Thread: Thandar TG2001
View Single Post
Old 21st May 2021, 6:46 am   #20
Radio Wrangler
Radio Wrangler's Avatar
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 17,857
Default Re: Thandar TG2001

In a case like this, the most precious thing you can have is certainty, the most damaging thing is confusion. If any of the replacement parts you buy are fake or faulty, then they soe confusion and can get you lost and keep you lost.

I restrict myself to buying parts only from main distributors who carefully manage their supply lines. Mouser, Digikey, Farnell/Newark, and RS.

You also seem to be trying to fix multiple dead function generators at once, and this can also create confusion as each may have different faults and it becomes difficult to keep them separate in your mind.

Something seems to be taking a lot of current. 8 regulator chips is an awful lot, however this is an opportunity. These things should not only act as voltage regulators, they are also implement current limits. They also act to limit their own temperature by backing off their current limit value.

To get one of these parts hot, you either have to provide it with too little heatsinking for the current it has to handle, or you have to try to take too much current from it. As these generators once worked, the heatsinking should be adequate, so long as replacement parts have been fitted with the right insulation washers and thermal compound. All of these devices have live tabs. They are either connected to the outputs or to the inputs. none of them should be grounded directly to grounded metalwork.

Eight regulators means you can look at the output voltages and see which ones are making their correct output voltage. This way you can spot a failed regulator shorting its input to its output and giving far too much output voltage. You can also spot a regulator feeding something taking far too much current and going into either current limiting or into thermal limiting.

This is where having so many regulators is a help. It breaks the unit up into smaller sections! So a bit of power supply voltage analysis will steer you to a smaller area.

For something taking a lot of current, there are several possibilities:

A shorted decoupling capacitor..... Tantalum electrolytics tend to fail short circuited. Ceramic capacitors can fail shorted as well. These could create heavy power rail currents anywhere in the instrument

The output amplifier stage is one of the highest power areas of the unit. It's very like an audio power amplifier, but with some differences because it has to work over a somewhat greater bansdwidth. These stages are the very devil to fix. One bad semiconductor can destroy other ones in the amplifier. Someone comes along and discovers Qxxx has failed, so they replace it. But Qyyy was also dead. Qyyy kills the new Qxxx. The repairer moves on, trusting Qxxx because it's new and finds Qyyy is dead, so they replace that next. But the dead Qxxx immediately kills the new Qyyy when power is next applied. This is the equivalent of the old joke of handing someone a card with "Please turn over" written on both sides! However in these amplifiers it isn't just two transistors in a mutually assured destruction pact, it's several, so the confusion generated is much greater.

The only way to deal with these circuits is to power them down, test every single semiconductor (and measure all the resistors if you're a suspicious devil like me) replace all the faulty ones, and only when you've done the lot dare you risk applying power to try it out.

Let's talk about protection. That output amplifier doesn't have any. If anything injects voltages back into it, it can be damaged, and the above horror scenario started.

Also, those three-terminal linear regulator chips in the power supply have aknown vulnerability. If something shorts or pulls down their input voltage quickly, the capacitors on their outputs can drive current BACKWARDS through the reg chip. This is known to destroy them. Wily designers know of this and fit diodes across input to output, reverse biased in normal operation, their job is to divert backwards current transients and limit the amount of collateral damage if there is a fault somewhere else. This instrument has none of these protection diodes. So either the designer wasn't aware of the risks, or his boss forced him to save every last penny.

The thing is also full of 4000 series CMOS logic parts. This is very very susceptible ot damage from electrostatic discharge, so you have to take precautions. It can go short circuit if damaged and take high current from power supplies.

The last bit on protection.... If you are using an instrument to drive something, think about what happens if that something is faulty or wrong. Can it feed power into the output of the instrument? This could wind up expensive! Even expensive function generators from Wavetek, HP or Tektronix can be killed this way. You need to think of some circuitry to protect your instrument to go between the instrument and your circuit. Getting this wrong can get very expensive very quickly.

Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done
Radio Wrangler is online now   Reply With Quote