Wonderful! From where and which version?
Important to know because I have found the usual errors and inconsistencies.
I have an undated version and also Rev 1. 7/2005 with more material.
Best wishes
David.
I think this may be the problem, or a substantial part.
Yesterday, late, when it didn't work I measured 260 V😱
My DMM is not calibrated but hard to believe it is that far out.
Today, in business hours with the extra load, I measure 248 V and it works.
That is consistent that it didn't work late the previous day either and then worked when I next tried it, in business hours by luck.
I assumed that the wait had reset some lock-up, instead it may be the coincidence that the mains had dropped.
At the moment the nominal D5V pin is at 5.6 volts.
I will repeat my tests later if the failure reoccurs, I don't trust my results from yesterday because it was late and I was tired.
Thanks everyone, I hoped for some general advice, that people actually trace the circuit is beyond expectations!
Best wishes
David
I don't yet understand the reset mechanism. It is not discussed in the service manual or the user handbook.
The +5V circuit appears to be OK, it measures about +5V. This is the standby circuit and the amp in standby will power up. .
The D5V circuit on pin 4 is a suspect. At the moment measures 5.6V and the amp is functional. This is the main Digital power circuit for the controllers.
Maybe excessive value would lock up the uController or cause a protection mode.
Best wishes
David
I suspect this is the correct track.
I did not even know that an entire product family of "Reset IC"s existed.
David
it's actually pretty simple... while the CPU is powering up and rail voltages are stabilizing, the /reset (the "/" meaning it's "active-low") is held low (a logic zero)... after a predetermined time interval (100mS or less usually) the /reset goes high, telling the CPU chip to begin it's startup sequence. if /reset goes high too early, while the rail voltages are still in a state of flux, the CPU can lock up. if /reset never goes high, the CPU never starts. a dried out cap in the reset circuit can cause the CPU to start too early, which may work sometimes, and not work sometimes.
It is indeed pretty simple - once you know the product exists!
Too low level for my profession in software and too binary for my hobby.
I read some application notes and learned, but your practical information is very helpful - Thanks.
Best wishes
David
D5V is pull_up voltage that keeps buttons input high (when pressed it is short to ground). Higher voltage could be polluted by some ac, should be checked.
I don't believe there is any problem here.
Once the uProcessor starts the display is OK.
When the uP is down then no display of course, but no LEDs illuminate either.
But I already checked yesterday because I wanted to be sure too😉
Some inconsistency in the manual about the correct value for some of FIP1.
Shown as 2.4 VAC on the power supply schematic and 8.4 VAC on the Front board schematic. Measures in the middle.
The DC supply is specified 44 V and measures about 42.
Best wishes
David
It seems to be the main power for uP IC74 too.
I checked this one also. Only a few millivolts of AC.
Best wishes
David
Last edited:
you actually should have about 2.5V AC on the filaments (measured on outside pins of FL display to ground but totalling about 5Vac measured from the ends of the tube)and anywhere from -30 to -45 as a DC voltage. this is the "cathode" bias of the tube. display data is sent as a serial stream from the CPU to the display chip (which does the level shifting internally) on some receivers there is a floating 5V supply (if the VFL appears as -35, for instance, the floating supply will appear as -30 when ground referenced), while on others it's a 5V or 3.3V supply referenced to ground. in any case there will be no display if the CPU sends no data, which is what seems to be happening from your earlier description of the problem..... if the reset circuit time constant is too short, this will happen: sometimes it works, and sometimes doesn't. the reset circuit is usually a transistor, an electrolytic cap (which may be THE problem component here) and a few resistors... the purpose is to hold /reset low for a specific time period.... if the circuit is not holding low long enough, the results are not predictable, since the CPU "wakes up" in an undefined state
i've even seen problems where /reset is noisy because the timing cap is open or dried out... it's usually a very simple circuit, as i said a cap, a couple of resistors, a diode and a transistor. the diode goes to the +5v rail, the transistor emitter to ground, the collector to the /reset line. in the simplest form of /reset circuit, a resistor, in series with the cap, goes to the base of the transistor from the +5V rail. when power is turned on, the turn-on "thump" passes through the cap, turning the transistor on, holding /reset low. as the cap charges up, the current through the base decreases until the transistor turns off, sending /reset high. then the microprocessor starts reading boot code (which if there was no noise on /reset, or the power supply rails already stable, the address pointer is at 0000). if the supply rails were not stable when /reset goes high, the address pointer may not be at 0000, but in some random location with random data in it's registers.
btw, replace ALL of the small electrolytic caps in the vicinity of 3-terminal regulators, as noise from any of these drying out can also account for the failure you are describing (primary one to focus on is the ones for the 5V standby supply)
i've even seen problems where /reset is noisy because the timing cap is open or dried out... it's usually a very simple circuit, as i said a cap, a couple of resistors, a diode and a transistor. the diode goes to the +5v rail, the transistor emitter to ground, the collector to the /reset line. in the simplest form of /reset circuit, a resistor, in series with the cap, goes to the base of the transistor from the +5V rail. when power is turned on, the turn-on "thump" passes through the cap, turning the transistor on, holding /reset low. as the cap charges up, the current through the base decreases until the transistor turns off, sending /reset high. then the microprocessor starts reading boot code (which if there was no noise on /reset, or the power supply rails already stable, the address pointer is at 0000). if the supply rails were not stable when /reset goes high, the address pointer may not be at 0000, but in some random location with random data in it's registers.
btw, replace ALL of the small electrolytic caps in the vicinity of 3-terminal regulators, as noise from any of these drying out can also account for the failure you are describing (primary one to focus on is the ones for the 5V standby supply)
Last edited:
I am pretty tired of it by now. To summarize the current status.
It may be that I had more than one problem and this has complicated the process.
Now that the faulty battery has been removed the unit will reliably come out of standby mode and stay powered up.
It does not need the backup battery and I have removed it for the tests until the new LIon cell arrives.
It still does not reliably work.
The audio section seems to be fine.
There appears to be a problem with one or both of the Front Board controllers.
Sometimes the display and other functions work immediately.
Sometimes they do work after a while.
Once they work I have not had a shut down since the board clean-up.
If the unit is placed back in standby then it may or may not work when re-powered.
When I have attempted to probe the amp while it was down then it has started to work. Since it sometimes starts anyway I don't know if this is coincidence yet.
Best wishes
David
Thanks, seems plausible but unfortunately it looks as if the reset only occurs when the unit is off standby and the backup battery is dead.
Because of the bad back-up battery the unit sat around for a while, a capacitor failure seems likely.
I suppose there is a faint hope that some capacitor may self heal or reform.
The only aspect that is clear is that the manual is bad.
Parts on the schematic but not the parts list.
Parts on the board but not in the manual.
At least it improves my skills.
Happy New Year to everyone.
David
Last edited:
This is repeatable. Touch the D5V pin with a DMMeter and the amp starts.
That narrows it down to a tiny area. Bad 5V supply probably.
7805 IC and a couple of capacitors. All easy to replace.
Are there common 7805 failure modes?
Finally feel confident about the problem.
Best wishes
David
Things have moved on since yesterday, looks like you are on to something. Yes, 5v regs do fail, mostly when fed from SMPS supplies but its not unknown. Check for dries too, and anywhere in the PSU. Don't overlook things like fuse holders, wiring in the mains plug etc. Anything that spikes a micro will corrupt it until the next hard reset.
David, I downloaded this one www.eserviceinfo.com - Harman/Kardon AVR230 Service Manual
It is in 5 parts...uncompressed is 8.656Mb...not sure which version.
Ok, exactly where please ?.
And DMM in Volt Or Resistance mode ?.
Dan.
Volt check on the board that powers the Front Board.
Away from my own machine so don't have the reference handy.
The actual 7805 IC and capacitors are on a separate board.
It seems even the small stray capacitance or conductance of the DMM is sufficient to trip the system on.
New Year celebration this side of the continent so no more tonite.
Happy New Year.
David
More results today.
The so called D5V supply is identified as 5V but measured 5.6 V and this >10% error worried me.
A closer look at the schematic shows it is intended to be 5.6V.
Finally found it marked as 5.6 V on the PCB, in very small print.
One more "problem" solved.
Now that I know what to check, it refuses to fail.
Will try later when the mains is up.
David
The so called D5V supply is identified as 5V but measured 5.6 V and this >10% error worried me.
A closer look at the schematic shows it is intended to be 5.6V.
Finally found it marked as 5.6 V on the PCB, in very small print.
One more "problem" solved.
Now that I know what to check, it refuses to fail.
Will try later when the mains is up.
David
Could there be a "race hazard" situation on power up whereby the rails are not appearing in the correct order, or rising/not rising at the correct rate etc due to dried out caps somewhere.
When faced with something totally intermittent it can also be worth just doing a blanket solder reflow on all possible areas of the circuitry that could affect this.
When faced with something totally intermittent it can also be worth just doing a blanket solder reflow on all possible areas of the circuitry that could affect this.
- Status
- Not open for further replies.