Larry Bachus
(a.k.a. “Pump Guy”)

Hello Larry,

I attended your private pump training seminar last year at the shipyard in Newport News, Va. I was trying to recall some information that was discussed regarding pump operation.

Our current procedure has centrifugal pumps started against closed discharge valves. If I remember correctly, you taught us that a pump should always be started against an open (or slightly open) discharge valve.

However, there were a few circumstances given in which it would be desirable to have the discharge valve shut during startup. Would you please provide me with the reasons as to when it is appropriate to start a pump with a shut discharge valve?

Chas B.


Hello Chas,

There is a misunderstanding, and I am guilty (with an assist from the English language). The misunderstanding comes from imprecise language. Striving for linguistic precision is a constant effort.

You say your current procedure is to start your centrifugal pumps against a closed discharge valve. Let’s consider why and how this may have become current procedure.

Maybe at some time in the past someone said the words “throttled valve.” The word “throttle” has a variety of meanings. Throttle can mean “closed.” Throttle can also mean “restricted” or “half-open.”

Checking my Webster’s Collegiate Dictionary, the word “throttle” is both a noun and a verb. As a verb, “throttle” means to control, obstruct, reduce, choke, strangle, or suffocate. As a noun, a “throttle” is a device (like a valve) that can control, obstruct, reduce, choke, strangle, or suffocate.

“Control” and “reduce” evoke certain meanings and images. “Choke,” “strangle,” and “suffocate” evoke totally different images.

With an industrial valve, the word “throttle” can be used to mean “shut” or “closed.” Other people might use the word “throttled” to mean “restricted” (but not closed).

While watching public TV this past weekend, I saw a documentary on the history of railroading and the days of the steam locomotives. An old engineer grabbed a handle he called the “throttle” and said the throttle governs the velocity of the steam locomotive. I can only imagine that the throttle either restricts, or releases the steam pressure to the pistons that power the wheels.

In the early days of the automobile, many people used the word “throttle” referring to the accelerator pedal (or handle). There really is no reason or logic to operate a standard centrifugal pump against a totally shut discharge valve. This action applies maximum stress to the pump, the mechanical seal, and the bearings. It is certainly counterproductive to start or operate a PD pump against a totally shut valve on the discharge pipe.

However, there are times when it is best to start a pump with a discharge control valve restricted about 80 percent (meaning 20 percent open).

  • When you start the pump in a totally new system, and the discharge pipes are empty.

  • After maintenance when the discharge pipes have been drained of liquid.

  • When the pump’s first duty is to fill an empty tank or vessel from the bottom.

In the above examples, it is best to start the pump with a discharge control valve restricted about 80 percent, meaning 20 percent open. This action (the restricted valve) provides some artificial resistance against the pump when there is no resistance in the pipes or tanks. The action holds the pump into the zone of higher efficiency on the curve.

When the discharge pipes are empty, there is no liquid volume, no liquid weight, and no resistance (friction) in the pipes. If the pump’s first duty is to fill the discharge pipes with liquid when there is no standing resistance, then the pump slams out to the far right end of the performance curve with no resistance. This can:

  • Damage the mechanical seal.

  • Damage and stress the bearings.

  • Cause tight tolerances in the pump to bump and scrape.

  • Send the pump into cavitation.

  • Possibly even break the pump shaft.

  • Possibly damage the electric motor.

For the above reasons, it really is totally counterproductive to start and operate a standard pump against a closed discharge valve, especially on an ocean-going vessel that can be 10 days from the next port, or 90 days from home base.

Once the motor is started and the operation has begun, slowly open the discharge control valve completely, or enough to allow the pump to run inside the sweet zone of high efficiency.

With differential gauges and/or a flowmeter and the pump curve, hold the pump (meaning “operate the pump”) in the zone of best efficiency.

Everything I have written to this point applies to standard pumps that deliver liquid from point A to point B. There are exceptions.

You might operate a pump against a shut discharge valve for a few seconds as a diagnostic procedure to verify the shut-off head. Don’t leave the valve shut for more than 15 or 20 seconds.

There are a few special-duty pumps that are designed to operate against a total resistance. The classic example is a “jockey pump” on a fire water system. Jockey pumps are an integral part of a complete firewater system in all office buildings, malls and warehouses. A firewater jockey pump is designed to run deadheaded (zero flow at maximum head) all year long.

There is another typical application rooted in history where pumps were started against a closed valve. Long before variable-speed motors, clutches and soft starts, the electrician would order to start large pumps/motors against a closed (throttled) discharge valve as a way to conserve the motor. The thought was: less flow = less kilowatt consumption. This practice from 50 years ago was the caveman’s soft start for the motor. The operator would open the valve after the motor came up to speed.

It is interesting how even 50 years ago, the electricians knew how to protect their motors by sacrificing the pumps on the MTBF alter. In most places I visit, the pump shop looks like “chaos-on-steroids” while the electric shop is neat and tidy. Now you know why.

Starting a pump against a closed valve also assumes you know the power consumption profile of the pump. Many pumps consume less power with less flow. But other pumps consume more power with less flow. If you don’t understand your pump curves, you might damage the pump and the electric motor. What about them apples?

If you haven’t re-written the operation manuals in 40 years, the instructions might still say to start large pumps against a throttled valve. You need to re-write the operation manuals.

I’m really glad you wrote to me with this misunderstanding. What if the chiefs and engineering officers give orders using vague words with more than one interpretation? An officer might order, “Throttle that valve!” meaning reduce the flow through the valve. And, what if the operator “shuts” the valve, following the order? This could be a big mistake with disastrous consequences.

Please spread the word to the other people in that pump class who might have misinterpreted my words.

Stay in touch,

Larry Bachus

I recently wrote an article about some engineering words and phrases that are imprecise and lead to misunderstanding. The article is printed in the Mar 2011 edition of Flow Control (page 36).

And strangely in that article, I mistakenly attributed research performed by Blaise Pascal to Daniel Bernoulli. And, I attributed Bernoulli’s research to Pascal. (Only one reader caught it. What does that mean?)

Larry Bachus, founder of pump services firm Bachus Company Inc., is a regular contributor to Flow Control magazine. He is a pump consultant, lecturer, and inventor based in Nashville, Tenn. Mr. Bachus is a retired member of ASME and lectures in both English and Spanish. He can be reached at