>Both the relays and solenoid valves, being inductive, will generate a
>surge of voltage when turned off.  Connecting any heavily inductive or
>capacitive load directly to a PIC pin is likely to cause trouble.  Don't
>expect the protection diodes in the PIC to carry heavy current and still
>maintain proper operation.  They are intended mostly to prevent permanent
>damage.  Use an external diode arrangement to prevent any voltage higher
>than Vdd from reaching the PIC.  The best way to do it is to use driver
>transistors for each relay.

I have transistor drivers for each relay and also have diodes in the coil of
the relay.

>Dumping the inductive kickback into the Vdd line can also cause trouble.
>If the circuit doesn't use a lot of power, the burst of energy from a
>relay coil can make Vdd jump well above 5V.  A large capacitor (1000 uF)
>from Vdd to ground can limit this voltage rise.  Such a capacitor should
>be standard in any design that's used in a noisy environment anyway.

Also  I have large cap in the power supply (+-1000 uF)

>If driving the solenoids via relays, the inductive "kickback" from the
>solenoids can make the relay contacts arc as they open.  The
>high-frequency energy from this arcing may couple into the PIC and
>disrupt it.  It also is likely to shorten the life of the relay.  Either
>an MOV or a series resistor-capacitor snubbing network can be connected
>across the solenoid coil or maybe the relay contacts to reduce this
>problem.  Just using a capacitor would make the relay contacts arc as
>they close.

I'll probe with the snubber (100 Ohm and .1 uF I think)

>If you still have problems, test without the solenoids connected or
>energized to see if it's coming from the relay coils or the solenoid
>coils.

The circuit has no problems without load. When I connect the solenoid I get
the reset
I still don't know if a power on reset or an mclr reset is the signal that
reset the circuit

thanks for the suggestion I'll probe it