Micro USB B vs barrel jacks

[codeboy2k]

I am considering all my future projects that need less than 5V @ 500ma (2.5W) to be powered only by USB micro B connectors, even if there is absolutely no USB functionality.

I find these connectors are becoming ubiquitous, and there are plenty of 5V phone chargers with captive cables ending in a USB micro B, or even 5V chargers with just a full size USB A receptacle. The the user would use a USB A plug to micro B plug cable. Also, that same cable can be used with a PC or laptop to power my appliance (provided it consumed less than 500mA)

If my appliance needed more than 500 mA, then it could potentially hardwire the signals like a mobile phone does to say it needs more (ala. the USB charging spec)

finally, again in the case of needing more than 500mA, I can always use a USB wall wart / plug pack that can supply 1A or 2A, etc... what ever I needed.  The caution I see here is that if my appliance consumed more than 500mA and was plugged into a standard USB port that was only able to supply 500mA, it might blow fuses on the port and make the consumer angry at my device.  I think this situation is real, and would have to be protected against, somehow, definitely like the phones do it.

This discussion could go the way of the previous forum discussion regarding Dave's uSupply and his thoughts on using USB to power the supply.  I am strictly saying that if my device needs less than 5V @ 500mA (2.5W) then I would like to use a micro USB B receptacle to power it. If I needed more than that, I would have to think carefully about using USB without also including the ability to negotiate power (ala the charging spec)

I just see that the bulky barrel connector is not really needed anymore, and the part is about 3 times the price of a usb micro B connector.

So any thoughts ? Is this a good or bad practice, a misuse of the connector, or any other reasons not to do this in the manner I've outlined it??

thanks!

[Dago]

Don't quote me on this but I think according to USB standard USB ports should be able to handle overcurrent situations in a reversible manner (ie. not with fuses but some active circuitry). Not sure how well computers conform to the standard though. At least computers tend to ignore the specification saying you should only get up to 100mA without handshaking and up to 500mA when requested, they usually just give out more than 100mA by default.

[Balaur]

I've built a rather compact headphone amplifier powered by a LiPo battery charged by a MAX1551. Works very nice.

Since the amplifier is meant to be used together with my smartphone or tablet, I've specifically selected a Micro USB connector for recharging using the same smartphone USB cable.

There is no sophistication on the USB side, just GND and +5V. Works OK with the PC or charger.

I admit that the following is not the best example, but, if you examine all the dumb and gimmicky USB-powered accessories such as lights, fans, coolers or whatever, they just connect on the 5V line.

Cheers,
Dan

PS: I forgot to say that the MAX1551 has two power inputs, drawing different currents on each input: USB - 100 mA and DC - 280 mA. I've used the MAX DC input connected to the USB +5V without any ill effects.

[jerry507]

You can use PTCs for overcurrent protection. Many people use a special switch which will give you a fault output and let you switch the port on and off. Many computers will use just a single switch for all the ports on a hub controller which is why you can pull well over 500mA per port which is the speced limit per port.

They've pretty much eliminated all the protections now except for short circuit and inrush current limit. You don't even have to provide protection against 5V on the data lines anymore.

MicroUSB connectors meeting the specs much be able to carry at least an amp on the D+, D- and ID pins and at least 1.5A on the VBUS and GND pins while carrying 0.5A on the data pins. You've got more than enough overhead in these cables for a LOT of projects, so I think you're absolutely going the right direction.

[Bored@Work]

If my appliance needed more than 500 mA, then it could potentially hardwire the signals like a mobile phone does to say it needs more (ala. the USB charging spec)

This is not how it works. The charger is hardwired. This indicates to the device it is allowed to consume more current if it desires to do so.

So the device needs special detection logic to correctly identify a high-power charger. ICs for this are available. An MCU can do it, too.

Second, even if Dave repeats it a thousand times, a normal USB port is not supposed to deliver more than 100 mA without negotiation. A device can negotiate up to 500 mA. Just because most USB crap doesn't negotiate and just suck the hell out of a USB port doesn't mean you can always do it successfully and without damaging the port.

To summarize, whatever kind of port you have (normal or one supporting the charging spec), you can only safely get more current out of a port if your device gets active.

[amyk]

I just see that the bulky barrel connector is not really needed anymore, and the part is about 3 times the price of a usb micro B connector.

The barrel connector is also a lot more robust, or to put it another way, the microUSB is quite fragile in comparison. The number of cases I've seen of a microUSB connector becoming detached from its mount on devices like phones is more than I can count...

Just because most USB crap doesn't negotiate and just suck the hell out of a USB port doesn't mean you can always do it successfully and without damaging the port

That's what the spec says and while true in theory, are there many cases in practice where ports have been damaged by just trying to get 500mA from them? I'd say it's rare...

[ve7xen]

Second, even if Dave repeats it a thousand times, a normal USB port is not supposed to deliver more than 100 mA without negotiation. A device can negotiate up to 500 mA. Just because most USB crap doesn't negotiate and just suck the hell out of a USB port doesn't mean you can always do it successfully and without damaging the port.

My reading of the spec is that as long as the port is a) not a bus-powered hub and b) not battery powered, it must supply 500mA. There is language like 'Systems that obtain operating power externally, either AC or DC, must
supply at least five unit loads to each port.' which is not qualified by the state of negotiation. Any up-to-spec powered port should work fine with 500mA drawn from it unnegotiated.

Devices are supposed to be well-behaved and negotiate properly to ensure the power budget is honoured, but should also tolerate misbehaving peripherals.

[NiHaoMike]

An easy way to detect a charger is to connect one data line to a high impedance (CMOS) input (with a 1k-100k series resistor if need be) along with a 1M or so pullup to the supply rail. A charger that shorts the data lines or one that leaves them open would let the input float high, but a USB host has a 15k or so impedance that would pull it down.

[jerry507]

USB power delivery is often very misunderstood. Most people seem to think "negotiation" means that the current limit is actively controlled. First, there are very few ICs out there that even support any current indication other than overcurrent on the entire IC .

USB current negotiation is mostly a gentlemans agreement. I say I need 300mA, you say you need 100mA and the computer knows we're under 500mA so it's fine. If I were to connect and say I need 400mA and then you connected asking for 200mA, the computer would not allow you to finish enumeration. If I were to say I needed 400mA and then proceed to use 500mA, the computer wouldn't know the difference.

The reality is that most manufacturers will use a single IC to handle fault detection for an entire hub controller of 4 ports. This means a current limit at 2A (500mA per port). This is why you can pull more than 500mA per port. The plugs and cords can easily handle the 2A so it's rarely an issue. Sometimes you'll find a laptop or something that will be more anal. Phones tend to work a bit differently and can cause confusion. The battery chargers will current limit themselves to 500mA even though they can suck down more. Special software negotiation happens on some platforms (Apple is most notable here) to allow charging at 2A. Some phones and tablets have software hacks that will let you set the charging limit higher.

ve7xen's language is all 100% correct as far as terminology and states go. NiHaoMike's methods are also pretty standard. Many manufacturers use resistor dividers on the data lines to detect different chargers and their capabilities. Apple has had several chargers over the years with different wattages, and use different dividers to indicate the different chargers. Other manufacturers use simpler methods to detect charger vs PC.

The battery charging spec is a different ballgame, and I don't know much about it. It's mostly intended to bring some order to the wild west of USB, but IMO it's a waste of time. Most everyone will follow it and life today really isn't that bad anyway.

[andersm]

I've also found Apple most stringent in monitoring the actual current consumption. On the same laptop, a USB device consuming close to 1A worked fine under Windows, but plugging it in under OS X immediately resulted in the port being disabled and an overcurrent warning popping up. (The OS then automatically re-enabled the port, generating another warning, re-enabled it again, etc. New warning windows were still popping up 10-20 seconds after I yanked the cable.)

[codeboy2k]

thanks everyone for all the good replies here... yeah, I was thinking more along the lines of what jerry said.

I guess I am also thinking like the Raspberry Pi folks here. They made their device USB powered and said you need at least 500ma to boot, and 700ma works best, and if you add USB peripherals, then its best to use a 1A or 1.5A USB power supply to get the most out of it.

So if I make a device that has no USB functionality, but needs 1.5A (!) I'd best put a warning not to plug it into a PC or Apple's USB port, and even better if I can detect when someone did that and make sure that I am current limited so as to not blow their ports. that's my real goal.

(no I am not making a competing uSupply here.. just thinking for the future, when I might not want to use barrel connectors anymore due to cost and size considerations).

Also, regarding the comment by amyk

The barrel connector is also a lot more robust, or to put it another way, the microUSB is quite fragile in comparison. The number of cases I've seen of a microUSB connector becoming detached from its mount on devices like phones is more than I can count...

Yeah, I can see that too. Some are just surface mounted, others are surface mount with PTH guides, does this help anyway??

[marshallh]

thanks everyone for all the good replies here... yeah, I was thinking more along the lines of what jerry said.

I guess I am also thinking like the Raspberry Pi folks here. They made their device USB powered and said you need at least 500ma to boot, and 700ma works best, and if you add USB peripherals, then its best to use a 1A or 1.5A USB power supply to get the most out of it.

So if I make a device that has no USB functionality, but needs 1.5A (!) I'd best put a warning not to plug it into a PC or Apple's USB port, and even better if I can detect when someone did that and make sure that I am current limited so as to not blow their ports. that's my real goal.

(no I am not making a competing uSupply here.. just thinking for the future, when I might not want to use barrel connectors anymore due to cost and size considerations).

Also, regarding the comment by amyk

The barrel connector is also a lot more robust, or to put it another way, the microUSB is quite fragile in comparison. The number of cases I've seen of a microUSB connector becoming detached from its mount on devices like phones is more than I can count...

Yeah, I can see that too. Some are just surface mounted, others are surface mount with PTH guides, does this help anyway??

The two tiny plastic mounting bosses will only increase the lateral shearing resistance. As soon as you yank the cable upwards it's all over. It simply wont beat a regular 2.1mm barrel jack.

[helloworld922]

Depending on your project you can probably get away with using a micro USB for power. However, while 2.5W can get you to the moon as far as most electronics goes, if you start working with motors, solenoids, and other electro-mechanical parts you can quickly exceed 2.5W.

Also, while yes you can use a boost converter to get higher voltages from a 5V input, there are a few reasons why you wouldn't want to:

1. Added parts to the BOM list
2. EMI considerations, especially for higher-frequency converters
3. Additional board space
4. Designer's time
5. Efficiency losses and max power limits as mentioned above

That's not to say that a USB with a boost converter is never the way to go, just be aware of the gotcha's if you do go that route. For example, the designer's time may not be as big a consideration for a hobbyist as it would be for a full-time engineer.

[DavidDLC]

Me as a user don't like the idea to have a USB connector without the USB functionality, it can be confusing, even if you put a label sometime users forget to read it everytime.

[jaqie]

I'm of two minds on this...

microUSB is one very well defined size and shape and single spec, unless it's some kind of cheap crap you know at least it'll be 5vdc and itll plug in to any other micro usb out there.

barrel jacks... oh gawd what a mess of sizes, specs, voltages, ampere ratings, proprietary versions, et al ad nauseum.

I personally think that if barrel jacks all had the retention ring and clip inside as well as a very specific set of jack dimension specifications for different voltage ratings which had to be adhered to, I would want them and nothing but...  as it is, it's just an unholy hell of a mess.

considering this, I kind of like microUSB for 5v up to 1A and barrel for everything above those needs.

As an aside, microUSB could have color specs, a red anodized for over 500uA draw and plain/metal for the standard ones used now... *shrug* but it's all already out there.

[codeboy2k]

Depending on your project you can probably get away with using a micro USB for power. However, while 2.5W can get you to the moon as far as most electronics goes, if you start working with motors, solenoids, and other electro-mechanical parts you can quickly exceed 2.5W.

Also, while yes you can use a boost converter to get higher voltages from a 5V input, there are a few reasons why you wouldn't want to:

1. Added parts to the BOM list
2. EMI considerations, especially for higher-frequency converters
3. Additional board space
4. Designer's time
5. Efficiency losses and max power limits as mentioned above

That's not to say that a USB with a boost converter is never the way to go, just be aware of the gotcha's if you do go that route. For example, the designer's time may not be as big a consideration for a hobbyist as it would be for a full-time engineer.

I am not likely to run any motors from USB. Solenoids, hmmm...probably not.  Possibly, maybe, some 5V relays on board if the design needed relays.  But with any of those electro-mechanical devices, I would have to do a very careful accounting of the power budget.  I'm no hobbyist anyways, and with the rare exception of relays, I don't use those parts very often.

Boost converters are  easy if I needed 12V or so for something on a design, I'm not fearful of putting one in, given that it can still pass EMC. However, it makes no sense to save $1.50 on a barrel plug, just to put a USB micro connector and only have 5V and need to add $1.50 in parts or more to boost up to 12V. For a situation like that, I'd start with a barrel plug and spec a 12V DC external SMPS and buck down from there to my power rails.

So I am probably just considering a USB micro connector for designs that need only 5V  and lower, and less than 2.5W (2.25W more likely)

I don't have any design in mind at the moment that would go this way, but I started this discussion just thinking about the possibility for the future, and see how other designers felt about using it for designs that meet that criteria.

It's been a good discussion so far !  Thanks for everyones input.

[codeboy2k]

The two tiny plastic mounting bosses will only increase the lateral shearing resistance. As soon as you yank the cable upwards it's all over. It simply wont beat a regular 2.1mm barrel jack.

No plastic. The one I saw in use on another board was all metal, and the tabs are soldered into plated holes, as well as having the 5 SMT pins on the topside for the USB and power

It seems like it might resist tearing off in vertical tension, and off course it has more lateral strength too.

[amyk]

The two tiny plastic mounting bosses will only increase the lateral shearing resistance. As soon as you yank the cable upwards it's all over. It simply wont beat a regular 2.1mm barrel jack.

No plastic. The one I saw in use on another board was all metal, and the tabs are soldered into plated holes, as well as having the 5 SMT pins on the topside for the USB and power

It seems like it might resist tearing off in vertical tension, and off course it has more lateral strength too.

The surface mount ones are definitely too easy to rip off, that's what phone manufacturers have been using.

Another thought I had is that the tight fit of microUSB connectors tends to make people want to wiggle them around when inserting or removing, and that's primarily what causes the most stress. The socket is also much smaller so the plug acts as a bigger lever in comparisoin. Barrel plugs people either pull/push straight or rotate around their axis if they're really tight, which does not cause as much stress to the connector.

[Harvs]

I hate to say it, but this is one area I think Apple really cracked it with the magsafe power connectors on their laptops.


It's a shame such a great idea isn't in the wider community.

[T4P]

Hur. Apple Magsafe jacks, 'till you use one and have it catch fire. The japanese had magnetic AC jacks like forever running several amps and not creating much heat if at all on kettles 

About the USB negotiation thing, if this was 2002 sure. But motherboard makers don't give a damn anymore since then, take a modern desktop (Please, none of the P4 crap) and see how much current can you draw with the negotiation thing!
And well, if anyone's been watching, at least on the new ones the USB ports aren't 4 by 4 hubs but rather individual data + power lines at least on the proper ones
Yes, you can draw more than 2.5A on those "turbo" ports on some cases but that's not it
Now you can get 2A polyfuses on single ports and even minimum 1.5A or possibly 3A USB currents from Gigabyte

[ejeffrey]

That's what the spec says and while true in theory, are there many cases in practice where ports have been damaged by just trying to get 500mA from them? I'd say it's rare...

I don't know about any cases where ports have been damaged by drawing only 500 mA.  It used to be common to have ports damaged by overcurrent when cheap mobo manufacturers didn't put polyfuses on the USB power.

Still, it is a bad thing to do.  You can't control whether the user (even if the user is you) will plug a 500 mA device into a port that is genuinely only capable of supplying 100 milliamps (i.e., but powered hub).  This can cause the whole bus powered hub to lose power, shutting off every device connected.  If one of those is a USB flash drive or a hard drive, you have a very real probability of data loss.  I actually don't give a rat's ass about hardware damage, hardware is easily replaceable.  But data loss is unacceptable.  Admittedly this is a bit of an unusual scenario, but it is definitely a real world possibility, not just theoretical musings on the USB spec.

So go ahead and use your fake USB coffee warmers and whatever, but you are just fooling yourself if you think that there it is guaranteed to work, or guaranteed not to cause problems for other parts of your system.

The new USB power delivery specification actually addresses this, allowing power to be negotiated separately from data, and to exceed the 500 mA limit.  It is essentially an extension of the "short the D pins together" standard that can work with normal host ports and negotiate different power levels.  However, I don't know if any devices in the wild actually implement it yet, and I don't know how hard it is to use.

[jerry507]

If you can short a USB port and damage it, they aren't compliant with USB specs anyway.

[rr100]

The surface mount ones are definitely too easy to rip off, that's what phone manufacturers have been using.

I never checked what's used in phones but I never come close to dislodging a phone microUSB connector, even if I top off all the smartphones I had like mad multiple times per day. And I had one cable that got suspiciously hard to pull from the phone and after A LOT of fighting I bothered to look at the connector and one of the metal things on the bottom of the connector was totally bent, keeping the connector in the phone much harder than designed.

On the other hand I've seen a few micro/mini usb connectors (surface mount) that were ripped of from hobbyist boards, after probably less than 10 cycles of "insert/remove".

[codeboy2k]

On the other hand I've seen a few micro/mini usb connectors (surface mount) that were ripped of from hobbyist boards, after probably less than 10 cycles of "insert/remove".

I don't have enough experience with phones (I don't repair them) to say one way or the other.  I do know that anything I make will have metal and metal solder tabs into plated through holes, because that's just the only way to keep them on I think. Especially in light of what you and other people say about how easy it is to rip them off.

Especially hobbyist boards, yes, I have seen too many hobbyist boards and even professional development boards with simply the surface mount USB connectors, and you'd think these types of boards especially would undergo more insert/remove cycles than average.