Milliways
Posts: 304
Joined: Fri Apr 25, 2014 12:18 am

Raspberry Pi power problems.

Thu May 17, 2018 10:16 am

I had been browsing viewtopic.php?f=66&t=210914&start=50#p1316511, which is obviously not going to reach any sensible resolution.
While I was typing this entry the post was locked, but rather than let my pearls of wisdom 😉 go to waste, I will propose a "solution".

I agree with the Foundation engineers that the firmware should not be modified to work around hardware problems which may compromise operation.

One thing which has not been addressed is that the Pi power system is inadequate. The original idea to use a microUSB was a good pragmatic decision, which has outlived its time. (The situation in 2012 was quite different, when 1A "chargers" were adequate.)

Even if you CAN find a power supply which will supply the specified voltage at the rated current (I have tested many, and only found one - the "official" from the recommended suppliers are better than most, but fail to meet spec).

The REAL problem most users have is NOT power supply related, but inadequate cables. I make my own cables, BUT it is difficult to terminate adequate cables on a microUSB plug, so these still have to be constrained to short connections.

It is not always feasible to use short cables - the solution (which may be applicable to the original questioner) is to power the Pi through the 5V pin on the expansion header (in conjunction with adequate cables) and/or install a 5V regulator in proximity to the Pi.

No engineer worth his salt would attempt to supply 5V power over long (here this means 1m) cables.

User avatar
Imperf3kt
Posts: 1108
Joined: Tue Jun 20, 2017 12:16 am
Location: Australia

Re: Raspberry Pi power problems.

Thu May 17, 2018 11:41 am

I don't think I can agree with you on that.
I have been testing dozens of power methods for my Pi over the last several months, and so far of all the things I've tried, only three actually power the Pi adequately.
I am using cables that are stated to be rated 2.4A @ 5.0v by a reputable brand (Energizer) so I can almost guarantee it's not the cable in the event of low power.

The problem is the power supply chosen, and more often than not, the cable additionally.
User experiencing technical difficulties.

User avatar
HawaiianPi
Posts: 2482
Joined: Mon Apr 08, 2013 4:53 am
Location: Aloha, Oregon USA

Re: Raspberry Pi power problems.

Thu May 17, 2018 12:31 pm

Milliways wrote:
Thu May 17, 2018 10:16 am
No engineer worth his salt would attempt to supply 5V power over long (here this means 1m) cables.
1M cables are fine if they are of good quality with fat power wires. I have 1M cables with 23 AWG OFC power wires, as well as 21, 20 and even 19 AWG power wires that all deliver reliable power to a Pi3B(+) under heavy loads from a good PSU.

Now if you are talking about the typical USB 2.0 cable with 28 AWG wires, sure (some micro USB cables are even thinner), but a blanket statement that all 1M cables are bad is just wrong. I'm not saying shorter isn't better, but 1M cables can and do work.

Milliways wrote:
Thu May 17, 2018 10:16 am
One thing which has not been addressed is that the Pi power system is inadequate. The original idea to use a microUSB was a good pragmatic decision, which has outlived its time.
The problem isn't the connector, it's the fact that most so called "5V" power supplies are actually phone chargers, and many of those have mediocre voltage regulation. I have tested "5V/2.4A" chargers that dipped well below 5V at only a 1A load. Combine that with the ubiquitous crappy micro USB cable and you have a problem, which has nothing to do with the connector.

This discussion has been beaten to death many times. The usual solution proposed is to use a barrel type jack connector, but that is a horrible idea on so many levels (there are tons of different versions, and no voltage to size standards whatsoever). At least when someone connects a micro USB PSU to a Pi we know it's going to output no more than 5V DC. Maybe not with adequate current, but it's not going to burn up the Pi like using a 12V PSU with a barrel jack that "fits" the system would (and let's not get into AC vs DC on those jacks).

So micro-USB is, in fact the best choice (at least until the price of USB-C components drop considerably). What's really needed is less crappy micro USB cables and power supplies on the market.

And the next generation of the Pi will almost certainly be built on a smaller process node, resulting in more efficient products that need less power than the current Pi3B(+) models, keeping the current micro USB solution viable for awhile longer.
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?

hippy
Posts: 3337
Joined: Fri Sep 09, 2011 10:34 pm
Location: UK

Re: Raspberry Pi power problems.

Thu May 17, 2018 12:51 pm

Milliways wrote:
Thu May 17, 2018 10:16 am
I agree with the Foundation engineers that the firmware should not be modified to work around hardware problems which may compromise operation.
The thing is, if one doesn't need any 5V out of the Pi, then it can be working perfectly with no risk of compromised operation even if the low-voltage warning appears.

For the Foundation that is deemed a critical error which should be reported as such and must be resolved. For others it is perfectly acceptable, doesn't need to be reported, doesn't need to be resolved.

What one sees as a "hardware problem" the other doesn't; is simply an informed decision taken to operate outside the Foundation's recommended envelope.

Both are right when the issue is considered from each respective viewpoint.

The Foundation is not going to change their firmware further than they have but the means to do so is available for those who want to.

That resolves the issue for both even though argument over whether it is a "hardware problem" or not will continue.

Milliways
Posts: 304
Joined: Fri Apr 25, 2014 12:18 am

Re: Raspberry Pi power problems.

Thu May 17, 2018 12:59 pm

HawaiianPi wrote:
Thu May 17, 2018 12:31 pm
1M cables are fine if they are of good quality with fat power wires. I have 1M cables with 23 AWG OFC power wires, as well as 21, 20 and even 19 AWG power wires that all deliver reliable power to a Pi3B(+) under heavy loads from a good PSU.

Now if you are talking about the typical USB 2.0 cable with 28 AWG wires, sure (some micro USB cables are even thinner), but a blanket statement that all 1M cables are bad is just wrong. I'm not saying shorter isn't better, but 1M cables can and do work.
We are largely in agreement (I should have been clearer and stated > 1m); my problem with microUSB is NOT the connector, but the difficulty of connecting decent sized cable (the connectors are also difficult to source in small quantities - probably because the size of the connections make them almost impossible for non-professionals to use).

It would be possible for a manufacturer to make good microUSB cables, but the only need is for the Pi - modern phones use a higher voltage so can get away with thin cables.

I am unfamiliar with AWG, but 23AWG seems small. I doubt most people who express opinions have even measured the cable resistance, which is technically challenging in itself. I normally use 23/.011mm cable which is suitable up to 1.5m when carrying ~1A

jamesh
Raspberry Pi Engineer & Forum Moderator
Raspberry Pi Engineer & Forum Moderator
Posts: 20059
Joined: Sat Jul 30, 2011 7:41 pm

Re: Raspberry Pi power problems.

Thu May 17, 2018 2:20 pm

hippy wrote:
Thu May 17, 2018 12:51 pm
Milliways wrote:
Thu May 17, 2018 10:16 am
I agree with the Foundation engineers that the firmware should not be modified to work around hardware problems which may compromise operation.
The thing is, if one doesn't need any 5V out of the Pi, then it can be working perfectly with no risk of compromised operation even if the low-voltage warning appears.
Apart from the dramatically increased probability of, for example, SD card corruption and networking dropouts.

But apart from that, everything is fine. Yay!
Principal Software Engineer at Raspberry Pi (Trading) Ltd.
Please direct all questions to the forum, I do not do support via PM.

jamesh
Raspberry Pi Engineer & Forum Moderator
Raspberry Pi Engineer & Forum Moderator
Posts: 20059
Joined: Sat Jul 30, 2011 7:41 pm

Re: Raspberry Pi power problems.

Thu May 17, 2018 2:21 pm

Milliways wrote:
Thu May 17, 2018 12:59 pm
HawaiianPi wrote:
Thu May 17, 2018 12:31 pm
1M cables are fine if they are of good quality with fat power wires. I have 1M cables with 23 AWG OFC power wires, as well as 21, 20 and even 19 AWG power wires that all deliver reliable power to a Pi3B(+) under heavy loads from a good PSU.

Now if you are talking about the typical USB 2.0 cable with 28 AWG wires, sure (some micro USB cables are even thinner), but a blanket statement that all 1M cables are bad is just wrong. I'm not saying shorter isn't better, but 1M cables can and do work.
We are largely in agreement (I should have been clearer and stated > 1m); my problem with microUSB is NOT the connector, but the difficulty of connecting decent sized cable (the connectors are also difficult to source in small quantities - probably because the size of the connections make them almost impossible for non-professionals to use).

It would be possible for a manufacturer to make good microUSB cables, but the only need is for the Pi - modern phones use a higher voltage so can get away with thin cables.

I am unfamiliar with AWG, but 23AWG seems small. I doubt most people who express opinions have even measured the cable resistance, which is technically challenging in itself. I normally use 23/.011mm cable which is suitable up to 1.5m when carrying ~1A
It is certainly possible for manufacturers to build and supply decent power supplies with decent cables. The official one, for example....
Principal Software Engineer at Raspberry Pi (Trading) Ltd.
Please direct all questions to the forum, I do not do support via PM.

hippy
Posts: 3337
Joined: Fri Sep 09, 2011 10:34 pm
Location: UK

Re: Raspberry Pi power problems.

Thu May 17, 2018 4:03 pm

jamesh wrote:
Thu May 17, 2018 2:20 pm
hippy wrote:
Thu May 17, 2018 12:51 pm
Milliways wrote:
Thu May 17, 2018 10:16 am
I agree with the Foundation engineers that the firmware should not be modified to work around hardware problems which may compromise operation.
The thing is, if one doesn't need any 5V out of the Pi, then it can be working perfectly with no risk of compromised operation even if the low-voltage warning appears.
Apart from the dramatically increased probability of, for example, SD card corruption and networking dropouts.

But apart from that, everything is fine. Yay!
The SD Card doesn't use 5V. The network doesn't use 5V.

If I design a board which uses a Power Management chip which reliably supplies 3V3, 1V8 and whatever other voltages are needed by that board and that works reliably right down to 4V then it's absolutely fine with any power supply down to 4V.

You can power that board with 4.2V with no problems. Everything works as expected. Nothing is being compromised.

If I choose to add a low voltage warning set at 4.6V that doesn't mean supplying with 4.2V is somehow now a risk, a hardware problem has been created, or operation is compromised in any way.

That's exactly how it is with the Pi 3B+ except the low-voltage warning has been included from the outset.

If the Power Management chip works reliably down to 4V, reliably puts out all voltages used by the board and its components down to 4V, how is anything compromised when the voltage drops below the low-voltage level but remains above that 4V ?

Nothing is compromised.

Yes, anything which could be connected and requires 5V can be compromised but if nothing is using that 5V then nothing can be.

I simply don't understand how it's so difficult to understand. It is exactly like a fuel level warning in a car. The light comes on well before the fuel runs out. It coming on doesn't mean it has to be stopped immediately and parked up, recovery called. It is simply a warning that fuel is below a certain level. As long as fuel is above the amount needed it is fine.

Unlike a car's fuel supply, a power supply can remain below the low voltage warning level but above the minimum level for reliable operation and not cause any problems indefinitely.

If everything which requires a voltage is getting the voltage it requires at the current it needs, how is running with the low-voltage warnings going to cause "dramatically increased probability of, for example, SD card corruption and networking dropouts" ?

Why would it ?

Return to “Raspbian”

Who is online

Users browsing this forum: No registered users and 12 guests