I'm not clear what your system looks like.  Unless the control valve is wide open, control valves don't 'take' a pressure drop in themselves, the pressure drop is imposed by the equipment in the system and the hydraulics you have.

Let's say you have a pump that is putting out 400 psig pumping to a vessel at 300 psig.  Let's also so the piping losses are 25 psig.  The dP for the control valve is 75 psi (400 - 300 - 25).  If you put in a bigger control valve or trim, the pressure drop is still 75 psi across the valve, the larger valve or trim will just operate more closed.  Sorry if this is repeating the obvious to you, I'm just not clear what you are asking to be able to comment.

Remember you also need to consider the upset conditions, can the downstream pressure increase very much?  Are there times when the flow rate will increase significantly?  Can the inlet pressure decrease for some reason?  Those all have to be considered when sizing a control valve if you have to be certain you can supplied the necessary water flow.

Your assumptions on pressure are pretty close.  Pump outlets are around 375, boilers are around 250.  But where we normally operate, one pump is not quite enough, so we have to run 2.  Two is way too much, and the pumps tend to overheat/cavitate/wear.  The Cv of the control valves is pretty small.  I was hoping that a control valve with a larger Cv would allow me to run one pump (lower head, higher flow on the pump curve)and maintain controlability.  But I lack experience in this matter, so I was wondering if it was typical to have a very small Cv on boiler feeds, perhaps for tight level control.

No, the Cv comes from the process conditions.  Designing the system with a high dP across the control valve just results in higher pump operating costs due to the additional head that is required from your pumps.  That's not necessarily wrong, it might be necessary to do that for upset conditions but there might be other options.

What is the dP across the control valve when you do run only one pump (I'm assuming from your description that the valve goes wide open)?  Could one pump meet the requirements if the control valve dP when wide open was reduced (eg. bigger valve or trim)?

Can you put in larger impellers so that one pump does meet your requirements?  That way, you wouldn't need to run both pumps.

What is overheating BTW?  I wouldn't think you would have a problem with temperature rise across the pumps on the water side.

Paragraph one)  that's just what I'm getting at.  It looks to me like the control valve was mis-sized.
Paragraph 2&3)  I had them shut down one just the other day.  dP does not come down much.  Just a few pounds.  I went back and rechecked my work, and then let a computer recheck it too.  Other restrictions are unlikely.  It would be asking a lot to expect enough gain to be able to run only one pump as is, but it then would be possible to step up the motor/impeller size enough to run one pump, which can not be done now.
Paragraph 4) The pumps have a small recirc line back to the suction.  Already hot water pumped at high head back to the pumps own bowl.

Is the valve running close to or full open?  

When you shut down one pump and the dP only drops a few pounds, it sounds like you are right back on the pump curve so that running two pumps produces very little additional head and that should mean the valve is close to fully open.

What's the flow through the valve and the installed Cv, I can do a quick dP check or have you already done that?  is there a bypass around the valve, have you tried slowly opening that and see if you can then get enough water to the boilers with only one pump?  Basically, part of the flow will go through the bypass and a reduced flow will have to go through your control valve which, if it's fully open, will reduce the dP needed.

Dang, TD2K, I hope you're not spending so much time on me that you'll have to stay late on a Friday, but thanks.
The dP corresponds with the Cv by my calculation, so I won't burden you with that.  The valves don't really run full open, but around 65-75%, which ought to be pretty good.
I'm embarrassed I didn't think of the bypass as a test.  I couldn't run that way, but I can try it out for a few minutes.  Still, my big concern is if there is a good reason for such a big design dP, or if it just crept up with capacity.  People here that might have that answer are out.  I myself am pretty new here.
Thanks again.

first of all, the code specifies that the pump must be able to provide enough pressure to feed water when the safety valves on the drum are lifting, plus other margins. If you are using a constant speed pump, then most of the code required margin will be spent as useless DP across the fedwater control valve.

Secondly, the pump for a drum boiler normally should be sized for supplying 115-120% of MCR flow at full load to account for upsets in drum water level . This extra margin also yields a pump with a performance curve that has extra DP during normal steady operation, and this is wasted at the water level feedwater control valve.

Finally, an extra 95# of pressure loss is nothing to complain about. In high pressure power boilers that use constant speed boiler feedpumps , there is as much as a 2000 psid pressure loss across the feedwater control valve during startup.

in order of your points
first of all... the pumps installed are well capable of that requirement, but I was only aware of it because of an earlier post that I think you made.

secondly...there are three installed pumps.  One is not capable of full load, but 2 are normally run and easily meet this target.

finally...these are not power generation boilers specifically but relatively small waste heat boilers.  The system does not match up with our needs so well anymore.  Something has to be changed, and some of that 95 pounds is what is keeping me off where I'd like to be on a pump curve to run just one pump and save both maintenance and utility costs, which would be a win/win as opposed to revamping the whole thing.

Thanks.

I'd look at resizing the valve.  Determine what your minimim sizing case is (low flow and therefore high head off the pump and low boiler pressure), normal operation (duh) and maximum sizing (maximum flow and therefore minimum pump head with maximum boiler operating pressure).  Include your line losses of course.  

Davefitz brings up another case (I don't work with boilers so I am not familar with the requirements), the boilers firing with the safeties relieving. Dave, is this at normal firing rates, maximum, etc?  If so, this is likely the maximum sizing requirement.  From the description of how your system is operating today, if you increased the boiler steam pressure to relieving pressure, I doubt you'd maintain design water rates today.

Once you have those cases, you know the Cv for each case.  Compare that to the installed valve and the answer will pretty much fall out if resizing the trim or the valve is the answer.

One last question, when you say the valve runs about 75% open, is this with both pumps online?  If you shut one pump down, does the valve go full open?  It should via the control system if one pump as you say is unable to supply enough water.

One last comment, you want to look at min/norm and max cases so that you pick a valve that will handle all three cases.  If you just look at the max sizing case, you could have problems with the valve being nearly fully closed under other operating conditions.  There may be some cases which you want to consider (such as Dave pointed out on startup where for some boilers, you have 2000 psi dP).  I'm not sure that I would size the valve for this case but I would want to convince myself that it would be able to operate at those conditions during the startup.

Just a comment on BFW valves. Part of the problem is that the boiler code requires that you be able to maintain drum level at full steam production with all safeties at full lift.

This forces the feed water valve to be larger than normal operation might suggest.

hacksaw:
The code only requires one to be able to supply water to the drum when the safeties lift, but does not mandate full water flow at that pressure head.

For slojim's main problem, I would look at asking the pump vendor if he can cut back the pump wheel and provide a different performance curve.

my experience has been that the safest condition is keeping the drum level within its design range. we've had to design for the worst case steaming condition with the safeties at full lift. It has to do with the connected steam load.

hacksaw:
for power boilers, B31.1 and ASME 1 require automatic cutback of fuel to the boiler if the main steam outlet header or transfer piping exceed its design pressure,which is usually the case when drum safety valves lift. The normal cases of lifting of drum safeties is a trip of the steam turbine, or a reheater "fast valving" incident, either case would normally realize a cutback in fuel firing rate immediately .  In these cases ( ie power boilers) it is not common practice to size the BFP for relief pressure at MCR flowrate, as this may kick the BFP ( and conected piping) into a higher pressure class.

With recovery boilers, you'll several hours of nearly full steam generation, even with all fuels cut-off. Only in the swing and power boilers could we follow the approach you've outlined.