Reducing a digital signal's rise-time from 2ns to 60ns

Hi Faisal, 

Can you share a schematic of the design? Also, what is the behavior of the TMUX6112 when you provide it with these fast rise-time inputs? As far as I'm aware there shouldn't be a maximum bandwidth limitation as long as you wait for the enable turn on/off time. 

Regards, 

Connor 

Hello Connor

Here is the schematic and output waveform ( Green) that is obtained after summing the outputs of the switch. Please disregard the yellow  wave.

I can see  Spikes at the transition of the SEL inputs. Also, the off channels also show spikes at the transition of SEL input of any other channel.  

Thank you

Faisal

Hello Bryan

I looked at the clock buffers. The output bandwidth still looks very high for all of the available chips. 

Thanks

Faisal

Thanks a lot Emrys. I was thinking of doing things in an unnecessarily difficult way. I will apply this in my design. Btw, can I apply the same technique to reduce the Bandwidth of an LVDS signal (250MHz) to match with a LVDS TTL receiver( DSLVDS1048 200MHz)?

Best

Faisal

Hey Faisal,

I'm not very familiar with the LVDS standard, but I've seen this method used across almost all signal types for numerous reasons. I would suspect that the edge rate of the LVDS signal would be less critical since it's likely going into a CMOS device rather than an analog switch.

Hello Emrys

So I followed the suggestions and made an other board. Here  the rise and fall time of the select signal is 150ns and the IC is powered by Vdd = 10V (single) I can still see the spikes on the Output pins as well as on the input pins with reduced amplitude on the out. I can also  see spikes during the switching even when there is no power provided to  the TMUX6112 IC. I am confused if it is for bypass/decoupling capacitor or some other reason it behind this. I have used decoupling capacitor of 0.1uF very close to Vdd and VS. Please advice.

Above picture represents the situation when the IC is turned of .Cyan color represents the output signal and the purple color represents the SEL  input signal. Can see spikes on the output pin when switching in SEL happens.

Above picture represents the situation when the IC is turned on . Green color represents the input signal, Cyan color represents the output signal and the purple color represents the SEL  input signal. Can see spikes at both input and output pins when switching in SEL happens.

Thanks

Faisal

Hi Faisal,

It looks to me like you're getting less than 50mV of switching noise -- which is extremely good and shouldn't have any impact on operation. What's the issue?

Hello Emrys

The problem is I am suppose to pass analog signals of amplitude of 50to 500mV. It is throwing that signal off and also reduces the amplitude of the signal as you can see from the picture. 

Thanks

Faisal

I see -- it sounds like you need to have a dedicated analog board and remove any digital circuitry which will cause noise.

Hello Emrys

I am getting the digital signal from an FPGA. So, it is separate already. I am just feeding the digital signal from the board to the input pins of the Tmux 6112 using SMA connectors of 50ohm. I  also using 50ohm controlled impedance in the traces.

Thanks 

Faisal

Hi Faisal,

Can you share an image of your setup?

1. Emrys
2. The board ports are as indicated. The output is fed to an Oscilloscope with 50ohm parallel termination.

Hey Faisal,

Thanks for the image - this helps me to understand your setup.

It looks to me like this mux is drawing power with it's switching circuitry that's coupling across to the analog lines. I don't know if this is inherent to the device or an issue with your specific setup... I also don't know if this is specified in our datasheet. I've notified the mux team to take a look at this thread to see if they have any comments.

I would recommend trying to identify the frequency of those spikes and see if you can improve your bypass circuit for those particular frequencies. Most likely the 0.1uF is good enough, but it can't hurt to add some smaller (0.01uF) and larger (1uF) caps in parallel (you can solder them right on top of the existing SMT capacitor).

Thanks Emrys. Can you tell me if you have any Evaluation board for this switch or for any analog switch? I can't find any. 

I really appreciate your effort.

Best

Faisal

Hello Bryan

Anyway of reducing the injection other than/i addition to using the decoupling capacitor? I will be feeding the outputs to a summing amplifier. But the switch produces spikes at all the outputs during switching so the spikes will get added by the summing circuit. Can you suggest me similar 1:1spst  switch with one channel only? Thanks 

Hi Bryan

Any effective way of reducing the spikes at the outputs, for example, using filter?

Thanks

Faisal 

Hi Bryan

Regarding single channel 1:1spst switch, how is TMUX1101?  I want to avoid  the spikes getting added to all outputs whenever switching occurs so thinking of using 4 single channel ICs instead of using a single 4-channel one.

Thanks

Faisal

Hello Bryan

I am planning to feed a summing amplifier/ fanin circuit with the outputs of the switch. At the moment, I can see spikes at the output of every channel whenever any one of the SEL inputs makes any transition. So for example, if one channel outputs 50mVpp spike then the total for 4channels would be 200mVpp which is huge. So I am thinking of using 4 single channel Switches like TMUX1101  to isolate switches and keep the total spike upto like 50mVpp. Can it be done using TMUX1101? Please let me know.

Thanks 

Faisal