Selecting a Multiplexer for LTC2380-24

Hello!

I need help selecting a MUX for using with LTC2380. At least 2 differential bipolar channels are required. Signal BW is about 100KHz. I've made a design with the help of Analog Precision ADC tool:

https://tools.analog.com/en/adcdriver/#adc=LTC2380-24&cext=3.3e-9&driver=AD8021&frequency=100000&gain=1&inputSettlingChart=2&multiplexed=true&nTab=1&rext=10&rf=0&sTab=3&samplingRate=800000&sourceType=2&tab=1&thd=200&topology=3&vSupplyNeg=-1.8&vSupplyPos=6.2&vref=4.096&wbNoise=0

The tool is suggesting ADG1209, but according to its datasheet it has (0.15% or -56.5dB THD) and that is much worse than the whole system at 100kHz.

I was also looking at ADG409 and it's datasheet doesn't mention anything about distortion. Should I believe it's so low it can be disregarded?

Best regards,

Vitor.

Parents
  • +1
    •  Analog Employees 
    on Jun 23, 2021 3:27 PM

    Hi Vitor,

    The amount of distortion introduced by a switch, or multiplexer, in a circuit is directly related to non-idealities introduced by the multiplexer and the components of your circuit. The THD introduced by the multiplexers is directly related to its on-resistance flatness, and inversely related to its total on-resistance, the multiplexer input resistance and the load resistance, as shown here: https://www.analog.com/media/en/technical-documentation/application-notes/AN-251.pdf

    To summarize what it says there, in a simplified way, THD is proportional to this ratio: Rflat / (Rin + Ron + Rload). This is a quick way of estimating how different switches compare when adding them to your circuit.

    In the design you sent from the ADC Driver Tool, the multiplexer will be connected directly to the input of the op amp. As the op amp has a very high input impedance, the multiplexer will introduce very low distortion in your circuit. You can see this in the "Noise & Distortion" tab in the tool, where the system overall THD @ 100 kHz is -93.3 dB.

    Another important point is that the distortion spec in our switches and multiplexers datasheets is in fact a THD+N parameter. This differs from the THD spec seen in the tool output, as THD+N has the noise floor added to the calculation.

    Based on the way your circuit is designed, I recommend using the ADG1209, as its low capacitance and charge injection will reduce the voltage glitch and speed up the settling time when switching your input.

    Even though the ADG1409 and ADG409 would have a better distortion performance, based on the equation above. You might risk having settling issues when you switch your input, and voltage glitches. However, if you decide to use one of these two, here are some links that can help you tackle the issues that might be introduced by them:

    https://www.analog.com/en/analog-dialogue/articles/demystifying-data-acquisition-systems.html 

    https://www.analog.com/en/analog-dialogue/articles/front-end-amp-and-rc-filter-design.html 

    I hope this helps!

    Yours sincerely,

    Igor Ono.

  • Thanks for the great answer, !

    I have another doubt. Could I use MAX4618 after the op amp, as it is not bipolar but has about 10 ohm Ron? Would that be better or worse than using ADG1209 before the op amp?

    Also, I hadn't paid attention to the Vcm before the multiplexer. Does it mean I need an extra summing amp in both differential inputs? Won't a positive offset in the upper op amp suffice? What is expected is that Vin(-) will be close to GND + pickup noise, and that Vin(+) lies between -2V/+2V + pickup noise (with respect to GND), using 4.096Vref.

  • +1
    •  Analog Employees 
    on Jun 29, 2021 10:18 AM in reply to bleudoutremer

    Hi Vitor,

    Let's look at the distortion introduced by these multiplexers. In theory, considering a purely resistive circuit, these would be the THD proportions of these two parts:

    • Common elements:
      • R_LOAD = 10 MΩ (AD8021)
      • R_IN = 0 Ω (No resistor before the multiplexer)
    • Multiplexers calc:
      • ADG1209
        • R_ON = 120 Ω
        • R_FLAT = 20 Ω
        • THD ∝ 20 / (120 + 10M + 0) = 1.999E-06
      • MAX4618
        • R_ON = 8 Ω
        • R_FLAT = 1 Ω
        • THD ∝ 1 / (8 + 10M + 0) = 9.999E-08

    Considering this basic calculation, the MAX4618 would have a better distortion performance than the ADG1209. However, there are some caveats to this approach.

    Firstly, it ignores any capacitance your circuit may have. Thus your estimate THD is a straight line across frequency. But this is not true in reality. The ADG1209 has lower capacitance, which could improve your THD at higher frequencies.

    The second issue is that it assumes a linear on-resistance curve.

    The third and last problem is that this is just a proportion of THD performance between two parts. The actual THD value is complex to calculate. As I mentioned before, the high input impedance of the Op Amp would probably make the distortion introduced by these multiplexers negligible.

    Now, let's take a look at the settling time and voltage glitches introduced by the multiplexers. These are related to other specs of the multiplexer, such as switching time, capacitance, charge injection and leakage.

    Switching time is faster on the MAX4618 than on the ADG1209. However, the ADG1209 has lower capacitance, leakage and charge injection, which should reduce the voltage glitches introduced and improve your settling time.

    I recommend that you test this in a SPICE simulator, like LTspice, and see how well your circuit will perform during the transient analysis.

    Lastly, it might be worth it to look at some other details between these two parts. The ADG1209 has a higher temperature range than the MAX4618. Besides, what would the supply voltages for the ADG1209? Are you looking for a single supply device? The MAX4618 seems to operate at 5 V, and its input range is from 0 to +5 V. While the ADG1209 can work with dual supplies.

    Regarding the VCM question, I believe that your IN+ and IN- signals need to have a phase difference of 180º. However, to be certain, you should direct this question to the Precision ADCs group.

    Yours sincerely,

    Igor

Reply
  • +1
    •  Analog Employees 
    on Jun 29, 2021 10:18 AM in reply to bleudoutremer

    Hi Vitor,

    Let's look at the distortion introduced by these multiplexers. In theory, considering a purely resistive circuit, these would be the THD proportions of these two parts:

    • Common elements:
      • R_LOAD = 10 MΩ (AD8021)
      • R_IN = 0 Ω (No resistor before the multiplexer)
    • Multiplexers calc:
      • ADG1209
        • R_ON = 120 Ω
        • R_FLAT = 20 Ω
        • THD ∝ 20 / (120 + 10M + 0) = 1.999E-06
      • MAX4618
        • R_ON = 8 Ω
        • R_FLAT = 1 Ω
        • THD ∝ 1 / (8 + 10M + 0) = 9.999E-08

    Considering this basic calculation, the MAX4618 would have a better distortion performance than the ADG1209. However, there are some caveats to this approach.

    Firstly, it ignores any capacitance your circuit may have. Thus your estimate THD is a straight line across frequency. But this is not true in reality. The ADG1209 has lower capacitance, which could improve your THD at higher frequencies.

    The second issue is that it assumes a linear on-resistance curve.

    The third and last problem is that this is just a proportion of THD performance between two parts. The actual THD value is complex to calculate. As I mentioned before, the high input impedance of the Op Amp would probably make the distortion introduced by these multiplexers negligible.

    Now, let's take a look at the settling time and voltage glitches introduced by the multiplexers. These are related to other specs of the multiplexer, such as switching time, capacitance, charge injection and leakage.

    Switching time is faster on the MAX4618 than on the ADG1209. However, the ADG1209 has lower capacitance, leakage and charge injection, which should reduce the voltage glitches introduced and improve your settling time.

    I recommend that you test this in a SPICE simulator, like LTspice, and see how well your circuit will perform during the transient analysis.

    Lastly, it might be worth it to look at some other details between these two parts. The ADG1209 has a higher temperature range than the MAX4618. Besides, what would the supply voltages for the ADG1209? Are you looking for a single supply device? The MAX4618 seems to operate at 5 V, and its input range is from 0 to +5 V. While the ADG1209 can work with dual supplies.

    Regarding the VCM question, I believe that your IN+ and IN- signals need to have a phase difference of 180º. However, to be certain, you should direct this question to the Precision ADCs group.

    Yours sincerely,

    Igor

Children