Other Faults

Telequipment DM63 (17K)Over the years, I have fixed many faults on this oscilloscope. Here are the ones I that can remember:

Hum on trace

This caused a lot of headscratching, as it was not surprisingly caused by multiple faults.

I started by doing a blanket replacement of all of the electrolytic capacitors in the instrument. This might seem a little extreme, but bear in mind that they were all dated 1974, and obviously well past their prime. It's important to realise that there is no point trying to apply logic until components that gradually degrade have been brought back to their original specification. Also, there isn't actually that many, so it wasn't too expensive or time-consuming.

This improved things, but it wasn't a complete cure. Looking back at the notes I took at the time, there was 350mV, 100mV and 150mV of ripple on the 105V, 24V and -24V rails respectively, and replacing the capacitors brought this down to 130mV, 25mV and 15mV. The 105V rail still seemed too bad, so I changed TR416 which brought it down to 50mV - it had obviously suffered a reduction in hfe after years of running hot. But, there was still visible ripple on the trace, and further changes made little difference.

After much thought, I realised what was going on, and as with all the best faults, it was totally obvious with hindsight! The waveform of the ripple had been confusing me because it wasn't the normal 100Hz sawtooth that you might expect, but actually, this was the massive clue that I should have taken notice of: it was pulses - positive and negative going - the classic full-wave rectification current waveform. This picture will hopefully explain all:

Simplified schematic of the PSU (4K)

The power supply connects to the mother board (PC175) via edge-connector SK802. The only earth connection is via pin 5 - nothing too worrying about that in this particular application... until you notice the bridge rectifier hidden away on the mother board!

With age, the contact resistance had risen. The capacitor charging current pulses were causing a voltage to appear across this resistance, which was being added to the DC voltages produced by the power supply. The answer was to remove the connections from that winding of the mains transformer, and install a tightly-twisted pair directly from the transformer, terminating in a Molex connector. A corresponding Molex connector has been soldered directly to the rectifier diodes on the mother board.

This bridge rectifier and smoothing capacitor is for the heater supply to the flood guns, so this problem shouldn't occur on the non-storage version of this scope (the D63).

No X-Y mode, no alt-triggering on RH plugin

The same fault caused 2 symptoms, although you can be forgiven for not immediately seeing the connection. It was complicated in my case by my Veroboard bodge - I couldn't remember when the X-Y mode stopped working, or even if it ever worked since my initial repair. Shows how much I use that facility!

After lots of measurements, I convinced myself that my Veroboard version of IC1 was doing the correct thing in X-Y mode, but as I wasn't really sure where to head next, I decided to think about the alt-triggering problem instead.

The scope mainframe presents the timebase module with 2 trigger sources - LH and RH. However, the mother board can obtain each of these trigger signals from two places, depending on the 'Alt-control' signal from the appropriate vertical plug-in (the voltage on pin 6 of each vertical plug-in edge connectors). There's a range of different modes available, depending on the installed vertical plug-in:

  • V1 - standard single plug-in

    Pin 6 'Alt-control' signal is high, so the mother board takes the trigger signal from the main Y signal (via TR805/806 or TR807/808).
  • V5 - single delayed plug-in

    The trigger signal needs to be taken from before the delay-line contained in the plug-in. In this instance the 'Alt-control' signal is low, and this causes the signal to be taken from pins 7 and 10 of the Y plug-in.
  • V4 - dual plug-in

    The 'Alt-control' signal depends on which trigger source you select. You have a choice of 3 modes - channel 1 or 2, or Alt triggering. The latter is used if you just want to trigger on the signal that is visible on the screen and don't need to take the trigger signal from a specific channel. Or it can be used to all you to trigger on two completely different signals at once when you are viewing them in 'Alt' mode (with varying degrees of success in practice). Alt trigger mode takes the trigger signal from the main Y-signal (pin 6 high). However, if you are triggering on channel 1 or 2 specifically, the scope needs to accept a trigger signal on pins 7 and 10 of the Y plug-in (pin 6 low).

To complicate matters even further, when the scope is switched to X-Y mode, the X-amp is fed with the RH trigger signal that is sent to the timebase plug-in. To ensure consistent operation, the switching is arranged to force the RH trigger source to be via TR807/808 instead of pins 7/10 of the RH plug-in. This overrides the 'Alt-control' voltage from the RH vertical plug-in.

The way all of this is achieved is slightly strange, and understanding it is not helped by the diagrams. But once you've worked it out from the rather terse circuit description, you begin to see how the two faults are related. TR807 (an MPS6518) was the culprit.

Reduction of trigger sensitivity on one channel

It was a simple matter to prove that the fault was in the V4 plug-in. I started trying to examine the trigger circuitry in the plug-in, but based on my experience of the above fault, I decided to change the obvious things before wasting time trying to work out the details. It was a good plan, as I found the fault almost instantly - another MPS6518 (TR758).

These transistors aren't easy to find, but they are broadly similar to the ubiquitous BC327-25 - according to Towers, the only spec that is worse with the BC327 is Ft, but that doesn't seem to cause a problem.

Reduction of deflection, beam present at minimum intensity

When these symptoms occurred together, it was easy to realise that the EHT had gone high. The supply is generated and regulated by a 3-transistor circuit in the metal EHT can at the back of the unit. TR302 conducting more has the effect of reducing the EHT, and it gets its base drive from R311 (120k). This resistor looked suspicious, having around 100V across it, and it was a good hunch - it had gone open circuit.

Variations in EHT voltage

This drove me mad while trying to calibrate the unit! Initially, I noticed that if you used the storage modes, the trace was a bit soft when you went back to normal (non-store) mode. At the same time it went slightly off-calibration. It took me a short while to notice that the two symptoms were connected, but once you've made the connection, it was obvious that the EHT was changing. It seemed that large changes in beam current caused by the operation of the storage modes would trigger it. Power-cycling the unit would fix it, until the next time...

From an access point of view, the EHT generator is the worst part of the scope to work on. So, before spending time thinking too hard about it, I decided I should take the easiest route first, and luckily replacing the 2 socketed BC109s fixed it...

DC BAL Mod

DC-Bal mod (16K)You've probably noticed how difficult it is to set up the front-panel DC-Bal control on these units. The control has far too much range, and as soon as you get it right, it drifts off again. This in turn affects the remaining DC balances inside the plug-in.

Looking at the schematic, you'd think it would be a simple matter to change R603/R604 (220K) to reduce the range of the control. Not so! Because of the way these things are put together, it's impossible to get to the resistor without unsoldering lots of other wires...

The easiest solution is to solder an additional resistor directly to the potentiometer - lift the centre (wiper) contact, and solder a 1M resistor in series. On one of the four channels (shown here), I couldn't quite get enough adjustment, so I had to add another 1M resistor, again directly on the pot, between the new resistor and the end of the track...

Having done this, I subsequently discovered that later models have addressed this issue by adding a 680 ohm resistor in parallel with R605/R606, reducing the value of these from 1K2 to around 430 ohms. This is better than increasing R603/R604 - as my mod does - because it's always good practice to minimise impedances wherever possible. This picture shows how Telequipment applied this mod:

"Official" DC-BAL mod (30K)

Conclusion

This thorough restoration has been well worth the effort; the DM63 is a really nice bit of gear for general workshop use. Having 4 channels is sometimes incredibly useful, as is the storage facility.

The only remaining problem with this example is the tube has aged, and the storage functions don't work quite as well as they should. Part of the calibration routine is to adjust the storage cathode current (there's separate heaters and guns for the storage part of the tube), and I couldn't get close to the required value... But it's quite usable for most tasks.