|
|
Post by gjaky on May 30, 2014 13:33:17 GMT -5
Just remember that the problem you mention as solder/bga problem was non existent before ROHS solders came in to play... The problem is that these solders are more rigid than the older leaded solders, meaning in practice that with heating up / cooling cycles these joints are getting weaker and weaker. So either you have to change your solder material or have to get rid of the heat that destroys your solder joints. It's really cool how you worked out the repair process, but one thing is fixing a broken part and an other thing is to find the real cause of the problem! You successfully passed stage 1, don't stop now!  I'd rather cut a passive heatsink, it doesn't have to have a fan, but it's quite clear that chip dissipates too much heat to leave it without cooling. Have you heard of planned obsolescence? What if they decided in the onkyo factory not put heatsink on that particular chip to shorten its life cycle? |
|
nashou
Unmoderated Off Topic
Tech in Training.....
Posts: 1,239 
|
Post by nashou on Jun 24, 2014 13:50:56 GMT -5
Gabor you mean like this?  |
|
howie
Junior Member
Posts: 88
|
Post by howie on Jun 24, 2014 21:54:32 GMT -5
I have been giving this "reflow" thing a bit of thought. This is what I will try the next time I have a "reflow" to try. I am going to take a small cube of aluminum and install a thermistor in it so I can monitor it's temperature. prep the chip with flux as usual. Put the cube on the chip and put the whole thing in a oven to about a hundred degrees below the solder melting temp. Take the board out and place it on a flat surface. Then use a small torch to raise the cubes temp to above the solders melting temp and let it cool down. My hope is the weight and temp of the cube will do the job better then a heat gun alone. The biggest issue with my plan is being sure there is no tension on the thermistor wire so the chip doesn't wander on the board while the solder is liquified. You will have exact control over the temp placed on the chip so you will know when the solder is liquified. They call me.......................Dr. Phibes |
|
|
|
Post by gjaky on Jun 25, 2014 1:35:33 GMT -5
Gabor you mean like this? Yes, those heatsinks helping much. |
|
|
|
Post by mastertech on Jun 25, 2014 15:41:13 GMT -5
I have been giving this "reflow" thing a bit of thought. This is what I will try the next time I have a "reflow" to try. I am going to take a small cube of aluminum and install a thermistor in it so I can monitor it's temperature. prep the chip with flux as usual. Put the cube on the chip and put the whole thing in a oven to about a hundred degrees below the solder melting temp. Take the board out and place it on a flat surface. Then use a small torch to raise the cubes temp to above the solders melting temp and let it cool down. My hope is the weight and temp of the cube will do the job better then a heat gun alone. The biggest issue with my plan is being sure there is no tension on the thermistor wire so the chip doesn't wander on the board while the solder is liquified. You will have exact control over the temp placed on the chip so you will know when the solder is liquified. They call me.......................Dr. Phibes This is an interesting thought. I would be curious of the results if you ever give it a try. I don't know if I would oven temp it very high though as this would cause premature failure of any electrolytic capacitors. My own experiences have seen caps vent instantly at certain temps in an oven. |
|
|
|
Post by Slaw on Jun 27, 2014 7:54:43 GMT -5
Hello guys,
i have to say that you reflow/rework won't give a long term solution. Problem with the units is not in solder balls but DSP chip is the culprit. It must be replaced, problem is - there is a huge problem with ability to buy it. Onkyo is replacing those chips with D830K013DZKB3 which is corrected version of the chip, premature aging free.
Regards.
|
|
|
|
Post by tjmotter on Jan 27, 2018 17:38:19 GMT -5
Hello guys, i have to say that you reflow/rework won't give a long term solution. Problem with the units is not in solder balls but DSP chip is the culprit. It must be replaced, problem is - there is a huge problem with ability to buy it. Onkyo is replacing those chips with D830K013DZKB3 which is corrected version of the chip, premature aging free. Regards. Yup, I have now fixed 2 NR708's where reflow and reball didn't work. I ended up buying 12 of the TI parts from Rochester Electronics (TI's authorized EOL Distributor). Problem is that I don't believe that TI ever went to production with the "D" rev part so the only option was to buy "B" rev. I suspect they offered it to Onkyo and others but didn't get any takers because when I reached out to TI, the pointed me to Rochester and insisted that the "B" rev was the latest part offered. One other observation is that many people use the Onkyo "Network Standby" function so that they can remotely turn the receiver on. The problem with this is that it keeps that chip running 24x7x365 which (my current theory) causes the part to wear out much faster than expected. I have played around with this a LOT and was interested to find that the TI part is running on an ARM based CPU using a Linux variant. You can actually connect into it from one of the diag ports, watch it boot and even login to the OS. Good job on the reflow btw. I went out and bought a board heater and a soldering station with a controllable heat gun. Rather than build the "box" like your setup, I use Kapton tape to mask the area off and then cut a hole in a piece of tin foil which I then tape over the areas I don't want to heat. I warm the bottom side up to 120 degrees (monitoring it with an oven thermometer probe) and then hit the chip with ~435 degrees off the heat wand. In 60 seconds it will pull right off. The probe creeps up to about 140 degrees which still gives me about 20 degrees to play with before the Ag based solder melts and causes the parts to fall off. After letting the board cool down, I then clean up the solder with some wick, clean the whole area with some isopropyl alcohol (90% pure) using a couple of Q-Tips. Next, I start reheating the board and set the new chip into place (using a pin to nudge it within the lines on the board). Next, I hit it with the 435 degree heat gun for 60 seconds. I have done this on NR708's, NR616's, NR818's, NR515's and some Integra units (using the newer 400Mhz chips for the newer boards which are easy to get off eBay) and everyone works like new again. HTH Todd I have a bunch of brand new |
|
|
|
Post by mastertech on Jan 27, 2018 17:49:17 GMT -5
Nice work Todd. Do you offer this repair as a service to repair for others?
|
|
|
|
Post by tibimakai on Jan 30, 2018 11:13:33 GMT -5
Hi Todd, I see that you are here as well now(from avsforum).
I posted there about the 609, that I have and you have offered to repair mine, if you remember.
I have replaced the TI chip, but it did not start up. It shows the same symptoms as before. Maybe my soldering wasn't good enough.
My first attempt, to this style of chip.
|
|
|
|
Post by tjmotter on Feb 2, 2018 10:13:53 GMT -5
Hi Todd, I see that you are here as well now(from avsforum). I posted there about the 609, that I have and you have offered to repair mine, if you remember. I have replaced the TI chip, but it did not start up. It shows the same symptoms as before. Maybe my soldering wasn't good enough. My first attempt, to this style of chip. The NR609 (and the NR509 and the Integra 20.3) often suffer from corruption of the DSP firmware. This has the same symptoms as a bad DSP chip but is often easier to fix. Onkyo changed their firmware update process to no longer allow network based firmware updates after they found this issue. The DSP is an ARM based design and it runs on a variant of Linux. To make this work, an Onkyo has to first initiate a "Flash aware" shell based on "YAFFS" (Yet Another Flash File System) called "U-boot". This filesystem boots before the OS and its job is to manage the wear on the Flash device but in doing this, it also sets the initial values for things like the Ethernet MAC address and the location of the Linux boot block. Unfortunately, the "network firmware update option" can corrupt these initial settings which means the DSP won't boot Linux (and thus, you get no audio and no network. I have found that this can often be repaired by connecting a PC to the diag port on the DSP/HDMI board. You can make this connection with a basic serial port (since it is Linux based). I use a "bus pirate" for this and set it to terminal mode. Once connected, when you turn the unit on you can watch the boot process and interrupt it. Once you have done this, you can rebuild the environment variables. To start, you first need to figure out what is missing. You can get a read out of the current environment by typing "printenv". The environment should show something like the following: baudrate=115200 bootfile="uImage" verify=n ethaddr=xx:xx:xx:xx:xx:xx bootargs=console=ttyS1,115200n8 root=/dev/mtdblock8 rw quiet lpj=999424 rootfstype=yaffs mem=56M ip=off setparam=ok stdin=serial stdout=serial stderr=serial ver=U-Boot 1.3.3-svn (Nov 29 2010 - 18:05:03) bootcmd=nboot,jffs2 0xc0200000 0 0x254000 autostart=yes bootdelay=1 dspboot=yes Most of the settings are probably ok but the ethaddr setting will be wrong. To fix this, use the "forceenv" command like this: "forceenv ethaddr=xx:xx:xx:xx:xx:xx (simply input the 6 HEX address printed on the label on the ethernet port instead of the xx's) Next, save the values (this command will also recalculate the proper checksum as well) "saveenv" if any other variables are missing, add them using the same process Once completed, the firmware should be fixed and the system should boot properly. If not, you may have to find a working system and copy the firmware chip. HTH Todd |
|
|
|
Post by tjmotter on Feb 2, 2018 10:15:54 GMT -5
Nice work Todd. Do you offer this repair as a service to repair for others? Not so much of a "service" but I have fixed other peoples systems for them for just the cost of the parts. This is just a hobby for me so I don't make any money at it. In fact, I get most of my dead systems off eBay, repair them and then donate them to charity. |
|
|
|
Post by tibimakai on Feb 2, 2018 10:57:51 GMT -5
Thanks Todd for that. I may come back to this one day, and most likely I will ask your help than.
Right now, I'm redoing my garage(insulate/drywall) and I can't work in there.
|
|
jrnn
New Member
Posts: 1
|
Post by jrnn on Feb 28, 2018 19:55:42 GMT -5
Im new here, excuse me for my bad English. I live in the Netherlands and my pioneer vsx-921-k has the all known ue22 error.. Already tried to heat the dsp chip up, but sadly after a few weeks the same ue22 returned flickering on the display. Is it an option to replace the current 'B' chip with the newer D810K013DZKB4?
|
|
|
|
Post by tjmotter on Mar 12, 2018 9:22:56 GMT -5
Im new here, excuse me for my bad English. I live in the Netherlands and my pioneer vsx-921-k has the all known ue22 error.. Already tried to heat the dsp chip up, but sadly after a few weeks the same ue22 returned flickering on the display. Is it an option to replace the current 'B' chip with the newer D810K013DZKB4? Yes, I have done this many times on Onkyo receivers. The "D" rev part is relatively easy to find on eBay and sells for ~$30 US. The "D" rev is electronically identical to the "B" part so it is a direct replacement. Most of the supply is based in China but I have never had a problem with these parts. Many people complain that China sells counterfeit parts and while this is true for transistors, the DSP chip is too advanced to be copied so I find that they are fine. The "masking off with heat resistant tape like Kapton" and aluminum foil stays the same. You should also plan to preheat the bottom of the board to 120 degrees with a board heater (like this one: www.circuitspecialists.com/csi835b.htmlOnce the board heater reaches temperature, I shut it off and apply the heat gun to the chip from the top. After 1 min at ~435 degrees you can use a suction device - many advanced soldering irons have this such as this one: www.ebay.com/itm/Aoyue-968A-4-in-1-Digital-Soldering-Iron-Hot-Air-Station-Complete-Kit-USA-NEW/322948008987?epid=2255917850&hash=item4b31338c1b:g:WPkAAOSwmhJZtpAaYou use the suction to remove the old chip. Next, let the board cool to room temperature and once cooled, squeeze some "no clean flux" over the pads and use solder wick to clean all of the excess solder off of the pads. Next, clean the area with isopropyl alcohol. Next, spread a very thin layer of "no clean flux" on the pads and align the new chip to fit perfectly within the white outline drawn on the board. Reheat the bottom of the board to 125 degrees (turning the board heater off once you get to this temperature) and apply the heat gun to the top of the chip again at ~435 degrees for 1 minute. Once the board cools down you should be able to visually inspect the chip to see if the solder balls have compressed. If so, you can re-install the board in the receiver and it should work. If not (if they still look round instead of "squished") you will want to try heating it up again. I should note that I use an oven probe to detect when then the bottom of the board reaches 125 degrees simply because a board heater places the board about 2cm above the heat so the reading on the front of the unit is not accurate. HTH Todd Update: I just fixed an Onkyo NR709 using the method above but found that the board was thicker than the NR708 that I based this write-up on. For the thicker board, I checked after 1 minute and the chip was still affixed. I ended up having to heat the chip for 1min 45 seconds. If you run into this you should plan to use the same time for installation as you witnessed for removal. One last note: Some vendors on eBay will offer the "XD830K013DZKB4". This part is different than the "D830K013DZKB4". I don't know what the exact difference is but I have installed 2 of the "X" chips onto Onkyo boards (1 x NR616 and 1 x NR709) and while the DSP boots, USB works properly, the Network works and the the systems will output sound as normal, Audio streaming will no longer function. I tested it with several of the preset streaming services that Onkyo sets up but also with my iphone and I will get about 1 second of music and then dead air. When I installed a new "D830K013DZKB4" everything works as it is supposed to. I suspect some form of timing difference between these two devices or, it is possible they have different core counts/core frequencies but if you are looking to return a Receiver to full functionality, only the "D830K013DZKB4" will do. If anyone is adventurous, this device actually runs on a flavor of Linux and I have used the diag port several times to open a Linux shell. Instructions on how to do this are available on the web (along with the "root" password). Once in, you could probably run some Linux commands (like "ps") or compare the info in /proc/cpuinfo to figure out what is going on and possibly fix it. ![]() |
|
|
|
Post by Admin on Mar 12, 2018 20:21:49 GMT -5
Very nice write up Todd. I have made a copy and pinned it in this section for others to see rather then get buried here. Thanks. |
|
