| Electronics > Projects, Designs, and Technical Stuff |
| Any reason I shouldn't use a TL431 as a virtual ground? |
| (1/9) > >> |
| bd139:
I am migrating a design from typical +/-15V rails to single 9V rail. Total signal swing is only 2v p-p so plenty of headroom. Plan is to keep the existing LM358's (cheap) in the design and split the 9V rail arbitrarily at 2.5V with a buffered TL431 (also cheap) giving swing of 1.5-3.5V. This gives plenty of headroom for LM358 inefficiencies plus battery drain. Any reasons I shouldn't do this or is there a better solution? |
| MarkF:
Without knowing anything about what you're doing, you might consider something like this |
| bd139:
I did consider something like that but it has three problems: 1. the virtual ground operating point shifts as battery voltage declines (or load changes) 2. the swing on the opamp isn't symmetrical here (0V to Vcc - 1.5V). 3. The power dissipation is too high in the transistors for battery operation. |
| StillTrying:
1-3 probably don't count. "1. the virtual ground operating point shifts as battery voltage declines (or load changes)" Or just the +4.5 and -4.5 rails reduce. "2. the swing on the opamp isn't symmetrical here (0V to Vcc - 1.5V)." To the opamp it's output stays around +4.5V +/-1V. "3. The power dissipation is too high in the transistors for battery operation." If the transistors are removed and the virtual ground current is that high, the dissipation just transfers into the op amp's output. If you can know the virtual ground currents quite accurately, I've used this with as little as 1.5mA wasted through the 431. |
| T3sl4co1l:
I would complain far more that you're burning as much power on the ref (~1mA) as the opamps (~0.3mA/each?). Why not change to TLV2372 say, and at least a TLV431 (0.1mA) or better still anything that's properly made for battery use? (All resistor values shift up accordingly, to keep AC current consumption low.) Or I may be reading into that "9V" supply too much. Tim |
| Navigation |
| Message Index |
| Next page |