2019, Alexander Ross maillist_arm-netbook@aross.me
wrote:
For me, Posts like these are far better than a date :)
The irony is that the updates become a canonical list of Everything That Could Go Wrong on an ooen hardware project...
also shows how your one of the few projects to get a lot of things right:
*Is possible to devolop, make stuff in china. Without years of experience. By being in china.
This is a really funny summary, to be fair to purism they are now a
Benefit
Corp, and they had some influence with Intel, their requests for switching off the ME by default actually resulted in internal buzz around a "ME-less BIOS".
Then the "NSA off switch" was discovered and they were able to do FW BIOS patches for their products, without Intel's help, permission or
involvement.
Yea i was being hard on them. I may well end up getting there librem phone. I do wonder, suspect the fuss made + there clout of money they raised and media press and high end extra well paid worker/biz person target market, may well have helped things to change.
Even as far back as 2011 I could have compromised, released some SBC design, made some cash, established a business, and funded EOMA68 with it.
Or made some other compromise, done something a la fairphone or etc
etc and
got up and running that way.
The risk was, getting sidetracked by running such a business, instead of focussing on EOMA68.
Or, being criticised for compromising, undermining the whole purpose
of the
exercise.
What can you do, eh?
yea its tricky. i see others that did those things and start to wonder but then i think, have they produced hw as free as what your work on? no. they may be about too, like librem phone but im waiting for final hardware before i believe it.
maybe doing some of what they did but a lesser of the eviles version would have helped raise more money, fund efforts, make things easier.
like pyra’s evil dragon shop selling gpd pocket laptops and other proprietary hardware of interest which helps to provide funds but it does take days out of dev time...
or as you said sbc etc...
IDK, hard to say whats right. There isn’t a right way but more annoying ways too see lessons leant :). I guess only do what ya can or think you should. As thats what happens any way :)
Then again with libre-riscv, maybe things are going to get a lot bigger for ya, being involved with that project. new floss, innovative and cheap, under cutting status quo, cpu and related parts. Big potential but long way to go but maybe you be laughing in 10years time? :D with a nice amount in bank account :)
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA256
On Fri, 21 Jun 2019 16:13:28 +0100 Luke Kenneth Casson Leighton lkcl@lkcl.net wrote:
On Friday, June 21, 2019, Alexander Ross maillist_arm-netbook@aross.me wrote:
Thanks for the update. your detailed updates are far better than most companies! Golden standard :)
Shame about the lack of meant to be done doc but i guess all those mailing list posts contain reminders of things to check :)
Preetty much yeh
For me, Posts like these are far better than a date :)
The irony is that the updates become a canonical list of Everything That Could Go Wrong on an ooen hardware project...
L.
And everyone who can go morally right. Like yourself. What it costs. What can be achieved. Why to do it. How to do it. But most notably, you show us what the industry of a failed method, the dishonest, avaricious, power hungry with a lack of solidarity method, is *really* like. That it's not all roses with rainbows and smiles like is depicted in the news, even the free speech online tech news. And thus, we are led to seek a better way...
I'll let your imagination finish my email as I must go to sleep soon. You probably know where I'm going.
David
PS: One of the things that can go wrong is misspelling the word "open". :D
On 6/21/19, Luke Kenneth Casson Leighton lkcl@lkcl.net wrote:
(joshua appreciate you're busy with teardown, let's make this one short)
Mike got back to me, his engineer completed the review of the BOM, and we managed to work out all of the information that had been lost by Mike's long standing manager quitting with no notice.
The BOM still had components in it from the RTC, Ethernet and ONFI NAND, and we had had to change 10uF 0805 to a pair of 0603 4.7uF - all 12 or so of them - when there was that shortage caused by Apple, last year.
All of this was *supposed* to be documented as part of the pre-production runs that cost USD 2500 a shot, so that the longer runs are far less risky because a short trial run is supposed to prepare the engineers for the longer one. We cannot keep doing that, it takes 2 to three months each time, so we have to take the risk and go straight to QTY 100.
This is just how it is.
Patience is required. Success happens by solving each unknown and unknowable issue that comes up. However, it would be appreciated if people would accept that this is not something that cannot be predicted, neither what might happen next nor how long it will take. Continuously asking "when is shipping going to happen" the answer is, always and will always be: until we have actual finished product ready to put into the hands of the shipping agents, we don't know and we can't know. It really is as simple as that.
L.
--
crowd-funded eco-conscious hardware: https://www.crowdsupply.com/eoma68 _______________________________________________ arm-netbook mailing list arm-netbook@lists.phcomp.co.uk http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook Send large attachments to arm-netbook@files.phcomp.co.uk
good to know, better a good product that's a bit late than a rushed dud.
On Fri, Jun 21, 2019 at 12:06:10PM +0100, Luke Kenneth Casson Leighton wrote:
...
All of this was *supposed* to be documented as part of the pre-production runs that cost USD 2500 a shot, so that the longer runs are far less risky because a short trial run is supposed to prepare the engineers for the longer one. We cannot keep doing that, it takes 2 to three months each time, so we have to take the risk and go straight to QTY 100.
...
Patience is required. Success happens by solving each unknown and unknowable issue that comes up. However, it would be appreciated if people would accept that this is not something that cannot be predicted, neither what might happen next nor how long it will take.
Thanks for the update. I have read every update so far, read most of the emails on this list and skim those not so interesting to me. Following questions bother me and I don't recall that they have been already answered: Is it still planned to send pre-production cards to those (including me) who signed up? What is the status of the last pre-production prototype cards? Do they work? When can we test / will we know if Richards HDMI-layout is working?
kind regards Pablo
On Saturday, June 22, 2019, Pablo Rath pablo@parobalth.org wrote:
On Fri, Jun 21, 2019 at 12:06:10PM +0100, Luke Kenneth Casson Leighton wrote:
...
All of this was *supposed* to be documented as part of the pre-production runs that cost USD 2500 a shot, so that the longer runs are far less
risky
because a short trial run is supposed to prepare the engineers for the longer one. We cannot keep doing that, it takes 2 to three months each time, so we have to take the risk and go straight to QTY 100.
...
Patience is required. Success happens by solving each unknown and unknowable issue that comes up. However, it would be appreciated if
people
would accept that this is not something that cannot be predicted, neither what might happen next nor how long it will take.
Thanks for the update. I have read every update so far, read most of the emails on this list and skim those not so interesting to me. Following questions bother me and I don't recall that they have been already answered: Is it still planned to send pre-production cards to those (including me) who signed up?
Yes, aiyaa I still have some at my mum's house. Keep forgetting.
What is the status of the last pre-production prototype cards? Do they work?
Yes. There are I think 3 available.
When can we test / will we know if Richards HDMI-layout is working?
When someone works out the darn fex file. It is so long in between test runs I have to work out each time what the board bringup process is, over and over again.
Sigh.
When someone works out the darn fex file.
Nowadays, you should just use a mainline linux kernel (with DT) rather than the old Allwinner kernel (with FEX). Of course, that just shifts the problem to "someone works out the darn DT file", but at least it's much better supported (and there's even an upcoming patch for that vanilla kernel which can finally lift the write speed limit on SATA from 50MB/s to 150MB/s or so).
Stefan
On Saturday, June 22, 2019, Stefan Monnier monnier@iro.umontreal.ca wrote:
When someone works out the darn fex file.
Nowadays, you should just use a mainline linux kernel (with DT) rather than the old Allwinner kernel (with FEX). Of course, that just shifts the problem to "someone works out the darn DT file", but at least it's much better supported
Full hardware support still is not complete. Also I simply do not have the time to redo months of work.
Any time I spend now basically means I get zero money from NLNet as the donations from them are based on *milestone* completion, *not* on "time spent".
4 weeks on OS rework equals ZERO PAYMENT from NLNet for that entire time.
(and there's even an upcoming patch for that
vanilla kernel which can finally lift the write speed limit on SATA from 50MB/s to 150MB/s or so).
Reason why 3.4.104+ has to be shipped was explained 18 or so months ago.
Entire OSes need to be redone otherwise.
L.
On 21 June 2019 17:13:02 GMT-04:00, Luke Kenneth Casson Leighton lkcl@lkcl.net wrote:
On Saturday, June 22, 2019, Stefan Monnier monnier@iro.umontreal.ca wrote:
When someone works out the darn fex file.
Nowadays, you should just use a mainline linux kernel (with DT)
rather
than the old Allwinner kernel (with FEX). Of course, that just
shifts
the problem to "someone works out the darn DT file", but at least
it's
much better supported
Full hardware support still is not complete. Also I simply do not have the time to redo months of work.
Any time I spend now basically means I get zero money from NLNet as the donations from them are based on *milestone* completion, *not* on "time spent".
4 weeks on OS rework equals ZERO PAYMENT from NLNet for that entire time.
(and there's even an upcoming patch for that
vanilla kernel which can finally lift the write speed limit on SATA
from
50MB/s to 150MB/s or so).
Reason why 3.4.104+ has to be shipped was explained 18 or so months ago.
Indeed. We have a working set of software. After shipping, motivated community members can look at upgrading to mainline. It's still libre software after all :).
Entire OSes need to be redone otherwise.
L.
--
crowd-funded eco-conscious hardware: https://www.crowdsupply.com/eoma68 _______________________________________________ arm-netbook mailing list arm-netbook@lists.phcomp.co.uk http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook Send large attachments to arm-netbook@files.phcomp.co.uk
On Friday 21. June 2019 22.13.02 Luke Kenneth Casson Leighton wrote:
Reason why 3.4.104+ has to be shipped was explained 18 or so months ago.
Entire OSes need to be redone otherwise.
I have tried to document some of this here:
http://rhombus-tech.net/crowdsupply/status/
Some other topics are also covered, such as changes to the Computer Card hardware during the realisation of the campaign.
I hope this summary proves to be helpful, anyway.
Paul
On Sat, Jun 29, 2019 at 10:55 PM Paul Boddie paul@boddie.org.uk wrote:
On Friday 21. June 2019 22.13.02 Luke Kenneth Casson Leighton wrote:
Reason why 3.4.104+ has to be shipped was explained 18 or so months ago.
Entire OSes need to be redone otherwise.
I have tried to document some of this here:
http://rhombus-tech.net/crowdsupply/status/
Some other topics are also covered, such as changes to the Computer Card hardware during the realisation of the campaign.
I hope this summary proves to be helpful, anyway.
it's fantastic, thank you paul.
arm-netbook@lists.phcomp.co.uk
-
Adam Van Ymeren -
Alexander Ross -
Allan Mwenda -
David Niklas -
Luke Kenneth Casson Leighton -
Pablo Rath -
Paul Boddie -
Stefan Monnier -
zap