maemo.org - Talk

maemo.org - Talk (https://talk.maemo.org/index.php)
-   The Other Half (https://talk.maemo.org/forumdisplay.php?f=59)
-   -   TOHKBD rev2 (https://talk.maemo.org/showthread.php?t=93156)

ilpianista 2015-04-01 08:49

Re: TOHKBD rev2
 
There was a page listing most sold combinations, where is it?

KylliOrvokki 2015-04-01 09:44

Re: TOHKBD rev2
 
Quote:

Originally Posted by ilpianista (Post 1465502)
There was a page listing most sold combinations, where is it?

Here is at least one post:
http://talk.maemo.org/showpost.php?p...&postcount=922

Whippler 2015-04-01 20:02

Re: TOHKBD rev2
 
or, just search twitter for the tweet...
https://twitter.com/dirkvanleersum/s...87791808483328

dirkvl 2015-04-02 17:00

Re: TOHKBD rev2
 
domesheet are done, shipped tomorrow to me, keypad demo pieces should be sent the 8th. when they are checked we can order all keypads..

robnas 2015-04-03 05:59

Re: TOHKBD rev2
 
So somewhere between the 8th and, say 15th, you can (if all is ok) order the keypads? How long will it take and what has to be done after that?

dirkvl 2015-04-03 09:06

Re: TOHKBD rev2
 
yes, ordering before the 15th. they do not give estimations of how long it woll take and i cannot make a guess of that either.

but when they are here we can go straight to final assembly and shipping. all other parts will be ready tested waiting

droll 2015-04-04 00:37

Re: TOHKBD rev2
 
i'll do a rough guesstimate that all units will be out sometime end of May. :P ok ok, i'm jumping the gun.

we've waited so long........muuusstttttt.....wwwaaiiiitttt....a.a. ...bbiiit...mooooreeee......arrrgh! :)

dirkvl 2015-04-05 12:31

Re: TOHKBD rev2
 
eagle files now also in github

bennypr0fane 2015-04-06 22:48

Re: TOHKBD rev2
 
Quote:

Originally Posted by dirkvl (Post 1465665)
yes, ordering before the 15th. they do not give estimations of how long it woll take and i cannot make a guess of that either.

but when they are here we can go straight to final assembly and shipping. all other parts will be ready tested waiting

Which means the original ETA of mid April, which was supposed to be generous, is not holding. Did not expect that... :(
Man, it's just all about waiting, with the Jolla apparatuses these days, isn't it?

juiceme 2015-04-07 04:45

Re: TOHKBD rev2
 
Quote:

Originally Posted by bennypr0fane (Post 1465921)
Man, it's just all about waiting, with the Jolla apparatuses these days, isn't it?

Waiting makes it worthwhile, right? :D:D

pichlo 2015-04-07 07:51

Re: TOHKBD rev2
 
Quote:

Originally Posted by bennypr0fane (Post 1465921)
Which means the original ETA of mid April, which was supposed to be generous, is not holding. Did not expect that... :(
Man, it's just all about waiting, with the Jolla apparatuses these days, isn't it?

What, you haven't got used to Jolla's definition of "soon" yet? ;)

ggabriel 2015-04-07 08:21

Re: TOHKBD rev2
 
Quote:

Originally Posted by bennypr0fane (Post 1465921)
Which means the original ETA of mid April, which was supposed to be generous, is not holding. Did not expect that... :(
Man, it's just all about waiting, with the Jolla apparatuses these days, isn't it?

I think context is important here... the mid April statement was made before over reaching the goal of the project and adding more colours/options, so things obviously take longer.

dirkvl 2015-04-07 09:17

Re: TOHKBD rev2
 
Quote:

Originally Posted by ggabriel (Post 1465934)
adding more colours/options, so things obviously take longer.

well no, keypad production is the bottleneck.. which is also the only component that we had no idea how long it would take.

now at workshop soldering keyboard pcbs

ggabriel 2015-04-07 09:21

Re: TOHKBD rev2
 
Quote:

Originally Posted by dirkvl (Post 1465939)
well no, keypad production is the bottleneck..

Thanks for clarifying :) Although I would have thought that with more units requested, this task would take longer :)

Quote:

Originally Posted by dirkvl (Post 1465939)
now at workshop soldering keyboard pcbs

Nice, I can't wait! :)

dirkvl 2015-04-08 08:36

Re: TOHKBD rev2
 
Currently 48 TOHKBDs left in the shop.

When the time comes to assemble I will go in priority order mentioned before, funkyoh orders last. Like every other project, this is a batch process. So after the batch, no more.

I will have lot of spares left though, which means I will be offering TOHKBD in kit form, casings not included, some soldering and assembling.

coderus 2015-04-08 09:12

Re: TOHKBD rev2
 
Quote:

Originally Posted by dirkvl (Post 1466065)
Currently 48 TOHKBDs left in the shop.

When the time comes to assemble I will go in priority order mentioned before, funkyoh orders last. Like every other project, this is a batch process. So after the batch, no more.

I will have lot of spares left though, which means I will be offering TOHKBD in kit form, casings not included, some soldering and assembling.

Do not forget about special developer and signature editions ;)

dirkvl 2015-04-08 09:19

Re: TOHKBD rev2
 
Quote:

Originally Posted by coderus (Post 1466070)
Do not forget about special developer and signature editions ;)

I never forget devs and signature edition looks awesome.

deryo 2015-04-08 14:36

Re: TOHKBD rev2
 
I just had a look at the eagle schematics and board files and some questions arose:
  • "TOH Technical Specifications" from Jolla define some antenna keep out areas. The document says: "avoid anything else than pure cover plastic on those areas to keep antenna performance intact." If I understand the board layout correctly there are two ground layers covering these areas. Don't these shield the antennas?
  • VDD, SDA, SCL, INT and GND are directly routed from the phone to the contact surfaces. What happens when I have my Jolla with some metal (e.g. my keys) in a bag without the keyboard attached? These contacts are quite exposed and could easily be shorted. 3.3V output is switchable but I suspect it is turned on as soon as the back cover is in place and the corresponding software is installed and executed. SDA, SCL and INT use 1.8V level. Wouldn't 3.3V on that pins damage the phone? At least the EEPROM used on the TOHKBD can only manage VCC+0.5V resulting in 2.3V applied to its input pins according to a data sheet I found. Wouldn't this be damaged if SDA or SCL were shorted with 3.3V? I could not find any reference to a protection circuit inside the phone.

@dirkvl: If you plan on drilling the pad holes manually into your self made PCBs you should use 'drill-aid.ulp' to have small holes in the center of the pads. The thickness of the copper layer helps to keep the drill centered while drilling.

dirkvl 2015-04-08 15:12

Re: TOHKBD rev2
 
Quote:

Originally Posted by deryo (Post 1466100)
[*]"TOH Technical Specifications" from Jolla define some antenna keep out areas. The document says: "avoid anything else than pure cover plastic on those areas to keep antenna performance intact." If I understand the board layout correctly there are two ground layers covering these areas. Don't these shield the antennas?

Yes this is something I realized after ordering and soldering all the pcbs.. But in my use here I have noticed no issues whatsoever in signal strength, so hopefully in the less signal-generous areas this will not cause issues. But again, no signal issues on this side.

Quote:

Originally Posted by deryo (Post 1466100)
[*]VDD, SDA, SCL, INT and GND are directly routed from the phone to the contact surfaces. What happens when I have my Jolla with some metal (e.g. my keys) in a bag without the keyboard attached? These contacts are quite exposed and could easily be shorted. 3.3V output is switchable but I suspect it is turned on as soon as the back cover is in place and the corresponding software is installed and executed. SDA, SCL and INT use 1.8V level. Wouldn't 3.3V on that pins damage the phone? At least the EEPROM used on the TOHKBD can only manage VCC+0.5V resulting in 2.3V applied to its input pins according to a data sheet I found. Wouldn't this be damaged if SDA or SCL were shorted with 3.3V? I could not find any reference to a protection circuit inside the phone.

Check the connection detector sequence on the github, power is cut off when the keyboard is not connected. Should be fine!

Quote:

Originally Posted by deryo (Post 1466100)
@dirkvl: If you plan on drilling the pad holes manually into your self made PCBs you should use 'drill-aid.ulp' to have small holes in the center of the pads. The thickness of the copper layer helps to keep the drill centered while drilling.

Yeah I am just getting started, but there is already some drilling stuff in the gerbers. :)

deryo 2015-04-08 18:55

Re: TOHKBD rev2
 
1 Attachment(s)
Quote:

Originally Posted by dirkvl (Post 1466104)
Yes this is something I realized after ordering and soldering all the pcbs..

Damn! Well, you can't think of everything. I hope I too won't see any less signal strength.

Quote:

Originally Posted by dirkvl (Post 1466104)
Check the connection detector sequence on the github, power is cut off when the keyboard is not connected. Should be fine!

Did that now. Only thing I can imagine to lead to a hardware fault would be to short GND to INT pad and also 3,3V to one of SDA/SCL at the same time. This could possibly damage your EEPROM and phone because 3,3V would be switched on for a short time when pulling INT to ground.

BTW: What PNP transistor do you use? I'd like to check the collector-emitter-voltage ;) (I ended up making a small "simulation" using falstads electronic circuit simulator. The attached file can be opened with it.)

Quote:

Originally Posted by dirkvl (Post 1466104)
Yeah I am just getting started, but there is already some drilling stuff in the gerbers. :)

The drills/holes in Gerber files do contain the diameter that fits the THT device's pins. For manual drilling (without CNC) it's handy to etch a 0,1-0,3 mm hole in the copper layer so the drill centers itself. See drill-aid.ulp ;)

Kabouik 2015-04-08 23:08

Re: TOHKBD rev2
 
Quote:

Originally Posted by dirkvl (Post 1466104)
Yes this is something I realized after ordering and soldering all the pcbs.. But in my use here I have noticed no issues whatsoever in signal strength, so hopefully in the less signal-generous areas this will not cause issues. But again, no signal issues on this side.

This remains a big concern for me, as I've noticed severe malfunction of the antennas with the diptoh and the solartoh, and so did Clovis, while we have absolutely no network issues with Jolla OHs or toholed. Same observation for me in several different cities. This is something I've reported in the diptoh's thread (or solartoh's or tohkbd2's, can't remember) a few months ago, but got no real answer or testing on the matter, except another TMO member who said he didn't notice such severe network losses.

As posted in the aforementioned message, I admit I'm worrying a bit if tohkbd2 is designed the same way/covers antenna areas. It seems that not all Jolla units and/or connection providers among countries are equal (Clovis and me are French, both couldn't use phone with diptoh and/or solartoh, and there are not many other active French members here who use these OHs, so we lack additional feedback). :[

P@t 2015-04-09 09:09

Re: TOHKBD rev2
 
Quote:

Originally Posted by Kabouik (Post 1466136)
This remains a big concern for me, as I've noticed severe malfunction of the antennas with the diptoh and the solartoh, and so did Clovis, while we have absolutely no network issues with Jolla OHs or toholed. Same observation for me in several different cities. This is something I've reported in the diptoh's thread (or solartoh's or tohkbd2's, can't remember) a few months ago, but got no real answer or testing on the matter, except another TMO member who said he didn't notice such severe network losses.

As posted in the aforementioned message, I admit I'm worrying a bit if tohkbd2 is designed the same way/covers antenna areas. It seems that not all Jolla units and/or connection providers among countries are equal (Clovis and me are French, both couldn't use phone with diptoh and/or solartoh, and there are not many other active French members here who use these OHs, so we lack additional feedback). :[

I am in France (using Free) and have not seen bad signal with solartoh. I live in a dense area so signal is always quite strong anyway so not sure that this is a good place to test.
Next time I am going to some place where signal is weak, I will try to think about testing with that toh...

HtheB 2015-04-09 09:19

Re: TOHKBD rev2
 
Quote:

Originally Posted by deryo (Post 1466125)
Damn! Well, you can't think of everything. I hope I too won't see any less signal strength.

I hope that you realize that the compass will also not be usable anymore because of the magnets :p

dirkvl 2015-04-09 09:57

Re: TOHKBD rev2
 
Quote:

Originally Posted by HtheB (Post 1466155)
I hope that you realize that the compass will also not be usable anymore because of the magnets :p

We have tested it and it works fine! #surprise

HtheB 2015-04-09 10:13

Re: TOHKBD rev2
 
Quote:

Originally Posted by dirkvl (Post 1466160)
We have tested it and it works fine! #surprise

Great news!!!
Now I can find my way and never get lost...!! :p

juiceme 2015-04-09 10:33

Re: TOHKBD rev2
 
Quote:

Originally Posted by dirkvl (Post 1466160)
We have tested it and it works fine! #surprise

I tried to fool the compass once with magnets but found out it does not affect the device at all. From this I concluded that it is not a magnetic (hall device) compass at all, but rather some inertial dingbat...

MartinK 2015-04-09 11:01

Re: TOHKBD rev2
 
Quote:

Originally Posted by juiceme (Post 1466163)
I tried to fool the compass once with magnets but found out it does not affect the device at all. From this I concluded that it is not a magnetic (hall device) compass at all, but rather some inertial dingbat...

That would not work - not on any reasonable time frame, when you turn the device off, etc. But it could indeed be combining magnetic field data with data from other sensors (gyroscope, accelerometer, etc.) to get a better reading and filter out interference (such as juicemes magnets ;-) ).

juiceme 2015-04-09 11:11

Re: TOHKBD rev2
 
Quote:

Originally Posted by MartinK (Post 1466167)
That would not work - not on any reasonable time frame, when you turn the device off, etc. But it could indeed be combining magnetic field data with data from other sensors (gyroscope, accelerometer, etc.) to get a better reading and filter out interference (such as juicemes magnets ;-) ).

Could be. Or else it is just GPS+inertial without anything magnetic...
(If it is something to do with GPS then it would require the device to move a little in order for the GPS to record the diretion it is going... then inertial/acceleration fix to determine which way is the north-pointing edge... fairly complex operation.)

pichlo 2015-04-09 12:12

Re: TOHKBD rev2
 
I just HAD to try it. My Jolla's compass did not react to the magnet at all. And I tried VERY hard, from all sorts of angles and positions. The compass in a colleague's smart watch went completely berserk from about 50cm distance.

dirkvl 2015-04-09 12:19

Re: TOHKBD rev2
 
Quote:

Originally Posted by pichlo (Post 1466180)
I just HAD to try it. My Jolla's compass did not react to the magnet at all. And I tried VERY hard, from all sorts of angles and positions. The compass in a colleague's smart watch went completely berserk from about 50cm distance.

The Blackphone resets when TOHKBD is near :p

No problemo for Jolla

nieldk 2015-04-09 12:30

Re: TOHKBD rev2
 
Quote:

Originally Posted by dirkvl (Post 1466181)
The Blackphone resets when TOHKBD is near :p

No problemo for Jolla

Report that as a vulnerability :D

Kabouik 2015-04-09 12:40

Re: TOHKBD rev2
 
Quote:

Originally Posted by P@t (Post 1466154)
I am in France (using Free) and have not seen bad signal with solaroh. I live in a dense area so signal is always quite strong anyway so not sure that this is a good place to test.
Next time I am going to some place where signal is weak, I will try to think about testing with that toh...

Thanks for your input. I'm using Free too. I tried from Orléans to Brittany (and the places in between, including large cities like Rennes) as well as several smaller cities where the signal is usually very good. I also tried areas with data signal from EDGE to 4G. The connection loss was dramatic: note that I did not "try" it, I just noticed that, with solaroh or diptoh, I couldn't get any signal most of the time. At all, not just poor signal. It was quite obvious. Reboot with another OH, signal. Reboot with solaroh or diptoh, no signal. I noticed it after I realized that I couldn't call anyone or be called when I needed to, and had to reboot to try to solve it. Everytime I could get signal again (by removing the diptoh and keeping the Jolla uncovered in my jean's), I received a bunch of SMS that accumulated during the day. To be honest, sometimes it was working properly, but it seems it was getting worse over time (no scientific test on that).

Clovis observed the same with his solaroh, before he had to send it back. I think he's using Free as well. I never sent back my diptoh and solaroh, but I am still very concerned if I have the same issues with the tohkbd2. It's just that this toh is so much anticipated, I would be so disappointed if it ended up being unusable on my Jolla.

Let me stress again that I never have network issues with other toh (toholed prototype, toholed, Jolla OHs), hence my worries.

MartinK 2015-04-09 12:52

Re: TOHKBD rev2
 
Quote:

Originally Posted by juiceme (Post 1466170)
Could be. Or else it is just GPS+inertial without anything magnetic...
(If it is something to do with GPS then it would require the device to move a little in order for the GPS to record the diretion it is going... then inertial/acceleration fix to determine which way is the north-pointing edge... fairly complex operation.)

Indeed - GPS only gives you you current coordinates and that's it. To get direction of travel (and you can infer where is north from that) you need to move at some speed - from my experience it needs to be at least regular walking speed.

But that would still not work - you can turn Jolla off, turn it back on and even with disabled GPS it will still show you where is north, even when being rotated repeatedly. So the compass in the Jolla is definitely using the Earth magnetic field for reference, with possible help from other sensors to get better results.

Zero Six 2015-04-09 13:14

Re: TOHKBD rev2
 
Quote:

Originally Posted by dirkvl (Post 1466065)
Currently 48 TOHKBDs left in the shop.

When the time comes to assemble I will go in priority order mentioned before, funkyoh orders last. Like every other project, this is a batch process. So after the batch, no more.

I will have lot of spares left though, which means I will be offering TOHKBD in kit form, casings not included, some soldering and assembling.

Since I don't have a credit card, I'm depending on my friend's help on this.
I hope I'm not too late for preordering tomorrow, a TOHKB to Indonesia.

pichlo 2015-04-09 13:17

Re: TOHKBD rev2
 
Quote:

Originally Posted by pichlo (Post 1466180)
I just HAD to try it. My Jolla's compass did not react to the magnet at all. And I tried VERY hard, from all sorts of angles and positions. The compass in a colleague's smart watch went completely berserk from about 50cm distance.

I take it back. I definitely fooled my Jolla when I tried calibrating the compass with a (small and not very strong) magnet attached. So much so that it took another two calibration attempts to get it back to normal. So it looks like it does use a magnetic sensor but may do something like a long-time average to compensate for fluctuations. This may be a bad news for TOHKBD with magnets permanently attached.

nieldk 2015-04-09 13:32

Re: TOHKBD rev2
 
Quote:

Originally Posted by MartinK (Post 1466187)
Indeed - GPS only gives you you current coordinates and that's it. To get direction of travel (and you can infer where is north from that) you need to move at some speed - from my experience it needs to be at least regular walking speed.

Yes, unless, they have more antennas, which would highly increase the time needed for calculating direction (still, GPS needs to be available)

Quote:

But that would still not work - you can turn Jolla off, turn it back on and even with disabled GPS it will still show you where is north, even when being rotated repeatedly. So the compass in the Jolla is definitely using the Earth magnetic field for reference, with possible help from other sensors to get better results.
I believe this is what is going on. It is using both a mangnetic sensor, together with the gyro sensor to give you direction. Perhaps even functionality of A-GPS/GPS if available.

MartinK 2015-04-09 14:37

Re: TOHKBD rev2
 
Quote:

Originally Posted by pichlo (Post 1466193)
I take it back. I definitely fooled my Jolla when I tried calibrating the compass with a (small and not very strong) magnet attached. So much so that it took another two calibration attempts to get it back to normal. So it looks like it does use a magnetic sensor but may do something like a long-time average to compensate for fluctuations. This may be a bad news for TOHKBD with magnets permanently attached.

I think it should still work - the magnetometers (the solid state magnetic compasses used in smartphones are apparently made from 3 differently oriented magnetometers and the result compued by trigonometry according to Wikipedia)) probably measure changes in the magnetic field, not absolute values. So as long as the magnets are not moving in regards to the magnetometers & as long as the magnetic field changes caused by the device moving in the Earths magnetic field can be measured, then even a couple of strong magnets in a close proximity might not be an issue.

So maybe try to tape the magnets to you Jolla next time to see if they also influence the compass accuracy or not ? :)

dirkvl 2015-04-09 17:57

Re: TOHKBD rev2
 
New keypads are here!!!

Keypress is glorious!!! Feels like heaven

pichlo 2015-04-09 18:42

Re: TOHKBD rev2
 
Quote:

Originally Posted by dirkvl (Post 1466242)
Keypress is glorious!!! Feels like heaven

That's good PR but not good for managing expectations ;)

sfb 2015-04-09 19:20

Re: TOHKBD rev2
 
> I take it back. I definitely fooled my Jolla when I tried calibrating the compass with a (small and not very strong) magnet attached.

I could not reproduce this. Position of Compass is very stabel even exposed to quite a strong magnet.
> New keypads are here!!!

> Keypress is glorious!!! Feels like heaven

I'm so excited. Thank you for keeping us updated Dirk.


All times are GMT. The time now is 05:55.

vBulletin® Version 3.8.8