Sony WM-DD33 speed issues (and more..)

Is it normal, even when in sync, to see impulses on Pin 9? Or should it be more like a constant value?

 

A little bit more information. This picture is the voltage at the motor when it is locked. Are these spikes normal operation? Could this be a symptom of some mechanical issue that makes the voltage rise because the feedback loop detects it is slowing? Could this then trigger the speed issues?

 

Today I've received and restored to kind of working state DD22 from ebay, so new discoveries are in the way.
In the next few days I'll check the waveforms, auto stop theory and everything else

But for now, I want to say I love how stable quartz lock playback is, especially on electronic music. Day and night compared to FX877, even with partially disintegrated rubber ring on disc.

 

OK, so I could't sleep without some experiments.
So here are a lot of waveforms for every pin of servo IC601 CXA1423, recorded from normally working Walkman DD22:

PIN1: Signal from FG coil, around 200mV P-P, 1.2V offset from GND
PIN2: 90mV P-P, 1.2V offset from GND
PIN3: FG amplifier output, limited to 1.6V
PIN4: Sawtooth, 0 to 1,26V. Stable, without peak fluctuations.
PIN5: VRef, very stable 1.62V
PIN6: Very small downpeak, somewhat similar frequency to pin 9
PIN7: Power on input, 2.7V
PIN8: Power, 2.8V
PIN9: Short impulses, 2.6V P-P, about 1kHz frequency, duty cycle depends on motor load.
PIN10: Short impulses, 600mV Peak, about 1kHz frequency, peak voltage depends on motor load.
PIN11, PIN12: some stair type signal, height depends on motor load
PIN13, PIN14: Motor current request amplifier, depends on motor load
PIN15: Motor transistor control output, depends on motor load, 2.1V offset from GND.

Also, pin5 of IC701 signal is similar to Pin4, not so stable, but also not going to 6V as on my DD33.

 

Also, signal on PIN9 highly correlates to signal on pin 4.

Hope I will not be banned because of too many images in 1 post

 

Also, measured voltage directly on the Motor+ output.
Correlates with Pin15 with the IC, no tiny peaks of voltage as on your graph. Maybe motor is getting old? Frequency is very high, does not look like interference from control circuits, and spike goes up, meaning lower impedance on motor at those moments (or also could be interference on the input of the transistor) . Try cleaning the shaft of the motor, or if it could be disassembled, also there will be a good thing to clean brushes and their contact surfaces.

BTW, are you using batteries or bench power supply to power the walkman during tests? Because 30us pulse period means about 33hKz frequency, which is similar to what some power supplies are using.

 

Amazing!! Awesome work!

I can’t thank you enough for all the work you have done.
I had a very busy week and could not properly follow this topic. Next week will be even worse, but I will try during this weekend to also do a complete capture of all the waveforms to compare with yours.

 

Hello again.
I have bad and OK news for us.
The shown symptoms unfortunately are from dead/dying servo IC CXA1423N. I've swapped them between my DD22 and DD33, and symptoms traveled with the chips...
After some tinkering, I've found a way to make it work half-decent. By half decent I mean we still have auto stop, low battery control, but loose quartz speed precision.
To make it work, you have to de-solder only pin 6 of Quartz IC (IC602), leaving it in the air. All other parts of it should be connected. Then raise speed of the motor turning the potentiometer until it will start normally every time, preferably with calibration tape. PLL jumper should be connected during this step, otherwise motor will be wobbly, or will not start.

What we just did by disconnecting pin 6 of quarz IC: We removed phase feedback loop between quatrz and servo (it is used to precisely and instantly (during every pulse from FG coil) react to minor capstan speed changes, and adjust motor current to compensate for it). So now quartz IC works in kind of standalone state, providing right control frequency to the servo IC, bypassing damaged sawtooth generator comparator. Speed control is still happening, instead now, not the quartz IC helps servo maintain capstan speed, but servo tries to match capstan speed to quartz output. So after that modification wow should be like dd quartz, flutter like dd without quartz, with slower start speed, also drifting a bit with ambient temperature change.

This is not a fix to restore walkman, just a workaround to make it work. And seeing how difficult it is to source original servo chips, this solution makes me feel good and sad at the same time.

After all the experiments, fully working servo IC now lives in my daily DD33, and DD22 in described state went to one of the local Walkman collection enthusiast.