I'm facing a problem with the ADG1609 MPX. We need to turn on and off the Vcc of the switch in our design. When the power pins are supplied with 10V and GND the device shows a significant spike at its outputs (around 1.5V). Is ther any solution to handle this problem and to eliminate this spike?
I attached a screenshot of our scope to show the behaviour.
Green = 10V supply
Yellow = Output S4A
Blue = Output S4B
I can measure the spike in any configuration. The shown outputs are only an example.
Alle CTRL-Pins are tied to GND or Vcc.
Thank you in advance
I think that your problem is caused by the internal MOSFET capacitances.
- Does the time constant changes with resistive loading (for example, did you tried with a 10 MOhm scope probe and a direct 1 MOhm direct connexion to the scope ?
- Does the amplitude of the spike changes with capacitive loading ? (see what changes when you add 100 pF to ground) ?
I such is the case, the only way i can see is ramping the power supply, i.e limiting the dV/dt of your +10V by some low pass filter .. But are you sure that this capacitive coupling is not external to the switch ?
you are right. The time constant changes with changes off the resistive load and capacitence. External capacitence is not applied to the device. I'll try the Vcc ramp up circuit asap.
The www told me that switches with higher Ron have smoother/smaller spikes. Can you confirm that from you expirience?
Thank you very much.
- In the same family of switches, a higher Ron means a smaller area of MOSFETS, so smaller capacitances. But it is very dependent of the technology: compatibles switches with different maximum supply voltage will not have the same parasitics parameters. But on this subject AD guys know more than myself, I suppose ...
- Couldn't you add a switch to ground the output when not in use ? But this switch will have to be powered ... The only switch which is "on" when not powered or biased is made of a discrete depletion JFET such as J109 or depletion MOSFET (not common), and the external control circuit is not very simple ...
As Eleuthere mentioned, the spike you measured on the switch pin when turning the power on/off is caused by parasitic capacitances between the switch pins and the supply rails.
Considering just the IC itself, there are three main categories of parasitic capacitances in a switch:
1. Gate-Source and Gate-Drain capacitances of the MOS transistors used to build the switch cell. These capacitances cause what is called charge injection and the amount of charge dumped on the switch lines depend mainly on the size of the devices and the internal architecture of the switch. Capacitance is proportional with the size and inverse proportional with ON resistance. Choosing a switch with lower charge injection (higher ON resistance) will reduce the contribution of the charge injection on the amplitude of the spike seen on the switch pins.
2. Parasitic capacitances of the ESD protection diodes connected between the supply rails and the switch pins. Since lower charge injection switches lead to smaller size MOS transistors, the overall die size is likely to be smaller which in turn may lead to smaller ESD diodes with smaller parasitic capacitances. However, this depends on the process used to fabricate the device and the actual ESD protection circuitry used by the device.
3. Other parasitic capacitances between the supply rails and the switch pins on the die itself which are dependent on the internal layout of the device.
Although choosing a switch with smaller charge injection does not guarantee that the spike will disappear, it should at least reduce its amplitude. I would recommend trying the ADG1209 instead of the ADG1609.