For many years now I've been documenting all sorts of issues with Logitech consumer products. But there's one which, despite the longevity, still continues to haunt their products today: the "high pitch noise" issue with their mice.
A skilled EE-savvy person mailed me back in November 2010 and provided me with a complete analysis of the problem, as well as a fix. We're still talking about the issue today. He describes the problem as follows:
The solution he proposed is to replace the capacitor in question with a tantalum capacitor instead. Quoting the same fellow:
He also provided me a picture of it (I think this is after he did the fix).
Last night, a colleague of mine I sent an old MX500 mouse to tried the fix with a random tantalum cap he had laying around, and sure enough it worked. Awesome! Finally a solution for this god-awful issue that Logitech told me verbatim "is not a hardware flaw" (ask me for the link to the MP3 file of the voice mail their Customer Service Manager in Fremont left me stating the above).
However, the MX500 mouse isn't sold any more... yet as I said, the problem continues to plague their present-day mice. So I went out and bought three MX518 ("version 2") mice and opened one of them up.
The PCB for the MX518 mouse is different, and the capacitor they're using is so tiny that I can't tell which side is positive or negative, and one of the capacitor leads looks like it goes no where (friends of mine tell me it's sitting on a pad, which presumably means its grounded, but a continuity test turns up nothing since my multimeter's probes are too physically large to test).
Last night I tried using hot glue to cover C11, the resistors, and parts of the opposite side of the board (where the switches mount to the PCB, a couple pins on the LED/IR, etc.) to see if that would mask the noise, and it does not. I guess 15kHz goes through hot glue w/out a problem.
So, given the above information, is someone here willing to help/work with me (I'm willing to pay you for your time if needed, and can send you brand new hardware) on figuring out what exact parts to order from Mouser, and how to actually install the replacement tantalum cap?
I'm not trying to "just fix one mouse and that's it", I'm literally trying to put together a document that explains how to fix this issue on all of Logitech's mice (MX500, MX510, MX518, MX518v2, G5, G7, etc. -- yeah, I have to buy one of each one, fun times), and document the exact procedure, plus give people links to the exact products/parts they'll need to purchase.
A skilled EE-savvy person mailed me back in November 2010 and provided me with a complete analysis of the problem, as well as a fix. We're still talking about the issue today. He describes the problem as follows:
Quote:
The problem isn't terribly uncommon - ceramic caps have an awful lot of good qualities. For a number of applications, properly rated ceramic caps are pretty much the ideal choice. The problem is that if there is enough voltage ripple on a ceramic cap, it will physically "ring". The real kicker here is that whoever designed this for Logitech decided to pulse the wheel encoder LED at 5KHz, and due to the magic of frequency multiplication, we end up with noise at the 3rd harmonic, 15KHz which is high enough to only enough to drive some small fraction of the general population nuts...
If they were to change the sampling frequency up by 20% or more, or change the capacitor type, or just stop pulsing the danged thing and eat the extra couple mW of power dissipation, this issue would go away.
If they were to change the sampling frequency up by 20% or more, or change the capacitor type, or just stop pulsing the danged thing and eat the extra couple mW of power dissipation, this issue would go away.
The solution he proposed is to replace the capacitor in question with a tantalum capacitor instead. Quoting the same fellow:
Quote:
I swapped out the capacitor in my old MX500 with a 0.1uF 35V A-sized tantalum cap, and it seems to get rid of all the noise.
A quick test using a 0805 16V ceramic reduced the noise a bit, but it was still audible. (It's annoying that a $40 mouse skimped on an $0.10 part...)
...
The tantalum cap I mentioned happened to be in my junk bin, but anything rated >= 0.1 microFarads, and >= 6.3V should be fine. Keep an eye on the polarity since tantalum's don't like going in backwards. The series resistor (upside down 331 => 330 ohms) will prevent any permanent damage, but it probably wouldn't work if put in backwards. Tantalums are usually striped on the positive terminal, so just keep that side towards the resistor to it's left.
The critical thing is to remove the capacitor that was originally there. Cutting either trace would work, but removal with a soldering iron is would be my recommendation. Just put a 3-4mm solder blob on the tip of your iron and just "wipe" the entire part off.
In my case I have a soldering iron with a <1mm tip, so I just heated up the sliver of pad below the yellow capacitor and just reflowed it's terminals. An easier way is to just solder the terminals to some other location - the switch's center terminal is a nice place to pull ground, and the LED's positive terminal is a nice big place to solder the other end of the capacitor.
A quick test using a 0805 16V ceramic reduced the noise a bit, but it was still audible. (It's annoying that a $40 mouse skimped on an $0.10 part...)
...
The tantalum cap I mentioned happened to be in my junk bin, but anything rated >= 0.1 microFarads, and >= 6.3V should be fine. Keep an eye on the polarity since tantalum's don't like going in backwards. The series resistor (upside down 331 => 330 ohms) will prevent any permanent damage, but it probably wouldn't work if put in backwards. Tantalums are usually striped on the positive terminal, so just keep that side towards the resistor to it's left.
The critical thing is to remove the capacitor that was originally there. Cutting either trace would work, but removal with a soldering iron is would be my recommendation. Just put a 3-4mm solder blob on the tip of your iron and just "wipe" the entire part off.
In my case I have a soldering iron with a <1mm tip, so I just heated up the sliver of pad below the yellow capacitor and just reflowed it's terminals. An easier way is to just solder the terminals to some other location - the switch's center terminal is a nice place to pull ground, and the LED's positive terminal is a nice big place to solder the other end of the capacitor.
He also provided me a picture of it (I think this is after he did the fix).
Last night, a colleague of mine I sent an old MX500 mouse to tried the fix with a random tantalum cap he had laying around, and sure enough it worked. Awesome! Finally a solution for this god-awful issue that Logitech told me verbatim "is not a hardware flaw" (ask me for the link to the MP3 file of the voice mail their Customer Service Manager in Fremont left me stating the above).
However, the MX500 mouse isn't sold any more... yet as I said, the problem continues to plague their present-day mice. So I went out and bought three MX518 ("version 2") mice and opened one of them up.
The PCB for the MX518 mouse is different, and the capacitor they're using is so tiny that I can't tell which side is positive or negative, and one of the capacitor leads looks like it goes no where (friends of mine tell me it's sitting on a pad, which presumably means its grounded, but a continuity test turns up nothing since my multimeter's probes are too physically large to test).
Last night I tried using hot glue to cover C11, the resistors, and parts of the opposite side of the board (where the switches mount to the PCB, a couple pins on the LED/IR, etc.) to see if that would mask the noise, and it does not. I guess 15kHz goes through hot glue w/out a problem.
So, given the above information, is someone here willing to help/work with me (I'm willing to pay you for your time if needed, and can send you brand new hardware) on figuring out what exact parts to order from Mouser, and how to actually install the replacement tantalum cap?
I'm not trying to "just fix one mouse and that's it", I'm literally trying to put together a document that explains how to fix this issue on all of Logitech's mice (MX500, MX510, MX518, MX518v2, G5, G7, etc. -- yeah, I have to buy one of each one, fun times), and document the exact procedure, plus give people links to the exact products/parts they'll need to purchase.