Hello elt,
Yes getting close. In the spirit of parts reduction:
Gate resisters r8, r16, r17 are not needed. Gate resisters are used to isolate the driver from the capacitive load in high speed circuits. We are not high speed here. Opamps will need decoupling from a capacitive load in a closed loop, since we are not closed loop, not necessary. The resister value normally used is near the output impedance of the opamp, the lower speed stuff your looking at 100 to 1k ohms is more in line. If you like, put one resister in, test with and without.
What are r6, r5 for? Not in the application notes for LM61?
Remember to put a 0.1uf cap across the supply at the remote temperature sensor. In circuit use another cap on sensor output for noise reduction. Consider using a shielded cable or twisted or ribbon cable.
Consider putting a LED across the LM61, if the remote sensor is unplugged battery sensing will pull high turning on the load dump. The LED must be a very low current one, it wont have much light out with 150ua. The LED will keep the voltage from rising much, preventing the load dump. Or consider using two diodes.
I would not use a pot (r13) in the battery sensing string, used fixed 1%. The set point can be adjusted in the Vref string. I think you put it there to adjust for the output impedance of LM61, the 400 ohms or so will cause less error than the pot. Over time I have had problems with pots, enough so I try not to use them. A good pot can be costly, so consider something like: http://www.analog.com/en/prod/0,,761_797_AD5228%2C00.html
Better than a mechanical pot (for me). Many variation on this, click on digital potentio... upper left for more.
Voltage regulator is much more accurate than the zener diode. Non logic FET 10v is good idea.
With Q1 off the set point is controlled only by zener diode, r15 will not adjust. The temperature sensor and set point are calculated for 5v, change the value of the set point will change the balance.
Use a voltage divider like shown in my schematic. Top resistor use a pot to adjust for the right setpoint.
You seem very worried about the FET turning on with a sagging battery voltage? Why? With that said what voltage regulator will you be using? Lets assume a one volt drop, with a 10v reg it will start to loose it at 11v. Nothing going on yet. As battery volts continue to drop sensing is 1/3 set point is 1/2. Vbat = 6, 2 vs 2.5, opamp is still off. Below the FET turn on.
Based on this, I do not see a problem with logic FETS, and 6v regulator. Write down each volt as the battery drops and verify no unsafe condition.
Cap on opamp plus input is good idea, 1-10nf should do it.
48v
LM317 voltage limit is 60v, this is the set point for charging. CANT use it.
Suggest a 15v zener resister and transistor pre-regulator. Transistor might need heat sink.
Good to have D3 (only needed of load is inductive) and C3 (no fast signals here so not critical), and of course C2 (verify Vreg output cap requirements). Might add cap to the Vreg node going to neg input on opamp.
Opamp
Operating voltage should be rated above supply. Output and input must be within design.
If Vcc = 6v, set point is 5v, the opamp must handle 1v input from supply rail. Output should be rail to rail. Many older opamps will not go to the positive rail. Please verify. LM358 input range is Vcc - 1.5v, output is Vcc - 2v, so with 10v supply only 8v gets to gates. Not a good choice.
I perceived you were interested in learning. I have not told you everything to give you a chance to think of it, your doing great. I even hesitated to send you a schematic, was taking me to long to talk it through!
I thank you for the kind offer of a PCB, no wind gen... no need!
Have fun,
Scott.