| Gopal Amlekar 01/16/10 02:01 Read: 157 times Pune India |
#172511 - RE: The input multiplexer is damaged Responding to: Kai Klaas's previous message |
if you connect a too high capacitance across the input of ADC, then a brief overvoltage condition can result at this input during power down, when this capacitance is still charged while Vcc is already down!!! So, the time constant of the input filter should be way less than the time constant of power down In some situations, it will be possible to have a diode to discharge large capacitors down to VCC, to limit the amount of over-voltage during power-off when VCC drops below the value of the capacitor. Type of diode will greatly affect the amount of overvoltage allowed, but will also affect the amount of leakage Yes, this sounds very practical and very much probable to me. I tried with 3/4 different ADC0809 with same result. Unfortunately, the ADC board design is customer's design and my job was to program the micro to read this ADC and do something further. Although I have finished the job, it was something interesting problem to learn from. I will try to convince my customer to implement this change. And yes the time constant by the way is too large. They have a 68K resistor in series with this 10uF cap. A 47K potentiometer is connected such that some part of it is in charging path while remaining part of it is in discharging path. So charging time constant is 680msec while discharging time constant may be 0 when the potentiometer is at one end or it may be 470msec when potentiometer is at the other end. Moreover this capacitor is charged from a +12V DC while the ADC is driven from a 5V regulator. That seemed a poor design but I wanted a hint on what exactly is wrong and yes your answer is very helpful. Thanks for that Kai and Per. |
| Topic | Author | Date |
| ADC0809 problem | Gopal Amlekar | 01/10/10 09:09 |
| Is the very same ADC chip in both cases? | Daniel Contarino | 01/10/10 09:51 |
| From time to time... | Kai Klaas | 01/10/10 14:22 |
| Datasheet read, same hardware | Gopal Amlekar | 01/11/10 02:53 |
| Define 1000000000% | Per Westermark | 01/11/10 06:48 |
| Thoughts... | Kai Klaas | 01/11/10 09:04 |
| Solved partially | Gopal Amlekar | 01/13/10 04:06 |
| So the cap did help? | Kai Klaas | 01/13/10 08:56 |
| RE: That cap did help? | Gopal Amlekar | 01/15/10 06:48 |
The input multiplexer is damaged!  | Kai Klaas | 01/15/10 08:49 |
| Discharge diode | Per Westermark | 01/15/10 10:14 |
| RE: The input multiplexer is damaged | Gopal Amlekar | 01/16/10 02:01 |
| Indeed, a very poor design!!! | Kai Klaas | 01/16/10 09:38 |
RE: Indeed, a very poor design!!! | Gopal Amlekar | 01/17/10 06:27 |