Cassette deck Multiprocessor issues- Need Opinions

[Rogersjapan]

Been trying to fix this.Sony TC-WR99ES for some time now and have it narrowed down to Main Controller not starting. On main controller chip all voltages seem fine except the voltages going from this chip to reset the Mech drive chip and the reset the display controller. Problem is power on -no display or any controls work. Only deck lights on and capstan motors are running. I know problem is on main board only and nothing in the display or front panel as I tried it on good deck it worked fine. Any ideas from any of the Guru's here. This chip is impossible to find by the way.

 

[Retrovert]

From years of building and debugging microprocessor equipment I have two suggestions to start:.

(1) Replace (not check, but replace) all of the electrolytic capacitors. I see you marked some as replaced. If any of the local decoupling capacitors, somewhere about 4.7 uF, are bad, this will drag down the voltage and prevent the processor from starting.

(2) Verify that all the clocks are running on the main uP board and on the remote control board, if present, and that the reset line is not being held down. The front panel may have capacitors to debounce a switch or to clock the remote control receiver chip, and one of those could be bad. I don't know your unit, but I've seen very complicated boards on the front panel. What did you see on the databus or address bus? Is the processor running at all?​

 

[restorer-john]

The two deck mode switches can be a source of this type of trouble. Have you inspected and confirmed their correct placement and functionality? Check for cracked gears or gears on the mode switches that have fallen off. (BTW, unrelated but useful- Sony DAT (55/75/57es) mode switch gears are now available as a CNC machined steel gear from a dude in the EU- I think he may make other metal replacements for sony gears).

Are you 100% sure the decks individually are good- did you swap them over in the other deck to confirm it isn't a single deck mode switch issue?

Sony used a bunch of dubious surface mount caps at the time the deck was made, I second Retro's advice on decoupling and reset line startup caps.

Also, just one tactile switch being flaky or held down/on can prevent the deck from initiating at all. Get a bag of them from China and replace the lot, they are really cheap. Have you checked the transport switch PCBs or just swapped around the display PCB?

 

[Rogersjapan]

I do have an operational TC-WR99es and I swapped the whole front control panel with display onto operation unit and it powered up fine and all the controls seemed to be working. So i'm positive its not in the display boards or control buttons causing this. I've also tested each transport unit on the serviceable unit and they played a tape fine. The problem is on main board. Power supply has all new caps and I've tested all transistors. The voltages are spot on out of power supply area. I have a freq counter coming in this week and I want to test to see if the clock crystal is running. It's a 4.00 mhz crystal How can use this to test if the main CPU is even running?.

 

[dr*audio]

You need an oscilloscope. Get one with at least 60MHz bandwidth. You can pick up a Tektronix 2213 for cheap nowadays. Make sure you get scope probes that have 100MHz bandwidth and 12pF or less so they don't load down signals you are trying to measure.
A microprocessor needs:
Vcc
A clock signal at 0V - Vcc
A reset signal. The pin will be marked on the schematic as Reset if it requires Vcc level to reset it, and it will have a bar over Reset if it's a low level reset. In either case in order for the microprocessor to operate, it needs a reset pulse when the power comes on, then the reset has to go to the opposite level. If it's a high reset, it needs a high pulse, then 0V. If it's a low level Reset, the Reset line should go to 0V briefly, then go to Vcc. There is usually a cap to ground on the reset line. If the cap is leaky it will hold the reset line low and the chip won't start.
Sony sometimes has a "Stop" input on the microprocessor. If that line goes low the Micro won't start.

 

[NAD80]

Place frequency counter/ scope across the pins for the crystals. If the unit is working the counter will show the frequency . A scope will show a square wave. Just seeing a dc voltage on a scope means no oscillation. Check both crystals for both micros. What dr*audio says is right to check. Good luck with it.

 

[dr*audio]

Place frequency counter/ scope across the pins for the crystals. If the unit is working the counter will show the frequency . A scope will show a square wave. Just seeing a dc voltage on a scope means no oscillation. Check both crystals for both micros. What dr*audio says is right to check. Good luck with it.

Just to clarify, do not connect the scope across the crystal, one lead is ground and will short the crystal to ground on one side. Connect the ground clip of the probe to chassis ground and then touch the probe tip to one side of the crystal at a time. You must have clock signal on both sides.

 

[NAD80]

Very Sorry. Used to scopes at work. My scopes are floating due to testing non-isolated AC temp controllers. If the scopes at work were grounded they would short AC to ground.

 

[Retrovert]

The best place to monitor the clock is at the microprocessor's clock input pin. That demonstrates whether or not the clock circuit is working as a whole.

The data bus and address bus must be doing something for the circuit to work.

An easy way to tell if the circuit is running is to build a simple logic probe which indicates, via an LED, whether a line is high or low. A changing state will pulse the LED. Google will turn up a DIY page. Search for "logic probe".

You may have a bad solder joint or an oxidized socket.

 

[Rogersjapan]

Got a good deal on O-scope/probes so I will wait for them. Also ordered a simple logic probe package so may build that while waiting.

 

[Rogersjapan]

Get to play with these in the morning.

 

[Rogersjapan]

1. Update. Pin 29 has good clock signal-Counter reads 3987khz and verified good looking signal with o-scope. Also checked clock signals on display and deck micro-controllers and they are running also. Used logic probe on pin 27 and it briefly goes low on power up then back to high so that is doing what its supposed to do. I used logic probe on Serviceable deck to compare logic states from pins 23 and 25 giving reset signals to Deck and FL controllers and they briefly flash opposite and go to their respective states. On Dead deck MD reset stays high and wont jump to low and FL stays low and wont jump to high. There is no data being sent on any of the lines. Pin 15 should have High but is stuck in low. I verified all this with serviceable deck. I Put the the 5VDC feed on the Scope and looks very smooth with no noise. I'm at loss if any other input could cause the MPU to freeze like this. I think 95% the MPU is defective.

 

[Retrovert]

If I correctly understand this, your reset is remaining low which is holding the CPU in a reset state so it won't start.

Check the power-on reset (POR) circuit which is an RC timer feeding a schmitt trigger. The C is something like 1 uF, and if that capacitor is bad the circuit won't reset. Make sure the POR is being properly generated. From what you say it sounds like it is not. That is far simpler than the CPU being cooked.

Look at the timing for the CPU on good deck and compare to the dead one.

Something else may have the ability to reset that logic. Look for it. Might be an input from some other circuit which delays the CPU until some other condition happens.

Also, if something is dragging down VCC the reset will never start. A bad capacitor can do that.

 

[Rogersjapan]

Conclusion: POR circuit is operating normal-Goes low on power on then high. CPU timing is spot on from external crystal. Main problem is Pin 15 of CPU as it should read 5VDC B+ (Power in). CPU pin 15 is reading 2.2VDC. When is I disconnect the pin from the circuit I have the 5VDC from the power on circuit sitting on it.. But once connected it drags down the voltage to the 2.2VDC. This is feeding the I/O port of P4 circuits of the chip. This leads me to believe that the chips I/O port is shorted internally and therefor will not send out Resets to the other controllers to start them. I will Scour the land for this Sony CPU but as of now all hits come up NLS. Will keep eye out for a parts unit but these are so rare.

 

[Retrovert]

So when the CPU is connected VCC on the entire board drops from 5 VDC to 2.2 VDC? That sounds like a bad CPU. So, yes, reasonable conclusion.

Check all of the authorized Sony repair centers and distributors. I have located ancient ICs this way. Deadstock lives in peculiar places.

 

[Rogersjapan]

Just an update. Both CPU's on TC-WR99ES and My TC-WR-901ES were bad. Got parts units and now proud to have 2 operational units. I keep searching for the CPU's But have come up empty finding the parts.

 

[jdurbin1]

Nice save. Glad my TC-WR99ES has only misbehaved to the tune of needing replacement belts, so far - hoping it stays that way.

John