Building a 50 ohm 150 watt RF dummy load

Building a 50 ohm 150 watt RF dummy load
Building a 50 ohm 150 watt
RF dummy load
Roderick Wall VK3YC
Do you use a dummy load when testing your transmitters. Not only is it a good idea, it’s a
requirement of your licence to use one. Plus you don’t want to cause interference on our bands.
Here’s a design that has good VSWR up to 500 MHz. It uses a Bourns™ 150 watt 50 ohm terminator
chip, only 6 mm x 9 mm x 1 mm
thick.
The need for a dummy load
I have a Yaesu FT-301 transceiver
that I’ve owned since new. This
transceiver, when introduced in 1976,
was said to be the first solid state
ham rig with a 100 watt output. My
FT-301, after being in storage for many
years, was put back into service with
a SG-237 Smartuner and long wire. It
has now been in use for around two
years.
As the transceiver is now 33 years
old, I decided that I should check
the alignment to ensure that it was
putting out a clean powerful signal.
First step was to download the user
manual from the internet and print
it. I decided to start with the carrier
frequency adjustments. On reading
the manual it says to inject a 1 kHz
audio signal and to adjust the RF
output to 50 watts. Then to inject a
300 Hz audio signal (same amplitude)
and adjust the carrier frequencies for
12.5 watts output on each sideband.
Photo 1: The 50 ohm 150 watt RF dummy load.
The manual did not say to adjust the
carrier frequencies to 8.9985 MHz
for USB and 9.0015 MHz for LSB. This
puzzled me so I placed the SSB filter
onto the N2PK VNA (vector network
analyser) and did an insertion loss
trace to determine the filter’s band
pass shape.
My VNA software that drives the N2PK
VNA has a matching tool that allows
you to measure the filter at 50 ohm
impedance, then to display the trace
as if the filter was matched with 500
ohms.
Figure 1 shows that the centre
frequency is not 9 MHz and that the
carrier frequencies need to be set to
suit each individual filter. This is why
a power meter and not a frequency
counter is used to adjust the carrier
frequencies.
8 AMATEUR RADIO DECEMBER 2010
Figure 1: The FT-301 SSB filter trace.
I don’t have a 50 ohm dummy
load
Power measurements require a
dummy load. Since I didn’t have one,
a hunt was on to source the parts to
build a dummy load.
A search on the Farnell™ website
found a Bourns™ 50 ohm 150 watt
terminator 6 mm x 9 mm x 1 mm
thick at a very reasonable price.
Specifications were 150 Watts, DC to
3 GHz VSWR 1.2:1, +/- 5%. According
to Bourns derating curve, it’s able to
dissipate 120% of rated power up to
100 °C, that is, 150 W x 120% = 180
watts . Farnell at that time were also
shipping online orders free, so an
order was placed. Refer Photo 2.
Flange and heat-sink
Now the fun had started. How was I
going to dissipate 150 watts through a
tiny 54 mm2 area (6 mm x 9 mm) and
into ambient air? Since copper is a
much better heat conductor than an
aluminium heat sink, then a flange
has to be made from copper and have
a large surface area, and be as thick as
possible.
Peter VK3BPN gave me a piece of
copper 120 mm x 60 mm x 6 mm thick.
I didn’t do any calculations. I decided
that I would derate the power
handling capability and/or duty cycle
depending on the performance of
the heat sink. Bourns specification
indicates a roll off after 100 °C. I
haven’t tried it, but it may be possible
to use a computer CPU cooler.
The Xigmatek HDT-S1283 CPU cooler
has a thermal resistance of 0.16 °C/W.
I decided to use a large heat sink,
0.55 °C/W, that I had purchased from
Dick Smith™. A piece of copper as
large as the heat sink would be better,
but could not be found. A possible
source of copper plate may be a scrap
metal dealer.
While designing the flange, I also
considered using water to cool the
50 ohm terminator. A small hose
attached to a water tap and to a series
of holes drilled into the flange is
possible.
Making the copper flange
Marking the position for the resistor
chip! The best position is in the
middle of the flange as this is where
total thermal resistance will be
lowest. The flange also needs to be
mounted onto the heat sink where
thermal resistance is lowest, for my
Photo 2: A close up view of the copper flange
heat sink this is in the middle. Mark
the hole positions and drill the holes
for the flange/heat sink M6 mounting
screws. Tap the holes in the heat sink
for M6 screws.
No measurements are given because
your flange and heat sink may be
different. I used twelve M6 screws to
ensure the heat sink compound is as
thin as possible to reduce its thermal
resistance. You may want to position
the holes between the fins of the heat
sink. This will then allow you to drill
through the heat sink and to use a
taper tap. I found that the bottoming
(plug) tap I used didn’t cut very well
in aluminium as it kept clogging up.
Hint, use tapping compound or if
none to hand, grease or oil at a pinch.
First try assembling the resistor
chip onto the flange
The copper flange was placed onto an
electric hotplate and heated to allow
the flange surface to be tinned. I then
placed the resistor chip onto the
flange. While the flange was cooling
down, a screwdriver was used to push
the chip onto the flange. The solder
between the flange and chip should
be as thin as possible to reduce
thermal resistance.
Four M3 holes were tapped into
the flange for standoffs to mount
the BNC connector. The next step
was to solder the centre pin of the
BNC connector onto the top pad of
the resistor chip. But this is where I
struck a problem. I couldn’t get the
top pad hot enough because the heat
was being conducted through the
chip body into the copper flange.
That’s good because when the chip
is dissipating 150 watts you want it to
be conducted into the flange and on
to the heat-sink. But I needed to first
make a connection.
Change the assembly procedure
There’s only one answer to this
problem. Due to the resistor
chip body being a good thermal
conductor, both connections would
have to be reflow soldered at the
same time. It’s not possible to do
them one at a time.
If you solder the wire onto the top
pad, the heat will run through the
chip body into the copper flange.
If you solder the chip resistor onto
the flange, the top connection will
get hot and fall off. The solution is
a spring loaded fixture that holds
both connections to allow them to
AMATEUR RADIO DECEMBER 2010
9
be reflow soldered at the same time.
Figure 2 shows a fixture holding the
two top wires onto the resistor chip,
and the resistor chip onto the flange.
Referring to Figure 3, a 6 mm x 9
mm piece of vero board is used as
a clamp to hold the two wires onto
the top pad of the resistor chip. Use
a triangle file to file a vee slot into
the vero board for the wires. The slot
should not be too deep, as the wires
must not be loose while being reflow
soldered. A spring and slotted cheese
head screw holds the parts together
while reflow soldering takes place.
Tightly twist the strands and tin the
ends of two wires as in Figure 4. Pliers
were used as a heat sink to stop the
solder from flowing up the multi
strand wire. The multi strand wire
between the resistor chip and BNC
connector must be free from solder
and must be flexible. Warning! If the
wire is stiff, it’s easy to pull the top
pad off the resistor chip. Be careful,
the pad appears to be the weakest
point on the resistor chip. How do I
know, because I now have a resistor
chip with a missing top pad!
When the flange has been tinned for
the resistor chip, assemble the wires
and resistor in the fixture. Place the
flange onto an electric hot plate. Heat
the flange until the parts are reflow
soldered. Wait for the flange to cool
and remove the fixture. Be careful
when removing the vero board from
the wires, you don’t want to pull the
top pad off. Make sure the wire hole
in the vero board is large enough for
the wires to easily slide through.
Connecting the BNC connector
Figure 2: The soldering fixture.
To make room for the wires between
the connections, I shortened the
insulator and centre pin on the back
side of the BNC connector. Then add
the standoffs and mount the BNC
connector. Solder the two parallel
wires onto the centre pin of the
connector. Use pliers as a heat sink
to stop the solder from flowing down
into the multi strand wires. Because
there isn’t much room under the
connector, I used a rubber band on
the handles of the pliers to clamp the
wires while soldering the connection.
Final assembly
Figure 3: The vero board top clamp.
The base of the copper flange should
be flat and free of burrs. Clean the
base with Brasso™ before coating it
with silicon heat transfer compound.
With a spring washer and flat washer
on each M6 screw, mount the flange
onto the heat sink.
For maximum heat sink efficiency, the
heat sink fins should be vertical and
clear of the top of the workbench.
Attach four stick-on rubber feet to
allow this. I also attached a derating
curve to remind me to derate the
power/duty cycle above 100°C. Copy
the label from Figure 5 or download
from the WIA website, refer to the
URL link listed below. I used clear
book covering to protect and hold the
label.
Figure 4: Tinning the wires.
10 AMATEUR RADIO DECEMBER 2010
Performance
Glenn VK3PE was kind enough to use a HP8753C VNA to
measure the return loss, and the results are shown in the
graph - refer Figure 6.
Frequency
Below 50
MHz
100 MHz
200 MHz
300 MHz
VSWR Frequency VSWR
1.01:1 500 MHz
1.07:1
1.02:1 1 GHz
1.03:1 2 GHz
1.04:1 3 GHz
For 36 Watts when ambient temperature is 23º C, thermal
resistance is: 1.8°C/W between resistor chip case and
ambient air.
A rough guide can be obtained for other ambient
temperatures by using the following formula.
Maximum continuous power dissipation if ambient
temperature is 25º C:
1.16:1
1.39:1
1.85:1
Continues over page
Bourns specify the resistor chip as 1.2:1
VSWR at 3 GHz. High frequency return
loss can be improved if the connector is
mounted as shown in Figure 7. Bourns
also have a 100 watt version that is
specified as 1.1:1 VSWR at 5GHz.
Determining maximum continuous
power rating
Maximum continuous power rating
depends on: (1) Ambient temperature
(Tamb), (2) Thermal resistance between
the resistor chip case and ambient air
(RθCA in °C/W), (3) Resistor chip maximum
temperature (TCmax), (4) and, of course,
it can’t be above the maximum power
rating of the resistor chip. Interesting
in that thermal resistance is similar to
series connected resistors. Total thermal
resistance is the sum of the series
connected thermal resistance.
RθCA = RθCF + RθFH + RθHA.
Also refer to the ‘Thermal Resistance’
URL links below.
Figure 5: The derating curve label
Thermal resistance between the resistor
chip case (TC) and ambient temperature
(Tamb) must first be determined. A test
was done by dissipating 36 watts in the
dummy load and then measuring both
TC and Tamb when the case temperature
has stabilised. I started at 9 watts and
increased the power until the case
temperature was 88º C; do not go over
100º C. Ideally this test should be as close
to 100º C as possible and with ambient
temperature as high as possible. My
results were:
Power
Watts
Ambient
Temperature
9W
23º C
20 ¼ W 23º C
36W
23º C
Resistor
case Temperature
35º C
52º C
88º C
Figure 6: The return loss graph
Thermal
Resistance
1.33 ºC/W
1.43 ºC/W
1.80 ºC/W
Using the following formula, thermal resistance is
calculated.
Figure 7: Bourns S-Parameter test
AMATEUR RADIO DECEMBER 2010
11
Note, only use this as a guide.
Because thermal resistance increases
when the temperature difference
between ambient and resistor chip
maximum rating (100º C) gets less.
As shown in the above table, thermal
resistance also increases with
increased power.
I don’t have an environment
chamber to do tests for different
power levels and ambient
temperatures. Warning! you need
to make sure the resistor chip
temperature never exceeds 100º C.
Heat sink efficiency also depends on
what the air movement is around the
heat sink. Adding a fan will reduce
thermal resistance and increase
the maximum continuous power
dissipation rating.
I did a series of calculations for
different ambient temperatures. A
text editor was then used to generate
a stick-on table that gives a guide
for maximum continuous power
dissipation for different ambient
temperatures. See table below.
Maximum power dissipation can be
higher than 41 watts if the ON duty
cycle is less than continuous. If the duty
Watt
Ambient
Temperature
55
0ºC
cycle is 50% ON and 50% OFF then
power dissipation can be increased to
82 watts. Further, if the duty cycle is 25%
ON and 75% OFF, then the full power of
150 W can be dissipated.
Maximum continuous power
dissipation is perhaps misleading.
Because 150 watts can be dissipated
for short ON duty cycles, you only
need to ensure the resistor chip case
never exceeds 100ºC.
I have only ever used my dummy load
for low duty test cycles. I found that I
can dissipate 100 watts, do any tuning
as required, switch off and the heat
sink only got warm.
Of course your thermal resistance
calculations will be different as your
flange/heat sink will not be the same.
Power measurement
The dummy load has been built, but
I need to do power measurements.
Since I was only going to use it for
HF from 1 MHz to 30 MHz. I used
an oscilloscope and calculator to
measure the power.
The oscilloscope 10 x probe was
connected to the BNC centre pin,
not to the pad of the resistor chip as I
Maximum Continuous Power Dissipation
52
50
47
44
41
5ºC
10ºC
15ºC
20ºC
25ºC
didn’t want to pull the pad off!
For power calculations, I use a
freeware program called Mini dBCalculator. It can be downloaded
from Softpedia’s website. Softpedia’s
URL is listed below. Mini dBCalculator is handy as it’s also able to
do other types of RF calculations.
References
Derating curve label: http://www.wia.
org.au/
*Bourns 50 ohm 150 W resistor chip:
Now available from http://au.mouser.
com/
Thermal resistance: http://
en.wikipedia.org/wiki/Thermal_
resistance_in_electronics
Thermal calculators: http://www.
novelconceptsinc.com/
Mini dB-Calculator: http://www.
softpedia.com/get/Science-CAD/minidB-Calculator.shtml
CPU cooler: http://www.xigmatek.com
N2PK VNA: http://n2pk.com
38
30ºC
VK3YC
36
33
35ºC
40ºC
Bill of materials
Reference
R1, 50 Ω 150
W
Flange
Heat sink
BNC connector
Connector
mount
Connector
mount
Part No
Bourns CHF3523CNT500LW
Quantity
1
Copper 120 mm x 60 mm x 6 mm 1
0.55 °C/W 200 mm x 75 mm x
48 mm
Silicon heat transfer compound
Multi strand copper wire
BNC panel mount connector
40 mm
1
Metal spacers M3 x 10 mm long
4
M3 screw 5 mm long
8
12 AMATEUR RADIO DECEMBER 2010
1
Purchased from/Comments
*Farnell - S/N1435944
Or a size to fit your heat
sink
Dick Smith Cat: H 3406
Dick Smith Cat: N 1205
Around 0.5 mm diameter
Or your favourite connector
Cut the head off four
screws
ar
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertising