Cant dial, audio distorted, MC14LC5480EN chip is defective!

magicman · Post by **magicman** » Thu Jun 05, 2008 10:23 am

After some troubleshooting, I figured out the reason I cannot dial out with my magicjack its the OpAmp (audio amplifier) inside the magicjack that has severe distortion on outgoing sound (from my phone mic to the person at the other end) I figured out this distortion is occurring in the opamp section of the MC14LC5480EN PCM chip, which handles the PCM data to Analog sound conversion.
After reading a bunch of post in this forum, I have found many other with audio distortion problem, either outgoing, or incoming, they have all narrowed the problem to the actual Magicjack device itself, hence I believe their MC14LC5480EN chips are also defective.
I downloaded the datasheet from freescale semiconductor (who makes the chip). From my troubleshooting and reverse engineering attempts, it looks like magicjack designers have a slightly modified copy of the example ckt in the chip datasheet.
Im not sure how the chip is being damaged, be it electrostatic, or possibly over current draw from having too many phones connected. Eithor way I know my chip is defective, and I guess I'm eithor going to have to order a new chip and replace it. Or convince magicjack to send me a new device because there design sucks.
Any other experts out there with the same problem as me want to offer any other suggestions?
datasheet link..
http://www.freescale.com/files/timing_i ... LC5480.pdf

Stewart · Post by **Stewart** » Thu Jun 05, 2008 2:11 pm

Interesting stuff. How close is the MJ to the reference design here: http://tjnet.com/chips/tiger560C.htm ?

If very similar, it would seem unlikely for ESD to damage the codec; there is lots of clamping, I would expect the SLIC section of the 560C to fail first, and the 15K resistors feeding TI+ and TI- should limit current to well below what the 5480's internal protection can withstand. DC overload shouldn't affect the 5480 either, as its inputs are capacitively coupled, unless you believe that the damage was caused by the entire unit overheating. (I'm not defending this design in general; I think it sucks, though I don't believe that it has the specific shortcoming that you mention.)

I noticed that the reference design uses a Winbond codec W681513S; see http://www.winbond-usa.com/products/win ... 13_A10.pdf .
This part is pin and function compatible with the 5480, though I don't know whether you substitute one for the other. If you decide that you can, and a 20-SOP package is used, you could buy one from Digikey for $2.60; see http://parts.digikey.com/1/parts/477507 ... 1513s.html . Otherwise, probably easiest to beg for a sample from Freescale.

How did you determine that the OpAmp section is bad? Perhaps one of the feedback components has failed, or the input is distorted because of a failure in the 560, etc.

mutedmajority · Post by **mutedmajority** » Sun Jun 15, 2008 6:59 pm

Great info. Maybe I will be able to source that chip. I have a spare MJ maybe I will sacrafice since I have no way to change the number to my local area code anyways. (It's not registered to me, I got it at an auction of lost shipments) Mouser electronics is usually a great source. My phone tech intuition tells me you're right, because there is a lot of distortion etc. I've seen some chips that were just weeeeaaaaak. you would have to buy four or five just to get two good ones. (other products) If I get a chance to play with it, I will post my findings. Thanks Stewart I'm rarely impressed with anyone but you are an exception.

mutedmajority · Post by **mutedmajority** » Tue Jun 17, 2008 11:50 am

sheese, I tried defluxing the board, no help, I even tried wrapping the mj in some aluminum foil I got off my tv antenna

no help. I found the chips on digi-key.com and some overseas chip companies websites but you have to buy like 100 of them at a time and "inquire" as to price. Kudos to magicman for some impressive hackery. THanks!