| Electronics > Projects, Designs, and Technical Stuff |
| ADS1115 Front End Protection |
| (1/2) > >> |
| jram:
So I have a project to measure the cell voltages on my LiFePO4 pack. 3P4S - 540 Ah pack. Its a real world power supply in daily use, charged from a variety of sources depending on circumstances. Charging sources include MPPT solar (up to 900W), alternators and wind. Loads include some really beefy DC (12V) motors the largest of which includes a 1400W motor. The ADC (ADS1115) front end voltage divider is attached but basically divides the voltages down to a common level 1.7 V nominal) so the ADC can be used at the 2.048 level. Voltages are recalculated and scaled in software. I've had 2 prototypes without protecting the front end and I have witnessed significant spikes in voltage when operating the motor which, on the latest prototype may have led to an ADC channel 'locking up' and requiring a power-down reboot. I intend to implement the scheme discussed at http://eecatalog.com/medical/2015/09/01/how-to-protect-adcs-from-input-overvoltage/ but I thought I'd throw my design out there for more general (constructive ;D ) criticism and comment as well as guidance on sizing of the protection diodes. The referenced solution mentions BAT 5x series diodes but would a TVS be more appropriate? So, please fire away and give me the benefit of your experience so I can refine this design. Thank you :) |
| T3sl4co1l:
Diode in series..? You probably want to avoid series diodes, and shunt shottky diodes, because of voltage drop, and leakage current, respectively. Typical is to use zener/TVS or clamp diode (PN junction, e.g. BAV99) to limit voltage without much leakage. Mind that zeners aren't available in sharp, accurate voltages; leakage goes up sharply at say 80 or 90% of rated breakdown, and low voltage ratings (under 6V) have terrible leakage period, so much so that they're awful even for voltage regulation purposes. What's up with the switched common ground? Is that to reduce current consumption? Current still flows between taps though... Tim |
| iMo:
I would use 1N4148, their leakage is <10nA, that is at least 10x better than any schottky. With your 62uV ADC resolution and say 5nA leakage and 2k/2k divider worst case you are well below 1LSB error due to the leakage, imho. |
| voltsandjolts:
It's better to provide the whole schematic when asking for help. Where is the ADC Ground? Is it connected to the source of U6? (BTW: MOSFET designators are generally Q or T not U which is for ICs) |
| Kleinstein:
The filtering caps should be more like behind the dividers. Even with the diodes to the supplies there should be another series resistor of some 1 K (depends on the ADC) just before each ADC. The application looks like it is not very high impedance or need for very high accuracy. So the 1N4148 should be OK. |
| Navigation |
| Message Index |
| Next page |