With my last function generator related blog being such a success I wanted to open up the TTi TG210 function generator it's replacing on my bench and have a look inside to see how it works. Judging from the date codes I found inside the unit it's from early 97, making it only a year older than the HP. The technology used inside isn't anything new for 1997, it's actually surprisingly old school and very analog in a rather beautiful way.
The simple 2 MHz function generator in it's plastic box with aluminum front and back panels.
A closer look inside
The main circuit board is mounted neatly in the top of the case. The transformer along with a IEC socket/fuse holder is mounted to the rear panel.
Starting by having a closer look at the power supply you might wonder if something is missing? In the pictures there's only a LM337T, which is used for generating the negative 15 Volt rail as you'd expect. However mounted to the other side of the heatsink is an LM317T that's generating the positive 15 Volt rail. Behind them both is a for generating the +5 Volt rail.
Starting from the right side of the board (as pictured previously) next to the frequency dial, we see a lot of unpopulated footprints telling us that this PCB was used in different version of the function generator. A quarter of the LM324N quad OP-amp is used as a summing amplifier for the frequency dial and the sweep input. While the rest is used to buffer and invert this voltage. The LF351N is a single JFET OP-amp which is following the output from the LM324N. Together with a transistor this OP-amp is making up a current source, which is controlled by the voltage from the LM324.
The LF351N in the center of this picture is following the inverse output from the LM324N. This OP-amp is also driving a transistor but is this time making up a current sink. The CD4049UB is driving the TTL/CMOS output through a 47 ohm resistor, using two of its six inverters. A third inverter is used to buffer the signal coming from one of the outputs of the DS75107 dual line receiver. This line receiver is used as a comparator to make up the square wave from a triangle wave (which is generated by the main oscillator). The second half of it is used as a comparator in the main oscillator circuit itself and I'll talk more about that later.
The big capacitor in the bottom of the image is one of the timing capacitors that are switched in and out of the circuit by the frequency range switches.
The CA3046 transistor array is used to make up a sine shaper, which is making use of the non-linear (logarithmic) properties of transistors to create a sine wave out of a triangle wave (notice the large number of trimmers around it).
A very nice symmetrical power amplifier stage along side a not so beautiful repair. At this point I will admit that I did cheat and had a look inside the service manual for the TG200 series function generators to get a better idea of how the circuit works.
The group of three range switches here are for selecting the waveshape and D13-D16 are used to shape a +/- 15 Volt square wave from the +/- 5 Volt output of the line driver (a comparator output). While SW12 is switching a -20 dB attenuator, which might explain the near by repair?
Where's the oscillator?
So did I miss the main oscillator altogether? The main oscillator generates a triangle wave which is then used to make the sine and square waves. I mentioned briefly that the range switches are selecting the timing capacitor in the oscillator circuit and that the two single OP-amps were acting as a current source/sink. When you charge a capacitor with a constant current you will get a linear slope as the voltage across it increase over time (integrating a constant), now that's half a triangle wave! However there's something missing to make all that into an oscillator, namely something that alternates between charging and discharging the capacitors to get the up and down slopes of the full triangle wave I've been promising.
If you look back to the image of the CA3046 (second to last) you'll see four diodes directly above the frequency range switches. These are used to steer the current that goes in and out of the timing capacitors from the current source/sink. These diodes are controlled by the second half of the DS75107 dual line receiver we saw previously when looking at the CMOS/TTL output. This line receiver is acting as a comparator which together with the current source/sinks is making up a relaxation oscillator, consisting of an integrator (charging capacitor) and comparator oscillator.
So let's recap. The frequency dial is used as a variable voltage source that's buffered by the LM324. This voltage is then used to control the LF351 current source/sink. The rate at which the capacitors charge depend on the current applied and their capacitance, so by adjusting the current we get a fine control over the time constant and by switching the capacitor we can get a larger variance. The output voltage across the capacitor is then feed into a comparator that's switching between charging or discharging the capacitor and we now have feedback (an oscillator)!
So while these are all very cheap parts they somehow come together and make up a nice function generator, something I had to look twice to see.
If you're interested, check my blog index for my other teardown blogs.