There should not be much difference. But you should check if the times of heating and cooling are sufficient to get the right temperatures in the whole mix. For instance, if you have steep heating and cooling rates, the annealing time for a 20ul mix can be very short (actually 0 sec), so that the inner of the mix really gets the correct temperature and the primers anneal. But this short time might not be sufficient in 50ul reactions, so here the annealing would be considerably less efficient and you might not observe an amplification.
There should not be much difference. But you should check if the times of heating and cooling are sufficient to get the right temperatures in the whole mix. For instance, if you have steep heating and cooling rates, the annealing time for a 20ul mix can be very short (actually 0 sec), so that the inner of the mix really gets the correct temperature and the primers anneal. But this short time might not be sufficient in 50ul reactions, so here the annealing would be considerably less efficient and you might not observe an amplification.
As Wilhelm said there are no many differences. Try to scale down or up the reagent volumes but if you try to scale up, consider a revision of the step, not only for annealing but also for extension, and in a minor extent for melting.
An easy option is to run several 20 uL reactions and pool them at the end (as they say... never mess up with a working protocol...). Another point is to use thin-walled plasticware as the thermal properties of your PCR plate or tubes will be key to successful scaling up. Lastly, try using a different thermocycler (different lab perhaps) in case your is out of calibration and the difference between 20 and 50 uL is pushing it below the PCR efficiency threshold.
One thing to consider when scaling up the volume of PCR is whether you are also scaling up the reagents in the PCR. For example, primers should be present at a concentration sufficient to allow enough copies to be present to find the target region. MgCl2 and dNTPs are also sensitive to volume and should be scaled up. This also includes the buffer concentration and possibly the Taq.
If you are scaling up the reagents to the 50 ul volume, then I would agree that the PCR may have difficulty obtaining the temperature for the 50 ul volume. Some PCR machines like the ABI 9700 also include a volume setting before initializing the PCR.
However, if it is not necessary to run a 50 ul PCR, Then you might consider what others have already suggested. Run the lesser working volume and combine the products.
Good point, Doug. I presumed that all concentrations (except the cDNA) are similar, independent of the volume. If one takes a 20µl reaction and simply addas 30ul water to make it a 50ul reaction, things might not work, because all components were diluted. It may be even worst just to scale some components and not others (e.g. use more buffer and enzyme but the same [small] amount of MgCl2...).
theoritically no diffference in the resultant product but practically what i have ovserved that the products are not with that efficiency as in 20/25 microliter reaction... may be this is due to the Thermal cycler which is not able to provide such conditions for 50 micro liter reaction...