By Bud Izen, CET/CSM
Technicians usually get hung up on "dogs" for one of three reasons: making an assumption they were not entitled to make, lack of understanding of a basic concept or failure to perform an analysis in a methodical way (sometimes a combination of those factors). It is rare, indeed, that the reason a technician finds a certain problem difficult to solve is because of a lack of advanced knowledge of circuit theory. No matter how many new circuits or combinations of circuits the engineers may dream up, they still are composed of fundamental building blocks that rely on basic theory to operate. If you have a solid understanding of fundamental knowledge, and review it a lot, you will encounter far fewer dogs than the average technician.
This is important to remember: Regardless of the types of electronic devices you end up working on in the future, the functional block most prone to failure is and will continue to be the power supply. This is because it handles all the power used by the device. Notice I said handles and not consumes.
----------- Figure 1. A defective capacitor in this power supply circuit
caused a reduction in the voltage supply to the vertical circuit in this Sanyo
TV set. One of the symptoms was the presence of 59.94Hz vertical pulses across
capacitor C509.
The first steps of methodical troubleshooting allow you to isolate the source of the symptom to a defective block. Once you have reached that point, the next step is to make sure that the power supply reaching the block is correct.
Correct means that the value of the voltage measures very close to the schematic value (certainly well within 10% for most solid-state sets, and usually within 5% for sets having electronically regulated supplies) and has an acceptable level of ripple. The value of the voltage is usually checked with a voltmeter. Digital meters are a necessity these days, because most solid-state sets using integrated circuits require the voltage to be correct to at least one or two decimal places.
The oscilloscope, however, is a more convenient tool for quick and accurate analysis. Using the scope, you can examine the do voltage level and ripple voltage at the same time. From time to time, someone will ask me how to tell how much ripple should be present on a power supply line when the schematic gives no indication. A good example of this is found if you look at Figure 1. This comes from Sams Photofact 1971-2.
Although this is a B&W set, which my class received as a donation in defective condition, its power supply and failure symptom are common enough to use as an example of the principle I am talking about.
Notice that the ripple waveform at point 1 is the only one given.
Here comes the part where you have to start using your brains and your knowledge of basic electronics. Remember, troubleshooting analysis should be composed mostly of thinking rather than just taking a bunch of measurements.
So let's analyze the circuit of Figure 1. First of all, let's start with what we are given (always the best place to start). This may seem simplistic, but I always am surprised at how many technicians want to start working based on what they assume to be true rather than what they know to be true.
At point 1, you can see that we expect a do voltage of 131V and a ripple voltage of 6V. Both of those voltages should be dropped somewhat across R702 and R703 when current is drawn; Ohm's law indicates as much. This is easily verified from the schematic that shows 90.7V at point 2 and 80.7V at point 3. Because the schematic offers no other ripple waveform, it might be correct to assume that elsewhere in the power supply the ripple should be too small to consider. Rather than just accept that, perhaps we should question if that assumption is true, and if it is, why it is. The do voltage dropped (for example) from 131V to 80.7V between points 1 and 3. That is a reduction of 50.3V amounting to about 38%. Because the ripple must pass through the same resistors, it should also be reduced by at least this same amount, due to Ohm's law. I say at least because there are other components that affect the ripple besides just the resistors. In any case, the maximum ripple should be no greater than 38% of 6V or 2.28Vp-p.
C709B is a 200µF capacitor that has a reactance of about 1212 at the line frequency according to the formula: Xc = 1/(2 pi fC) = 1/(2 x 3.14 x 60 x 0.000220). We all know that reactance and resistance cannot be added directly. However, because the values are so small, only a slight error will be introduced if we do so. For the sake of rapid analysis, therefore, let's just do a voltage divider formula treating Xc like R. This means that at point 85, as far as the ripple goes, there should be no more than 12/162 (Xc divided by the approximate total of Xc plus the values of R702 and R703) of the ripple present at point 1. This would be about 0.44Vp-p maximum. Not very much indeed. This is a lot larger reduction than our previous estimate of 38%. It is more like a reduction of 93%! Let's do the same to calculate the maximum ripple at point 3. Using the same formula, the reactance of C509 should be about 26512. This creates an approximate voltage divider between R705 and C509 of 265/825 so that no more than 0.14Vp-p of the original ripple should remain at point 3. See how that basic theory helps.
If you didn't go through that analysis, you might have the same thing happen to you that happened to one of the technicians I supervise. When the set came in, it was assigned to one of the students who turned it on, examined the symptoms and turned the controls.
These are the first steps that any good technician would take as part of a preliminary diagnosis. He then noted that although the picture was stable, the vertical hold control was at one end of its range.
All of us make assumptions when we troubleshoot. How good a technician we are is often determined by how few assumptions we make. It would be reasonable to assume that the problem in this case was a defective control, a resistor or capacitor in the frequency-setting time constant circuit, a leaky transistor, a defective capacitor in the control circuit, or a cold solder joint. After all, these are all viable possibilities.
As stated earlier, you should never assume that the power supply cannot be related to the heart of the problem. It is good practice to check the power supply first, once the source of the symptom has been isolated, regardless of the symptom, but check the power supply voltage at the circuit under suspicion, not at the power supply itself. Why? Remember cold solder joints, bad connections, and secondary dropping resistors and filters, illustrated as part of the circuit but in function actually part of the power supply, certainly can cause problems.
In this case, the first thing the student did was to whip out the voltmeter only to find that the voltage was about 70V instead of 80. Because he was not yet familiar with the scope, he could go no further. So he referred it to the senior technicians.
The CET confirmed the student's diagnosis, but then started to make assumptions. Is it reasonable to assume that if the power supply to the vertical output circuit is low, that the symptom would be insufficient height? Well, the answer is maybe. As in maybe it would be unless the technician (or customer) already has turned the height control to make up for it. However, the tech made the mistake of assuming that this would be the only symptom which the reduced power supply voltage could cause. Then he made a worse assumption, based on the first.
Without thinking about the fact that the set was not drawing any more current than usual, he assumed that the vertical circuits were loading down the power supply, and that this was related to the control not working correctly.
I became aware of the problem after the tech had been working on the set for the better part of an hour. He came to me with questions about the Miller capacitor and so forth. I told him that I thought the problem might be a bit more basic, and it was.
I asked him if he had measured the power supply. He said yes, and that it was 10V low. I asked him if the power supply was "clean" as in pure dc. He said that he did not know. We then scoped the waveform that was present at point 3 and got what you see in Photo 1.
I asked him what he thought that was. He said, again making an incorrect assumption, that it was ripple. Although the scope was set to display the line frequency, that oddball waveform certainly did not look anything like what we all call ripple, which gets its name from its waveshape. The amplitude of that messy pulse was about 20V. I asked him if he thought that this was an acceptable level of ripple and he said that he did not know, that a value was not given on the schematic. I asked him where ripple came from, and he correctly indicated that it originated from pulsating do created by the rectification of the incoming ac. I asked him to scope point 85, where he got the waveform of Photo 2. He still did not get it. Finally, I indicated that it was impossible to get something for nothing, and could he just guess at where the waveform of Photo 1 was coming from. He said that it was coming from the vertical output, but did not sound confident enough to argue the point.
--- The waveform at Test Point 3 in Figure 1, which should have been almost pure
dc, exhibited these 20V pulses.
I asked him to carefully examine the waveform and see that indeed there were pulses on it. He agreed that it did appear to come from the vertical circuit. I asked him if he thought that the capacitor (C509) could tell the difference between the 60 Hz ripple and the 59.94Hz vertical output pulses, and he said no. So I asked him then, what did he think was causing the problem.
The light went on, and in less than five minutes, a new C509 cured the symptom, and the control would vary the vertical frequency faster and slower than normal from the approximate center of rotation.
If you're a basic theory oriented technician, you'll rarely overlook little things like this. Always scope that power supply first.
-------- The presence of these 0.4Vp-p pulses at Test Point 85 confirmed
that the pulses at Test Point 3 were not ripple, but were coming from somewhere
else.
When the B+ voltage of the set with schematics shown in Figures 2 and 3 was properly adjusted and the height correct, nonlinearity was exhibited.
---------- Figure 2. A check of the 110V voltage source on this Panasonic
set revealed a voltage that exceeded the expected voltage by 15V. When the
B+ adjustment was adjusted to return this voltage to its correct value, it
was impossible to achieve both proper height and linearity.
Here's another problem tailor-made to illustrate the exact troubleshooting methods I have been talking about. A Panasonic set using the ETA-12 chassis (Sams Photofact 1499-1) came in for a tuner-related problem. After we fixed the original complaint, we performed the standard set-up adjustments, the first of which was to check and set the 13+voltage.
Refer to Figure 3. The main power supply voltage, measured at point 1, should have been 11OV. We found it about 15V higher than normal. Because we always use a variable ac power supply to hold the input ac constant at the manufacturer's recommended , voltage (120Vac in this case), all we had to do was turn the B + adjustment (R817) and the proper do output was obtained.
Unfortunately, when the main do voltage was adjusted back to normal and we proceeded with the set-up adjustments, it was impossible to obtain proper height and linearity. If we were willing to put up with exceedingly poor linearity then we could get enough height, but if the linearity control was adjusted for a linear vertical scan, we could not get enough height. (See Photo 3.) I let the technicians work on this for over a week, an hour or so at a time. I figured it would be more educational for them to go through it themselves until they acknowledged defeat. (I was trying to get them to see that if they were not hot on the trail within 30 minutes, that it was likely that they would be no closer after several hours.
Such turned out to be the case.) Finally, they came to me for assistance. I made my own fatal assumption: that at sometime during the past week they had checked the power supply feeding the vertical circuits. Refer to Figure 4. It turns out that the vertical circuits are fed by both power supplies 1 and 2. They told me that both measured all right (with a meter) where they came into the vertical circuits.
We then took voltage and waveform measurements throughout the circuits and found that all was fine until we reached the input waveforms to the output devices.
When adjusted for proper linearity, the waveforms had insufficient amplitude but not as much as expected from viewing the symptom.
I like asking for help with nonstandard failures, so I called a local dealer friend of mine who probably sees more Japanese color sets in two weeks than we see at school all year. When I described the symptoms and what we had done, he said that he had seen nothing like this, but had we checked the power supply? I said that we had, but this started me thinking along the lines of the analysis I presented above.
Notice that there are two places where power supply 1 feeds the vertical circuits at which its voltage is reduced by a dropping resistor, and the result is filtered by a relatively small-value electrolytic. This occurs in the collector circuits of TR453, the vertical driver, and TR454, the lower vertical output transistor. It appeared worth the time to measure and scope the do at the collectors of both devices. The voltage at TR453 was clean and correct.
However, the voltage at the collector of the output device was a lot lower than expected and had a significant pulse on it (definitely not ripple). Unlike the previous example, the pulse was not radiated back through the power supply to any other circuits, which perhaps made it harder to find.
As you can imagine, replacement of the capacitor restored the correct do level and completely remedied the symptom. Proper height and linearity were restored.
It was obvious that the technician who had previously worked on the set had cranked up the B + to hide a problem he could not remedy.
The unfortunate part of that was that he caused a hidden safety defect, because increasing the B + also increased the high voltage beyond the recommended level.
(The set obviously had been worked on before--several components in the vertical circuits had been changed and removed from the circuit; the solder work was not factory quality. On its own, the B+ adjustment could not have drifted so far out of adjustment.
The moral of the story is that it is very important both to scope and measure the do power supply.
A secondary lesson that should be learned from this is to scope and measure any voltage that is reduced and filtered to feed any particular circuit to which the suspected cause of a symptom has been localized. Once this analysis becomes second nature to you, you will never worry about the question: Is it the power supply or is it the circuit?
------------Figure 3. A defective electrolytic capacitor in the vertical circuitry was causing the problem in the Panasonic set. Replacement remedied the problem.