PCM3793A: RRTC is longer than the set value.

Hi Jeff,

Recovery time is measured with an AC voltmeter.
I have confirmed that the level of the system slowly increases over a period of 20 to 30 seconds, whereas it would normally take 3.4 seconds.
I have confirmed that the recovery time is 3.4 seconds.

A register map is attached.

0x47	0x04
0x4D	0x1F
0x53	0x13
0x54	0x01
0x55	0x00
0x59	0x77
0x5C	0x00
0x5D	0x00
0x5E	0x00
0x5F	0x00
0x60	0x00
0x61	0x69
0x62	0x3D
0x63	0x0A
0x64	0x0D
0x65	0xF1
0x66	0x3A
0x67	0x66
0x68	0x00
0x7D	0x1C
0x40	0x40
0x41	0x33
0x42	0x14
0x43	0x40
0x44	0x3F
0x45	0x39
0x46	0x01
0x4B	0x00
0x51	0x40
0x5A	0x01
0x49	0x0A
0x56	0x50
0x48	0x03
0x58	0x45
0x49	0xBC
0x4C	0x03
0x4C	0x00
0x4A	0x41
0x52	0x33
0x57	0x32
0x4F	0x0C
0x50	0X00
sleep	500
0x49	0xBD
0x57	0x31
0x4F	0x0C
0x53	0x83

Best Regards,
Ryu.

Hi Ryu-san,

Thank you. Can you tell if there is a difference in the recovery time if the Expansion Level Control is at other values, such as default 0dB, 6dB "01", or 12dB "10"?

Also what are the clocks being provided to the codec?

Thanks,
Jeff McPherson

Hi Jeff-san,

Longer times in all settings.
Here are the measurement results.

　　＜1＞ ALC-Exp：24dB、PG3：0dB　→　Recovery Time：30s
　　＜2＞ ALC-Exp：14dB、PG3：10dB　→　Recovery Time：28s
　　＜3＞ ALC-Exp：6dB、PG3：18dB　→　Recovery Time：20s
　　＜4＞ ALC-Exp：0dB、PG3：24dB　→　Recovery Time：6s

Can you please confirm if this is the case there as well?
I suspect that there is a problem with our control.

I understand the following about the operation of ALC.
Is this correct?

I understand that ALC is a function to prevent waveform clipping (distortion) by lowering Gain to compress when a signal above a certain level is input.
For example, if ALC-Exp is 24dB, ALC operates by decreasing 24dB when the signal input exceeds a certain level.
On the other hand, if ALC-Exp is set to 6dB, the input range over which ALC operates is 6dB, and if ALC-Exp is 0dB, the waveform will clip when the signal input exceeds a certain level.

Best Regards,
Ryu.

Hi Yamashita-san,

Can you share the waveform plots/recordings of those measurement results? It will help be better understand the difference.

One additional test would be helpful is if the recovery time is measured with different step sizes. For example a signal going from -6dB to -18dB step compared to --6dB to -32dB step. This codec is very old so I am not familiar with it, but I am wondering if the recovery time settings acts more like a time constant instead of a fixed amount of time.

Your understanding is mostly correct, but ALC-Comp is the parameter that reduces large level signals to prevent clipping. If ALC-Comp is -6dB, then the maximum signal level will be reduced by -6dB. Signals above a certain threshold will be reduced by some ratio depending on the expansion level, where the 0dB input signal will result in -6dB output. If the ALC-Exp is 24dB, then signals below a certain threshold will be boosted by 24dB. 

Best regards,
Jeff McPherson

Hi Jeff-san,

Waveforms are attached.
What parameter should be used as the step increment with step size?
It would be helpful if you could tell me the setting values.
Also, is it possible to reproduce it at your company?

Best Regards,
Ryu.

Hi Yamashita-san,

The parameter is step function amplitude delta. -6dB steps should be enough starting from 0dB. Keep the expansion and compression setting the same.

So the functions should be 0dB -> -6dB, 0dB -> -12dB, 0dB -> -18dB, etc. to -36dB at the lowest.

Is the plot you shared using a single tone sine wave? The time scale is so large that I can't tell. If not, please use single tone sine wave.

Unfortunately this EVM has been long obsolete, so I don't have the ability to double check here.

Thank you and best regards,
Jeff McPherson

Hi Jeff-san,

It uses a 1KHz monophonic sine wave.
We have confirmed that the phenomenon follows only ALC-exp.
Changing other gains such as PG3 did not change it.

Best Regards,
Ryu.

Hi Ryu,

Jeff is Out of office and will update the thread by next week,

Hi Yamashita-san,

So a sinewave that steps from 0dB to -24dB will have the same ALC exp recovery time as 0dB to -36dB? The step size does not matter, it is always the same?

I also forgot I asked earlier, what are the clock values being provided to the codec?

Thanks,
Jeff McPherson