Yamaha A-1000 DSBG

[ic-racer]

Amp was perfect for a few days then, today it acted up again. Specifically a loud crackle or pop through both speakers just after projection relay clicked in and then immediately protection clicked out again. I turned it off-and-on again and the protection stayed in protection mode. After a few minutes I tried it again and it came out of protection in the usual time and works fine.
This will be hard to solve, I think. It needs to get worse or happen more frequently to track it down. I'm kind of looking for a reason to get the new Yamaha amp.

 

[Nashou66]

Did you change the relay?

Maybe it needs to be swapped out.

Nashou

 

[ic-racer]

Pretty sure the relay is fine, though I maybe have not detailed all the symptoms as well as possible. Rather than a gentle 'pop' the sound is more like a phono needle being dragged across the record. A real horrible sound. Also similar to the sound of a 9v battery across the speaker terminals to check polarity. That is why I think there is an intermittent DC leak which is appropriately being detected by the protection circuit.

This is going to be hard to catch because the amp has been playing all day and I have turned it on and off a few times throughout the day today and it is fine.

Anyway one idea just came to me. I bought a little digital VOM for my kids science fair project and he won't be needing it. One way to catch it in the act is to bring some leads out from the output before the relay and hook to the little VOM and leave it there. I have it hooked to some cheap speakers. Next time when it gets 'stuck' in protection, I can turn on the VOM and see if indeed there is spurious DC.

 

[OMGCat!]

Could it be a failing bias pot cause the issue? I don't know how often one would be intermittently bad but if the bias voltage is jumping then I can see the outputs being none too happy.

 

[ic-racer]

More clues today. With speakers disconnected, when turned on, after appropriate delay the relay clicks ON and then shortly clicked back OFF. For the next 5 minutes I can't get the relay to click back on, it stays off when I re-start the amplifier. After five minutes of rest (off), when turned on the relay will click on and after about 2 seconds click back off again. I got this whole cylce to repeat three times.
Fingers are crossed that when I get it over to my workshop bench it will do the same thing and stay in protection so I can measure some voltages.

 

[ic-racer]

I got the unit up on the workbench and found the pins to the protection relay from the bottom.

These results were similar on both Right and Left sides:

When powering on the pre-relay speaker output voltage goes off the millivolt scale and may settle at around -10mV in which case the protection relay clicks ON. Usually, however, the voltage stays high and wanders all over the place. I actually had to go to the Volt scale and found around -5 Volts. This negative 5 volts stays on the pin for a while even after the power switch is turned off. The protection relay does not come on or if it does come on it clicks back off almost immediately.

Since the voltage is wondering all over the place, I can't tell if tapping on the PC board is causing anything.

 

[ic-racer]

I THOUGHT I SOLVED THE PROBLEM.

The voltage instability was traced to the bias circuit. The three pins on the transistors mounted up on top of the cooling fins (TR171 tiny PC Board #3) should read 1.8v, -1.2v, -1.8v. I measured whacky voltages up to -8v.
I de-oxed the bias trimmers on each side and this fixed the problem. I then set the bias with the trimmers to the recommended 11mV.

This worked until the next day when it acted up again.

 

[ic-racer]

Next thing to do is trace the power supply to the Bias circuit (correction, the bias circuit on each side looks to feed of the power rail on each side.).

One explanation that twisting the bias pots may have temporarily cured it is that the increased current through the bias system as I twisted the pots made the yet undiscovered, partially conducting cold joint or failing component, temporarily function again.

 

[ic-racer]

More clues. It started working for about an hour and I banged on the housing and that caused the crackle and triggered the protection circuit. Almost certainly a cold solder joint. Still need to find it, though.

 

[ic-racer]

More clues. I can reliably cause the fault (-6v on the speaker output) by pressing certain areas of the PC board. Mostly near two of the 2SC3370 output transistors. I have re-flowed all the transistors soldered leads.
At this point I'm not totally convinced it is a power transistor failure because I unscrewed the 2SC3370 from the heatsink so it is suspended in the air by its leads. I can still cause the fault by pressing on the PC board in the area of the transistor's leads. Need more investigation at this time....

 

[OMGCat!]

Magnifying glass and trace following in order? Sounds like you've almost got it now that you can reproduce it reliably.

 

[ic-racer]

At this point it seem as if it would be possible to narrow the location of the fault based on some oddball things:

1) Pressing the Right circuit side or Left circuit side of the Main Board near the transistors causes the fault in both Right and Left channels together. (I hooked VOMs to both channels and the abnormal readings occur at the same time on both channels.)
2) The fault is always -6V measured from positive to negative speaker terminal.It wavers up and down but never goes positive or lower than -6V.
3) The fault developed 6 months ago with the amp just sitting in the same location for over 3 years. In its lifetime it has never been "Shipped" other than to the original owner. I'm the second owner, it was local and picked it up in my car ten years ago.
4) In contrast to some other Yamahas, every solder joint in this amp looks perfect. Plenty of gobs of solder around each large connection. None of the PC solder joints look the slightest bit suspicious.
5) The amp has never been run hard. When my M-45 is hooked up in its place, the graphic meter display only bounces to the lowest two indicators at my usual 'office background music' listening levels.

 

[ic-racer]

Only having a short period of time to work with this each day. Yesterday I determined that twisting the chassis reproduces the fault (top and bottom panels are off). Flex clockwise eliminates the fault and flex counterclockwise causes the fault (-6v on both outputs) every time. I feel like the answer is close, but still not evident.

One conclusion from the items listed in #32 above is the may be a fault with the ground. In this amp the ground is common between the Right and Left channel. If something were to raise the ground +6v without affecting the positive pole this would give the symptoms.

 

[ic-racer]

Today's project was to systematically un-solder the twelve transistors on the main heat sink and test. Even after all the output transistors have been disconnected, the fault is still present and still reproducible by twisting the chassis.

So the problem does not seem to be due to leaking or faulty output transistors.

Next step is to consider the negative feedback loop as there are not many circuits left that connect to the output.

 

[ic-racer]

To recap, with all the output transistors removed from the circuit there were still three circuits connected to the positive speaker output. One is the ClassA-AB switching TURBO circuit. This connects to the output via a transistor on each channel. I hate removing potentially working components, but I removed and protected them. This did not solve the problem.

The only other thing (besides the protection/sensing circuit) connected to the output is the negative feedback loop. On each side (R and L) I found a jumper to remove to disconnect the NFB circuit from the positive output.

LEFT SIDE NFB DISCONNECTED: zero to 50ma DC on the LEFT positive output
RIGHT SIDE NFB DISCONNECTED: 20V!! DC on the positive RIGHT output

This is the first thing that I have done that yielded different DC voltages on the RIGHT and LEFT outputs.

I interpret this as some rogue connection between a 20V source and the RIGHT output before the relay.

 

[ic-racer]

At this point I suspect spurious voltage from some leaky component. Not a transistor as all transistors connected to the output are out.
Next step will be to systematically remove or lift every connection that leads to the relay until I get rid of the 20 volts.

 

[ic-racer]

As I removed components from the positive speaker output trace the voltage just went up. Voltages below are between the negative and positive RIGHT speaker leads at the base of the protection relay.
TR163 removed --- 20V
TR164 removed -- 20V
R194 removed --20V
R264, R266 removed --20V
R268 removed -- 30V
R280 lift -- 30v
3 jumpers removed -- 30v
D153 lift -- 27V
R250 lift -- 28V
R248 lift -- 29V

The only things left are three capacitors. They all have high meg resistance measured in circuit. At this point they all have to come out.

 

[ic-racer]

The last two capacitors to take out were the two new Muse 330uf/25V capacitors that I had recently placed. I removed them and it fixed the over voltage. One of them was passing 30V. I'm not sure the capacitors are to blame as one of them is seeing 40V (they are only rated for 25V) compared to the other channel where I measured a correct 17V.

So, one problem solved only to start a whole new search for the reason to the over voltage on the capacitor.

 

[ic-racer]

Things are not so simple. As I re-attached components (removed in post #37 above) to the main positive speaker output trace, the abnormal voltage came back. So it is leaking through from multiple components. At least two resistors in addition to the leaky over-voltage electrolytic cap. So, beyond those resistors there is a major over-voltage problem that is very difficult to solve. Basically, it seems as if almost every remaining small transistor has abnormal voltages. One leading to multiple others; it is almost impossible to know where the abnormality is starting.

 

[Nashou66]

Voltage must be traveling from someplace else. Are you sure there is not a solder bridge anywhere?

Nashou