I've built the cassette player with speed control. Everything is fine, except the significant noise in audio output generated by the DC motor. The motor’s wires are as short as possible and shielded well, the motor itself has a separated power supply (see attached circuit).
It would make sense if it had one psu for audio and motor shield pcbs, but there is no electrical connections between those two pcb.
How the motor can induce noise this way? Any suggestions to reduce it?

 

I am going to assume that the noise is coming from the motor itself, not the power supply.
Small DC motors use a split ring commutator and brushes.




The split ring interrupts the DC current and reverses the current two times on every rotation of the coil. Every time this happens you can observe electrical sparks between the brushes and ring. This generates noise.

Try putting a 0.1μF capacitor across the motor terminals.

 

Try putting a 0.1μF capacitor across the motor terminals.

And, if the motor has a metal case, another two 100nF between each terminal and the case.
All of these should right by the motor.

 

In your schematic, you appear to have your motor connections shorted. See D1 and the connection from its cathode to its anode.

C1 is great for bulk storage, but it will help to put a small ceramic capacitor there as well. 0.1uF is usually sufficient.

Also, ensure your power supply leads are twisted together and your motor leads as well. You want minimal space between the supply current and the return current paths.

 

Try a lower PWM frequency, just above the frequency range of your audio circuits. 20Khz is often used. It will not radiate from the wiring and motor as much as 122Khz does.

 

The circuit may be working but that schematic won't, Q1 creates a direct short across the DC Motor Supply.

 

A PWM speed control for a motor in an audio device is a very poor choice because it will certainly put noise into the system. That is why they use a linear analog speed control. No, it is not as efficient, some times efficiency is not the most important thing, and low noise is.

 

I am going to assume that the noise is coming from the motor itself, not the power supply.
Small DC motors use a split ring commutator and brushes.
The split ring interrupts the DC current and reverses the current two times on every rotation of the coil. Every time this happens you can observe electrical sparks between the brushes and ring. This generates noise.

The motor itself produces an audible noise, but it wouldn't be a problem if it didn't induce noise the audio lines.

Try putting a 0.1μF capacitor across the motor terminals.

As you can see on the schematic, I had made some precautions putting C2 across the terminals and C3, C4 from each terminal to the motor case. 0.1μF caps didn't help a lot, and I used 680nF caps that made some significant noise suppression, but it didn't effect on that "whirling" sound at all.

 

And, if the motor has a metal case, another two 100nF between each terminal and the case.
All of these should right by the motor.

Yep, I have them all, but 680nF for all three caps. I had experimented a lot with values and 680nF did the job, but didn't solve my general issue.

 

In your schematic, you appear to have your motor connections shorted. See D1 and the connection from its cathode to its anode.

This diode used to eliminate voltage spikes and it is put in parallel with terminals with opposite polarity.

C1 is great for bulk storage, but it will help to put a small ceramic capacitor there as well. 0.1uF is usually sufficient.

I even have a space for it on my pcb, but since it didn't have any effect I didn't use it. Anyway I think I will put it back.

 

Try a lower PWM frequency, just above the frequency range of your audio circuits. 20Khz is often used. It will not radiate from the wiring and motor as much as 122Khz does.

You know, 7.8 Khz works for my motor well (much less audible and electric noise), but I can't get the desirable speed rage with this frequency. So I switch to 122Khz and decide to separate power supplies to reduce the effects of motor noise on audio and other power lines. But I don't reduce anything in this way.

 

The circuit may be working but that schematic won't, Q1 creates a direct short across the DC Motor Supply.

Yep, you're right, thanks to notice. This is my circuit drawing problem, hopefully the schematic doesn't have this issue. I've edited the attached circuit.

 

Here is another possible path for the noise, which may be the transformer action of the motor magnetic field into the tape head itself. Given the very small magnetic field change from the tape, there must be a lot of amplifier gain, and tape heads are known for magnetic noise pickup. That is why some tape decks have several layers of magnetic shielding around the motors. And that is why most of them do not use PWM speed control.