Repair of TEAC T-H380DNT Tuner standby circuit

[Uup]

That black spot inside the VFD is not a failure, rather it is a remnant of the manufacturing process. After the VFD glass is evacuated and sealed, some air molecules inevitably remain inside. To remove the remaining air molecules, and to create a very high vacuum level, a chemical process is used.

A common method to achieve this is by including a barium ring inside the VFD (called a getter). After the sealing process the getter is quickly heated up via an induction coil, where the barium will react and burn up with any trapped air inside the VFD, resulting in a very high vacuum.

You will notice that some VFDs have a small or even an absence of a black spot, and a nice unburnt ring inside it. Others, such as yours, have a relatively large and noticeable black spot. The size of the black spot is proportional to the amount of trapped air that was inside the VFD after it was sealed.

As for diagnosing the issue, there are only a few external signals to the MCU that could cause the symptoms that you are noticing. If you can rule out those signals as being the cause then the failure would likely be with the MCU (IC43), and to a lesser degree, depending on how EXT failure is handled in firmware, could be corrupt/bad EEPROM (IC42) or RTC (IC41).

Also assuming that you properly checked voltage rails and clock signals, as mentioned in previous posts. Seems like you confirmed that the voltage rails were good and carried out many of the first order obvious troubleshooting steps.

KEY0, KEY1, SCROL_UP and SCROL_DWN are the inputs to the MCU for the front panel buttons. These are an analogue input and their voltage level indicates which button, or combination of buttons, are pressed. EXT-INT signal (output) is used if more than one button on KEY0 line is pressed.

Check KEY0 and KEY1 signals (input to MCU). With no buttons pressed there should be a steady 3.2V (or VREFH) present on both.

If the voltage on the KEY0 or KEY1 is less than 3.2V (or less than VREFH) then check the buttons and relevant components to determine the cause.

If the inputs are as expected and are OK, then try the following to narrow down where the issue is.

Since the power/stand-by LED is changing state, from stand-by to ON, or vice versa, you can use this as a trigger event to see what is occurring in that moment.

Connect your scope as follows:

CH1: to ST-BY-LED signal
CH2: to P_DWN signal
CH3: to RESET signal
CH4: to PWR_ON signal

Trigger from CH1
Change the trigger mode to Normal, so the screen only updates on a trigger.
Adjust Horizontal time base as needed so you can see what’s occurring on trigger.

You can then use a one-shot trigger, for a closer look at anything that looks of interest.

When you identify a signal of interest, then you can substitute a (now) redundant signal on one of the channels with a more relevant one.

For example, if the RESET signal is going low, then you can trigger off RESET (CH3 + falling edge) and then substitute ST-BY-LED signal on CH1 with VCC (AVCC/DVCC) in order to narrow down the cause.
 

[solarbot]

That's good news, thanks - the hunt goes on :-)

I've now checked over buttons and solder joints on the front panel PCB and all seems in order - the resistance between FL1 and FL2 is about 5 ohms.

Just rolling back to the start to check voltages at the transformer:

PINS 3-4 = 5.7VAC
PINS 5-6 = 33VAC
PINS 7-8 = 11.1VAC
PINS 8-9 = 11.1VAC
PINS 7-9 = 22.2VAC
PINS 10-11 = 4VAC

Voltages then measured at the associated power rails:

DAB+5V RAIL measuring at D916 = 0VDC
AC measurements @ D901 D902 D903 D904 = 4.6VAC

VCC+12V RAIL measuring at D914 = 0VDC
AC measurements @ D905 D906 = 16.8VAC

MCU_+3.3V measuring at D913 = 3.3VDC
AC measurements @ D907 D908 = 11.1VAC

(display rail ?) measuring at D919 = 4.1VDC

I guess this all looks in order whilst the unit is in standby?

[Uup]

In that case then continue and set up your scope to see what is occurring during the transition of the stand-by LED.

If you don't have a 4 channel scope, then you can still do it with a 2 channel scope, with extra steps.

For example:

CH1: ST-BY-LED signal
CH2: RESET signal

Trigger off CH1

If reset looks normal then move on and swap CH2 signal to another signal, mentioned in my previous post.

[solarbot]

Thanks Uup,  That's very interesting to hear how the vacuum is pulled in the VFD - I know what I'm looking at next time, thanks for explaining.

Our messages crossed but good to know the readings are good, you have given me a good path to explore which I will report back on soon - I have a 2 channel scope but good to know its still possible.

Cheers :-)

[solarbot]

* I've just seen that I got the ST-BY-LED resistor wrong! - this should be R410 so I will try again and report back.

Hi,

So here is some updated info following on from Uup's help:

KEY0 measures 3.44V at R468
KEY1 measures 3.44V at R401
SCROL-UP measures 3.44V at R456
SCROL-DWN measures 3.44V at R455
EXT-INT measures 3.32V at R413

The scope gives steady and clean voltages at all these locations.

ST-BY-LED measures 3.32V at R409
P=DWN measures 4V at R412
RESET measures 3.6V at C406
PWR-ON measures 0V at R408

The scope gives steady and clean voltages at all these locations with no triggers provided at ST-BY-LED

If I scope the LED directly I get the attached signals - yellow is the orange LED and blue is the Blue LED.

I'm also noticing that Q901 is producing some heat relative to the rest of the board - measured with an IR camera.

I've had a close look at the MCU which shows no visual signs of a 'burn'

I've attached a photo of the area around the NIMH battery which died, presumably causing the surrounding damage - a bunch of resistors on the other side needed replacing which has been done as well as replacing the battery.

[solarbot]

Hello,

So, measuring ST-BY-LED at the correct resistor does yield a trigger! but alas nothing visible on P=DWN, RESET or PWR-ON.  I've attached a scope screen to show what is seen.

I've also attached an IR image of the warm/hot Q901 transistor which looks suspicious to my eyes but perhaps this is normal?

[GraemeG]

The power dissipation of Q901 might be normal but it's hard to say. It is a simple regulator to bring the rectified output of transformer pins 7/8/9 (no idea but possibly 10-12V) down to UNSWITCH+5V at about 5.4V. This rail is used in several places and also feeds IC95 so we don't know the current draw. But I remember you replaced Q901 and a quick check shows the original transistor C3198Y has a 625mW rating while the replacement C1815 is 400mW. This could vary depending on the manufacturer too.

If you still have it and it tests OK I would put the original C3198Y back as a precaution. But whether the heat it is generating indicates a problem, or is normal, is hard to say. The current it is providing is going somewhere. Do you see any other warm spots on the board that relate to UNSWITCH+5V? It's worth a look around but this could be unrelated to the problem.

I'd continue with Uup's suggestion. Trigger off the ST-BY-LED signal and use channel 2 to see if you can find something else that is jumping or glitching in time with it. Check all the power supply rails and maybe just work your way around the pins of the MCU. Did you try a gentle tap test on the crystals X401/2/3?

A few more random thoughts. Is there any hint that the fault is heat related? Is it better or worse when you first turn the unit on?

The corrosion under the battery is odd given that it is on the other side of the board. Have you had a very close look for track damage, bad solder joints, etc. You mentioned that several components around the battery were replaced. Do you have a full list of what was bad? Were they damaged by corrosion, cooked from heat? Could there be other components nearby that were not replaced but may have hidden damage?

[solarbot]

Hi GraemeG,

Thanks again. As Q901 tested OK and replacing it didn't change anything I put the original back - I thought this was an NPN rather than a regulator but perhaps an NPN acting as part of a regulator?

I've included an IR overview below along with a zoom in on 2 resistors in the same region as Q901 which stand out in terms of heat.  These are R904 (3.3K) which connects VIN of IC93 to GND (part of UNSWITCH+5V) and R915 (22K) which connects the collector of Q906 to GND (part of DIS+35V).

I've also included an original shot of the resistors by the failed battery which all showed signs of corrosion, were very brittle and have been replaced.

I can't see any signs of this being a heat related fault but when I'm probing the board sometimes the LED turns a solid RED or BLUE but hard to replicate this - the unit still doesn't turn on though.

I've cleaned the damaged tracks around the battery and tested to make sure the connections are still good.

All very mysterious to me but I don't have much experience..... good for learning though 

[GraemeG]

Yes Q901 is an NPN bipolar transistor configured as a simple voltage regulator. The 6.2V zener D924 and the parallel capacitors give a fairly stable voltage on the base. The transistor begins passing current from C to E until the emitter rises in voltage. It all balances out when the B-E voltage reaches stability at about 0.6-0.7 volts. If the load current drops then VBE drops (technically the base current drops) and the transistor begins to turn off. If the load increases, the base current increases (with small increase in VBE) which turns the transistor on more and passes more current to the load. Not as effective as a regulator IC but still pretty good when small variations don't matter.

I don't think R904 or R915 are a problem. They exist to provide a small load to their unregulated voltages when their regulators are turned off and to discharge the parallel caps when power is turned off.

Having the before and after photos of the battery area is good. I should have paid more attention to this area, I can see in the before photo some of the damage and the lead of R463 is obviously detached! I recommend you go back to this area and resolder and recheck every component and link in the leakage area. I can see a dark area around and inside each of the nearby component holes in the PCB and whatever leaked was clearly corrosive so every component and link with a dark ring around the wire is suspect. It's possible for the corrosion to creep down between the wire and solder and cause a bad connection even if the solder appears ok. Make sure the solder is "wetting" and flowing well on each lead when you resolder the joins. Check at least one component beyond the edge of the leakage. Go out at least as far as J121 and J217 to the right, R436/7/8 to the top and J216 and J510 to the bottom. And on the bottom of the board have a really close look at IC41 and IC42 for damage.

Any component in this area could be damaged even if it seems to test ok. The crystal is a worry as it's a fragile device mounted in a tin can. Be suspicious of everything and look carefully for signs of how far the corrosive liquid spread, it may have gone further than I can see in the photos.

[solarbot]

Thanks GraemeG,

Having an explanation is very much appreciated.  I've cleaned up the copper side of the board and will rework the connections soon.  I tool the opportunity to hook up an arduino to the I2C bus on the RTC and Eprom which it can see with the correct addresses but I've not managed to get any info back from the RTC other than it's presence.  I also took a look at some of the transistors in the same area of corrosion which come under two types: hvtkra107mt and hvtkrc107mt but I can't find what these are of indeed recognise the symbols - may I ask what they are (see below)?

[Uunoctium]

These are so called "digital transistors" with integrated base- and base-emitter resistors. Can be driven without using any further components direct from logic circuits or MCU's.

[GraemeG]

These transistors seem to be more commonly known as KRA107M/KRC107M or even just A107/C107 and are available online if you need replacements. Don't mix them up, they are not interchangeable. The A107 is a PNP while the C107 is an NPN.

I have never had to test one so I am not sure how they will behave in a transistor tester. You could watch the input to the base and monitor the output with the 2 channels of your scope. If the input pulses and there is no change on the output then the device is probably dead. For example put one channel on R408 and the other channel on the connection between Q403 and Q407. If/when the signal on R408 goes high then the output should drop from 4.5-5V down to almost 0V. If you see that happen then move the second channel to the PWR_ON line and this should go high in line with the signal on R408. Of course if the MPC never switches the signal on R408 you won't see anything!

[solarbot]

Thanks again to you all for teaching and helping - I've learnt a lot and now have a working tuner thanks to yo'all.

I guess this is all about logic and method at the end of the day.  I went in expecting this to be down to dry caps and then lacked experience to hunt the problem down with a more structured workflow.  Whilst I did focus on the damaged area to start with I didn't follow this up with some simple checks to make sure the circuit was complete, especially given the schematic provided here.

So, with your patience and guidance, I eventually ended up back at the start and found a tiny trace between R460 and R463 was damaged!  A good clean and more careful look with the schematic/multimeter in hand revealed the break so a link wire across the two resistors was all that was needed.  Powering up immediately gave a stable orange LED and before even pushing the ON button I was sure this the missing link!

It's very satisfying to fix these things and I hope to be able to do more.  In the meantime, the problem with this particular TEAC tuner seems to be very common and I expect the fix is likely to be the same: a good clean, replace the NIMH battery, replace the damaged resistors and check back against the schematic.

Hopefully this helps others fix this problem and learn as much as I have.

Thanks again to you all, very much appreciated 

[GraemeG]

Well done! It's been a long process but it is so satisfying when you finally pin something like this down.

R460 is the pull-up resistor for the RDS_DATA signal line. It's purpose is to make sure there is a "high" signal on that line unless something drives it "low". With R460 disconnected, the data line will be floating and probably giving random data to the MCU. Do you know what RDS refers to? I don't know what data the MCU expects on that line but it was obviously very unhappy with what it was seeing.

You are right about the symptoms looking power related and we all suspect electrolytics when there is anything wrong in the power supply. This is a good reminder that they are not always guilty!

[solarbot]

Thanks GraemeG.... for explaining the resistors in question and for refocussing me back on the damaged area    RDS I believe stands for Radio Data Service and provides text information on each broadcast channel, odd that this was enough to prevent start up, perhaps noise across more some of the other lines?

Hopefully next time I will be able to take a more structured approach using what I have learnt here :-)