Ramsey Electronics COMPUTEMP CT255 User manual

CT255 •1
COMPUTEMP 255
BINARY THERMOMETER
KIT
Ramsey Electronics Model No. CT255
A really neat educational kit to learn the principles of binary
math, how simple digital to analog converters work, and how
instrumentation amplifiers may be used to measure the
environment.
•Visibly counts up to the current temperature in binary numbers !
•Eye catching display, great for home or office
•Super accurate readout of +-1 degree Celsius, no calibration
required !
•Precision sensor, designed with remote sensing option
•Resurrected Ramsey kit from the past; build something your Dad
built 20 years ago!
•Jumper settings for Fahrenheit or Celsius temperature scales
•Range of 0-127.5 degrees Celsius in 1/2 degree steps. (32—261.5 F)
•Runs from 7-15VDC.

CT255 •2
RAMSEY TRANSMITTER KITS
• FM100B Professional FM Stereo Transmitter
• FM25B Synthesized Stereo FM Transmitter
• MR6 Model Rocket Tracking Transmitter
• TV6 Television Transmitter
RAMSEY RECEIVER KITS
• FR1 FM Broadcast Receiver
• AR1 Aircraft Band Receiver
• SR1 Shortwave Receiver
• SC1 Shortwave Converter
RAMSEY HOBBY KITS
• SG7 Personal Speed Radar
• SS70A Speech Scrambler
• BS1 “Bullshooter” Digital Voice Storage Unit
• AVS10 Automatic Sequential Video Switcher
• WCT20 Cable Wizard Cable Tracer
• ECG1 Heart Monitor Kit
• PG13 Plasma Generator Kit
• IG7 Ion Generator Kit
• R2XL1 Balanced to Unbalanced Audio Converter Kit
• LC1 Inductance-Capacitance Meter
RAMSEY AMATEUR RADIO KITS
• HR Series HF All Mode Receivers
• QRP Series HF CW Transmitters
• CW7 CW Keyer
• CPO3 Code Practice Oscillator
• QRP Power Amplifiers
RAMSEY MINI-KITS
Many other kits are available for hobby, school, Scouts and just plain FUN. New
kits are always under development. Write or call for our free Ramsey catalog.
COMPUTEMP 255 KIT INSTRUCTION MANUAL
Ramsey Electronics publication No. MCT255 Rev 1.1a
First printing: November 2001
COPYRIGHT 2001 by Ramsey Electronics, Inc. 590 Fishers Station Drive, Victor, New York
14564. All rights reserved. No portion of this publication may be copied or duplicated without the
written permission of Ramsey Electronics, Inc. Printed in the United States of America.

CT255 •3
COMPUTEMP 255 BINARY
THERMOMETER
KIT
Ramsey Publication No. MCT255
Price $5.00
TABLE OF CONTENTS
Introduction...................................... 4
Circuit Description............................ 5
Schematic Diagram.......................... 13
Parts Layout Diagram...................... 14
Parts List.......................................... 15
Assembly Instructions...................... 17
Case Assembly................................ 23
Troubleshooting Guide..................... 25
Conclusion....................................... 26
Warranty .......................................... 27
KITASSEMBLY
AND INSTRUCTION MANUAL FOR
RAMSEY ELECTRONICS, INC.
590 Fishers Station Drive
Victor, New York 14564
Phone [phone]
Fax [phone]
www.ramseykits.com

CT255 •4
INTRODUCTION
Years ago Ramsey Electronics had a really neat little kit called the
CompuTemp 127. This was a favorite kit purchased by many who were into the
electronics hobby in the ‘70s and ‘80s. It was unfortunately retired to make
space for a growing list of other popular kits being added to the catalog at the
time.
This is an original drawing (touched up) from the original 6 page manual
showing the stylish kit case that John Ramsey had spent his hard-earned
money on the die for when he was fresh out of college. We have come a long
way since then!
At the time, this case was one of the only good looking extrusion cases that a
hobbyist could use for his projects. Now there are a huge variety to choose
from. Hobbyists have no idea how good they have it!
A very popular kit at the time, we have brought it back for your enjoyment. It
makes a great conversational piece too!
LENS
FILTER
P.C. BOARD
CR4 CR10
U3 FRONT VIEW OF
PCB BOARD IN CASE

CT255 •5
CIRCUIT DESCRIPTION
There’s a lot of circuitry in the CT255 that is in use to this day in a variety of
circuits. We will begin with the temperature sensor itself, detailing how we did
this 20 years ago and how we do it now.
The actual temperature sensor is the LM35DZ. It is a specialized part that is
pre-calibrated at the factory to output 10mV per degree Celsius. For example if
we connected the part to power and measured the output at 0 degrees Celsius,
the output would be 0mV. If the temperature were then raised to 50 degrees
Celsius, the output would become 500mV, which is 10mV * 50C. Really simple
isn’t it?
In the past we used a diode junction to do the sensing, because a diode
junction has a very predictable change of voltage drop versus temperature. In
the case of silicon, this drop works out to be 2.1mV for every degree Celsius.
Unfortunately using diodes caused us to use calibration to compensate for
offset and gain, which was a real pain.
Now it comes down to needing to display this output in a form our eyes can
understand, and this means driving a display. In our case we will be displaying
the value in a way that both humans and computers can understand: binary.
How do we do this easily? Hang on, here we go…
The binary counter
U2:A and U2:B form a cascading ripple counter which is simply designed to
count pulses on the clock pin (pin 1), and output the count as binary. The high
order bit of U2:A (pin 6) is then tied to the clock input of U2:B to make an 8 bit
counter in total.
A binary counter is a very simple set of flip-flops that are fed from one to
another in sequence. The lowest order (bit 0) is tied to the input of the next (bit
1) and the output of bit 1 is tied to the input of bit 2 and so on. For every two
pulses on the input to the flip flop, the output switches once. When the flip flops
are cascaded, they make a “ripple” counter, meaning all outputs are effected by
the single input, but all in succession, not simultaneously. When we put the kit
together, this count sequence will make much more sense since you can see it
occur!
The binary count is then tied to a resistor “ladder”. Believe it or not, this makes
a very simple digital to analog converter! As the count increases on the outputs
of U2, the voltage increases in a linear step across R38 and C5. This is set up
here to be an eight bit digital to analog converter. Many digital to analog
converters use a ladder style resistor network just like this. To increase number
of bits, they just use more precision on the resistors, and increase the count.

CT255 •6
The DAC (Digital to Analog Converter)
The resistor values in this ladder are chosen to give us nice even steps in the
voltage range we are able to work with. The outputs of the counter can’t
actually achieve 5 volts with the LEDs on the output, but really close at 4.67V.
We then figure what the voltage will be across R38 when all outputs of the
counter are at 4.67V and we will have the maximum output of our DAC. In this
case it is 3.10V. Of course the minimum works out to be 0V. To find our
minimum resolution we find the number of steps the counter has, which is in
this case 2^8 which is 256 possibilities. 0 is one of them, so take 3.10V and
divide by 255 to come up with 12.16mV per step or count. This means this is
the smallest change in voltage that we can measure with our DAC.
When the counter is counting up, you can take a volt meter, probe R38 and see
the voltage “ramp” up with the count. On an oscilloscope, you can not only see
the voltage ramp up, but also take 255 steps on the way up to 3.10V, each step
being 12.16mV! It’s worth it just to check this.
The ADC (Analog to Digital Converter? In a sense…)
Well, you may wonder how this ramp from our counters and resistor ladder can
be used to display temperature. This will be done in a way many analog to
digital converters used to make conversions. ADCs use much more
complicated methods now to increase conversion speed, but that is beyond this
manual. Now you know that each count is equal to 12.16mV, so if you have a
count of 100, you should have an output of 1.216V from our DAC. Well this
means that the display should also be indicating 100 degrees Celsius, right?
This also means that our temperature sensor should be putting out 1.216V at
100 degrees Celsius to get this count of 100, but it doesn’t since it only outputs
at 10mV per degree Celsius, which would only be 1.00V. This means we need
to scale the output of the temperature sensor to meet the needs of our DAC so
the count display and the voltages match.
Scaling to Celsius
To scale the output, we just have to figure out what it will take to make 1.00V
become 1.216V. Not too tough here, we just multiply our 1.00V by 1.216. But
wait, we’re going to make things complicated. We want to display temperature
in 1/2 degree increments. This means that we have to multiply the sensor
output by 2 to multiply our reading by 2. In this case we need 1.216 x 2 or a
gain of 2.422. Since the display has a potential count of 255, and we will never
find an air temperature of 255 degrees Celsius (and like it), we can divide the
display by 2 to scale it to the range of 0-127.5 and use the last LED for 1/2
degree steps.
U1:A is set up as a non-inverting amplifier, and the output of the temperature
sensor is sent to the non-inverting input (pin 3) to be amplified. R39 and C6 are

CT255 •7
in series with the input to reduce any noise that may be present on the
temperature sensor to increase accuracy of your reading. C3 in the feedback
branch is also used to reduce noise in the reading. In the Celsius jumper
setting, R37 and R27 are used to adjust the gain of our temperature sensor for
a Celsius reading. The gain of a non-inverting amplifier is calculated by the
formula :
A = 1 + Rf / Ri. Where Rf if the feedback resistor R27, and Ri is the input
resistor R37. (Even though the input isn’t at R37, it is considered that for
mathematical purposes).
So looking at the values in the circuit:
A = 1 + 1430 / 1000 or A = 2.43
As you can see this is very close to our requirement of 2.422, which is more
than close enough for +-1 degree C resolution.
More on the ADC
To stop the count when the output voltage of our temperature scaling amplifier
matches the count of our ripple counter DAC, we use a circuit called a
comparator. A simple op-amp makes an excellent comparator, as shown by
U1:B. When the voltage at pin 5 is above the voltage at pin 6, the output at pin
7 goes as high as possible, in this case about 3.9V. When the voltage at pin 5
is below the voltage at pin 6, the output goes as low as possible, or zero volts.
It is called a comparator because it is in reality comparing two voltages, and
giving you an output based on the difference. We employ this circuit to stop
the count in our ripple counter as soon as its output voltage surpasses the
output of our temperature sensor circuit. This allows us to take a reading from
the display, and this is what a computer would see as a value for a complete
Analog to Digital conversion!
Oh that extra count!
Ok, maybe you’re smart and figured out that for the count to stop, the output
voltage of the DAC has to surpass the output of the temperature sensing
circuit. This means that if your temperature is zero, and the count is at zero,
the output of the DAC does not quite surpass the output of the temperature
sensor. This means that the temperature reading will have to increment at
least once to have a reading. This means that your temperature readings will
at the most be high 0.5 degrees F or 0.5 degrees C depending the range you
are in. However, on average it will only be 0.25 degrees high. All ADCs have
this problem, and it is related to the bit resolution (in our case 8 bits). This is
still within +-1C accuracy.

CT255 •8
How the counter oscillator works
U1:D is set up as a simple oscillator to generate our clock pulses that drive the
ripple counter. When the comparator U1:B output is low, this circuit will run
normally, generating pulses for our counter. When the U1:B comparator
output is high, the oscillator stops. This is how the count is “held” for display.
This oscillator works through the same principle as the comparator, by
comparing the charge voltage across C1 charged through R1 to the voltage at
pin 12. When the output of U1:D is our max of 5V and the output of U1:B is 0V
(running mode), the voltage at pin 12 should be about 3V (R4, R2 and R3 act
as a voltage divider). At this point C1 is charging through R1 over a period of
time. Eventually its charge will surpass our 3V at pin 12, where the output of
the opamp comparator will switch to low. This makes our voltage at pin 12
about 1.6V. Now the capacitor C1 discharges until it surpasses 1.6V in the
opposite direction, and the output switches back to high. This repeats over
and over and is the basis of our oscillation. The reason why the oscillator
stops when the output of U1:B goes high is the voltage at pin 12 becomes a
level that the charge on C1 can never reach. It is greater than 5.0V due to R15
limiting the voltage on C1 to 3.4V, so the comparator function is disabled.
Restarting the conversion cycle
Now our count stops at a certain voltage, but how do we make the count reset
every so often so we can get a new reading? Simple. We have another opamp
stage that is set to be yet another comparator. In this case we are comparing
the voltage between R30 and R34, in this case 2.5V, to the charge on C2. C2
charges through R24 in about 5-6 seconds. This capacitor charges while the
output of U1:B is in the disable count mode with the output being high. Once
the charge on the capacitor surpasses the 2.5V threshold, the output goes
high, clearing the count on the ripple DAC.
This in turn re-enables the oscillator of U1:D since U1:B compare values are
reset, and C2 then discharges quickly through D9 to get ready for the next
round starting with the first clock pulse in U1:D. C2 does not charge until the
count is complete, so the reset cycle will hold the finished count long enough
to be seen.
For the complicated Fahrenheit design
Skip this section if you don’t want a bad headache.
Now how do we switch to Fahrenheit? Not a simple as it sounds. For one our
Celsius to Fahrenheit conversion is listed as:
F = 9/5 * C + 32
We see by this formula that not only do we have to scale by 9/5 we also have

CT255 •9
to add 32. We do this by switching in R22 into the circuit, which combined with
R23 adds in the 32 offset that we need. The problem however is this acts like
a voltage divider on the output of U1:A our scaling amplifier, so now we have
to adjust the non-inverting amplifier’s gain to compensate.
First off we need to set the zero point of the sensor to a count of 32. This
requires us to have a voltage on the opamp pin 6 of 12.16mV * 32 * 2. (Don’t
forget to scale by 2 for our 1/2 degree steps!) This means we need pin 6 to be
778mV when the output of U1:A is 0.0V.
We do this with our voltage divider of R22 and R23. We know our supply
voltage is close to 5.0V from our regulator VR1, so we choose a random
resistor for R22 within the 1K to 10K range. In this case we choose 10K. Now
we can find R23 using the formula:
Vout = Vin * R23 / (R22 + R23) or 0.778 = 5.0 * R23 / (10000 + R23)
Rearranging we get:
R23 = Vout * R22 / (Vin—Vout) or R23 = 0.778 * 10000 / (5.0—0.778)
So R23 = 1.84K. The closest 1% tolerant resistor is then 1.82K.
Now we have to find out what it takes for 100.0 degrees Fahrenheit to have a
proper count of 100.0 on the display. For this we need a voltage of 12.16mV *
100 * 2 or 2.432 volts. This will have to be at pin 6 of U1:B, not pin 1 of U1:A.
So how do we get that?
First we have to work backwards from pin 6 of U1:B. Here we will insert our
2.432 volts. This means the voltage at pin 1 of U1:A has to be found by taking
the 5.0 volts on R22 on the header side and the 2.432 volts at the pin 6 side
into account to find what the voltage needs to be at the pin 1 of U1:A side
needs to be to make the numbers work. This can be found by looking at the
current through R22, which is found through Ohms law by:
I(R22) = (5.0V—2.432V) / 10000 =
So I(R22) = 0.256mA
Then using this current to find the voltage across R23.
So V(R23) = 1.82K * 0.256mA or 0.467 volts across R23.
This does not take into account current into pin 6 of U1:B, which could effect
these levels considerably if it is large enough. We have designed in a nice rail
to rail opamp to reduce this to a minimum. This means the output of U1:A
needs to be 2.432—0.467 or 1.965 volts at 100.0 degrees Fahrenheit.

CT255 •10
At 100.0 degrees Fahrenheit the temperature sensor will have an output of a
value related to Celsius, so to make life easier, we will convert to Celsius first.
So:
C = 5/9(F-32) or C = 5/9(100—32) which is 37.78 degrees Celsius.
This means the output of the sensor should be 377.8mV at 100.0 F.
So to convert 377.8mV to 1.965 volts we need to multiply by:
1.965 / 0.3778 or 5.201. This will be the gain we need from the non-inverting
amplifier.
We go back to our old non-inverting amplifier formula of:
A = 1 + Rf / Ri, which is the gain of the non-inverting opamp.
We already have Ri = 1.00K and A of 5.201 so we have to find Rf.
Rearranging we get Rf = (A-1) * Ri so Rf = (5.201—1) * 1000
So Rf = 4.201K. Our closest 1% value is 4.22K.
Binary math
Oh no! Your brain hurts and doesn’t want any more theory; on to the kit
building! Ok, go ahead and build the kit, but come back here later to review
how to read the display properly, and why it works that way.
Normally we would only deal with integer binary math, but we at Ramsey like
to mix things up a bit, and throw in fractional binary math so you learn
something new! It is the basis of how all math works on computers, including
that pesky floating point stuff, especially on DSPs. We will also learn why
computers are NOT perfectly accurate.
A binary number is either 1 or 0, on or off, meaning it is only two states. Two
LEDs will have 4 potential states, 3 will have 8 states, and 4 will have 16. The
relationship is 2 to the power of the number of bits. This is seen as 2^n when
n is the number of bits. In the case of our display we have 8 bits, so we have
2^8 or 256 potential states.
As the value counts up, you will see the right most bit (LED) toggle the fastest.
This LED is considered the least significant bit, meaning it has the least value
associated with it. In our case it is worth a measly 0.5, but normally would be
worth 1 if we didn’t throw in a decimal point to confuse you. The LED on the
far left is called the most significant bit, because it is worth the most. In our
case it is worth 64 but normally 128. This is because the .5 LED had to flash