Post Go back to editing

ADG728 Analog MUX interference between channels

Hello,

i am struggling a problem on ADG728. I am trying to measure optical density over a photodiode array of 8 photodiodes(You can view the image i attached). My problem started with my realization of false readings over channels. When i connect the Analog MUX's S1 pin to D output(via I2C command) and open circuited all other S(2-8) pins on both ADG728 itself(via I2C command) and also on the hardware(picking up the header connections existing between photodiodes and AMUX's S(2-8) pins) i make a correct reading for channel 1. But, if i connect physically a second channel pin(that is already in OFF state) to the photodiode array, (it does not make any change for which channel to use), the reading i am getting from channel 1(the only ON channel) changes dramatically.

After experimenting this behaviour from various channels, i made some measurements with my Fluke 289 multimeter. I disconnected all photodiode header(physical) connections and tried to measure the resistance values between channels on some conditions.

For example, after powering up(+5VDC) while all channels are in the OFF state, there is approximately 60-70 Mohm between each channel. This value is satisfying i guess(compared to the infine value(open circuit) that it should be in ideal MUX). The problem is that, when you turned ON only one channel(via i2c command), for example connect S1 pin to D, the resistance between that turned ON channel and any other OFF channel decreases to approximately 200-300 Kohm. In fact, that value can be measured with multimeter in one direction. That means when i connect the red(positive) probe of the multimeter to the ON channel(S1) and the black(negative) probe to the other one of OFF pins(S2-8). If i change the measurement nodes(black probe -> S1 , red probe -> other OFF Sx pin), i measure infinite (open circuit) value that it supposed to be.

I think this 200K value between ON and OFF channel causes the incorrect measurement i mentioned above. And i guess, this behaviour is unchangeable since it is the nature of ADG728. Can you advice us any other parts which does not have this drawback?? We still prefer i2c communication and small ON resistance(approximately 2.5ohm or as well).

Thanks in advance. Looking forward to hearing your comments and advices...

Tankut ACAR

Design Engineer

tankut.acar@diagnosistem.com

Diagnosistem

Message was edited by: TANKUT ACAR

photodiode_array_measurement_circuit_schematics.pdf
  • Hello Tankut,

    Can you confirm that the 60-70MOhm that you measured is between an Sx pin and the D pin, or is it between any two Sx pins when all the channels are off?

    Regards,

    Sean

  • Hello Sean,

    when i saw your question i repeated the measurements. Below is a small list of measurement results with respect to the multimeter probe connections. There is no physical connections with the Channels(Sx) and the photodiodes in these measurements.

              Red Probe                    Black Probe                    Condition                         Resistance Value

    1.       S1                                S8                              All Sx OFF state                 Approx. 70Mohm

    2.       S8                                S1                              All Sx OFF state                 Approx. 70Mohm

    3.       S2                                S6                              All Sx OFF state                 Approx. 70Mohm

    4.       S6                                S2                              All Sx OFF state                 Approx. 70Mohm

    5.       S1                                D                                All Sx OFF state                 Shows Open Circuit

    6.       D                                  S1                              All Sx OFF state                 Approx. 310Kohms

    7.       S2                                D                                All Sx OFF state                 Shows Open Circuit

    8.       D                                  S2                              All Sx OFF state                 Approx. 310Kohms

    9.       S6                                D                                All Sx OFF state                 Shows Open Circuit

    10.     D                                  S6                              All Sx OFF state                 Approx. 310Kohms

    11.     S1                                D                                S1 ON, All other OFF          Approx. 4ohms

    12.     D                                  S1                              S1 ON, All other OFF          Approx. 4ohms

    13.     S2                                D                                S1 ON, All other OFF          Shows Open Circuit

    14.     D                                  S2                              S1 ON, All other OFF          Approx. 310Kohms

    15.     S1                                S2                              S1 ON, All other OFF          Approx. 310Kohms

    16.     S2                                S1                              S1 ON, All other OFF          Shows Open Circuit

    17.     S7                                D                                S1 ON, All other OFF          Shows Open Circuit

    18.     D                                  S7                              S1 ON, All other OFF          Approx. 310Kohms

    19.     S1                                S7                              S1 ON, All other OFF          Approx. 310Kohms

    20.     S7                                S1                              S1 ON, All other OFF          Shows Open Circuit

    21.     S3                                S7                              S1 ON, All other OFF          Approx. 70Mohm

    22.     S7                                S3                              S1 ON, All other OFF          Approx. 70Mohm

    I also retried the measurements with a new ADG728 soldered and the results were exactly same.

    I guess, you will find these measurements helpful for my question. If you examine the circuit diagram that i posted before, you can understand that only one channel will be turned ON in one time. After setting a channel ON, sufficient amount of time is being waited, for the output become stable. I guess that the resistance measured in 14th experiment above interfere with the S1 result making me incorrect read of channel 1. And i guess that this reading error will be higher with all other channels' interference too. Do you have any idea for this failure to be corrected?? And maybe you can advice me some other circuit topologies including Analog MUX. I do not want to use 8 times of the opamps and digital pots existing on the output of the MUX. Maybe you can advice me any other Analog MUX that is more reliable for my situation and which is more likely behave like an ideal Analog MUX.

    Thanks again. Your helps will be greatly appreciated.

    Tankut ACAR

    Design Engineer

    tankut.acar@diagnosistem.com

    Diagnosistem

  • Hello Tankut,

    I don't believe the Fluke handheld meter is an appropriate device for measuring the resistance in this case; tests 5 & 6, 7 & 8, 9 & 10 should show open circuit in both directions. The discrepancy in readings would indicate that the multimeter is reading the nA of OFF leakage differently in both cases. Also, the consistent 310kΩ reading in your other tests indicates that either ranging of the meter is occurring or is an effect of the ON leakage of S1.

    What are the physical characteristics of your diode array? In terms of output voltage and current.

    Regards,

    Sean

  • Dear Sean,

    i do not know why you believe the Multimeter is not appropriate but i think it is appropriate since it can measure the open circuit conditions and 70Mohm values well. Anyway, i made the photodiode readings without using Fluke. I made the measurements via ADC of my MCU connected to the output node of my circuit labelled "OD_voltage(Optical_Density voltage)". And as i told you before, with just one photodiode connected to the S1 pin(all other Sx pins are not connected to the photodiode array - open circuit) and S1 pin routed(connected) to D output, the measurements i made was correct. In this connection state, while reading just one channel, the result is OK. That means, for example if no light is directed on the photodiode, i am reading just a few millivolts, and with increasing light intensity, the readings are also increasing proportionally. If i suddenly connect a second photodiode to, for example S2 pin( that is always in OFF state during my tests - disconnected from D ), my reading from Channel 1 changes dramatically. For example, while reading zero light on channel 1(all other Sx pins open circuit), if i connect the header coming from second photodiode which has some light on, the reading of channel 1 changes to 700-800mV. This single behaviour shows that channel 2 (S2) interferes with channel 1(S1) making the reading result of channel 1 incorrect. And i should say that my communication protocol (i2c) is perfect and i am sure that i am sending the correct code that is 0b00000001 for just making S1 to be turned on. You can believe me that i am doing my job well.

    i want to send you a few photos of the system and photodiode array. The photodiode array is a module supplied from China and fairly i do not know much about its specifications. The only thing i know is that i could make some readings with one photodiode connected each time and the readings in this condition were logical.

    Regards,

    Tankut

    Images.rar.zip
  • Hello Tankut,

    What is the maximum voltage of the photo diode array?

    Is there any optical isolation between light from light source 2 and Photo diode 2? I am asking if the increase in voltage seen on photo diode 1 is not an ambient increase in light?

    Regards,

    Sean

  • Hello Sean,

    i measure a voltage of -0.3 ~ -0.4 V approximately on each photodiode in open circuit condition. The voltage is negative since the photodiodes should be used in reverse direction and the photodiode array module has common anode connection topology, that is anodes are connected each other and this connection is grounded. ADG728 is supplied with +5VDC single supply. Do you think that this is the cause of the failure ?? Should i supply the ic with +/- 2.5 V ??

    i think no optical isolation is needed in our system. I guess, you mean that light coming from channel 2 falls on photodiode 1 and affects the result on photodiode 1. If so, i should say that all similar systems have similar measurement system and they work well. The affect of this, if there is any, is very very small and can be ignored. And also, if such a behaviour exists, it is there all the time. The failure happens when i connect a second channel to an OFF Sx pin.

    Thanks for your interest on the subject, and spending your valuable time. I look forward to hearing your comments especially on voltage specifications mentioned above.

    Best Regards,

    Tankut

    I forgot to tell you that i measure -0.3 ~ -0.4 V approximately on each photodiode in open circuit condition. When i connect the photodiode to an OFF Sx pin, the voltage slightly decreases like -0.25 ~ -0.35V. This can also be another failure for us since it can change the measurement's real value to something else which is incorrect ...

    Message was edited by: TANKUT ACAR

  • Hello Tankut,

    Your signal voltage is outside the supply rails for the part. Operating the device in this way is not recommended and we cannot guarantee any specifications.Degradation of performance and lifetime are likely to occur.

    There are internal ESD diodes that will be turned on by your signal and this is will affect both your "ON" switch and "OFF" switch.

    If you can't bias your signal from the diodes at an appropriate voltage, you can use ±2.5V supplies, but you will need to ensure that your digital signals are appropriate to the device ground reference: i.e. VINH = (2V - 2.5V) = -.05V; VINL = (0.4V - 2.5V) = -2.1V.

    Regards,

    Sean

  • Dear Sean,

    Thanks a lot. This was the first time i had used an Analog Mux and i thought that the Analog Mux would behave like an ideal Switch controlled digitally. Anyway, i think i am going to multiple the output OPAMPs and digital switches in my next design. This will be simpler. Thanks again.

    Have a nice day.

    Best Regards,

    Tankut