Electronics > Projects, Designs, and Technical Stuff
switching gain with 4052 multiplexer on audio output
(1/1)
jmibk:
I'm designing a dsp board and like to have a switchable gain output, schematic for two outputs in the link down there:

http://www.audiowerk.at/downloads/temp/eev/dsp_output_stage/Schematic_ADAU1701_4x4_2020-06-06_09-39-04.pdf

VOUT0 and VOUT1 are voltage outputs of an DAC, 0 - 3,3V, zero line is on 1,65V.
MPXo1_0 and MPXo1_1 are coming from an 74HC595 shift register that is controlled by an mcu.

Now the problem - maybe I oversee something (just mention channel 1, other have the same problem):

- feeding the system with a 1kHz sinewave
- CHo1_0 views the sinewave at it should
- CHo1_-10, CHo1_-20 and CHo1_+10 are either clamped on positive or negative rail dc, often fragments of the sinewave are visible, but mainly there is this huge dc offset.
- If i short CHo1_0 to ground and open it after that, everything works as aspected.
- changing the value of R80 to a smaller value like 200 ohms doesnt change anything.

If I remove the 74HCT4052D (U5) the offset is gone, lines CHo1_-10, CHo1_-20 and CHo1_+10 are zero volts.
Why does the 4052 pull the line to + or - 12V?

I know that it is not ideal drive the OPs with 12V and the 4052 with 5V but it should do right if the signal isnt greather than 5V.

Thanks for any hint or solution!

Kleinstein:
The 4052 supplied with +-5 V should not cause clamping to 12 V levels.

One may have to take care of the other channel - to high a voltage at channel 2 may also effect channel 1.
CMOS MUX tend to be sensitive to latchup, if the voltage goes beyond the supplies. Transient on turn on could go beyond the planed +-5 V signal level.   As a protection I would have resistors in the +-5 V supplies. The static supply current is very low, so some 10 K would not matter, but can limit the worst case current and the chip should recover from a latch up.

There are a few not directly related points:
The TL074 OPs don't like driving capacitive loads - even 2 m of cable can be too much. So there should be some series resistance (e.g. 50-100 Ohms) at the outputs.
The circuit is rather low impedance. R80 is more like not needed at all. R118 - R120 already provides enough DC path to ground. R5 and R10 give quite some load to the switch. The switch on resistance is not that linear at higher current. So R5 and R10 should be higher value, more like >10 K each. Even than it would still effect the -10 dB setting a little (more like -11 dB).
Navigation
Message Index
There was an error while thanking
Thanking...

Go to full version
Powered by SMFPacks Advanced Attachments Uploader Mod