Multiplexer output does not match the input

I attached the PDF file of the schematic...

Hi Harry

Thanks for your advices. To your questions:

-- To determine that the ADC is not causing any side effects on the MUX I did disconnect it. But the effects are all the same as the MUX is connected. So the scope pictures are valid for ADC connected and disconnected.

-- I first had a small cap at the INPOS input pin as a charge reservoir but removed it. The cap at the input of the ADC caused signal changes at the input of the MUX after switching the channel. To avoid this, I removed it and increased the cap at the MUX input. C116, C117, C6, C119 are the charge reservoirs to load the internal 59pF cap of the ADC without any changes in the signal.

-- You are right, the 1k resistor at INNEG is a copy paste misstake which I already changed on the board but not in the schematic. But the MUX effect is still the same.

-- The time scale for the scope picture in my first post is Microseconds. The target is to get the signal stable after changing MUX channel in about 200ns.

Pascal

I did some further test with another device from Analog Devices. I used a AGD508A multiplexer with the following setup:

Input 1 is tied to 3V and input 2 to GND. I'm toggling between these two inputs with 5kHz frequency and observe what voltage I can see at the output. With this device I get the following results:

Is it a general phenomenon of multiplexers that they take some time until they reach the real voltage level at the output according to the input?

Hi Pascal,

Would you be able to remove the ADG758 from the system and test it on it's own? That would be useful as we would be able to see if the issue is being caused by the part or the other devices in the system.

Also in the case of the ADG508, It appears VSS pin has been left floating by looking at the schematic. Is that the case?

Regards,

Stephen

PB,

1. Why would you need 0.1uF capacitors at unused S1 and S8? Or, can you at least change them to something like 1000pF?

2. What is the width of PCB traces for 5VA1 rail (and for Ground)@? If it is not a plane but thin trace, you may want to add 10uF tantalum or aluminum capacitor. It would not hurt to have one such capacitor per rail even if there is a plane.

3. As you already disconnected outputs and performance has not changed, I do not see the need removing the ADG758 from the circuit and testing it separately.