Electronic load advice

[Richard Henry]

I need to test some battery packs. We predict the load of the target
product will cycle from high power (80 W for 10 ms) to low power (5 W
for 240 ms) during operation. I would like to simulate that load
including the dynamics in the battery test. I have found a few loads
for rent at Electrorent, but before I commit, does anyone have some
suggestions? Precautions?

 

[D from BC]

I need to test some battery packs. We predict the load of the target
product will cycle from high power (80 W for 10 ms) to low power (5 W
for 240 ms) during operation. I would like to simulate that load
including the dynamics in the battery test. I have found a few loads
for rent at Electrorent, but before I commit, does anyone have some
suggestions? Precautions?

I liquid cool ordinary tube type power resistors.
Cost: $5.00 from electrical surplus..

Light bulbs are ok if you account for the dynamic resistance.

Perhaps get some nichrome wire and wind your own test load.
22AWG is about 1ohm/ft
Stabilize it's temperature to control R drift.


D from BC
British Columbia
Canada.

 

[no_one]

I need to test some battery packs. We predict the load of the target
product will cycle from high power (80 W for 10 ms) to low power (5 W
for 240 ms) during operation. I would like to simulate that load
including the dynamics in the battery test. I have found a few loads
for rent at Electrorent, but before I commit, does anyone have some
suggestions? Precautions?

We have used both Elgar and HP at work. They can do up to 5kw and can
switch dynamically.

 

[Leon]

I need to test some battery packs. We predict the load of the target
product will cycle from high power (80 W for 10 ms) to low power (5 W
for 240 ms) during operation. I would like to simulate that load
including the dynamics in the battery test. I have found a few loads
for rent at Electrorent, but before I commit, does anyone have some
suggestions? Precautions?

When I needed something like that for similar loads I just built my
own using a power transistor on a heat sink. I was only testing for
constant power, you'd need to switch the transistor with the output
from an MCU for the appropriate time.

Leon

 

[mkaras]

I need to test some battery packs. We predict the load of the target
product will cycle from high power (80 W for 10 ms) to low power (5 W
for 240 ms) during operation. I would like to simulate that load
including the dynamics in the battery test. I have found a few loads
for rent at Electrorent, but before I commit, does anyone have some
suggestions? Precautions?

You can make a nice load by placing a power N-FET on a heat sink. Wire
the source to ground through a resistor with a small ohms value. Drive
the gate of the N-FET with an opamp and the inputs from a voltage
source (IN+) and the other input (IN-) from the feedback at the
resistor source connection. This makes a nice voltage controlled
current sink that sinks from the drain of the N-FET. If the opamp is a
dual it is easy to use the other half to make a voltage scaler such
that control voltage can be a larger range than the actual voltage
applied to the IN+ of the first opamp. For higher power applications
put a heat sink fan over the N-FET.

ciscodsl

 

[gearhead]

I need to test some battery packs.  We predict the load of the target
product will cycle from high power (80 W for 10 ms) to low power (5 W
for 240 ms) during operation.  I would like to simulate that load
including the dynamics in the battery test.  I have found a few loads
for rent at Electrorent, but before I commit, does anyone have some
suggestions?  Precautions?

What criterion will you use to judge battery capacity?
I'd like to know about those rented "loads." Can you show us a link
to them?
You didn't say what kind of battery pack your product uses.

 

[Richard Henry]

What criterion will you use to judge battery capacity?
I'd like to know about those rented "loads."  Can you show us a link
to them?
You didn't say what kind of battery pack your product uses.

The test is to compare different packs in our simulated application.

Electrorent page:
http://www.electrorent.com/products...=AT-6050A&desc=HPIB+Electronic+Load+Mainframe

The most promising technology appears to be lithium ion stacked to
provide 12-14 volts.

 

[gearhead]

The test is to compare different packs in our simulated application.

Electrorent page:http://www.electrorent.com/products/search/detail.aspx?mfr=Agilent&mf...

The most promising technology appears to be lithium ion stacked to
provide 12-14 volts.

Yeah, I was wondering why you need to rent a load. That thing weighs
almost 50 pounds.
I'm not going to offer advice about the rent-a-load, because I
wouldn't go that way. I'd build the load. If you're developing a
product, surely someone there is competent to build a load that cycles
at 4 Hz between 80 watts and 5 watts, with a 4% duty cycle on the
higher power. A hobbyist could do it with a 555 timer, a couple of
power resistors and mosfets, and spend no more than a few dollars
doing it. Actually, you only need one mosfet. Keep the 5 watt load
on constantly, and cycle a 75 watt load at 4% duty cycle and 4 Hz,
using power resistors calculated to dissipate 5 and 75 watts according
to the battery voltage.
Or if you want constant current loading, that's easy enough, with only
a slightly more complicated circuit.
Then you'll have some criterion, such as voltage, by which to measure
the battery life.
So which would you prefer, a fixed resistance, or constant current
load?

 

[[email protected]]

Yeah, I was wondering why you need to rent a load. That thing weighs
almost 50 pounds.
I'm not going to offer advice about the rent-a-load, because I
wouldn't go that way. I'd build the load. If you're developing a
product, surely someone there is competent to build a load that cycles
at 4 Hz between 80 watts and 5 watts, with a 4% duty cycle on the
higher power. A hobbyist could do it with a 555 timer, a couple of
power resistors and mosfets, and spend no more than a few dollars
doing it. Actually, you only need one mosfet. Keep the 5 watt load
on constantly, and cycle a 75 watt load at 4% duty cycle and 4 Hz,
using power resistors calculated to dissipate 5 and 75 watts according
to the battery voltage.
Or if you want constant current loading, that's easy enough, with only
a slightly more complicated circuit.
Then you'll have some criterion, such as voltage, by which to measure
the battery life.
So which would you prefer, a fixed resistance, or constant current
load?

I suppose for battery testing it isn't an issue, but for LDO testing,
the homebrew load is the way to go. I've seen some really crappy
loads, i.e. hardly looking like a resistor. For the homebrew loads,
I've gone to a local store with "free resistors" and just paralleled
them using scrap PCB. It makes a very low inductance resistor.

 

[Richard Henry]

Yeah, I was wondering why you need to rent a load.  That thing weighs
almost 50 pounds.
I'm not going to offer advice about the rent-a-load, because I
wouldn't go that way.  I'd build the load.  If you're developing a
product, surely someone there is competent to build a load that cycles
at 4 Hz between 80 watts and 5 watts, with a 4% duty cycle on the
higher power.  A hobbyist could do it with a 555 timer, a couple of
power resistors and mosfets, and spend no more than a few dollars
doing it.  Actually, you only need one mosfet.  Keep the 5 watt load
on constantly, and cycle a 75 watt load at 4% duty cycle and 4 Hz,
using power resistors calculated to dissipate 5 and 75 watts according
to the battery voltage.
Or if you want constant current loading, that's easy enough, with only
a slightly more complicated circuit.
Then you'll have some criterion, such as voltage, by which to measure
the battery life.
So which would you prefer, a fixed resistance, or constant current
load?- Hide quoted text -

- Show quoted text -

Two constant power modes, switching between them as described.

I have an old PIC development kit that could be lashed up with
resistors and DC/DC converters to do what I need, but I don't have the
time or budget to pay someone to build something I can rent.

 

[Ross Herbert]

:I need to test some battery packs. We predict the load of the target
roduct will cycle from high power (80 W for 10 ms) to low power (5 W
:for 240 ms) during operation. I would like to simulate that load
:including the dynamics in the battery test. I have found a few loads
:for rent at Electrorent, but before I commit, does anyone have some
:suggestions? Precautions?


I have a Kikusui PLZ152W electronic load
http://db3.kikusui.co.jp/fmi/xsl/kai/detail.xsl?&IdFamily=0029

(read the operations manual for load switching to see how it operates.)

This electronic load can handle 2 loads A and B, with potentiometer adjustable
switching times such that the total load cycle period (Load A + load B) is 200mS
(not quite what you want but I am sure it would be close enough for practical
purposes). The load can be set to CC or CR depending on what you want. It also
has a BNC connector so you can monitor the load current duty cycle on an
oscilloscope.

Other loads may require connection to a PC for load programming but the PLZ152W
is totally manual in standard form and is a nice little performer.

 

[Fred Bloggs]

I need to test some battery packs. We predict the load of the target
product will cycle from high power (80 W for 10 ms) to low power (5 W
for 240 ms) during operation. I would like to simulate that load
including the dynamics in the battery test. I have found a few loads
for rent at Electrorent, but before I commit, does anyone have some
suggestions? Precautions?

You can buy one for not too much:
http://store.cd4power.com/cgi-bin/cd4power.storefront/47a72971011cc9c0271d0c9f894206aa/Catalog/1241

 

[neon]

light bulbs are not linear devices so that is out . 80w for 10 ms is not hard to come up for a dynamic load. You will a load of 80w and just turn on a transistor or an fet for that time as required that is notm hard at all to come up with a few parts

 

[Joseph2k]

I need to test some battery packs. We predict theloadof the target
product will cycle from high power (80 W for 10 ms) to low power (5 W
for 240 ms) during operation. I would like to simulate thatload
including the dynamics in the battery test. I have found a few loads
for rent at Electrorent, but before I commit, does anyone have some
suggestions? Precautions?

As an old test engineer myself i would just build one.

The nearest similar thing that i have done is a current pulse test
set.
i could produce accurate repeatable pulses from sub millisecond to
sub microsecond with fast edges. The most difficult pulse was
300 A at 300 ns and 10 ns edges. All done with discretes and a
few one shot chips. First transistor provided voltage and current
gain from logic to 12 V the second drove the big FET output. That
was about 25 years ago. I had to contend with 12 foot lead
lengths to get in and out of the vacuum chamber.

Do all your wiring with shielded twisted pairs. It will prevent a
lot of pain. It worked for me.