Notable audio electronics designer, Roger Modjeski of Music Reference and RAM Tube Works, founded the RAM Audio School to educate students on understanding and designing audio circuits based on practical application and experience. Roger wants to share his vast knowledge and experience in audio electronics theory and practice.
Courses progress from basic content into more detailed subjects. They can be taken as a series in the order listed or separately.
Roger Modjeski – Founder and Director
My first knowledge of the insides of a tube amplifier came in 1956 when, at age five, I watched my father’s every move as he built his Heathkit Mono hi-fi system. <!–more–> The amplifier was a “Williamson-Type” model W4-M, the preamp a WA-P2, and the FM tuner an FM-3A. I still have all three, in working condition, including the manuals that instructed my father, under my watchful eye, to <code></code>build them. The manuals are dated 1954, just two years after Heath released its first hi-fi kit. Heath was the largest and most professional of the companies to offer kits for audio, amateur radio, test equipment and other interesting electronic products from 1952-1991. I built a dozen in my youth, including the dual trace oscilloscope (1965) that we still use in our service department and the IM Distortion analyzer (1963) used to test your RM-10. The scope, which is all solid state and built on printed circuit boards, has been less that totally reliable and is hard to service. The older IM Analyzer is built on a chassis with tubes and has never failed. (Reliability Lesson One!)
Dad’s hi-fi served us well for eight years and then developed a noise in the preamp that my Dad could not fix–nor could Dr. Joe, our ham radio friend, nor could I, at first. I must have traced that problem off and on for months, trying everyone’s suggestions: noisy tube, noisy (leaky) capacitor and, finally, noisy resistor. Of course, the culprit turned out to be the high quality, precision part which we least suspected. And there I learned that the most expensive parts are not always the best. I continue to see this when Kavi Alexander of Water Lily Acoustics brings me his tube mics and preamps that have had their resistors replaced by expensive “audiophile” brands, only to have the noise rise above the level of the removed cheapies. After installing my selected brand of low noise resistors (the same ones I use in the critical locations in your RM-10) the noise faded into oblivion. Many of the current audiophile resistors have been chosen on the basis of their high price or exotic materials–often by people who do not understand noise mechanisms. Their makers are often concerned with precision and other characteristics that non-audio applications require. I have not found low noise to be one of them. Lesson Two: Expensive resistors are not always the best.
Other than that one resistor, the Heath preamp was trouble-free. I wish I could say the same of the power amp. After eight years of flawless operation, a noxious smell came from the cabinet in which the amplifier lived. It was the smell of cooking fluorescent ballasts which we all remember from our classrooms. The power transformer, in its smooth, military-looking drawn metal can was hot and oozing smelly tar all over the chassis. I was alarmed at the failure of this major part, which had the appearance of infinite longevity. However, Heath was there for us, and $13.50 later I had a new one to install. Only a few years ago, I learned from a maker of fluorescent ballasts that the tar or pitch used in the transformer industry was one of the most unreliable potting compounds. Batches often contained moisture and contaminants that caused oxidation of the internal wires, causing opens or overheating shorts leading to that characteristic smell of burning tar. Thankfully, the insulation industry has improved several orders of magnitude in the intervening years, so that your transformers are wound with double coated, long-life polymer insulated wire and sealed in modern resins. That original amplifier, by the way, still works and resides in my hi fi museum with its original tubes, which still give full rated performance. Lesson Three: Never pot transformers on tar, wax or any other moisture absorbing material.
Incidentally, the FM tuner was probably the most time-consuming to build and never worked well. It drifted as it warmed up, had too many tubes and no style. When it comes to tuners, we should be thankful for solid state. Properly repairing and aligning tube tuners is beyond the expertise of most anyone and best avoided.
My design career began at 11, when I built my first 3-tube, class A single-ended amplifier. I still have the little thing, which is capable of about two watts. It’s description reads much like that of a $5000 amplifier that the high-end magazines might cover today. (Should I send it in for review?)
As I grew, I wanted more style than I could produce in my dad’s basement shop, so I bought my first stereo amplifier kit in 1964 for $29.95. Carefully perusing the catalogs, I found the best bet to be a Knight (by Allied Radio) KG-240. It had two-line level inputs, passive tone controls, and a class A output stage capable of nine watts per channel. (Does this sound like something you just read about in the latest magazine or some manufacturer’s literature?) The truth is that the little Knight had too much hum–which I fixed right away by adding filter capacitance–and had mediocre sound due to its poor transformers and poorly tuned circuit. But for $29.95, what could I expect? It wasn’t until many years later that I learned the meaning of “eating the menu instead of the meal.” Modern advertising depends on just this principle. It’s the difference between the mouth-watering description of the latest fast food delicacy and the experience of eating it. We also see it in the recommendations of magazines to buy the flashy chromed amplifiers that have less output and higher distortion that your RM-10. Will the joy of polishing all that chrome make-up for the poor sound these amplifiers produce under difficult musical conditions?
Around 1964, my interest and the industry’s turned to the new “miracle” transistors. I, in my basement shop, and the giants of the industry all did our best to design good sounding amplifiers with these new devices, and we all failed. I even invested $79 in a compact Eico Cortina 70-watt (Total IHF music power rated) integrated amplifier because I wanted more power, and because I wanted to graduate from a ceramic cartridge on my cheap Knight turntable to the magnetic Empire 888 I had installed in my dad’s Garrard RC-80. By now, my father’s system had melded into mine and I was the Hi-Fi Meister of Melbourne Drive. I had to convert the Garrard to stereo, and went through a lot of cartridge damage before I realized it was too rough on the Empire (or any modern cartridge of the late ’60s). I later made the biggest expenditure of my teen life on a Dual 1019 turntable, which is still in active service as a dedicated 78 RPM player for my jazz records.
The Eico had its flaws, including hum and severe distortion on the phono when the mono switch was used on a stereo record. Although I don’t remember measuring its power output then, I did recently; it’s about 20 watts per channel into 8 ohms. My dream of becoming a “Kit Engineer” for Knight or Heath was rapidly being eroded by the quality of their designs. When a 15 year-old has to redesign flaws out of a professional product, something is wrong; or maybe something right was inspiring me to make better designs. Of course, I also had no idea at that time what a manufacturer has to go through to produce a product at an attractive price with good service (I had to become one myself to get that lesson). That was the last amplifier I ever bought. From then on I built my own, continuing to follow the progress of the transistor in the pages of Popular Electronics and Radio Electronics through the 60’s and 70’s. I experimented with many transistor circuits and invented a few of my own to explore the world of transistors in full-complementary, class A, no feedback circuits.
In 1969, I went to the University of Virginia to get my degree in Electrical Engineering, where I was the only student in my E.E. program who had built his own amplifiers. I wasn’t expecting to be so lonely! I couldn’t even find a hi-fi buff on the faculty. So I spent a lot of time in the music department, studying music history and theory, teaching fellow students how to use the ARP Synthesizer and recording concerts on their Revox A-77. I had to go up on the catwalk in the skylight above the stage to insert and remove the batteries in the Sony condenser mics for each recording. Because the adrenaline rush of potentially falling 50 feet through the skylight onto the Steinway below kept me alert for the session, I quickly began reading about phantom powering those mics. Being low on funds, I made up some phantom power boxes with some old matching transformers I found lying around and a couple of batteries. Now I only had to go up into the loft for fun!
As a pianist, I was fascinated by the inoperative Skinner pipe organ in the hall and wanted to play it. Finding the keyboard 50% inoperative due to broken wires, I took on the task getting it back in action for my midnight “Phantom of the Opera” sessions, where I would play Toccata and Fugue in D Minor for my friends. Amid all this fooling around, I got my Engineering degree with High Distinction and headed to the world of industry.
I landed at IBM, and quickly saw this “great opportunity” was a dead-end. I then opened an audio repair shop in the basement of Atlantis Sound, in my home town of Richmond, Virginia. Here I learned about the reliability (and lack of same) in audio equipment. I also learned to swear at the names of those companies who made it hard to get in and out of things. I charged on a flat rate and didn’t have time for poor serviceability. I was averaging a unit an hour, start to finish, including writing the ticket. I saw things like the Sony STR 7065 receiver that failed like clockwork after a few years (totally hour- and heat-dependent) because they used a 1/4 watt resistor at 1/4 watt. Didn’t they know, I mused, about derating? Did it slip by some engineer’s check? Sadly, almost every unit failed, and it cost many customers $35 to $50 (in 1974 dollars) to have those one-cent resistors replaced. Many years later I learned the same hard lesson at Beveridge, when a properly derated, high quality resistor failed repeatedly in the RM-1 power supply. I finally decided that, even though the resistor was also within its voltage rating, the manufacturer’s rating was too ambitious, and the best option was putting two resistors of half the value in series. After that there was never another failure of that part. Luckily, I learned that lesson in 1978, and that’s why the plate resistor of V1 in your RM-10 is split in two.
For two full years, I repaired an average of 10 pieces a day, three days a week, for a total of about 2,000 audio units, from Audio Research to SAE. One day a pair of Futtermans (early OTL amplifiers) came in, which kept me busy for a while. When the store owner learned I could make an A-77 sound better than new, he offered a $50 tune-up special. I must have repaired and aligned 100 Revox A-77’s. I still have a few in my museum and find them to be aging poorly. All those wonderful dipped tantalum capacitors are opening up as the years go by. I figure it would take several hours to replace them all because Revox hid them everywhere. They were easy to get to before the mother board was set deep in the machine. Getting the mother board out is another time-consuming issue. There I leaned about products that are designed for manufacturing efficiency with little regard for repair efficiency.
Cars are similar; some starters you can replace in 1/2 hour and some take three. You know that if you keep the car long enough, you will replace the starter and the water pump and the alternator, among other things. We don’t throw the car away when the starter quits, but we are tempted to junk the Revox when 30 hard-to-reach capacitors are all failing. I don’t use my Revox anymore, I just look at it. I still use reel-to-reel every day, an equally old (1971) Sony whose capacitors are still good because they are aluminum electrolytics, like the ones in your RM-10. As you see, I’ve made a lifetime a study of the reliability of components. Strangely, if we look at the manufacturers’ life rating for many components, we find things lasting far longer than some ratings would lead us to believe. The capacitors in the Sony are rated for 1,000 hours, but are now 23 years old and, given my usage, have probably seen 20,000 to 40,000 hours. There are about 60 of them in there and not one has failed.
By 1975 I had my fill of fixing things and was itching to become a teacher. To that end, I went to Stanford University for a Master’s degree. After one year of classes and student teaching, where I became known as Mr. Parts because I had a desk full of this and that, I saw that teaching in a university was not going to be what I had hoped for. I then went back to Virginia where I opened the High End store, Audio Art. There I spent several years in sales and service, getting to know the audio customer and his needs.
Although I first met Harold Beveridge while at Stanford, I got to know him and his product as his dealer in Virginia. In 1978, Bev hired me as a consultant, later to become his Chief Engineer in Santa Barbara. I went to work for virtually no salary to be the student of this very creative, well-seasoned electrical engineer. In the same tradition, I have taken on a few apprentices and always have room for those who burn with the desire to understand and make fine audio equipment.
I did what I could to promote reliability at HBI. I was a real pain about it, but Bev was in denial about reliability and I left after three years. While there I designed the RM-1 preamp and its RM-2 power supply. In the early versions I got caught by not derating the 6DJ8 tubes enough. Although they were run at 1/2 their rated dissipation, I later found that it took derating to 1/4 to get them to last the years I wanted them to. The later RM-1’s were darn reliable in their “low-power” version, especially when one considers that they had over 1,500 components. I also did a solid state crossover for Bev which was very reliable, except that the pots would break loose from the circuit board if the unit was tossed around in shipping. Although they could be easily re-soldered, I learned another lesson about reliability.
By 1981, the time was ripe for me to start my own company, Music Reference. My first product was the RM-4 Vacuum Tube Head Amp. I designed it out of the challenge that a tube headamp for a moving coil cartridge was an “impossibility.” I had been using the 6DJ8 for three years and knew that quiet samples could be found. I developed the equipment to measure the “RAM Factor,” at first by hand and later by computer, as we do to this day. If you read the hi-fi magazines you have probably noted reviewers solving noise problems with RAM tubes. We could have never made the large number of RM-4’s and, later, RM-5’s at their extremely low noise levels without the RAM Computer Tube Tester.
Those first RM-5’s were delivered in 1981 and are still in service. From my experience of repairing other brands of tube preamps, I would say the RM-5 is the most reliable in the field. Part of that reliability is due to its simplicity, part to construction and to its short circuit-proof power supply. As it turns out, many tube preamps fail at the power supply. These often fail at turn-on, thus giving false credence to the myth that it is better to leave tubes on. They also fail when an owner or technician slips with the voltmeter probe and briefly shorts the high voltage. After the smoke clears, the technician is faced with repairing the power supply before getting back to the original problem. This exercise is then repeated until the unit is repaired or the technician is carried off to the funny farm. I experienced this very problem fixing a popular preamp of the late 70’s which later went through several revisions disguising the fact that, in the original design, the pass transistor in the regulated power supply was over-stressed by charging the filter capacitors. That experience and others committed me to short circuit-proof power supplies. You should be aware that, to this day, there are products being designed and produced with power supplies that are failures waiting to happen. In terms of reliability, I cannot overstress the importance of a reliable power supply. If you have that, and reasonably competent design in the audio stages, your troubles will be limited to tubes alone.
This isn’t to say you’re in the clear of designers who choose inappropriate tubes for the application, (e.g., 12AT7’s and 12AU7’s for input tubes) or who operate tubes too hot or at too high voltages. The chapter on the design of your RM-10 will tell you how these were chosen.
By 1984 I had added power tube testing to the RAM Computer. The process is entirely different from testing for low noise preamp tubes. It requires extreme precision to achieve accurate matching. RAM Tube Works was the first to offer a two parameter match of both bias and transconductance. I realized that bias matching alone was worthless unless the matching was done at the precise voltages and currents of the specified amplifier in which the tubes were to be installed. That realization made conventional matching a joke. Even today, the concept of matching is one of the most misunderstood and abused added-value services in the tube industry. Yet bias and transconductance matching is the only way to get an output stage to perform optimally.
My interaction with these new power tube customers gave me a great deal of information about tube life in the various amplifiers and the attitudes of their makers. One manufacturer of a popular low-cost amplifier was quite satisfied with his 1,000 hour tube life. I had discussed lifetime with tube design engineers at Sylvania, GE and Gold Lion and found that the design life of 10,000 hours could easily be achieved by reasonable derating. The challenge was to design an amplifier with good sound at reduced tube dissipation and safe voltages. On the bench I saw what every amplifier designer before me has seen: crank up the tube current and the distortion goes down. Sadly, <i>the tube life goes down faster than the distortion</i>. This led to the challenge of the RM-9: to design a cool running amplifier (tubes at 1/2 rated dissipation) with good sound. Seven years later many of our customers are still enjoying their original tubes with full performance. The RM-9 was the first modern amplifier to employ B+ fusing to the plates and eliminate the screen resistors that often fail with shorted tubes. Tube shorts are a fact of life and the RM-9 has proven that smoking resistors are unnecessary. Proper fusing also saves tubes from damage due to intermittent lint-shorts which destroy unfused tubes.