SN74CBT3384A: Audible signal through the switch during off-state?

Hi Bo,

Welcome to E2E. 

Just to double-check, what are your control signal voltage high levels (VIH)? Could you measure both /OE pins to confirm they are above the VIH threshold of 2V?

Additionally, could you measure the output "bleeding signal" current when the switch is confirmed "OFF"?

To confirm, do you need this exact configuration? (10 bits with each set of 5-bits controlled separately)

Thanks!
Kate

The control signal to the ~OE pins goes to about 4.3V when high, so definitely above the 2V threshold.

I need a bus switch with at least a 5-bit-wide bus, it doesn't really matter beyond that. A 10 bit bus with 5 bits controlled separately is ideal, and the 24 SOIC package that's pin-compatible with the SN74CBT3384ADWR is ideal because my prototypes are not terribly simple to build.

I'm not exactly sure how to measure the current of the bleeding signal in any quantitative way as I have no signal generating hardware. What I could do is route audio through the switch and into the microphone jack of my computer and record the sound that bleeds through.

Hi Bo,

If you could, I would appreciate the recording of the audio through the switch while "ON" and the recording of the audio you are hearing while the switch is toggled "OFF". Could you include a description of the test conditions as well?

To be clear, are you noticing this on the output of the same "OFF" 5-bit bus as the input? Or, is the other 5-bit bus where you are hearing the "bleeding"?

Do you also happen to have a schematic of this portion of the system? This could be very helpful as I assist the debug.

Thank you!
Kate

I have created a recording of the audio and I have created a simplified schematic to illustrate my issue.

In the attached .mp3 file you will hear about ten seconds of Viva La Vida by Coldplay. For the first 4.5 seconds of this recording, SW1 is closed and SW2 is open. For the remaining duration of the recording, both switches are open, leaving both ~OE pins high and the bus should be, as far as I understand, in its "off" state and no signal should travel from PC_AUDIO_OUTPUT to PC_MIC_INPUT. As you will notice, a lot of signal bleeds through.

To answer your other question, if I connect PC_MIC_INPUT to any other output of the bus switch, I get no sound at all (the expected result).
simpleschematic.pdf

Hi Bo,

Thanks for sending the recording and the schematic. Do you experience this same issue when you try the input/output on the other bus controlled by SW2?

Best regards,
Kate

Hi Kate, yes. I have purchased many of these SN74CBT3384A chips and used them in this design, and the signal bleeding I experience always occurs. Input x of switch y always bleeds to output x of switch y when switch y is in the off state.

Hi Bo,

Another double check you can try is to short the PC_MIC_INPUT to PC_AUDIO_OUPUT and see what that sounds like.

Thanks,
Kate

Hi Kate,

When I short PC_MIC_INPUT to PC_AUDIO_OUTPUT, the sound is always clear and not effected by the state of the switch.

Hi Bo,

Thanks. In order to best help you debug, I would need to see the input and output waveforms with the switch on and off. Do you have any instruments to measure the signals? Could you share the full schematic as well?

Best regards,
Kate

Here's the schematic. I've already posted the requested waveform; see my mp3 file. If you want to see it instead of hearing it, I suggest you download it and open it in Audacity.schematic.pdf

Hi Bo,

Thanks for sending the schematic. I will review it and get back to you soon.

Best regards,
Kate

Hi Bo,

Thanks for your patience.

Could you fill out the attached table by testing your full system? SN74CBT3384A troubleshooting.xlsx

For each of the four logic combination states, record the actual voltage and the control pin node and the sound observation at the output of each bus.

This test may help us narrow down in exactly which state we hear unexpected results. If the device is in fact "bleeding signal", I am wondering why you heard the expected sound even with one bus in the "OFF" state.

Additionally, I would like to confirm what component is symbolized by the yellow boxes?

Thank you!
Kate

Please review the messages I have posted. Please review the simplified schematic I posted. The SN74CBT3384A is drawn, in the simplified schematic, as three distinct pieces.

• Once piece is solely the power supply and ground.
• One piece is one of the 5 bit bus switches.
• One piece is the other 5 bit bus switch.

Please review only the uppermost of the three distinct pieces drawn in the simplified schematic.

According to the datasheet, when /OE is high, AX should be disconnected from BX for any X between 1 and 5 inclusive.

The problem that this thread is all about is the fact that when my /OE pin is high, measured at about 4.3v (which is well above the logic threshold mentioned in the datasheet), my AX appears to be connected to my BX for any X between 1 and 5 inclusive.

When my /OE pin is low, the behavior which is described in the datasheet appears to be correct, as my AX is connected to my BX for any X between 1 and 5 inclusive.

The problem I describe as signal bleeding is that, for example, when I have an audio source connected to A4 and an audio receiver at B4 and /OE for this bus switch is high at about 4.3v, the audio receiver picks up large amounts of signal from A4 which it should not have access to.

Please review the audio that I posted to see what I mean. During this audio, I first play music through the switch for a few seconds while the /OE pin is low and everything sounds the way I would expect it. When my /OE pin goes high roughly 4.5 seconds into the sound file, you hear a very awful noise which is somewhat similar to the song that should be blocked out by the now inactive bus switch.

Kate Dickson said:

If the device is in fact "bleeding signal", I am wondering why you heard the expected sound even with one bus in the "OFF" state.

This is an incorrect assessment of the situation. Suppose that all busses are in the OFF state. This would mean that all inputs are disconnected from all outputs, right? This would mean that any audio playing into the inputs should not appear at all on the outputs, right? The sound I expect to receive from the bus switch's outputs right now is no sound at all: it should be silent. However, I very clearly hear pieces of the signal which is at one of the bus inputs. This is what I refer to as 'bleeding signal'. Please review the messages I have posted and the diagrams I have shown.

Kate Dickson said:

Additionally, I would like to confirm what component is symbolized by the yellow boxes?

This is a kicad standard symbol for jumpers. The input panel is a removable piece connected through a male and female 20 pin jumper. The second pair of jumpers is for connecting this instance of this circuit to other instances of the same circuit. All of the information pertaining to the jumpers and the serial input selection device is fully irrelevant to the problem at hand. Please direct your attention to the simplified schematic I've posted and please review my messages and audio file.

If you would like any of this explained in more detail, please let me know immediately.

Thank you.

Hi Bo,

I appreciate the thorough response and I understand the concern.

It appears that this may be related to the off-isolation of the device. In order to test this, could you provide the following information?

What is the maximum voltage amplitude of the signal?
What are the loading conditions on the output, as this information is not available on the schematic?
What is the source impedance?

In addition, I know you have tried testing a short across the device. Could you observe/send the results of the audio playback with the device completely removed?

Thanks!
Kate