I have revised my thoughts since post #59. It might be that the outputs are going into tri-state, which is expected behaviour. There are clearly defined logic levels at the 1:51 time in the video. It all hangs on whether the address data bus is valid when the /OE line goes low. Outside a LOW /OE (pin 28), the data read from the ROM (within a small safety margin) is 'don't care', and the ROM's data bus is placed in tri-state mode.
To prove this point, if the 'scope is dual-trace, the /OE line can be displayed simultaneously with the data bus. As long as the data bus has a clearly defined state just after the falling edges of /OE, then all is well.
Last edited:
Folks, thanks a lot for sharing your ideas. I'm happy that you keep suggesting me things to try/verify. That is great! Cause I really had no clue what to do here
I'm thinking now about the source for +MB signal... I've checked a lot af newer Yamaha receiver fixes when replacing of a film capacitor was enough to bring a unit back to life. In those cases, voltages were ok, but probaly additional ripple made uP not working properly... My question - is that normal that output AC voltage from Sub Power transformer is changing and a bit higher than defined in spec?
Well... I've found supercap actually was replaced or at least removed and placed back (I see fresh solder on its pins). There's some slightly sticky flux-alike coating in supercap area (cleaned that out), but nothing really points to electrolyte leakage.
That made the signal to be flat meander-like on D14.tauro0221 said:
Too bad, it's not dual-trace 🙁currentflow said:
I'm thinking now about the source for +MB signal... I've checked a lot af newer Yamaha receiver fixes when replacing of a film capacitor was enough to bring a unit back to life. In those cases, voltages were ok, but probaly additional ripple made uP not working properly... My question - is that normal that output AC voltage from Sub Power transformer is changing and a bit higher than defined in spec?
Why not just get a new digital board for it? It's not nice , but all the common fixes aren't working and you can,'t reload the software or fix the chip...
Hi,
I have a suggestion. Now, that you have a scope Can you check the relay enable signal? To make sure that the pin never get a pulse?
I have a suggestion. Now, that you have a scope Can you check the relay enable signal? To make sure that the pin never get a pulse?
The AC voltage measured is directly related to your local mains voltage as well as transformer loading / transformer regulation and shouldn't be cause for concern from your readings. It is normally fairly well controlled within the electricity provider's tolerances. There will be some fluctuations as the power grid becomes loaded, but that is normal.
A few more thoughts -
Can you enter DIAG mode at all? (Service manual page 8)
If this is successful, page 22 of the service manual shows expected values for THM, PRD and PRV. Any one of these protection monitors being out of range will turn off the power via the CPU. If you cannot enter DIAG mode, can you measure the voltages on the CPU pins 138, 139 and 141 please. Edit: It's safer to measure these values at the board connectors as there is less risk of causing shorts! Let me know if you need help finding these.
THM = heatsink temperature
PRD = DC detection
PRV = power voltage
The main CPU uses its A-D (analogue to digital) inputs to monitor these lines, so they need to be read as exact voltages referenced to 0V.
Last edited:
Absolutely no pulse...
DIAG is unavailable/unreachable. But! Here's what I have after measuring (DMM, minus connected to the chasis):currentflow said:
THM 0.18V
PRD 0.86V
PRV 0.87V
So each three is out of spec, and per service manual, indicating that the unit is in protection mode?
I checked thermal sensor and found that it doesn't receive "+5V"! Also measured several "+5V" pins found on ribbon cables - all have about 0.18V (circled in red on below pic).
At the same time, couple of "VCC" pins I checked (R701, R702, C673) were ok with around 5V reading (circled in green).
Hi,
I checked the PRV signal and the signal it is coming from the power supply. If you do not close the power relay the PRV signal will be 0 volt. The problem with troubleshoot an equipment with micro that the most important parts in the system are the cristal, the micro and the system program loaded in the memory to control the equipment. One easy way to troubleshoot it is by replacing the micro and the memory. Here the problem it is that the micro it is not in a socket that you can easily replace it since they are soldered and it is not easy for the troubleshoot. So you must eliminate all possible scenario that the problem it is not the memory or the micro. Also here I think you are not getting any messages in the display that let you know that the micro it is alive. I have one question when you scoped the clock, the clock keep oscillating or it stopped and goes to +4.8? I think you mentioned it but I am not sure.
I checked the PRV signal and the signal it is coming from the power supply. If you do not close the power relay the PRV signal will be 0 volt. The problem with troubleshoot an equipment with micro that the most important parts in the system are the cristal, the micro and the system program loaded in the memory to control the equipment. One easy way to troubleshoot it is by replacing the micro and the memory. Here the problem it is that the micro it is not in a socket that you can easily replace it since they are soldered and it is not easy for the troubleshoot. So you must eliminate all possible scenario that the problem it is not the memory or the micro. Also here I think you are not getting any messages in the display that let you know that the micro it is alive. I have one question when you scoped the clock, the clock keep oscillating or it stopped and goes to +4.8? I think you mentioned it but I am not sure.
Following the 5V rails backwards on the schematic shows that the protection values are all dependant on the main supply. Therefore the CPU must activate the power relay before it can act upon any of the protection status lines. The CPU refuses to switch the power relay, yet it appears to be accessing its ROM normally. This is looking more like a corrupted ROM, which is only offering invalid code to run the CPU.
The ROM receives a power reset from the +5M1 super-cap rail. Could the super-cap be faulty and not holding a charge? +5M2 has 3000uf of storage capacitance and supplies the CPU VREF input and an array of pull-up resistors. Could +5M2 be dropping too quickly at power off?
Could you please measure the +5M1 and +5M2 rails (separately) to see how fast and how much they decay as soon as the power is turned off.
For +5M1, measure at 1 minute intervals up to 5 minutes.
For +5M2, measure how many seconds it takes to fall to 2V.
No oscillation at all when it goes to +4.8tauro0221 said:
currentflow said:
+5M1
0s 4.92
20s 3.13
1m 20s 2.43
2m 20s 2.24
3m 20s 2.13
4m 20s 2.04
5m 20s 1.95
Manual says supercap is 0.047F/5.5V while I see 1.0F/5.5V - another indicator that it was replaced previously
+5M2
It takes about 2 seconds to drain down to 2V from initial 5V and about 10 seconds to zero. Probing doesn't seem to affect this times significantly.
Hi,
Now that currentflow mentioned the +5M1 and the +5M2 voltages. Did we checked the voltages with the scope and see if it they are clean?
Now that currentflow mentioned the +5M1 and the +5M2 voltages. Did we checked the voltages with the scope and see if it they are clean?
The backup supply from the 0.047F (or 1F) cap should retain its voltage for weeks and months. If the cap is good and the voltage is falling away quickly then that indicates the uP is drawing excess current from the backup line.
Typically a unit will work OK with the cap removed, it just doesn't remember anything when you pull the plug.
It still might be worth working on the principle the uP is stuck and this might sound a bit weird but its worked many times...
Firstly you discharge the supercap and the other supplies around the uP. We've gone to that point already. Now measure and make sure there is no significant voltage present on the uP (a few 10's of millivolts is fine). Whizz your meter round all the pins making sure.
Now press a piece of tinfoil onto the PCB covering the uP and all its pins. You are aiming to short every pin to every other.
Now heat the uP with either a hairdryer or a not fully up to temperature soldering iron. You are aiming to get it hot... very hot. Around 80C. It needs to be like that for quite a few seconds. That increases internal leakage within the chip and helps any charge on the 1000's of internal FET junctions that the chip is made up of dissipate.
Now try it.
If its still no good, and having exhausted all other possibilities then and only then is it time to point a finger at the uP.
Typically a unit will work OK with the cap removed, it just doesn't remember anything when you pull the plug.
It still might be worth working on the principle the uP is stuck and this might sound a bit weird but its worked many times...
Firstly you discharge the supercap and the other supplies around the uP. We've gone to that point already. Now measure and make sure there is no significant voltage present on the uP (a few 10's of millivolts is fine). Whizz your meter round all the pins making sure.
Now press a piece of tinfoil onto the PCB covering the uP and all its pins. You are aiming to short every pin to every other.
Now heat the uP with either a hairdryer or a not fully up to temperature soldering iron. You are aiming to get it hot... very hot. Around 80C. It needs to be like that for quite a few seconds. That increases internal leakage within the chip and helps any charge on the 1000's of internal FET junctions that the chip is made up of dissipate.
Now try it.
If its still no good, and having exhausted all other possibilities then and only then is it time to point a finger at the uP.
As you say, if the super-cap is good, it could be being discharged by the CPU. The +5M1 rail also powers the ROM and a C-R reset for the ROM. There is also a number bypass capacitors (100nF) around both the CPU and the ROM, of which one or more may be suspect.
If the super-cap is faulty (and holding a partial charge), assuming the CPU and ROM are both good, there may or may not be a clean reset taking place on the ROM or CPU in order to get the CPU into a known state at power-up. Removing the super-cap completely and discharging the rails thoroughly before start-up would help rule out this possibility.
Good idea - but perhaps remove the super-cap first?
This may be due to invalid code being run, as the CPU has the ability to stop its own clock.
One hopes that the previous repair was successful at that time and that just the supercap had failed.+5M1
0s 4.92
20s 3.13
1m 20s 2.43
2m 20s 2.24
3m 20s 2.13
4m 20s 2.04
5m 20s 1.95
Manual says supercap is 0.047F/5.5V while I see 1.0F/5.5V - another indicator that it was replaced previously
You could remove the supercap, short the +5M1 rail to 0V with a wire link for a few 10s of seconds (no resistor) and try starting up. This is relatively quick to perform. If no behaviourial change, then I would follow Mooly's shorting and heating suggestion.
I've found in practice that some uP's don't reset when power is fully removed... its as if internal charge still remains, hence the 'tinfoil and heat' tips.
A popular Sony remote suffers from this. It totally locks up and will not respond to any button presses. Removing the batteries for a few hours doesn't work either. Leave it for 10 to 14 days unpowered and after that it works perfectly. Heating the chip by placing the PCB on a hot radiator fixes it in under an hour.
A popular Sony remote suffers from this. It totally locks up and will not respond to any button presses. Removing the batteries for a few hours doesn't work either. Leave it for 10 to 14 days unpowered and after that it works perfectly. Heating the chip by placing the PCB on a hot radiator fixes it in under an hour.
Very interesting. Although I've not personally experienced this degree of 'stubborness' in microprocessors, this looks like an attractive option!
Well it used to be the day job fixing stuff like this, with the 'chewy' faults 😀 usually ending up on my bench eventually.
Hi,
On the reset normally if you waited 10 to 30 seconds after the power is applied you can reset the micro by grounding the reset pin and the micro should come alive. This is what you call a hard reset. If the micro was lockup by the software problem it should come alive. The schematic showed that the micro it is powered all the time and it is possible that it goes to sleep. Remembered that the power switch it is tied to an interrupt pin. That it is the only way you that you can bring the micro back from sleep. Normally by grounding the pin the micro should come up. If the micro does not come up then something it is wrong in the programming or a bad micro. Here we know that the micro address the memory and stopped. So my suspicious it is the memory.
On the reset normally if you waited 10 to 30 seconds after the power is applied you can reset the micro by grounding the reset pin and the micro should come alive. This is what you call a hard reset. If the micro was lockup by the software problem it should come alive. The schematic showed that the micro it is powered all the time and it is possible that it goes to sleep. Remembered that the power switch it is tied to an interrupt pin. That it is the only way you that you can bring the micro back from sleep. Normally by grounding the pin the micro should come up. If the micro does not come up then something it is wrong in the programming or a bad micro. Here we know that the micro address the memory and stopped. So my suspicious it is the memory.
While following Mooly's tinfoil suggestion after supercap removing gave no result, I found strange thing: PRY (power relay) signal is shorted to ground! I tried to remove PRY +MB MG connector from sub power board and then saw only 60 Ohm between MG and PRY on mainboard (MG is not connected to chasis in this case) and no connection between MG and PRY on sub power board. I don't see where PRY line gets connected to MG outside of uP, so probably not suspicious...
Hi,
With the connector out check the resistance between PRV pin and the mg pin. This will check if the base of the transistor Q311 it is shorted to ground or something it is grounding the signal. Also you can do the check at the connector.
With the connector out check the resistance between PRV pin and the mg pin. This will check if the base of the transistor Q311 it is shorted to ground or something it is grounding the signal. Also you can do the check at the connector.
- Status
- Not open for further replies.