Tektronix 475 Channel Selection and Channel 1 Problems

Channel Selection

So now the scope is sweeping but not triggering. I also noticed that no matter what channel switch I select it always shows channel 2 and decided to tackle this next.

The circuit is thankfully mostly digital and based on 74 series logic. The only quirk was that the VSS line was at -5V and VCC was at ground. So logic high is zero and logic low is -4.8V

There is a dedicated channel MUX IC that switches between the different channel input and then the logic decides which channel to display. There is logic to switch between channel 1 and two on the sync pulse for alternate mode and there is a 2MHz oscillator made up from a couple of capacitors and some NAND gates that swaps between channel 1 and 2 in 'chop' mode.

There is a transistor configured as a voltage regulator to generate the -5V and -5.7V used by the logic from the -8V line. The output was a little off and the -5 was more like -5.7 and the -5.7 was closer to -6.4. I tried swapping the transistor but it didn't change.

The outputs of the flip-flops control the channel selection lines on the MUX. These really didn't look right - instead of being near 0 or around -5V they were at around -1.5V. I checked all the resistors to look for a short but no luck.

74 series logic is cheap and easy to come by so the next day I bought replacements for the 7474 and the 7400s. After replacing the 7474 I still didn't have channel selection but the outputs of the flip-flops now looked like sensible logic levels. I decided the output of the 7400 in the chop/alt circuit was off so I just replaced all the 7400s. Thankfully I noticed that U330 is upside down compared with the other chips on this PCB (pin 1 is on the right). Now I can change channels! The other interesting thing was that now the -5V and -5.7 volt lines were the correct voltage.

The alt mode also worked and it displayed two waveforms. Chop however didn't. I looked at the input to U330B and there was no signal from the oscillator. I checked the voltage at the input to U330A and I noticed the -5.7 was present. Weirdly, just touching the legs of U330 with the DMM caused the oscillator to start running. 

Looking at the circuit I actually don't understand how this works. The resistor divider network on pin 12 I thought would set it to low (-5V). When the chop switch is closed pin 1 and 2 go low which brings pin 3 high which brings pin 13 high. But as 12 is low this won't effect 11 and so won't start oscillation.

As I changed both 7400s together I thought I would take a punt and put one back in and see what happens. The first one I put in was the one that used to be U340 and it didn't work at all - the outputs were at -1.5V (not high and not low). When I put the one that was originally U330 it worked! Alt, chop and everything now worked. I still don't know why. I have to assume something about it being 74LS changed things but again I don't see how it works in the first place.

Channel 1

So now when I move the vertical position of channel 1 and 2 the right trace moves and if I am in channel 1 then the channel 2 position has no effect etc. If I connect a signal to channel 2 I can see it on the scope! Its a mess as trigger doesn't work but it is definitely progress. I get nothing from channel 1 however no matter what I do with the knobs.

There are two metal cans at the front of the scope that contain input attenuators. The white boxes below are the attenuators. Apparently they are a ceramic hybrid and are frequency compensated so they attenuate by the same amount across the scopes frequency range. There are a series of tiny gold switch contacts which switch in or out each attenuator and the switches are engaged by pins that poke up from this barrel attached to the volts/div knob below. In the photo below you can just see the first two switches although these ones are the AC/DC/GND switch contacts. The silver input capacitor is visible at the top.

I knew channel two was working so I attached a scope to the output of the attenuator block of each channel and compared. The output from channel 1 was very low. The attenuators are socketed so I carefully removed each of them one at a time and used a piece of paper soaked in Isopropyl Alcohol to clean the switch contact. Essentially you turn the knob until the switch opens, put the paper under the switch and then turn the knob to close the switch. Then you carefully(!) move the paper around to clean the contacts and pull it out. This got the signal shown on the test scope back to something reasonable and now the signal didn't change so much when I moved the knobs.

When I turned the scope on the trace for channel 1 was still a straight line. I then assumed I had an electronic problem. The vertical pre-amplifier is implemented using two custom, hybrid 'cascode' amplifier ICs. A cascode amplifier is two transistor arrangement that offers good bandwith, isolation and slew rate.

As channel 2 worked I could compare channel 1 and 2 to find the difference. I could see the signal at inputs of the second cascode amplifier. I couldn't see any signal at the ouputs. I searched for dead capactors, bad resistors, I tried removing the two transistors following the second cascode to see if the signal would appear but nothing. I even swapped the cascode amplifiers between the two stages but interesting this had no effect - channel 2 still worked and channel 1 still did not. Hmm.. I looked at the biasing of the second amplifier and noticed this was a little different. I noticed that if I set channel 1 to the lowest setting (5mV/division) that I did get a signal although a distorted one.

It turns out there is a second set of switches *underneath* the barrel that changes the volts/div. These ones set the gain for the second cascode amplifier (see below)

I found that I could do the paper trick but from the other side and clean the contacts of the switches underneath. Then I turned it on a again and yeah! A signal!

But still no trigger.

Till next time....