Tivo Series 3: Green blinking light of death

[SluggerB]

Anyone familiar with these old Tivos and the blinking green light issue they can develop?

Someone sent me one that is exhibiting the issue, and I've been taking reading around the board and comparing them to a known good Series 3 Tivo motherboard I have, and so far the only different I can find is a difference in diode reading on a voltage regulator. However when I swap over the voltage regulator from the good Tivo to bad Tivo, it shows the diode readings of the original component, so something else is causing the issue

Power supply is good so I know it's not that.

There is a "hairdryer trick" that is supposed to let these boot, and so I thought a reflow in the area the video tells you to blue the hair dryer would fix it, but I did a reflow, tested all the components, checked all the connections, and still nothing.

Anyone familiar with this issue and know what is faulty on the board?

[SluggerB]

Re: Tivo Series 3: Green blinking light of death

No one with any experience on these devices?

I can post pics, test results, etc

[CapLeaker]

Re: Tivo Series 3: Green blinking light of death

if heating a certain area on the main board made it boot and work ok, there are 2 options.
Either a bad solder connection (reflow) or my bet is that these SMD capacitors dried out and need replacing. Look at the voltages on each capacitor with a oscilloscope and you will see which ones are bad. Or you check them with an capacitor ESR tester.

Probably that guy went over a bunch of small SMD caps on the main board with that hair dryer, right?

[SluggerB]

Re: Tivo Series 3: Green blinking light of death

if heating a certain area on the main board made it boot and work ok, there are 2 options.
Either a bad solder connection (reflow) or my bet is that these SMD capacitors dried out and need replacing. Look at the voltages on each capacitor with a oscilloscope and you will see which ones are bad. Or you check them with an capacitor ESR tester.

Probably that guy went over a bunch of small SMD caps on the main board with that hair dryer, right?

So the first thing I tried was using my hot air station to reflow everything in that area, but it didn't help.

I happen to have a spare old Tivo HD that works fine, it's a non lifetime unit so I have no real need for it, but I figured I can use it to make like for like measurements between the two units.

I spend about 2 hours checking voltages on everything in the area that the hair dryer trick was performed, I checked every capacitor (yes there are several SMD Capacitors there), resistor, chip leg, everything, and it all matched between the two boxes. Voltage wise there is no problem in that area of the board.

I checked capacitance on a few capacitors I removed and they checked out. but I did not check all of them


I then measured the CPU test points. On the good unit, they all tested at 1.29 - 1.3v. But, on the bad uni, there was one test point, test point J, that consistently tested 0.33 volts. A difference!!

The first thing that test point connects to is what I think is a part of a
resistor network, there are a bunch of them they have 8 legs, are tiny, and have "220" on them. I checked for resistance and got 22 ohms. I kept tracing via continuity, and the next stop was a via that read 1.29 volts on the good board, but on the bad board, it read 0 volts. The via connects to another resistor network component leg on the top side of the board labeled RN850, on the good board that leg shows 1.29 volts, on the bad board, 0 volts.

I need to take the board out of the tivo and see what is on the other side of that via. From other posts on this issue, the leading candidates for the issue seem to be a small chip labelled 3003, a crystal, or the RAM chips. I swapped over the 3003 chip and the crystal and no help so I know its not those. I haven't swapped RAM chips (yet)

[CapLeaker]

Re: Tivo Series 3: Green blinking light of death

Yeah, so this one isn’t the 95% of the cause being the capacitors. If the Chrystal is dead, she won’t do anything at all and the board acts stone dead. Does anything get hot in there? Btw those test points you read zero volts on…. They aren’t shorted to gnd and just missing voltage?

[SluggerB]

Re: Tivo Series 3: Green blinking light of death

Yeah, so this one isn't the 95% of the cause being the capacitors. If the Chrystal is dead, she won't do anything at all and the board acts stone dead. Does anything get hot in there? Btw those test points you read zero volts on…. They aren't shorted to gnd and just missing voltage?

Well, All I can say is the capacitors in that area read the correct voltages, and the 5 I have removed tested correct capacitance on my multi meter. I have not removed all of them and don't have an ESR so you could be 100% correct in your initial assessment. Turns out removing SMD caps is not a specialty of mine and traces were getting torn so I stopped.

The board does come to life, the fan spins, the HDD spins, the CPU heatsink gets to the normal warmth. But it gives that blinking green light error and doesn't complete the boot.

I will probe that via and the leg of that component again more carefully to make sure I got it, it seems weird to me that they would read 0V but downstream there is at least 0.33v (even though it should be 1.3v)

I was thinking, should I try using my bench power supply to inject 1.3v on that CPU test point and see if it boots? If it does that would prove thats the issue and I'm not just going down a rabbit trail. But, the CPU is a BGA chip and I don't have the tools to swap that out so I am wary of doing anything that could damage the CPU. If I tried it what do you think would be a safe Amp setting to use on the bench power supply?

[SluggerB]

Re: Tivo Series 3: Green blinking light of death

One other thing, when I first did the CPU test point measurements, test point D was also showing 0.33v, upon inspecting the board I found a bridge on two legs of the first memory chip that must have happened when I reflowed them. I wicked away that bridge and D started showing the correct 1.30v after that. So it could be a RAM issue?

[CapLeaker]

Re: Tivo Series 3: Green blinking light of death

Could be anything. Could be RAM, could be corrupted firmware, could be something with the main CPU, could be a bad data line between cpu and eeprom / SPI or whatever, could be that wherever the firmware is stored on like a SPI is bad, could be a bad hdmi or modem IC, could be a bad non working crystal, could be a drifted out of spec crystal, a broken trace and the list goes on. I you want to troubleshoot this, using an oscilloscope is pretty much a must.
See if the ram chips get hot. Put a bunch of 99% IPA on them and plug that rig in. When one chip gets faster hit than the others, the alcohol will be gone there first.

[SluggerB]

Re: Tivo Series 3: Green blinking light of death

Could be anything. Could be RAM, could be corrupted firmware, could be something with the main CPU, could be a bad data line between cpu and eeprom / SPI or whatever, could be that wherever the firmware is stored on like a SPI is bad, could be a bad hdmi or modem IC, could be a bad non working crystal, could be a drifted out of spec crystal, a broken trace and the list goes on. I you want to troubleshoot this, using an oscilloscope is pretty much a must.
See if the ram chips get hot. Put a bunch of 99% IPA on them and plug that rig in. When one chip gets faster hit than the others, the alcohol will be gone there first.

I'll keep probing

I have a can of freezing spray to find overheated chips, I will give it a shot. I may just blind swap the RAM chips to, there are only 4 of them and I need the hot air practice.

I will mention that the Tivo is not mine, and not essential, so if it is unrepairable it's no loss, but I would love to get it working again.

Thought on the 1.3v injection to that CPU pad?

[SluggerB]

Re: Tivo Series 3: Green blinking light of death

Think you may have been onto something with the AL SMD caps.. I was looking at the caps and getting the values, and in my working Tivo, ALL the caps are Panasonic, in the other about 1/2 are Panasonic, the other half are not, I'm not sure who the manufacturer is.

Also, there is a bank of 5 large capacitors that on the bad Tivo are rated for 6.3 volts. On my Tivo they are rated at 25 volts. And those 5 capacitors are right where the hair dryer was blowing. AND Test Point J is connecting to one of them. So that's a lot of evidence pointing to a failed capacitor(s). The fact Tivo at some point switched to all Panasonic caps and upped the voltage indicates there may have been a design flaw in whatever model the old one was that was rectified by switching to bigger/better caps at a later date when I got mine.

I ordered up a new set of caps for the board and will be swapping them out soon. I will probably do the whole board while I am in there, at least every non Panasonic cap. There are 30 total caps on the board.

[CapLeaker]

Re: Tivo Series 3: Green blinking light of death

Caps can look brand new, but yet be dry. You can’t tell until you tested them with an ESR meter. Just look across each cap with the oscilloscope. If there is a ripple on it you will see.

[SluggerB]

Re: Tivo Series 3: Green blinking light of death

Caps can look brand new, but yet be dry. You can't tell until you tested them with an ESR meter. Just look across each cap with the oscilloscope. If there is a ripple on it you will see.

I don't have a scope, and my son dropped his dumbell on my ESR meter and I haven't gotten a new one yet

I'm just going to recap the whole board, it's only 30 caps and they will be here tomm. Let's see if it works

[SluggerB]

Re: Tivo Series 3: Green blinking light of death

Well, I recapped the board where the hot air was being blown, no dice
I also jumpered from a known 1.37v source to Test Point J, still no dice.

I did find a chip that was shorting across several legs, and in my board it was not, I swapped my chip over to the bad board but the shorts were still there, so I am going to follow all the traces to everything that chip is connected to and check it out. This is a real challnege to fault find

[CapLeaker]

Re: Tivo Series 3: Green blinking light of death

It does take time to figure sh*t out. I would stop at the moment and take a breather. You changed so many things, that you got to make sure you have the same problem and didn’t add any new ones. So I am hoping this thing is still turning on when heated in a certain area?

[SluggerB]

Re: Tivo Series 3: Green blinking light of death

It does take time to figure sh*t out. I would stop at the moment and take a breather. You changed so many things, that you got to make sure you have the same problem and didn’t add any new ones. So I am hoping this thing is still turning on when heated in a certain area?

No, the hot air "trick" never actually did work for me. The owner of the board says it worked for him. Here's what I have done so far

Swapped the 3003 chip & 3002 chip, both are chips near where the hot air was applied
Swapped both large SMD caps in between those chips
Recapped all the AL SMD the entire area (14 caps) and put in 25v caps to replace the old 6.3v ones
Swapped the voltage regulator in the area
Swapped another chip labeled U11 in the area (that several of the caps connect to)
Swapped a diode that looked like it had some cold joints on one side
Swapped all three of the crystals
Jumpered 1.3 volts to Test Point J on the CPU directly from a known 1.3v source

And none of any of those made any difference. Oddly enough, as I take components off the working Tivo to swap over, I will try and turn that one on, and that one never has a blinking green light, even with components off the board.


Even though they read the same voltage and diode wise I'm thinking of swapping over the RAM chips since Test Point J shows up on a bunch of those pins, but blowing a hair dryer back on the corner of the board wouldn't do anything to RAM chips 3 inches away. I can't think of anything blowing a hair dryer on that area of the board would do, it's not hot enough to reflow solder, and if it was expanding the board to restore a cold joint that should be fixed as I have reflowed every component in that area, if it was kickstarting some dead caps that should be resolved since I recapped that entire area. The only consistent difference I get is Test Point J on the CPU shows 0.33v vs 1.27v on the working board

I'm going to take a bunch of pictures of the board and post them here so I can start showing where things are and what I done. But yeah I think I am going to take a bit of a break.