A novel idea to fix the USB port
 

Partially because I didn't have a suitable micro-sized replacement, and partially because I wanted to be able to plug USB thumbdrives in the phone without having to cart an adapter around with me, I did this Frankeinsteinian arrangement. I took pictures of the whole process, but my camera promptly and helpfully corrupted the files, so I took a few pics of the completed job.

http://img407.imageshack.us/img407/1...sbfix01.th.jpg

http://img32.imageshack.us/img32/918...sbfix02.th.jpg

I soldered wires to the usual points on the motherboard (except ground, which I got from the mounting point of the deceased microUSB socket), and I let them outside from the old receptacle of the port; in order to do so I had to carve away a significant part of the top enclosure using a dremel.

http://img204.imageshack.us/img204/9...sbfix03.th.jpg

I also removed a small part of the metal shield inside the top casing. I grabbed this picture online.

http://img43.imageshack.us/img43/248...sbfix04.th.jpg

The red part is what I cut - there's a sort of "wall" that interferes with the passage of the wires from the board, and it had to go. I sanded the result so it wouldn't cut the wires and cause a short.

I then affixed the type-A USB socket to the phone itself using a large glop of melt glue; I did not have much available space to glue to, so I did the most of what I had - melt glue being the only type of glue that works better the more of it you apply. So far it seems stable; should the glue eventually let go I'll try two-part epoxy instead.

Then I just had to solder the four wires, and cover the solder points in more melt glue, and this is the result. Works perfectly for plugging in stuff, and all I had to do to connect the phone to the computer/charger was hack up a male-to-male USB cable. Coincidentally, if you ever need to do that, use chunky cables and keep them short; my first attempt with the cheapest cables I could find got the N900 complaining it wasn't getting enough power to charge.

Re: A novel idea to fix the USB port
 

that's insane. :eek:

Re: A novel idea to fix the USB port
 

You do realise that a microUSB is more hard wearing than a standard size port.

There was talk on mating cycles on IRC recently.

If you charge your device once a day; a microUSB port has a life of ~27 years, a miniUSB is ~11 years and a standard USB is ~4 years.

Since you now have a USB port sticking out, it maybe worth considering a DC power connection too, to take the load off the USB.

Re: A novel idea to fix the USB port
 

So, you've traded carrying female-to-female adapter, for carrying male-to-male adapter? :)

Nevertheless, good work. It's kind of a demolition, but, if it's OK for Your use case, why not? Congratulation on finishing it, as repairing port via cable method isn't as easy as it may sound.

/Estel

Re: A novel idea to fix the USB port
 

you asre very fun!!
thankss

Re: A novel idea to fix the USB port
 

I find these numbers questionable. My main computer is about four years old, and its ports routinely see more than one connection per day; as a computer tech I'm always using USB stuff of one kind or another, and I can easily average 5 or 6 connections daily. No ports have failed yet.

Even if that were true, though, as much as I'm loving this phone I don't think I'll still be using it four years from now. I hope they'll put out something else Linux-powered in all that time, anyway; I used to like Android (I still think it's better than the common alternatives), but after using Maemo for only a few days it feels inelegant and cumbersome.

As for the charging-once-a-day issue, I have another... thing... planned. It involves more melt-glue. When I do it I'll post it here.

The female-to-female would have to be carried wherever I might need to interface with USB, and that is essentially everywhere. The male-to-male in contrast only needs to stay where I charge the phone, typically at home.
Also, it's easier to find cabled male plugs than female sockets (extension cords being just about the only source of the latter); in an emergency I could get a pair of scissors, gut the first two USB interface wires I found, tape the result and have a charging cable ready in a few minutes.

I gave a shot to resoldering the original port before I started this whole thing, but it proved entirely unfeasible. Every other method would have required wiring of contacts anyway.

Re: A novel idea to fix the USB port
 

That's absolutely true. I don't know where USB-alliance got those numbers from, but it seems like totally made-up, IMO.

The main reason for microUSB was to "save wear on ports", and put it on cables, that are much more cheap to replace. Good idea, but fatal in practice - from day one of microUSB presence, mobile service centers are literally full of broken-port repairs. At the same time, it was much less of a concern in times of microUSB.

Maybe it's not entirely USB alliance to be blamed (no idea) - as it's possible, that internal construction of ports make them less prone to wear. Hoever, on 99% of devices, microUSB port is soldered only to surface of PCB, where roughly 70% (from practice, not scientifically measured) of miniUSB ports were soldered through holes in PCB, and definitely 100% of normal-sized USB ports are made this way.

This sole fact defeat whole purpose of wear-prone ports, as they fall apart like flies, and we need to do tricks like soldering them down.
---

As for cables vs. "normal" repair, I never meant that You did it "wrong". It's absolutely understandable, that You decided to go cable way, as repairing it on pads is PITA, and once pads were damaged 1st time, you can't tell for sure if they won't fell apart completely, even despite soldering port down etc. Using testpads method is only sane way of repairing it. I just meant, that it isn't as easy as some people may think, judging by pictures only - soldering to pads is OK, but creating free space for cables may be PITA, again.

In fact, it's the "demolition" you have made inside case - under camera space - that shocked me more, than a "outside look" of Your device. Not much choice, though, for having cables reliable placed. Another approach would be to cut some of unnecessary plastic from under secondary board (one with camera flashlight and microSD), and run cables through there.
---

After all, despite looking terribly, good job :) Just keep in mind, that You don't have 5V fuse, and external antii-ESD protection for USB (but, USB chip itself have internal protection components, rated for 8,8 kV, so unless You want to catch lightning-strikes by your USB cable, You should be OK).

/Estel

Ps.

You're planning to add power-mat to device, for wireless charging? :) It was covered in some thread, here, on TMO.

Re: A novel idea to fix the USB port
 

Fact

The microUSB was specially designed so the male side takes the most of the stress and wear from reinserting.
The other types of USB the female side has a spring to push the male connection in place. This means the female part takes more wear.

The idea is the cable is easily replacable, the port not so.

Hence rated for 10000 mating cycles on microUSB over 1500 on standard USB.

I wouldn't hold your breath.

Re: A novel idea to fix the USB port
 

Those numbers hardly keep their ground in practice.

Sure - idea of putting D+,D-,+5V and ground springs to male plug, and leaving solid pads on female side is OK, and I'm all with it. port's case springs ("tooths) are irrelevant, though - in "big" USB, they were present on female ports cases, and I yet have to see single spring of this type broken (unless someone bend it on purpose).

The problem is - as I've stated in previous post - way used to mount microUSB ports to PCB. I'm 99% sure, that, while testing connecting cycles, they've checked sole port and plug, not port soldered to PCB via surface-only, like it's done in 99% of microUSB electronic. Basically, nothing to add, above what was written in my last post - this part of design just sucks, and we know it "personally".

Maybe there was some good will, due due to this single flaw, every serviceman I've ever meet, claimed to have *awesome* number of broken microUSB port repairs, if compared to miniUSB and normal USB era. I don't argue, that microUSB port/plug design isn't superior, but way of mounting it to PCB is horrible and wrong from foundation.

/Estel

Re: A novel idea to fix the USB port
 

Do remember though I am refering to wear and mating cycles, these figure reflect that and I can believe them.
I can't say that a big USB would get used to the extent of 1500 mating cycles so how can you compare.
This isn't a port design issue, I feel the port is a lot better design; for the reasons above.

The problem as you say is the way the port has been surface mounted to PCB's.
I do agree that these tests must have only included the port internals and not the mounting to the device.
I feel that the methods behind mounting the USB port was not throughly tested before being forced on to the market as a compulsary requirement; so the most cost effective was used.

Re: A novel idea to fix the USB port
 

Micro USB is so small with the tiniest pins for 1A or more to flow through.

I just want to say the tolerances on the micro USB ports I've used (nokia ones) are very close. The fit is very snug. So, maybe in those mating cycle tests there is the understanding the mechanical purchase one might have on the socket is taken up by the socket rather than the interface as is the case with bog standard full-size USB ports. You know, the full-sized plugs float around in the socket. The snug-ness of the micro USB connection might prevent damage to the actual pins? Just a thought..

Re: A novel idea to fix the USB port
 

Well, in our case, tooths and plug itself is filled a little (if user care about port safety), so plug is "moving" inside socket pretty easily. Despite this, no problem with 1250 mA going through those small pins, completely 0 problems.

It would be ideal, if not for stupid, *****ic way of mounting it to PCB. Sure, I can understand, that they may not want holes in PCB, for various reasons. But, if yes, why another industry standards - like one used for audio-video jacks, that have 4 pins already - isn't used?

Yes, i'm talking about this spring-pad solution, that made audio-video port on N900 virtually indestructible. Same apply for speakers connections. Basically, for freaking everything inside N900, except for microUSB port.

Sadly, it's also the case for most other devices. sure, they ports may be stronger (or weaker, in some cases), but overall, surface-soldering for thing like port can *never* be trusted.

/Estel

Re: A novel idea to fix the USB port
 

From what I've seen, the fact that full-size ports aren't super-snug is a big plus. If everything is completely snug every little movement of the plug gets transmitted to the socket; if it's floating around a bit, there's some leeway before the socket mounting points have to withstand significant force.
As for the solidity of the electric connection, USB type-A is surprisingly resilient; the size of the contact pads has a lot to do with this.

I concur. It's a good idea in theory, but it's useless unless the ports are soldered through-hole. I wonder who the hell thought it was a good idea to replace through-hole miniUSB with anything surface-mount, regardless of the shape and mechanics in the port itself.
Also, I don't really see the necessity of moving away from miniUSB; the shape is small enough for everything that can mount microUSB and the keying of the port is unmistakable, whereas microUSB keying can still be mistaken if one is not careful.

I had no other choice. I tried routing the cables around the camera module, but the casing wouldn't fit. I tried putting them under it, but then the module itself wouldn't stay in place, and again the casing wouldn't fit. It soon became clear that some plastic had to go.

There is less risk of damaging stuff by cutting off casing plastic, which is essentially inert, than by removing material under a board that must stay in its position if the camera is to work properly. I could have placed some tape or something where I cut the casing plastic away, but I don't particularly care how the device looks when the back cover is removed. Or indeed when it isn't, otherwise I wouldn't have a big glop of melt-glue on the side of it. :p

Not at all, what I have in mind concerns the battery. It's a surprise. :p

Ok then, let me put it this way: I'm hoping someone comes up with a way to install Linux-based systems in devices originally meant to run Android.
We've only got marketing to thank for the impossibility of doing this today, as the devices themselves are more than capable if the locks are defeated and drivers are available. I'm hoping enough lock cracking and homemade driver writing happens to allow this to work. I'd gladly help if I had any programming abilities whatsoever, but I don't, so I hope. And as they say, hope springs eternal. :)

Re: A novel idea to fix the USB port
 

Plastic under secondary board can be easily removed completely - and indeed, once I've fixed secondary board via two-side adhesive. worked great, and created much space for relatively "big" (in diameter) cables.

Of course, if Your way works for You, it's absolutely OK - I've posted above, just as tip for others, who may want to follow "full-sized-port way, without deforming N900's case too much (or, particularly, want to do cable's method. It seems to me, that if someone is determined enough, it's possible to avoid cutting any plastic at all (other than removing plastic thing'ie from secondary board, as mentioned earlier).

/Estel

Re: A novel idea to fix the USB port
 

Hmm. If I get any more N900s with broken ports I'll give your way a try. :)

Re: A novel idea to fix the USB port
 

Just a little aside. I measured my Nokia AC-10u 1.2A charger. Slit open the cable--stuck a Fluke meter in between. My charger maxed out at 650mA. I just had a feeling it wasn't actually charging a li-ion battery at or near 1C almost exactly 0.5C. I made sure my battery was well discharged before I measured.

Hmm...

Re: A novel idea to fix the USB port
 

That will be the charging controller in the device, not the charger.

Re: A novel idea to fix the USB port
 

Why? When under the same or similar circumstances I am seeing 800mA from my own home-built solar charger.

The way I understand the charging process is the charger is tested by the phone along the way. First second or two, to see if it can supply more than a couple of hundred mA. The charge is kept at approx 500mA for a few seconds, I haven't timed it but it seems like around 5 seconds, maybe 10. There seems to be another check after this to see if the charger can supply more.

There are kernel messages seen in dmesg | tail at this time too.

The only way I can see this charger is supplying 650mA is because that is all it is able- / limited- to supply

1C is quite a significant charging rate.

Re: A novel idea to fix the USB port
 

Are you sure the multimeter isn't adding resistance? Best way to measure this would be to gut a USB cable, short the data pins, connect the power ones to a laboratory power supply set to 5V and see what the current meter says.
I might do that if I can muster up enough willpower to start tearing apart USB cables again.

Re: A novel idea to fix the USB port
 

When talking about 1250 mA charging, I use custom 5V 2A charger, and I've tested before, that it actually can supply so much power. no idea about stock N900 charger, but it shouldn't be *that* lame (650 mA?!). Maybe this certain unit is malfunctioning?

Can't say for sure, as mine died (erm, helped it a little by screwing modification) some time ago, and I don't remember if I measured charge while using it. Since that, I'm always charging via mentioned 2A unit, or 50A modified ATX power supply (made from desktop PC ATX PS unit).

/Estel

Re: A novel idea to fix the USB port
 

What about thinking about it this way?

If it was charging a li-ion battery (which is what? 97% efficient at taking a charge? It's one of the most efficient charging chemistries I know that much anyway..(wikipedia says 99.8% charge/discharge efficient)) at 1C. We'd all have full batteries in just over an hour. It always takes my battery from flat a good couple of hours plus (2.5? guessing..) to be fully charged? That wouldn't be happening if it was charging at 1C, 0.5C yes.

>> added info:

Since writing this post I did a little experiment. When it's happy my phone idles at 8- 9- and 12-13mA through the powermeter.sh script.

Using a Fluke 111 to measure the current through the AC10u charger I am getting 642-648mA. With the current I am getting around the same. Taking in the 12mA standby current there is about 20-30mA difference between the two readings. That is, between what the current reading from the powermeter.sh script says and what the Fluke 111 offers up. Incidentally the built in current monitor appears more finely grained and is reading those 20-30mA more than the Fluke meter. An aside, I altered the powermeter.sh script slightly to reflect the bq27x00_battery kernel appears to reading out correctly in micro amps now (uA).

The current measured by the bq27x00_battery module does not change more than 4mA or so when the Fluke meter is present. So the current shunt is very low resistance.

>>> MY CHARGER IS FAkE :eek:

I just tried the other AC10u charger we have here and it provides a more healthy 830mA at the State of Charge the battery was in. The current was tapering down so indeed the original OEM Nokia chargers could possibly provide a full 1A when the battery is fully flat

Cellphoneforever (a notorious ebay seller):
So-called NEW (actually refurbs) N900s with the following:
- Send the phone back with no tracking when the phone needs repair (then they say it got lost in the post)
- Keyboard is fake (non-OEM) The keypad is wears out.
- The battery fake
- And now, found a fake charger! There is very little to distinguish between the fake charger and the real one other than the lead is slightly more heavy duty and slightly lighter in colour and more "rubbery"

Re: A novel idea to fix the USB port
 

Not exactly. Li-Ion's are not accepting same current due whole charging process. Basically, nominal "C" value of charge (for example, 1C) is only happening during ~half of charging process. Then, battery is "reporting" ~4.2 V, and charger starts to gradually decrease current, keeping battery at no more than 4.2 V (less in case of N900, as it's charging and using power at the same time), during so-called "saturation" stage. BTW, the higher initial current, the faster "saturation" charge happens. This sole fact defeat purpose of charging LiIon's to full with more than ~0.8C, as time saving is negligible (due to longer saturation phase).

OTOH, there is method called "fast charging". Basic principle of it, is to charge only as long, as time/power ratio is optimal (i.e., only to about 70%-80% of maximal capacity). This way, we benefit from using higher initial current, thus charge time decrease dramatically. Of course, drawback here, is that we don't use full potential of battery, but it have added benefit of extending overall battery life (Li-Ions lifetime benefit from keeping it between 20%-80% of State of Charge [SoC] - this is how they're used in "heavy duty" environments, like electric trains, where cost of replacing huge battery's arrays is higher, than benefits from charging them to 100%).
---

As for Your discovery about fake charger - it explains mysterious 650 mA max. How lame, 50% of declared max current :(

/Estel

Re: A novel idea to fix the USB port
 

That's the only reason I'm guessing it's fake, because of the amount of current compared to the other charger. Otherwise the charger looks as good as the original--very minor cosmetic differences only between the two.
However, the genuine AC10u charger isn't a stellar performer either. I have a no-name charger that outperforms both in terms of output. So the genuine AC10u isn't current limited by the SoC of the battery.

Re: A novel idea to fix the USB port
 

Not entirely correct.
First up, the CC/CV (constant current / constant voltage) charging process switches to CV somewhere at 80% capacity, not 50% --edit-- I got doubtful about this and I'm currently looking for relevant information --end edit--
Secondly, the N900 definitely does charge its LiIon cell to 4.2V; the negligible power it's using to keep itself running in the meantime is irrelevant.

Coulda been worse.

Re: A novel idea to fix the USB port
 

Of course, it was also a "mental shortcut" from my side - when it switched to CV, depends on current during CC phase. Higher current in CC (as % of battery C) - the sooner CV will happen. So, charging standard bl-5jwith 1250 mA it happens near 50% SoC, but charged with 950 or 800 mA, CV will jump in around 80% of SoC.

As for N900 LiIon being charged to 4.2, I can't confirm it. Every measurement I took - both from in-device (bq27200) and using external components - shows little less than 4.2V as absolutely max. IIRC, it was something like 4150mV-4170mV at most.

Genuine N900 charger was charging at full 950 mA during CC, after all (just tested with one borrowed from a friend) - so it's limited by chip. Haven't tried to charge with 1250 mA - it isn't my charger, and I don't want to fry it during experiment ;)

/Estel

Re: A novel idea to fix the USB port
 

I thought I'd update this. I've been using the N900 ever since I fixed its port, and it's still running fine. The port did come unglued at some point, but after regluing it on with two-part epoxy instead of the melt-glue I'd used originally it's been behaving.

As foreseen, it's been useful several times to use thumbdrives on the fly.

I now wish more phones came with a full-sized USB port.