Anyone here use DC load instruments?

Thread Starter

someonesdad

Joined Jul 7, 2009
1,583
I'm curious if any of you folks find a use for a DC load instrument, such as those from Agilent, Chroma, or B&K. If so, please let me know what you do with it. I wound up getting one in a trade, but so far about the only thing I've done with it is to program it to measure the internal resistance of batteries or to look at the transient behavior of a couple of power supplies.

Considering that the thing is capable of dissipating 300 W and it's worth around $1k, I suspect I traded for something I'm not going to use much.
 

beenthere

Joined Apr 20, 2004
15,819
Just an idea - you might find the gear handy for testing/diagnosing battery problems in hybrid cars, if you can up the power dissipation. If not you, work up a presentation and sell it to a garage.
 

Thread Starter

someonesdad

Joined Jul 7, 2009
1,583
I've seen loads that go up to 10 kW and I think there are others that are larger (I think most of them just use lots of large FETs, heat sinks, and fans).

They're pretty pricey for a garage, not to mention probably outside the technical comfort zone of a mechanic. I'd imagine they do their work with carbon pile resistors or have to use special tools supplied by the car's manufacturer.

I suppose I'll keep mine -- some day someone may want some consulting measurements and one or two jobs ought to pay for it.
 

John Luciani

Joined Apr 3, 2007
475
The electronic load is used to test and burnin power electronics. Unless you
are testing a lot of power electronics you probably will not use it much.

An electronic load is easy to build with FETs. Since a burn-in system requires
a large number of loads it was less expensive to build the load.
For automatic test equipment it was less expensive to buy loads. IIRC The loads
from Agilent (formerly HP) can run in constant current, constant resistance or
constant power.

Unfortunately the current measurement in the electronic loads is not accurate.
We always required an external shunt to calculate efficiency. It would be nice
if (1) a more accurate meter was in the load or (2) access to an accurate internal
shunt was provided.

(* jcl *)
 

studiot

Joined Nov 9, 2007
4,998
I can think of lots of uses for an electronic load, but have never been able to justify the cost of a commercial one.

Because of the currents involved they are good for anyone who want to properly test computer power supplies, as I do.

I recently posted on another forum about the difference in US ($25 + $180 postage to UK) and UK (£250 + £25 postage within UK) prices for second hand ones.

I don't think I am allowed to mention the (commercial) site here.
 

John Luciani

Joined Apr 3, 2007
475
I can think of lots of uses for an electronic load, but have never been able to justify the cost of a commercial one.
They are very easy to build. I have an example schematic
at http://tinyurl.com/6cbn6h (down towards the bottom of the page).

A PC power supply is a little tougher since you really want to load multiple
independent supplies. If you had separate control voltages for the individual load
cells you could use them independently or in parallel.

I am building a current sink board that has four independent output channels.
20A per channel. The design is for high current LEDs put would work for an
electronic load (with proper heatsinking).

(* jcl *)
 

studiot

Joined Nov 9, 2007
4,998
Thanks, John, nice to have someone else do the design donkey work.

I did build a bidirectional one with 3055/2955 a few years ago.
 

Thread Starter

someonesdad

Joined Jul 7, 2009
1,583
I agree with the statement that they are a bit pricey. I have a new one, but just don't see lots of uses for it; maybe it would be best to sell it off and use the money for something else.

I've opened mine up and it is primarily some heat sinks with fans and the power FETs on the heat sinks. If one had the parts, it would definitely make sense to build rather than buy.

The load I have can run in constant current, constant voltage, constant resistance, and constant power. I've checked its ammeter with an external shunt (whose resistance I measured with my trusty HP 3456A) and it's spot on. I found that the load could be used as a poor-man's digital voltmeter, which can be handy.

I suppose the only area where I could justify having a load would be to do repetitive testing with complicated load characteristics. For example, one can do things such as consume constant current from a source, then every 100 ms draw 5 times as much current for 5 ms.
 

John Luciani

Joined Apr 3, 2007
475
Replacing a load with a constant current-sink (which works in a lot of applications) is
not difficult. If you need the constant resistance, constant power, constant voltage
modes (which are handy for automated test) then the control loop is more complicated.
Stability to transients and fault protection adds more complication. Measurement
and computer control adds more complication.

(* jcl *)
 

studiot

Joined Nov 9, 2007
4,998
load I have can run in constant current, constant voltage, constant resistance, and constant power
What exactly is a load running in constant voltage mode please?

Sounds like a glorified zener.
 

John Luciani

Joined Apr 3, 2007
475
What exactly is a load running in constant voltage mode please?
The load will attempt to sink enough current to get to the programmed voltage.
This is useful when trying to find the current at which the power supply you are
testing goes out of regulation.

(* jcl *)
 

John Luciani

Joined Apr 3, 2007
475
So are you suggesting the 'load' never achieves the 'set' voltage?
If the load is properly specified for the test the voltage will be achieved. But is not necessarily true.

For example say you want to determine the foldback current of a 5V/10A supply and
you have a 5A load. You set the load voltage to 4.5V. You do not have sufficient current
to ever reach that point (which will be at a current greater than 10A).

Also the nonsensical condition of setting the load to a voltage above the power
supply setpoint will not be achieved.

(* jcl *)
 
Top