Andrew Sempere wrote:

> I am attempting to build a transistor h-bridge to replace the DPDT
> relay I'm using at the moment to switch the direction of a motor.

>        +5v
>  C1-----|------C2
> B1              B2
>  E1-|       |- E2
>     |---M---|
>  C3-|       |- C4
> B3              B4
>  E3-----|------E4
>         |
>        gnd

  OK, your circuit is fine, but there are a few things you should know
about it, which are addressed in the published designs such as the
classic: http://www.bobblick.com/bob/projects/hbridge/index.html

  For *your* design, we presume you are using all NPN transistors
(sorry, don't recognise the types you quoted).  Now, the "low-side"
transistors must be switched on through resistors, that is B3 and B4,
while you need no resistors for B1 and B2. If you connect the same drive
to both diametrically opposite bases together, the "low" transistor will
be biassed at the expense of the "high" transistor.

  Also, if you use all NPN transistors, you really need to switch on
each of the upper pair with a voltage significantly greater than 5V (in
which case they also will need resistors), as there is an emitter-base
voltage drop of 0.7 to 1.2V even when the transistor is fully switched
on and only 0.2V drop between emitter and collector.

  If you can't provide the "high-side drive", you may need to use PNP
transistors as suggested, and as Bob uses in the design above.  He uses
darlingtons, but you pay the same voltage drop penalty for using those
as you do for simply using NPN transistors on the "high side.
--
  Cheers,
        Paul B.