You say you have 7.75V on the input number 5.
That's 250mV above the 7.5V, which should be enough. But for a silly case where the voltage can be off a little.
If your supply is off by 10.3% (15.5V), the half-supply with perfect 0% resistors will be 7.75V.
If one of the resistors is off, or just mis-picked from a 90k batch, your center setpoint will be off enough. It puts you in a whole new realm of debugging. And I'll tell you for why:
If you have a scope probe set at 1x or a normal low budget digital multimeter the chances are good they are only 1MOhm. 1MOhm next to 100kOhm makes 91kOhm. Which is also off by a good 9%, dropping your V- on the second op-amp several hundred milivolt if you measure there with the 1M probe or DMM relative to GND/V-.
In the case of the scope-probe, you should really set it to "10x", since that will make it 10MOhm total, which reduces effect of the parasitic resistance your schematic will see. If you already had that, well I could just be dead wrong.
Point is, every measurement device has an internal resistance, for a voltage measurement device it's high, but not infinite. My best Multimeter still has 100MOhm as maximum setting, and that's a very very high-end one and measuring something will have you put that resistor in parallel.
Another option is that you connect the ground clip to a wrong terminal by accident, shorting a signal to ground. Remember that your oscilloscope is hard-connected to power earth in most cases, and so might your board's GND or V- be. You need to be sure about which it is before you clip on the probe's ground. A digital multimeter that's battery powered will be floating and you can then put either probe anywhere and you will only have the internal resistance to think about (if not known, assume 1MOhm for cheap, 10MOhm for mid range, more than 10M if you've burnt some serious cash on it ;-) )
That's about all I can think of to look for in the debugging process with the information currently available, just before I move towards my bed.