Raspberry Pi Power Supplies

[Neil Higgins]

I now have 5 (and counting) Pis, and the most tiresome aspect of them by far is the power supply issue. The Pi Foundation obviously decided that part of the solution to designing a cheap computer is to ignore power supply standards; hence the ridiculous (and I mean ridiculous) 5.2(?) volt Raspberry Pi "standard". Yes I can just go and buy another Pi-certified wall wart each time, but I'm getting tired of that. I have seen mention of alternative solutions in these pages - repurposing old PC supplies, etc. Any other innovative / compact / economical ideas out there?

[Jonathan Trites]

Not sure about compact, you could try usb 6 port power supplies. Might be overkill but probably can offer a fairly good guarantee of enough power and consolidate the wall wart needs if you have a lot of separate usb powered devices.

I have a couple of these around my place:

https://www.amazon.com/Anker-6-Port-Charger-PowerPort-iPhone/dp/B00P936188/ref=sr_1_1?ie=UTF8&qid=1530973981&sr=8-1&keywords=Anker+6+port&dpID=41LvMKC6FKL&preST=_SY300_QL70_&dpSrc=srch

[Romilly Cocking]

I use the Anker 6-ports for Pi clusters; they work very well.

--
You received this message because you are subscribed to the Google Groups "[PiDP-11]" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pidp-11+unsubscribe@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/pidp-11/f83705ea-c4e5-4878-a296-65354df2073f%40googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

--

personal:@romillyc work:@rareblog skype:romilly.cocking web:http://blog.rareschool.com/

[Whit Turner]

Agree - it is a pain - not to mention the silly micro usb connectors. I just built a custom linear supply for a Pi-based audio player and it works fairly well with LT1083 regulator boards. One thing I ran into is that the Pi runs in performance mode when booting (with the distro I am using) and can suck in almost 2 amps. I had to re-do the wiring with 18 awg to avoid a 0.1 volt drop at the board (a couple of feet of wiring between the regulator and the Pi). Power issues have plagued Pi users forever.

I have one of these running my PiDP-11/70: https://www.amazon.com/gp/product/B01N336XEU/ref=oh_aui_detailpage_o02_s00?ie=UTF8&psc=1

They claim 3.0 amps and it has an in-line switch. Pretty cost effective.

Whit

[Whit Turner]

A lot of people use Meanwell supplies - 5 volt, 3A like these: https://www.mouser.com/ProductDetail/MEAN-WELL/RS-15-5?qs=hPu8TEI4q5cBcSse2JHxJA%3D%3D&gclid=Cj0KCQjwyYHaBRDvARIsAHkAXctJw75IjfSdUUQt4T2M5HXfvmCZ-JMXe95JX0nkdTC8Kim7TOYm5XQaAh-sEALw_wcB

Whit

[Randy Mongenel]

One of the most frustrating things about powering a Pi isn't a Pi-specific problem, but a cable problem. Voltage drop on cables, even short ones, varies greatly. Most USB cables weren't designed to pass 2-3A of current and some cables with 24-28ga wires can have 1-2V drops over relatively short cable lengths. I bought some inline USB voltage monitors and some of the cables I have are just absolute shit when the current kicks up over 1A. Even better is that the Pi detects this and can throttle itself to prevent resets. So much ... fun. My PiDP-8 has a high-quality 10-foot cable (for shelf mounting) that barely has a .1V drop while running. Many 5V, high-amperage (2A+) chargers output 5.1-5.3V to help compensate for the voltage drop at high currents. The Pi itself has a regulatory compliance specification of 5.0V.  /2cents

--

You received this message because you are subscribed to the Google Groups "[PiDP-11]" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pidp-11+unsubscribe@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/pidp-11/9c5570d1-7b8a-42a4-a686-2c5a7d317b59%40googlegroups.com.

[Neil Higgins]

Well, you live by the sword and die by the sword. Design a computer to run off cheap, low-spec power supplies and you’ll get substandard performance, including (for example) spontaneous re-boots that correlate with GPIO usage. That is indeed a hallmark of the Raspberry Pi.

[oscarv]

There is a reason that the 'official' Pi power supply delivers 5.35V (officially, it's 5.1V but all I measured are 5.35V): the power circuit on the Pi is pretty - economical. On the Pi 2 and 3 it is most problematic, whilst the Pi Zero has no problems at all. 

I found that even a decent lab power supply set at 5.00V would give problems on my Pi 2's. Indeed, thin cables are an issue. But it's the design of the power circuit on the Pi that makes it unreasonably sensitive. The Official Pi Power Supply is pretty cheap, and solved my problems with corrupted SD card once and for all.

Kind regards,

Oscar.

[Colin Little]

Just out of interest Dr Baker did some interesting tests on USB power cable voltage drops - http://www.smbaker.com/measuring-usb-power-cable-voltage-drop-with-my-dc-load

To unsubscribe from this group and stop receiving emails from it, send an email to pidp-11+u...@googlegroups.com.

[Whit Turner]

Completely agree on the cables often being the culprit, and also cheap connectors. The best power supply in the world can't compensate for cable drops unless it has remote sensing over separate leads.

The Pi Zero only draws a couple of hundred milliamps, so almost any random cell phone charger will be fine. Of course, performance is a fraction of the Pi 3's.

Whit

[Jonathan Trites]

I didn't specifically say it, but I will say that I have had zero issues with the Anker 6 port and the Pi 3B running the pidp8i stuff. No SD card corruption encountered so far. 

Along with Romilly's report on the use of them in his Pi clusters, that probably is the easiest alternate power supply option that I can see apart from more official Pi wall warts.

One of the advantages of that kind of 6 port power supply is that its power cable only takes up one regular power port on a power strip, rather than a wall wart which can take up the space of three. Granted, it is a separate brick possibly laying on the ground. Otherwise, I have to resort to solutions like one foot extension cables with wall warts so I can still use all of the ports on my power strips. 

[Louis Mamakos]

You can get USB charge-only cables that have larger power conductors (20 AWG), for example: https://www.amazon.com/PortaPow-Specialised-Blackberry-Transformer-Charging/dp/B00RQ5AZL6 that will have lower voltage drop over the cable.

[Whit Turner]

Good catch - there are not a lot of cables which actually spec the AWG size. I am thinking 20 AWG would be a minimum for Pi's over a couple of feet.

[Chris Smith]

For small devices, the Micro USB connectors mean you get that nice profile. But for something larger - like a PiDP11 -- or like a printer -- would there be any advantage in using the USB Type B jack?

The classic Type B jack always struck me as much more robust. It is still seen on printers, likely because printers rarely need to economise on back panel space. There are Type B panel jacks, and it could be independently wired with nice conductors to the Pi's wiring port, rather than just into the USB connector. I have heard (never tested) that the direct power port gets around some of the power distribution problems within the Pi.

The other part of this is that it seems possible to get heavier gauge printer cables. I think I can find 3ft USB printer cables (A-male, B-male) with 20AWG power lines, which should make for about 0.03 ohm each way over 3 feet, so 0.06ohm * 3A = 0.18V drop at most -- far better than smaller cables.

Finally - if I was going to make my own USB power cables, the USB 2.0 Type B connectors look to be bigger and so less finicky to solder.

[oscarv]

Chris,

Instead of USB, you can also hook up a 5V power supply to the PiDP board itself, avoiding USB connectors altogether. The power traces are extra wide, and the back panel has a round push-out slot for a normal round DC power connector.

Kind regards,

Oscar.

[Neil Higgins]

Incidentally, you'll be a better person than I am, if you can find any official (i.e. on the raspberrypi.org web site) hardware specifications for any RPi model. I don't mean just a feature list; I mean nominal values AND tolerances for operational parameters.I see lots of people asking, and no official answers. It's an obvious way to keep a product cheap - don't have any performance specifications! That is why I tell all of my acquaintances that using a Raspberry Pi in any commercial application is a HUGE risk, and I advise them against it. A number of them have ignored my advice on the basis that a $25 Linux machine is just too good a deal to pass over. Good luck to them.

[Nigel Johnson]

On Saturday, July 7, 2018 at 9:24:33 AM UTC-4, Neil Higgins wrote:

I now have 5 (and counting) Pis, and the most tiresome aspect of them by far is the power supply issue. The Pi Foundation obviously decided that part of the solution to designing a cheap computer is to ignore power supply standards; hence the ridiculous (and I mean ridiculous) 5.2(?) volt Raspberry Pi "standard". Yes I can just go and buy another Pi-certified wall wart each time, but I'm getting tired of that. I have seen mention of alternative solutions in these pages - repurposing old PC supplies, etc. Any other innovative / compact / economical ideas out there?

I have never had a problem, maybe I am lucky.  My raspbery pies are on the roof of my condo. 110V goes through two Omron industrial grade power supplies which feed a DC-DC converter with  a 7AH gel cell floating.  The cabinet is full of radio equipment that has differing loads so from reading this I am surprised.  In fact, the only problem I have had is three out of the four SDR dongles I have plugged into the pies recently went offline until I rebooted during a heat wave in Toronto an dtheequipment is in a boiler room on the roof!

I have never heard of the 5.2 Volt spec and have pretty close to 5V out of the DCDC converters.  Maybe it is just a matter of providing sufficient current, a low impedance source?

However, I teach a course in systems integration and teach my students never to use raspberry pies or arduinos in production systems.  They are OK for protoyping or hobby. (I am using them for Flightaware.com, liveatc.net, marine traffic.com and ham radio APRS.

cheers,

Nigel Johnson

[Neil Higgins]

The “standard” is like all RPi doco - unofficial. If you read their advice on how to power a RPi, they say that not all 5V USB supplies are created equal and they recommend a RPi “certified” supply. If you then go to a supplier and look at the spec for such a supply, you find that they explicitly state a nominal voltage of 5.1 - 5.2, and note that it is “suitable” for a RPi. You and others are right in stating that load regulation (and probably also transient response) is the culprit; their solution is unprofessional but cheap. It sounds like your rooftop installation takes a professional approach, which would explain why you have had no problems.