An Idiots Guide to Oscilloscopes
Many Internet tube ‘gurus’ will vicariously demonstrate their prowess servicing a vintage tube guitar amplifier utilizing nothing more than a Fluke DMM. They make no mention of using old-fashioned test equipment like a VTVM, a dummy load, or even a signal generator. Mention the word ‘oscilloscope’ to any of them, and you are answered with a sneer. Mention anything about biasing your output tubes utilizing an oscilloscope, and the rest of your day will be spent listening to complaints about how ‘wrong’ it is to do so. Ignoring all the canard, let’s see what little I can ‘teach’ you about what an oscilloscope is, and how to use it to your benefit. Towards the end of this ‘Lesson’, I will demonstrate one method of biasing your output stage, using a number of analog oscilloscopes (and other test equipment) I purchased during the last year specifically for this purpose. While I do not have the deepest of pockets, I did spend considerable amounts of cash over the previous twelve months just for this ‘Lesson’. I spent a lot of time getting acquainted with my ‘new’ test equipment, and I believe I now have the minimal required experience and insight few ‘gurus’ possess. I will now attempt to pass along some of this acquired knowledge. The oscilloscopes used will be of varying ages and qualities; the point will be to show that the quality and condition of your oscilloscope will definitely make a difference in the outcome of your ‘work’. With that out of the way, let’s get to it! The oscilloscope is of course based on the cathode-ray tube, which displays the electrical signals in graphic form. It is probably the most widely-used test instrument because it can be used to observe waveforms as well as measure voltage, time, frequency, and phase angle. Below is a simple drawing demonstrating a basic oscilloscope cathode-ray tube.
There are opposing vertical and horizontal deflection plates inside the CRT, which is not evident in the drawing above. The electron beam is centered, and without any voltage is applied to the electron gun, we end up with a simple ‘dot’ on the viewing screen. The drawing below should make this a little clearer.
One vertical deflection plate and one horizontal plate are typically ‘grounded’, and the deflection of the electron beam itself should obviously be dependant on the applied voltage to the electron gun and the opposing deflection plates. Any signal to be observed is typically applied to the vertical deflection plates, and a sawtooth waveform (to be explained later) is applied to the horizontal deflection plates, in order to accomplish ‘sweep’. The drawing below will demonstrate that quite nicely.
If you are still following along, so far so good. We are almost done with the dry theory part. In normal oscilloscope operation, the beam is horizontally deflected from left to right across the screen at a certain rate. This sweeping action produces a horizontal line or trace across the screen. The actual sweeping of the beam is accomplished by the application of a sawtooth voltage across the horizontal plates. The rate at which the sawtooth goes from negative to positive is determined by its frequency. This in turn establishes the sweep rate of the beam. All oscilloscopes have provisions for selecting various sweep rates. A pictorial example of how a sawtooth voltage results in horizontal sweep is seen below.
Now that we have dealt with the mundane (but necessary) theory, let’s continue, and learn the controls on a typical oscilloscope. Keep in mind that while a wide variety of oscilloscopes are available, all have certain operational features in common. Whether it is a relatively simple instrument with limited capabilities, or a sophisticated model that provides a variety of optional functions and precise measurements, these ‘most common’ front panel controls operate identically. How more advanced features may help us will be discussed later.
This control varies the brightness of the trace on the screen. Caution should be used so that the ‘Intensity’ is not left too high for an extended period of time. Damage to the screen can result from excessive ‘Intensity’.
This control focuses the beam so that it converges to a fine point on the screen. An ‘out-of-focus’ condition results in a trace that is not sharp, but rather fuzzy.
This control positions the neutral horizontal deflection of the beam. This can make for more convenient viewing or measurement of a voltage.
This control positions the ‘crossover’ point (or ‘reference’ point, or ‘zero’ point) for easier measurement or observation of the waveform. Utilizing a dual-trace oscilloscope, for example, we could not clearly note each waveform if they both referenced to the same ‘zero point’.
This sets the horizontal sweep rate. This is also called the time base control, and selects the time interval that is to be represented by each horizontal division in seconds, milliseconds, or microseconds.
This sets the number of volts to be represented by each division on the vertical scale. This allows most waveforms, regardless of their voltage, to be displayed conveniently within the entire screen.
This switch allows the input signal to be AC coupled, DC coupled, or Grounded. The AC coupling eliminates any DC component of the input signal. The DC coupling permits DC values (or ‘offsets’) to be displayed. The Ground position allows for a zero-volt reference to be established on the screen.
This allows the beam to be triggered from various selected sources. This triggering of the beam causes it to begin its sweep across the screen. It can be triggered from an internally generated signal derived from an input signal, or from an externally applied trigger signal. The modes of triggering are most often Auto, Normal, and TV. In the Auto mode, sweep will still occur in the absence of an adequate trigger signal. In the Normal mode, a trigger signal must be present for the sweep to occur. The TV mode provides triggering of a frequency that is a harmonic of a Television sync signal. There will be a Slope switch, which will allow the sweep to begin on either the positive or negative slope of the trigger waveform. The Level control sets the voltage level on the trigger signal at which the triggering will occur. With all that out of the way, what makes one oscilloscope superior or inferior to any other? This will all depend on the application. The first ‘spec’ everyone knows to ask about is the bandwidth. In the good old days, the oscilloscopes were crude, and had a limited bandwidth. There was no need for anything fancy, until Television came along. Here we need to observe certain waveforms that have a frequency of approximately 4.5MHz. Before this, the highest typical frequency to observe was the 455KHz IF in the radio receivers of the day. Therefore, it is very common to find really old oscilloscopes with a bandwidth of 500KHz or 1MHz. With the advent of Television, oscilloscopes now had a typical bandwidth of 4.5MHz or 5MHz. Eventually, 10MHz oscilloscopes became the most prevalent. For audio work, almost any oscilloscope will suffice, but the higher bandwidth models usually also came with a few other ‘upgrades’ that makes them far more preferable. My advice is not to settle for anything less than a 20MHz bandwidth oscilloscope to use as your ‘#1’. The reasons are as follows.
- Older oscilloscopes almost always have a ‘single trace’. While this isn’t important to many, it does make it almost impossible to compare phase angles between the input and output of a coupling circuit, as an example. I say ‘almost’ because you may be able to find an electronic switcher, which was devised for just such a purpose. Still, a dual-trace oscilloscope is just infinitely more versatile.
- Older oscilloscopes usually do not calibrate their vertical amplifier or their horizontal time base. Many real ‘dinosaurs’ do not even feature a graticule on the CRT. This makes it hard to measure voltage or time. While this isn’t really important, it can come in handy.
- Typically, all 5MHz (and less) oscilloscopes will utilize vacuum tubes. While this may seem ‘cool’, these oscilloscopes will always have a set of quirks and headaches waiting for you. I haven’t seen many 10MHz oscilloscopes that used vacuum tubes. One notable exception will be the ‘laboratory grade’ Tektronix models. More on these later. But for now, vacuum tube oscilloscopes will have leaky capacitors (that affect the sawtooth waveform), as well as resistors and tubes that drift. It can be fun to restore these classic pieces, but the upstart technician is better off to take on these ‘fun’ projects later on during his education. Until you learn how to interpret the waveform presented to you, new oscilloscopes will often have a waveform that is sharper, clearer, and more accurate because the components inside aren’t fifty-years old, and are much more stable.
- Older oscilloscopes routinely utilized ‘banana jacks’ for their input connectors. The ‘average’ technician will not be able to use a low capacitance shielded cable for his measurements, and this can add some instability or noise to the signal being measured. Also, it is most likely that the capacitance of your ‘home brew’ test leads will seriously affect the waveform you see1. The astute technician knows that there are adaptors available today that allow a modern probe with a BNC connector to be used in these instances. The banana jacks are almost always spaced 3/4″ apart, and the people who make the adaptors know this. The adaptors are available to convert from banana jacks to a BNC connector, an RCA jack, or even a 1/4″ female jack (perfect for signal generators). Those pesky banana jacks are on a lot of older test equipment, so please have a few adaptors handy. Using unshielded cable for the signal generator and the oscilloscope is just asking for a lot of ‘hash’ on top of your reading.
- Along with the solid-state technology, the ‘modern’ oscilloscopes will also often have a higher accelerating voltage for the CRT. This yields a trace with superior intensity and sharpness. They will also be much more sensitive, giving more accurate waveform readings. As a digital analogy, think of this as comparing 8-bit to 16-bit or 32-bit. A complaint common to users of older B&K oscilloscopes is their rather ‘ambiguous’ trace at higher Television frequencies. My manual for the B&K 1470 oscilloscope lists the accelerating voltage as 1600VDC. I had one old-timer mention that he preferred to use his B&K oscilloscope with the room lights off, so he could really observe the trace as he was checking TV waveforms! My advice is avoid B&K and (those rebranded) Kenwood oscilloscopes as your #1, because of this quirk. Unless, of course, the price is just too irresistible. They do make an adequate back-up unit, and work well for using as a curve tracer. For comparison; cheaper Tenma oscilloscopes list their accelerating voltage as 2KVDC, some cheaper Telequipment oscilloscopes list their accelerating voltage as 3.5KVDC, the ‘intermediate’ Tenma oscilloscopes list their accelerating voltage as 12KVDC, and some Tektronix oscilloscopes list their accelerating voltage as 32KVDC! Which do you think has the brighter, more accurate trace?
- Along with higher accelerating voltages, you can also expect a faster edge speed or rise time. Lesser quality oscilloscopes could not keep up to faster edge speeds, because these generally require higher currents to produce them. Higher currents tend to cause ‘ground bounce’, especially on wide busses in which many signals switch at once. Moreover, higher current increases the amount of radiated magnetic energy and with it, crosstalk2. This means a re-design of the circuit boards, and not just slapping more B+ into the circuit. It also lead to the much-improved digital operating system for oscilloscopes (DSO). For analog oscilloscopes, higher accelerating voltages means clearer, more accurate waveforms. This did not go unnoticed until higher accelerating voltages were ‘discovered’. Many magazine articles and books were written, advising the technician how to interpret wave forms, including those that did not appear exactly as expected. Other ‘culprits’ can include lack of calibration, as well as the probe itself.
Interpreting oscilloscope waveforms is an art in itself, especially when a simple probe is not so simple.
- Much like an automobile, newer oscilloscopes simply have less mileage on them. This means they should last for quite some time yet. Of course, there are ‘lemons’ out there, just as with automobiles. There is nothing wrong with owning any oscilloscope, but be aware that until you learn its quirks and idiosyncrasies, your readings may not be accurate or meaningful. Much like the venerable tube tester, the astute technician learns how to interpret what he believes the oscilloscope seems to be telling him. You are essentially buying an electronic Ouija board, until you shell out the serious dollars for ‘laboratory-grade’ test equipment. Until you get a good feel for using test equipment, start small, and work your way up to the ‘big-league’ items.
There are many brands of oscilloscopes available, and it quickly becomes a jungle out there differentiating between any two. Here is a quick synopsis of a few of the brands I see at Ham Radio Swap Meets and on eBay.
- B&K should be avoided unless relatively new or the price is extremely attractive. As mentioned before, they usually have a lower accelerating voltage, and therefore an ‘ambiguous’ trace at higher frequencies. For simple audio sine-wave viewing, they will suffice nicely. They do make an adequate backup unit, and seem to hold up to constant usage. The older Model 1470 is very popular, and is a good, basic dual-trace 10Meg oscilloscope. The Model 1564 is much newer, has a 60MHz bandwidth, and is more preferred. Other models include the Model 1541 (40MHz), and the Model 1580 (100MHz). Avoid the Model 1450, the Model 1465, and the Model 1466; unless you can get them for about $40. When they do break down, they can make great door-stops, thanks to obscure Japanese transistors that are difficult to replace. Lastly, older B&K oscilloscopes (including the 1470) ‘feature’ a UHF connector (properly called a PL259) for the probe. This makes it hard to use just any old probe you may have, unless of course you have the handy-dandy adaptor. These adaptors are available, and they work just swell, so this isn’t a major concern. It is still something to consider.
- Bell & Howell oscilloscopes (made by Heathkit) are best describes with three words; (1) avoid, (2) avoid, and (3) avoid. Vacuum tube units like the Model 34 have a trace that makes a sine wave look almost like a sine wave. These were mostly school-training units, and I wonder how many students got discouraged about electronics after using one of these oscilloscopes. For the nostalgia, or for $5 or $10, you can have one of these as a conversation piece; to show how the ‘good-old-days’ weren’t always so good!
- Eico is one of those ‘oldies’ names you’ll see frequently at the Ham Radio Swap Meets. They made everything from battery eliminators to stereo receivers to tube testers, so finding an oscilloscope really isn’t out of the ordinary. Their very common Model 460 (4.5MHz) isn’t anything to write home about, but does work reasonably well. Of course, this is only after you’ve replaced every capacitor and resistor inside. As a beginner’s oscilloscope, it will do, providing you do not pay more than about $30 for it.
- Heathkit is another one of the more popular ‘oldies’ to find. Quite a few vacuum tube oscilloscopes like the OM-3, the 10-12, or the 10-30 out there. You know my feelings about any oscilloscope that uses vacuum tubes; if you can get it for cheap, go ahead. But, be prepared to spend your days getting it into tip-top shape. And, when all is said and done, the 900VDC accelerating voltage can still make you squint until you get a headache. A ‘step up’ is the solid-state single-channel IO-4530 (10Meg), but this is still just a ‘so-so’ beginner’s oscilloscope. You can do a lot better, and for about the same amount of cash to lay out.
- Kenwood are usually rebranded B&K oscilloscopes. The CS-5350 is a good oscilloscope (50MHz), but for $1,300 (list price today) I would want something that could also butter my toast and sign my kid’s report card.
- RCA is another of the ‘oldies’. The old vacuum tube based WO-33A or WO-91B may look like a classic, but getting one to work ‘properly’ can be a chore. They look impressive on your vintage workbench when friends drop by, but as an ‘everyday’ oscilloscope, you can do much better. I own a WO-33A, and it isn’t my ‘go-to’ oscilloscope, by any stretch of the imagination. Aside from the 4.5MHz bandwidth, the teeny 3″ screen, and the non-calibrated graticule, the built-in clip-lead probe simply doesn’t inspire confidence.
- Tektronix would have to be the Hickok of the oscilloscope world. They were very expensive in their day, and command high prices on eBay today. There are very old units using vacuum tubes, and more modern solid-state units. Good vacuum tube units include the Model 310 (very useless for anything other than audio work because of a 3.5MHz bandwidth, but it is still a solid unit) or the Model 506 (an excellent 23MHz unit, although it uses plug-in modules that may be hard to locate), while the Model 453 (50MHz) is one of the best ‘basic’ solid-state units out there. Also look for the Model 465 (100MHz), the Model 454 (150MHz), the Model 475 (200MHz), and the absolute ‘Cadillac’ in the Tektronix line of analog oscilloscopes; the Model 485 (350MHz). Most Tektronix oscilloscopes have a high accelerating voltage, and a very nice trace that is clear and accurate. However, even Tektronix made a lemon or two, so check out your prospective purchase very carefully.
- Telequipment is simply the British Tektronix. What is strange is the Telequipment oscilloscopes I have or have seen seem like they were made from a kit. They are not made to a heavy-duty standard like the Tektronix, but are still around, and they usually still work fine. The Model D52 and Model S51e are very good and very basic, but aren’t my first choice; least of all because of their banana-plug inputs. The Model D61a is a better bet; still a basic 10MHz unit, but it seems to work well. The trace appears fine, partly due to its 3.5KVDC accelerating voltage. And better still is the Model 67a. This is a 25MHz oscilloscope, with a 10KVDC accelerating voltage. The trace is very nice, indeed.
I have left out many brands of oscilloscopes here. These would include Hameg, Hewlett-Packard, Hitatchi, Leader, and a few others. This does not mean they aren’t worth considering. What this means is that I simply do not own more than one, nor do I have the resources to ‘try a few out’ before I pass any opinions along. Owning one or even two Hitatchi oscilloscopes (which I do) is not enough to form a valid opinion and pass it along to you. Once I have used the aforementioned oscilloscope enough to understand their quirks, only then will I gladly pass along any observations. The opinions listed above are ‘first impressions’, and are subject to change at the drop of a hat.
Kicking the tyres
What I hope to do here is show you a handful of oscilloscopes I recently purchased. I will demonstrate these oscilloscopes ‘in action’, as I attempt to bias a few different vintage tube guitar amplifiers. Let’s meet the oscilloscopes, and the amplifiers. Also, while it should seem obvious, I’ll also have to reveal the rest of the test setup. These items include the following.
- Audio generators. This a whole ‘Lesson’ in and of itself, but for now I chose to use either an RCA WA-504A or a B&K E310B. The RCA unit is a very well-made unit using MOSFET’s, with a THD (Total Harmonic Distortion; a relative indicator of how ‘pure’ the sine-wave is, or the ratio of distortion to sine-wave present in the signal) figure of less than 0.25%. The B&K unit is another excellent audio generator, with a very clean and accurate waveform. You may have an older tube unit like the Heathkit AG10 or an Eico 377. If the unit is well maintained, whatever you have will suffice. While the square wave function is ideal for testing the frequency response, distortion, and tone control operation, I simply used the sine-wave function. I chose a 1kHz signal, with a measured output of 250mVAC. This simulates the output of an ‘average’ single coil pickup, and at a frequency somewhere in the middle of the range to the combination of the guitar/amplifier/speaker you’d probably be using. The test parameters were not chosen to try and obtain a predetermined outcome; I’m a little older than that, OK?
A selection of good, reliable signal generators isn’t necessary, but it makes experimenting a little more fun.
- Dummy loads. You have to use one, so how do you choose? Do I have to go with a non-inductive load? Since I wasn’t sweeping the audio range, this does not matter. Do I go with 7.5-ohms or exactly 8-ohms? Since the rated speaker impedance is a nominal figure, the value I use is not as important as some may argue. I have tried various setups; two 15-ohm resistors paralleled for 7.5-ohms, and two 4-ohm resistors in series for a total load of 8-ohms. The results were identical. Therefore, I won’t mention exactly what I use, simply because it isn’t important. Use what you can get your hands on, as long as the power rating is about four-times the rated power of your amplifier. Often I used 150-watt resistors, available from people like Mouser. They’ll still get warm, but not as hot as using a 50-watt resistor to test your Bassman. Make sure your resistors are mounted to a solid ‘base’ and have adequate ventilation. If I plan to leave the amplifier running into a dummy load for extended periods of time, I will also have a small fan mounted near the resistors.
- Volt-Ohm-Meters. You need to measure the bias range, and the output power of your amplifier. I use digital VOM’s and VTVM’s. Use what you have, as long as you are cognizant of the fact that your readings will only be as accurate as your meters. However, the exact reading isn’t crucial. We are more interested in knowing that the output power we have, and our bias voltage/plate current, seem correct, rather than the exact value down to the microvolt.
- Tube testers. I checked the NOS tubes, and the new tubes, for ‘balance’ in the 12AX7 halves. I also chose output tubes with a gm reading that read within 5% of each other. The tube testers used were a Hickok 752A, a Triplett 3423, and a B&K 747A. I chose these testers thinking that they were popular on eBay, and very likely to be similar to what the aspiring technician is using.
- Tubes. I decided to bias up the amplifiers with at least two different sets of output tubes, to compare if optimum bias settings were much different with different tube brands. I could have chosen new tubes from Sovtek, Svetlana, Groove Tubes, or JJ, as well as NOS American tubes I had in my collection. Output tubes were not ‘burned in’ before the testing, but the amplifier was left idling, played for some time, and the bias was checked again. I understand that after the amplifier has been ‘out in the field’ for a while, the bias will possibly need to be tweaked, but this is not the issue at hand today. We are interested in setting the bias as quickly as we can, in an effort to keep the ‘professional repair shop’ moving along efficiently. For those of you at home, with all the time in the world, you can perform this procedure as you see fit. Burning in the tubes is not a bad idea, and it will save time later.
Lastly, while it should seem obvious, to some it may not be as plain to ‘see’ the proper setup arrangement for the whole enchilada. That would be shown below, in my best artwork. This is a simplified setup, and geared towards your particular amplifier. I’ll discuss an ‘advanced’ version in a second or two.
Test setup is not elaborate, nor complicated. This is a good thing, because I wouldn’t have been able to explain it.
The ‘key’ points are to keep the leads as short as possible, and to have coupling ‘losses’ at a minimum. This is achieved by using the proper cables, and the necessary adaptors where needed. If you plan to do a lot of bench work, a semi-permanent setup is advised. I use a shelf above the bench itself, with a panel utilizing all the jacks and switches I need. From there, it is one cable from the signal generator to the amplifier, and one cable from the amplifier to the panel. Everything else is ‘hard-wired’ behind the panel, to keep any losses as insignificant as possible. This is far neater than having a bench top that resembles a plate of spaghetti. I have multiple dummy loads, that can be switched in or out of the circuit, and an AC VTVM hardwired to the dummy load switch! This lets me ‘see’ at a glance the output power of the amplifier, or the gain factor of any modification/extra stage/FX Loop/etc. I may be experimenting with. However, I really don’t think many of you are going to go that far.
The amplifiers I chose are listed below, and the dedicated ‘page’ can be accessed by clicking on the appropriate ‘link’. This was done to keep loading times as brief as possible. Select the amplifier most similar to what you are interested in learning how I dealt with it. There is nothing more to add, so buckle up, and pay attention. AA1164 PRINCETON REVERB. This combo was chosen for numerous reasons. There is also an anticipation that this will be a very interesting exercise. For starters, there is not a lot of choices when shopping for a 6V6 today. Secondly, there is no real ‘adjustment’ possible for the bias voltage setting! There is only a very narrow range of resistors to choose from, and hopefully have them act as a voltage divider and supply the ‘proper’ bias voltage. In other words; expect a lot of surprises on this trip. AB165 BASSMAN HEAD. This head was chosen for a few reasons. It is a popular/common amplifier, and I had a chance to try a few mod ideas seen elsewhere on this website. However, for the purpose of biasing and demonstrating the results, the amplifier was left stock except for replacing the bias ‘balance’ circuit with a conventional bias circuit. WARNING I demonstrate a lot of different oscilloscopes, so it will take a few minutes for all of the images to load. Please be patient, and you will be rewarded! 50-WATT MARSHALLl JCM800 HEAD. This head was chosen to demonstrate two things. First, biasing any Marshall is a matter of personal taste. Secondly, biasing many EL34 amplifiers is a lot more complicated than biasing 6L6 amplifiers.
Where To Learn More About Oscilloscopes
Here are a few websites you can visit, and learn a little more about oscilloscopes. There are also of course vintage electronics text books, but I’ll concentrate specifically on the websites (for now). There are actually quite a few oscilloscope websites out there, but many caused my computer to crash(!), or were erratic in loading. The websites listed below all passed my ‘no hassle’ test. CATHODE-RAY OSCILLOSCOPE TRAINER is absolutely recommended for the fact that you can play with an analog oscilloscope ‘on-line’! Adjust controls, and see how they affect the waveform. Hook up different signals, use the ‘Chop’ function, etc. This is for beginners that feel intimidated in asking a live person how to use an oscilloscope. OSCILLOSCOPE-TUTORIALS.COM has a name that says it all. You will get a lot of dry theory, but you will learn, damn it. USING AN OSCILLOSCOPE is an interesting British website. You even get a few little experiments thrown in for good measure. The Hameg oscilloscope used for demonstrative purposes won’t be too common on this side of the big pond, and there is no ‘virtual’ training, but it is still a pretty nifty site. You should understand all of the oscilloscope controls a little better after a good visit or two. VIRTUAL OSCILLOSCOPE is one of the cooler virtual oscilloscope ‘trainers’. You have to download the program, and have ‘Flash Player’, but it is definitely worth it. After you ‘click-and-drag’ various signal sources to be hooked up to the oscilloscope trainer, you set controls, and watch the display. Another ‘beginner’ website, but still one of the most explicit learning experiences available. Get a glimpse into the way things really went down over at the Fender laboratory! If you have any curiosity at all about what test equipment Fender felt was ‘good enough’ for them, CLICK HERE!
If it’s good enough for Fender
Get a glimpse of the way things were done by the Fender technicians in their very own laboratory. If you wondered what test equipment they used and believed was good enough for them, SEE HERE.
1) Probes For Test Instruments, by Bruno Zucconi and Martin Clifford, 1957. 2) The XYZ’s of Oscilloscopes, published by Tektronix, 1985.