jdb
Raspberry Pi Engineer & Forum Moderator
Raspberry Pi Engineer & Forum Moderator
Posts: 2165
Joined: Thu Jul 11, 2013 2:37 pm

Re: Timeframe for USB C power board revision

Thu Oct 24, 2019 7:33 pm

jcyr wrote:
Thu Oct 24, 2019 4:45 pm

But we are flogging a dead horse. The issue is understood and there are simple workarounds... The only thing left folks can bitch about is the lack of timeframe.
Which, in a roundabout way, validates my original point.

Timeframe/availability information is a necessary input for any company wanting to do volume manufacture.

Now, apply this to the irate requests for "WE NEED A TIMELINE!!!one" - consider a hypothetical company that wants to ship 30000 widgets that have a Pi 4 inside and, for some completely inexplicable reason, have prototyped and tested an arrangement that includes an e-marked cable/supply that *doesn't work* with Pi 4.

Also, consider the fact that the set of power supply assemblies that *don't* work with Pi 4 require extra silicon in the form of a SOP controller chip.

What would any reasonable company do in this situation? Rage at RPTL for the fact that they need their broken prototype to work in volume, or omit parts of the power supply design that stop it from working?
Rockets are loud.
https://astro-pi.org

User avatar
Imperf3kt
Posts: 2993
Joined: Tue Jun 20, 2017 12:16 am
Location: Australia

Re: Timeframe for USB C power board revision

Thu Oct 24, 2019 8:24 pm

HawaiianPi wrote:
Thu Oct 24, 2019 6:07 pm
rpdom wrote:
Thu Oct 24, 2019 5:12 pm
I'm slightly surprised that no one has started marketing a cheap adaptor that gets around the issue. A simple USB-C socket, two resistors and USB-C plug.
I think you'd just need to connect a single resistor to ground in the cable/adapter (which would be even simpler).
Not that simple, reading near the start of the thread it seems you need to cut a trace which one person achieved by lifting the physical pin of the connector off the board.
An external adaptor can't do that.
55:55:44:44:4C
52:4C:52:42:41

User avatar
davidcoton
Posts: 4269
Joined: Mon Sep 01, 2014 2:37 pm
Location: Cambridge, UK

Re: Timeframe for USB C power board revision

Thu Oct 24, 2019 11:40 pm

Imperf3kt wrote:
Thu Oct 24, 2019 8:24 pm
An external adaptor can't do that.
Isn't that equivalent to leaving the pin in the plug disconnected? Then put a resistor from the (disconnected) wire to ground in the plug.
It's all in the link you listed....
Signature retired

drgeoff
Posts: 9933
Joined: Wed Jan 25, 2012 6:39 pm

Re: Timeframe for USB C power board revision

Fri Oct 25, 2019 12:14 am

davidcoton wrote:
Thu Oct 24, 2019 11:40 pm
Imperf3kt wrote:
Thu Oct 24, 2019 8:24 pm
An external adaptor can't do that.
Isn't that equivalent to leaving the pin in the plug disconnected? Then put a resistor from the (disconnected) wire to ground in the plug.
It's all in the link you listed....
Not according to my reading of https://hackaday.com/2019/07/16/explori ... -in-depth/

Currently the RPi4B has joined the two CC lines and put a common resistor to ground. A smart PSU expects to see the two CC lines separate and each with their own resistor to ground. Only way to do that is to disconnect both CC lines from the pins of the USB-C plug and put a resistor from each line to ground, ie two resistors.

User avatar
HawaiianPi
Posts: 4888
Joined: Mon Apr 08, 2013 4:53 am
Location: Aloha, Oregon USA

Re: Timeframe for USB C power board revision

Fri Oct 25, 2019 12:35 am

drgeoff wrote:
Fri Oct 25, 2019 12:14 am
Only way to do that is to disconnect both CC lines from the pins of the USB-C plug and put a resistor from each line to ground, ie two resistors.
Nope, as it's already been done with just one pin lifted and a single resistor to ground. Here's a pic from that thread.
RPi4B-USBC-MOD-sm.jpg
RPi4B-USBC-MOD-sm.jpg (70.58 KiB) Viewed 1377 times

I was speculating that it could be done externally with an adapter that adds the second resistor, and doesn't pass that connection through to the Pi itself. So the PSU would see both resistors (one in the Pi and the other in the adapter) as separate lines, as desired. Although I'll admit it was a spur-of-the-moment thought, and I haven't looked into it too deeply yet.
My mind is like a browser. 27 tabs are open, 9 aren't responding,
lots of pop-ups...and where is that annoying music coming from?

User avatar
Imperf3kt
Posts: 2993
Joined: Tue Jun 20, 2017 12:16 am
Location: Australia

Re: Timeframe for USB C power board revision

Fri Oct 25, 2019 2:47 am

It could be done but not with only one resistor - you'd need both and both would need to be not connected to the Pi4 USB-C socket (as that would short them together, rendering the adaptor useless)

It's all written in the post with the picture.
55:55:44:44:4C
52:4C:52:42:41

User avatar
rpdom
Posts: 15618
Joined: Sun May 06, 2012 5:17 am
Location: Chelmsford, Essex, UK

Re: Timeframe for USB C power board revision

Fri Oct 25, 2019 4:08 am

Imperf3kt wrote:
Fri Oct 25, 2019 2:47 am
It could be done but not with only one resistor - you'd need both and both would need to be not connected to the Pi4 USB-C socket (as that would short them together, rendering the adaptor useless)
You wouldn't need to do that with an external adaptor. Just cut one of the two wires in the USB and connect that via the one resistor to ground instead. That way it won't matter if both are connected together on the Pi, as one of them is no longer connected to the cable.

User avatar
davidcoton
Posts: 4269
Joined: Mon Sep 01, 2014 2:37 pm
Location: Cambridge, UK

Re: Timeframe for USB C power board revision

Fri Oct 25, 2019 8:26 am

The only problem with the one resistor adaptor is that the other CC line will still see two resistors in parallel -- so half value. I have not checked to spec to see if that is within tolerance. If not, it is possible that the cable would work with the plug inserted one way round only. Nightmare!
Signature retired

User avatar
clicky
Posts: 416
Joined: Thu Oct 25, 2012 7:34 am

Re: Timeframe for USB C power board revision

Fri Oct 25, 2019 2:24 pm

0.5g is not small amount of salt. Come on - we need USB boot fixed! Oh, sorry - wrong topic!

Or, should I start with 'take it all with pinch of salt'?

Now seriously:
rpdom wrote: You wouldn't need to do that with an external adaptor. Just cut one of the two wires in the USB and connect that via the one resistor to ground instead. That way it won't matter if both are connected together on the Pi, as one of them is no longer connected to the cable.
That seems quite a good idea. Original PSU cannot get two of SSDs powered at the same time and I got an itch to try to rsync one to another and thus kickstart another Pi4 machine (not 100% sure what for, but I have it and why not?!).

Would someone be able to make some post/blog with pictures of how-to? Diagram - which pins (wires) to be cut so such resistor can be added. I have spare Apple adapter for that, one off operation (two SSDs on two USB3 ports on one Pi4). I am happy to source and sacrifice cheap, short USB-C to USB-C cable for operation...

drgeoff
Posts: 9933
Joined: Wed Jan 25, 2012 6:39 pm

Re: Timeframe for USB C power board revision

Fri Oct 25, 2019 4:39 pm

clicky wrote:
Fri Oct 25, 2019 2:24 pm
Would someone be able to make some post/blog with pictures of how-to? Diagram - which pins (wires) to be cut so such resistor can be added. I have spare Apple adapter for that, one off operation (two SSDs on two USB3 ports on one Pi4). I am happy to source and sacrifice cheap, short USB-C to USB-C cable for operation...
Figure it out for yourself. Top left corner of https://www.raspberrypi.org/documentati ... educed.pdf and https://en.wikipedia.org/wiki/USB-C

jcyr
Posts: 505
Joined: Sun Apr 23, 2017 1:31 pm
Location: Atlanta

Re: Timeframe for USB C power board revision

Fri Oct 25, 2019 5:12 pm

clicky wrote:
Fri Oct 25, 2019 2:24 pm
Would someone be able to make some post/blog with pictures of how-to? Diagram - which pins (wires) to be cut so such resistor can be added. I have spare Apple adapter for that, one off operation (two SSDs on two USB3 ports on one Pi4). I am happy to source and sacrifice cheap, short USB-C to USB-C cable for operation...
Here's a schematic. If you're using a premade male to female USB-C cable, only one wire needs to be cut. Cut either CC1 or CC2 and connect the female side to ground through a 5.1 KOhm resistor.

Personally I'd go with the onboard mod pictured above, rather than mangling a perfectly good USB-C cable.
Attachments
USB-C_to_PI4.png
USB-C_to_PI4.png (10.77 KiB) Viewed 1152 times
It's um...uh...well it's kinda like...and it's got a bit of...

chwe
Posts: 130
Joined: Tue Jul 31, 2018 1:35 pm

Re: Timeframe for USB C power board revision

Fri Oct 25, 2019 5:59 pm

clicky wrote:
Fri Oct 25, 2019 2:24 pm
That seems quite a good idea. Original PSU cannot get two of SSDs powered at the same time and I got an itch to try to rsync one to another and thus kickstart another Pi4 machine (not 100% sure what for, but I have it and why not?!).

Would someone be able to make some post/blog with pictures of how-to? Diagram - which pins (wires) to be cut so such resistor can be added. I have spare Apple adapter for that, one off operation (two SSDs on two USB3 ports on one Pi4). I am happy to source and sacrifice cheap, short USB-C to USB-C cable for operation...
and you think that's gonna change something when you use a 'apple PSU' (it's not only apple, they're just the famous ones)? USB-C is rated up to 3A at 5V, even the 'apple charger' wont change that. They just go to higher voltage when delivering more juice which would need a power management IC (PMIC) capbable of this. As far as I know, currently only one SBC largely available can handle higher voltages. So you won't gain much (tbh I don't know if the RPi PSU provideds proper 5+ V at 3 amps - it powers the RPi headless under all my usecases so I'm fine with it - for NAS/mass storage applications I always opted for other SBCs than the RPi)..
jdb wrote:
Thu Oct 24, 2019 7:33 pm
jcyr wrote:
Thu Oct 24, 2019 4:45 pm

But we are flogging a dead horse. The issue is understood and there are simple workarounds... The only thing left folks can bitch about is the lack of timeframe.
Which, in a roundabout way, validates my original point.

Timeframe/availability information is a necessary input for any company wanting to do volume manufacture.

Now, apply this to the irate requests for "WE NEED A TIMELINE!!!one" - consider a hypothetical company that wants to ship 30000 widgets that have a Pi 4 inside and, for some completely inexplicable reason, have prototyped and tested an arrangement that includes an e-marked cable/supply that *doesn't work* with Pi 4.

Also, consider the fact that the set of power supply assemblies that *don't* work with Pi 4 require extra silicon in the form of a SOP controller chip.
first you limit it to 'only companies request such an information' to build up your case so that the workaround is easy for them. IMO this is likely an issue which affects casual users more (e.g. your RPi sits on the bench and is powered a few ours/day/week etc and you're not willing to have multiple USB-C PSUs on your desk or in a mobile setup). They might want to know when a fixed version might be available.

jdb wrote:
Thu Oct 24, 2019 7:33 pm
What would any reasonable company do in this situation? Rage at RPTL for the fact that they need their broken prototype to work in volume, or omit parts of the power supply design that stop it from working?
their hypothetical prototype is not broken, your PD implementation is. Cause the fact that the RPi doesn't work in such a configuration came out quite early it's likely that this won't happen anyway. But here's a third option just omit the RPi and opt for a board with a proper USB-C PD implementation or any other sane powering. It's not that there are no other ARM SBCs on the market.

User avatar
clicky
Posts: 416
Joined: Thu Oct 25, 2012 7:34 am

Re: Timeframe for USB C power board revision

Fri Oct 25, 2019 8:06 pm

chwe wrote:
Fri Oct 25, 2019 5:59 pm
and you think that's gonna change something when you use a 'apple PSU' (it's not only apple, they're just the famous ones)? USB-C is rated up to 3A at 5V, even the 'apple charger' wont change that. They just go to higher voltage when delivering more juice which would need a power management IC (PMIC) capbable of this. As far as I know, currently only one SBC largely available can handle higher voltages. So you won't gain much (tbh I don't know if the RPi PSU provideds proper 5+ V at 3 amps - it powers the RPi headless under all my usecases so I'm fine with it - for NAS/mass storage applications I always opted for other SBCs than the RPi)..
LOL. I don't think that's going to work - I'm just optimist that hopes it would :roll: . And it is not Apple because it is famous - but because I have two laying around (one I used at work - now use as travel charger and one I used at home - now spare/'power supply when I need it in the kitchen'). I see how easily it can be proven wrong idea. Thanks for pointing that out :)

Maybe the right option is just to wire 5V from old PC power supply directly to pins.
drgeoff wrote: Figure it out for yourself. Top left corner of https://www.raspberrypi.org/documentati ... educed.pdf and https://en.wikipedia.org/wiki/USB-C
Yeah, but, using words my 16 year old would say: "It requires effort" - and also is prone to errors from misinterpretation what I read. :oops:
jcyr wrote: Here's a schematic. If you're using a premade male to female USB-C cable, only one wire needs to be cut. Cut either CC1 or CC2 and connect the female side to ground through a 5.1 KOhm resistor.

Personally I'd go with the onboard mod pictured above, rather than mangling a perfectly good USB-C cable.
Thanks - that's first step! :) I don't have male to female - but just 'ordinary' USB-C to USB-C (beat me if I know is it male or female!) cable. I was thinking just identifying two wires there (GND and SS1 as on your schematic) - cut them (which is probably easier than trying to strip them on a small section) and solder them back with resistor between them.

jcyr
Posts: 505
Joined: Sun Apr 23, 2017 1:31 pm
Location: Atlanta

Re: Timeframe for USB C power board revision

Fri Oct 25, 2019 9:00 pm

clicky wrote:
Fri Oct 25, 2019 8:06 pm
Thanks - that's first step! :) I don't have male to female - but just 'ordinary' USB-C to USB-C (beat me if I know is it male or female!) cable. I was thinking just identifying two wires there (GND and SS1 as on your schematic) - cut them (which is probably easier than trying to strip them on a small section) and solder them back with resistor between them.
Wouldn't you need something like this: https://www.amazon.com/AmazonBasics-Typ ... B01GGKYYT0 (male to female, though this one is USB C to USB A... I think?) if you intend to use between your Apple supply and the Pi?

You could also modify the the Apple supply cable, rendering it useless for anything else!


[edit] Disregard all of that. I didn't realize the Apple supply has a female connector. Plain old USB-C Male to Male will do.
It's um...uh...well it's kinda like...and it's got a bit of...

glenk
Posts: 5
Joined: Sat Aug 17, 2019 1:33 am

Re: Timeframe for USB C power board revision

Fri Oct 25, 2019 10:04 pm

ehem wrote:
Wed Oct 23, 2019 9:46 pm
The official announcement was June 24th. Depending on which dates you consider comparable, 5 months would put revision 2 out in November, while 7 months would put revision 2 in January. This seems reasonably plausible.



Now glenk, you've got to weigh your options. What is your cost of waiting an additional 1-3 months? Does that cost outweigh the benefit of getting one with a proper USB-C connector? I see distinctly less hassle from not having to worry about non-interchangeable USB-C power supplies, yet as I'm likely to use the official power supply that benefit is distinctly dubious. Meanwhile waiting means delaying the project which I have in mind which has a significant cost.
Thanks for that reply eham, I appreciate the time you have taken the think about it and respond. The projects I want to do are all personal hobby related so a wait of a few months is no big deal.

Return to “General discussion”