Stop the sirring noise!

[Grench]

Pyra is a handheld gaming console that doubles as a computer.

I have been searching for a truly open e-ink reader. Having one that is connected and doubles as a phone would be awesome.

Ideally for me... I want @hns to build me a device with all the subtle styling of a street paving brick.
6" Carta e-ink screen.
Resistive digitizer layer, stylus silo & stylus.
Double Pyra shoulder buttons on all 4 long side corners. 8 total.
Pyra battery in the center of the case for balance.
microUSB 3.0 OTG, dual USB A, microHDMI, headphone/mic jack ports on 1 long side.
SoC & charging & speakers & dual exposed microSDXC slots on the other long side.
Up to 2cm thick OK. Thinner better.
Debian Linux (or close) with a phoneDE launcher.
WiFi, BT, 4G

As far as I can tell nobody is currently making an eink reader platform free from crippling store DRM restrictions and OS captive environments. So, please make an open phone that doubles as an ereader?

 

[rygD]

Pyra is a handheld gaming console that doubles as a computer.

I have been searching for a truly open e-ink reader. Having one that is connected and doubles as a phone would be awesome.

Ideally for me... I want @hns to build me a device with all the subtle styling of a street paving brick.
6" Carta e-ink screen.
Resistive digitizer layer, stylus silo & stylus.
Double Pyra shoulder buttons on all 4 long side corners. 8 total.
Pyra battery in the center of the case for balance.
microUSB 3.0 OTG, dual USB A, microHDMI, headphone/mic jack ports on 1 long side.
SoC & charging & speakers & dual exposed microSDXC slots on the other long side.
Up to 2cm thick OK. Thinner better.
Debian Linux (or close) with a phoneDE launcher.
WiFi, BT, 4G

As far as I can tell nobody is currently making an eink reader platform free from crippling store DRM restrictions and OS captive environments. So, please make an open phone that doubles as an ereader?

That makes two of us.
[doublepost=1517896872,1517896661][/doublepost]

The '90s called (on a Phyre), they want their xtreme spellings back.

No, Fyre, because Pyra+Fone. Don't judge me!

 

[pmprog]

6" Carta e-ink screen.

A 6" phone is just too big to fit comfortably in my pocket. Really needs to be 5" at tops, but I'd be happy with an eink screen

 

[Askarus]

The Pyra is already a phone, so what?

I don't buy a phone without a keyboard.

 

[pmprog]

The Pyra is already a phone, so what?

I'm still waiting to actually get my Pyra to see how workable it would be as a phone. Where does the mic sit? Do you need to open the Pyra to talk? Then it's a case of where and how you hold it. Also, IIRC the Pyra's a little bit bigger than the Pandora, and the Pandora doesn't fit in my pocket (comfortably), so a version with an e-Ink screen (longer battery life) and a more common phone factor (specifically a bit thinner) would certainly be of interest to me

 

[PokeParadox]

I'm still waiting to actually get my Pyra to see how workable it would be as a phone. Where does the mic sit? Do you need to open the Pyra to talk? Then it's a case of where and how you hold it. Also, IIRC the Pyra's a little bit bigger than the Pandora, and the Pandora doesn't fit in my pocket (comfortably), so a version with an e-Ink screen (longer battery life) and a more common phone factor (specifically a bit thinner) would certainly be of interest to me

I think the main idea for practical "Pyra as phone" scenarios was to have paired to a bluetooth handset with a button to accept calls.

 

[Eight Bit]

Or just wired in-ear plugs with built in mic

 

[Pocak]

I have been so busy with my studies I totally missed/forgot the schematics. Are the ones attached to this thread the latest public version?

Yep.

The only thing I've found so far is that LDOIN can be connected through a 4.7µF cap to PGND instead of tied to AUX if we don't want to use the 3v3 LDO (the ENLDO pin is tied to ground, so it's permanently disabled anyway). What we currently have isn't wrong, it's just an optimization. It's mentioned in section 8.3.8 on page 11 in the TPS2505 datasheet.

Why add a component if you don't have to? The exisiting solution needs no additional capacitor.

Which capacitor is making the noise?

Good question for @hns. Do all five output caps need replacement? Does the input cap join the ensemble?

Which capacitor is making the noise? Oh and C601 is 100µF while the datasheet recommends 10µF unless the application requires more, is it a typo or to we really need 100µF?

Interesting, we don't seem to fall into any of the cases listed in section 8.3.21.2. Maybe it's to fight the inductance of the ferrite bead? Or to move the resonance peak below 1 MHz (of the LC circuit formed by the bead and the cap)? Less input ripple has some other advantage? EMI?

Just an idea; does the noise go away if we wire the input to VBATT endpoint instead of VSYS? I mean it's a boost converter and we enable it in software from a GPIO pin anyway... Efficiency will probably suffer a bit when the ports are enabled but it will change operating mode from no switching (essentially passthrough) to fixed 1MHz PWM. I'll leave it up to more competent people to determine if it's a good idea but for a test it would be interesting to see if it has any impact (if C601 is squeaking then it definitely should).

You seem confused about VSYS; it's not a 5V rail. VSYS is almost the same as VBATT, only separated by BATFET inside the charger chip. You'll see a significant difference only when the battery is missing . (And when the battery voltage is very low.)

Where does the mic sit?

Near the dpad and the 1 key. You can see the vent hole on some pictures.

 

[Djhg2000]

Why add a component if you don't have to? The exisiting solution needs no additional capacitor.

That is indeed a fair point, I just thought I'd mention it for completeness.

You seem confused about VSYS; it's not a 5V rail. VSYS is almost the same as VBATT, only separated by BATFET inside the charger chip. You'll see a significant difference only when the battery is missing . (And when the battery voltage is very low.)

I checked the schematic and we do in fact have have essentially the same voltage on VSYS and VBATT. Yesterday I could've sworn VSYS was a 5v bus but now I see it's definitely not.

We do however have an extra 4u7 inductor there which I didn't account for when resolving the resonance frequencies so I'll check that when I get home.

Edit:
I tried to have a look at the resonance frequency with the inductor and first of all it's 1u, not 4u7. Must've mixed up the values with the capacitor next to it when I looked at it on my phone.

The combined calculated capacitance not hidden behind a filter on VSYS is 92.898 nF. The filter inductor going into the TPS2505 circuit has a calculated value of 95.493 nH. Then there is the 100 nF capacitor between the filter and the IN pin. This is when I realized actually calculating the equivalent LC circuit goes beyond my current education. Hopefully these values are of use to somebody but I'm at my own limit. I'll see if I can ask my favorite engineer about it.

I only just now realized we're PWM driving LED-PWR, LED-L and LED-R (rated max 10 mA per color channel) straight from VSYS. The PWM frequency isn't mentioned in the IS31FL3199 datasheet but a few kHz isn't unlikely, right? I know we have input filtering to the TPS2505 but it's scarily close to the ripple being fed back from the SW pin.

The IS31FL3199 can sink up to 20 mA per output at full blast but the 680 Ohm resistors limits that to around 6 mA at full battery charge, so it's possible we're only hearing the issue on units where the LEDs aren't running at full brightness. Can anyone else look into this please?

Edit 2:
Just to be clear I still think the TPS2505 ripple, possibly combined with the 100 uF input filter capacitor, is the most likely explanation. Now I'm just trying to resolve every possible source of interference.

 

[Grench]

A 6" phone is just too big to fit comfortably in my pocket. Really needs to be 5" at tops, but I'd be happy with an eink screen

Phone is not the primary function of the above envisioned device. eReader is.
Since e-ink screens seem to require thick bezels, this would 'feel' more like a 7" phone.
If you're unable to pocket a 6" smartphone (Galaxy Note series, iPhone 7XL), then this isn't going to fit comfortably in your pocket. Neither is the Pyra. Neither is the GPD Win or GPD Win2 or GPD Pocket or Planet Computers Gemini or anything else that is going to have novel or new capabilities beyond the 4.5" smartphone you're currently holding. Your pocket is not a primary criteria.

Heck, I'm even willing to give up the 'phone' bits entirely and just have a truly open e-ink tablet running Debian with WiFi. Right now there is nothing on the market that fits that. Having a phone built in would simply be a bonus.

 

[pmprog]

Phone is not the primary function of the above envisioned device. eReader is.
Since e-ink screens seem to require thick bezels, this would 'feel' more like a 7" phone.
If you're unable to pocket a 6" smartphone (Galaxy Note series, iPhone 7XL), then this isn't going to fit comfortably in your pocket. Neither is the Pyra. Neither is the GPD Win or GPD Win2 or GPD Pocket or Planet Computers Gemini or anything else that is going to have novel or new capabilities beyond the 4.5" smartphone you're currently holding. Your pocket is not a primary criteria.

Heck, I'm even willing to give up the 'phone' bits entirely and just have a truly open e-ink tablet running Debian with WiFi. Right now there is nothing on the market that fits that. Having a phone built in would simply be a bonus.

When talking about a "Pyra Phone", I'm kinda not willing to give up the phone bits - after all, that's what I'm interested in... and as you rightly say (and I had previously mentioned), the Pyra isn't going to sit comfortably in my pocket. My current "smartphone" has a 5" display, and I think that's pushing the boundaries of my pockets.

It doesn't have to be truely unique (though the eInk screen would be a nice touch), just a non-Android Linux based phone would be nice... My Pyra and GPDPocket will cover my developer and gaming needs.

 

[rygD]

Heck, I'm even willing to give up the 'phone' bits entirely and just have a truly open e-ink tablet running Debian with WiFi. Right now there is nothing on the market that fits that. Having a phone built in would simply be a bonus.

I wouldn't think it would be much to ask for. The lack of a device like this tells me it is. The companies probably want to nickel and dime you, so something like Debian isn't a consideration.
[doublepost=1517933195,1517933058][/doublepost]Did we ever figure out what "sirring"sounds like? Is it a searing sound?

 

[ible]

i would prefer a large e-ink screen (that you can also draw on), with a mic, which is capable of being used as sheet music; software should be added to listen to keywords "next" and "back" (to navigate pages), and eventually based on the actual audio-in it should be able to change the page when necessary. although i suppose you could also wire up a usb pedal button to navigate, which might be even easier at the start.

i'd also prefer it to have 64 colors like the pebble watch, so that i can doodle in mypaint on it, but i suppose that's not a screen on the market right now.

 

[levi]

Did we ever figure out what "sirring"sounds like? Is it a searing sound?

I have a working theory that it's been said it's a misspelling of a 'whirring' sound, but I can't remember where I heard that, so I just keep using the quote as is but in quotation marks. In terms of quietening the noise down as long as we know what's causing it, we can resolve that without needing to know what it sounds like. It's only important if we were still considering living with it, which I don't think we are any more.

 

[Kippykip]

I have a working theory that it's been said it's a misspelling of a 'whirring' sound, but I can't remember where I heard that, so I just keep using the quote as is but in quotation marks. In terms of quietening the noise down as long as we know what's causing it, we can resolve that without needing to know what it sounds like. It's only important if we were still considering living with it, which I don't think we are any more.

Basically the high pitch noise that cheap phones and chargers make. I have so many cheapy electronics and CRTs around, high pitch noises don't even bother me anymore haha

 

[hns]

Back to topic: we will try to split the function of the tps2505 into two chips.
Fortunately, such chips exist:

• tps2561a for the USB-power switch and current limiter and
• tps630250rnd for the buck/boost converter to make 5V out of the battery voltage.

Interestingly the chips are smaller so that the total area they need is approx. the same as for the tps2505,
The tps630250rnd uses a different scheme for low-power-consumption clients than the tps2505.
It can run in PWM mode at 2.5GHz while the tps2505 did reduce the recharging frequency until it became audible by the piezo effect of the capacitors.
So I am quite confident that we can solve this issue while still using the planned MLCC.

 

[Grench]

Back to topic: we will try to split the function of the tps2505 into two chips.
Fortunately, such chips exist:

• tps2561a for the USB-power switch and current limiter and
• tps630250rnd for the buck/boost converter to make 5V out of the battery voltage.

Interestingly the chips are smaller so that the total area they need is approx. the same as for the tps2505,
The tps630250rnd uses a different scheme for low-power-consumption clients than the tps2505.
It can run in PWM mode at 2.5GHz while the tps2505 did reduce the recharging frequency until it became audible by the piezo effect of the capacitors.
So I am quite confident that we can solve this issue while still using the planned MLCC.

Will this require a board revision? This is on the main board with the keyboard contacts, right?

 

[CyrusDreams]

Regarding the Tantalum issue, I recommend two options:
1. go for Niobium-Oxide. Same performance but the stuff is mined mostly in Brasil (see wikipedia). Although Niobium is sometimes also listed as a conflict material; this is because is occurs in Coltan, the source ore of Tantalum. But for me, Niobium-Oxide is conflict free enough. (also doesn't go BANG)
2. go with KEMET, I know that KEMET has its own mines and its own smelters. And since it is an American company, it is closely monitored for conflict materials. Very extensive documentation about the issue on their webpage, too.

As a side note, I would recommend Tantalum-Polymer instead of Tantalum capacitors, as they go "bang - open" instead of "bang - burn" in case of a failure.

Best regards,
CyrusDreams

 

[levi]

Will this require a board revision? This is on the main board with the keyboard contacts, right?

That's my understanding yes. But a board revision isn't a major concern at the moment as anything you order now will likely happen after chinese new year anyway. And the patched up CPU boards aren't being etched and laminated till after cny anyway.

 

[ashcloud]

It can run in PWM mode at 2.5GHz

I guess you mean MHz there - A boost converter with a 2.5GHz pwm would be a little too crazy.