PAM8016 haptic driver analog input

Thread Starter

ahov90

Joined Feb 26, 2024
7
Hi! Does anybody have an experience with PAM8016AKR? I assembled simple evaluation board: Vdd = SD = 3V3, vibro between OUT+ and OUT- and tried to supply to the inputs analog voltage, not PWM: Vin_minus = 0V, Vin_plus 0..3V3, and discovered that smallest Vin_plus= 3mV even produces full-scale output (Vout_plus - Vout_minus) = 3V3 almost.
I produced resistors system (see schema) to the inputs and now for gradually increased Vin_plus =0V..2V6 I receive (Vout_plus - Vout_minus) = 0V..3V3 gradually increased also which is exactly what I want. The question: why there is nothing about that in datasheet? Or I don't see something obvious?
1708965252621.png
 

Thread Starter

ahov90

Joined Feb 26, 2024
7
Yes capacitors is another confusing part.
But datasheet says:

"The PAM8016 support flexible inputs with differential, single-ended or PWM inputs to ease the host microcontroller in driving the haptic elements."

I guess single-ended input is exactly what I have played with.
 

Irving

Joined Jan 30, 2016
3,890
What do you have connected at the output while testing?

The datasheet is quite clear on the need for input capacitors on page 6.

"The application diagram shown in Figure 1 illustrates the connection between the application processor and PAM8016. For nominal-differential or single-ended analog-control signals generated by the application processor only the typical input decoupling capacitors, Cb1 and Cb2, are required for screening out unintended noises to PAM8016 (as shown in the left of Figure 1).

However, if the PWM signals are provided, the low-pass RC filter is suggested to filter out the high-frequency noises induced by PWM driving signal (as shown in the right of Figure 1). Because typical haptic elements operate in the range of a few hundred hertz (~300Hz or below), it is suggested to set the cut-off frequency for the low-pass RC filter network (R1 C1 and R2C2) to 10Khz or less to eliminate the untended noise impact while driving the haptic element."

Also later on page 6:

"Input Capacitors (Ci )
In the typical application an input capacitor, Ci, is required to allow the amplifier to bias the input signal to the proper DC level for optimum operation. "
 
Last edited:

Thread Starter

ahov90

Joined Feb 26, 2024
7
What do you have connected at the output while testing?

The datasheet is quite clear on the need for input capacitors on page 6.

"The application diagram shown in Figure 1 illustrates the connection between the application processor and PAM8016. For nominal-differential or single-ended analog-control signals generated by the application processor only the typical input decoupling capacitors, Cb1 and Cb2, are required for screening out unintended noises to PAM8016 (as shown in the left of Figure 1).

However, if the PWM signals are provided, the low-pass RC filter is suggested to filter out the high-frequency noises induced by PWM driving signal (as shown in the right of Figure 1). Because typical haptic elements operate in the range of a few hundred hertz (~300Hz or below), it is suggested to set the cut-off frequency for the low-pass RC filter network (R1 C1 and R2C2) to 10Khz or less to eliminate the untended noise impact while driving the haptic element."

Also later on page 6:

"Input Capacitors (Ci )
In the typical application an input capacitor, Ci, is required to allow the amplifier to bias the input signal to the proper DC level for optimum operation. "
Yes, exactly, it is the point. The datasheet says:
"The PAM8016 support flexible inputs with differential, single-ended or PWM inputs to ease the host microcontroller in driving the haptic elements."

What is single-ended input in that case? Microcontroller can provide or DC plus or GND or pulses = PWM.
PWM schema is specified separately.

So for what is nominal-differential or single-ended analog-control signals for which the typical input decoupling capacitors, Cb1 and Cb2, are required. Because of capacitors - they can not be DC. So what are they?
 

Irving

Joined Jan 30, 2016
3,890
Single-ended or differential can apply equally to digital and analog signals. The examples given, reproduced below are for single ended analog (left) and single-ended digital (right). The capacitors are required to ensure the internal bias point for the input stage is not affected by external circuitry. The capacitors, in conjunction with the chip input impedance and source output impedance creates a low-pass filter to remove noise (analog) or filter high frequency switching noise.

Later on page 6 it actually clarifies that the inputs are biassed to VDD/2 so the capacitors are required for virtually any input configuration.

1709137425563.png
 

Thread Starter

ahov90

Joined Feb 26, 2024
7
Single-ended or differential can apply equally to digital and analog signals. The examples given, reproduced below are for single ended analog (left) and single-ended digital (right). The capacitors are required to ensure the internal bias point for the input stage is not affected by external circuitry. The capacitors, in conjunction with the chip input impedance and source output impedance creates a low-pass filter to remove noise (analog) or filter high frequency switching noise.

Later on page 6 it actually clarifies that the inputs are biassed to VDD/2 so the capacitors are required for virtually any input configuration.

View attachment 316426
First of all, thank you for your patience!
Would you please tell, how this schema can create low-pass filter?
1709137425563.png
Schema on the right is for PWM input, OK. But what kind of signals supposed to be supplied to the inputs of the left schema?
 

Irving

Joined Jan 30, 2016
3,890
Would you please tell, how this schema can create low-pass filter?
My bad, typo, its a high-pass filter.

1709146507619.png

Schema on the right is for PWM input, OK. But what kind of signals supposed to be supplied to the inputs of the left schema?
Analog, as per the caption. Typically a varying voltage that represents speed (frequency) or amplitude (for fixed speed vibrators).
 

Thread Starter

ahov90

Joined Feb 26, 2024
7
My bad, typo, its a high-pass filter.

View attachment 316446


Analog, as per the caption. Typically a varying voltage that represents speed (frequency) or amplitude (for fixed speed vibrators).
is it suggested equivalent schema ? If so - we can remove input capacitors and supply DC to the inputs, receiving relevant DC output, right?
So I continue to wonder why in datasheet they do insist on input capacitors and don't mention a possibility of DC input. Taking into account they mention microcontroller output, and it is highly unlikely to be varying analog voltage (I understand we can get varying analog voltage from microcontroller's internal DAC but it is quite strange to consider the option and don't consider simplest microcontroller's DC digital).
 

Irving

Joined Jan 30, 2016
3,890
is it suggested equivalent schema ? If so - we can remove input capacitors and supply DC to the inputs, receiving relevant DC output, right?
So I continue to wonder why in datasheet they do insist on input capacitors
Not easily as DC operating point is internally set and directly applying DC to the input would upset that. You'd probably have to ensure that both inputs are biassed to around VDD/2 and that operational signalling maintained that DC offset from ground level. Since that's likely to compromise the input stage they use capacitors to isolate the internal DC bias. The data-sheet doesn't specify common-mode rejection so I suspect the input stage's ability to handle common mode signals is poor , another reason to use capacitors

Remember this is not an op-amp. The output is a 400kHz PWM which synthesises the appropriate drive signal for the vibratory device, not a pure static voltage. Its more akin to a class-D audio amplifier (which other devices in their range, eg 8013/8014, are).
 

Thread Starter

ahov90

Joined Feb 26, 2024
7
yeah. I see. My schema which I assembled by trials and errors looks like something like what you suggested.
Anyway it is strange that input biasing around VDD/2 is mentioned between the lines almost, in the chapter about capacitor polarity choosing.
Thank you for your help!
 
Top