Feb. 12, 1994

Dear John,

I was torn between writing a "letter to the editor" or to you personally. Upon reading the staff page I see that you are one and the same. I read your magazine religiously always starting with the "Letters" section. It is a window to the public that I enjoy. I admire your publishing the good, the bad, and the ugly letters that come your way. Include this one if you like.

I enjoyed your introduction to the YBA Amplifier in the January issue. My Maserati is turbo charged and there is indeed a throttle lag. That is not the least of my problems as I rarely get to drive it other than to visit my mechanic to have something else replaced. I am studying its unreliability much as I have studied the unreliability of many amplifiers that are often compared to high-performance race cars. So far, my theory is that neither inherently must be unreliable to perform. What the Maserati lacks is the engineers desire to make it reliable. Given that Maserati is a racing marque and that racing engines are torn down after every race I am not surprised at the frequency at which parts must be replaced. The Maserati Club which I joined to obtain parts and service tips recommends replacing the timing belt at 15,000 miles. While you are at it might as well do the water pump (which rides on the timing belt) and the tensioner bearing. I learned the hard way after having a belt fail at 26,000 miles (then the recommended interval was 30,000) and soon after having the water pump start leaking. I must say that my driving pleasure is decreased after experiencing a timing belt failure accompanied by 18 bent valves and a major head job. Had I known all this ahead of time I probably would have kept my Honda Accord, a car obviously designed with reliability as a high objective. I also see that Honda has recently combined reliability and performance.

So now that you have read my car story, I will talk amplifiers. First, $5.00 worry free, tight tolerance premium resistors are not all they are cracked up to be. Ask our friend Kavi Alexander of Water Lilly Acoustics who put the premium Caddock resistors in his microphone amps only to have me take them out because they were noisier than the cheap carbon films that Tim De Paravicini admitted to Kavi he installed "in a pinch". I replaced them with $0.08 Roderstein, the best thing going. We are lucky that Roderstein cares about noise where many manufacturers do not. I have tested many brands of premium resistors and talked with their makers who agree that lowest noise is not their main consideration. Certainly, they provide and maintain their high tolerance over time and temperature. That is what you are paying $5.00 for. An experienced engineer knows that in the face of transistors that vary over 200% in some parameters and tubes that vary almost as much that a 5% resistor is far from "sloppy". In a feedback loop where it sets the gain of the amplifier a 1% resistor will give 0.1 dB accuracy, but tube and transistor variations will often reduce that accuracy to 0.5 dB. In the "old days" you paid for resistor accuracy over 10%. Now, manufacturing has gotten so good that a 1% resistor is as inexpensive as anything.

I have also seen quality metal film resistors go way up in value at a fraction of their voltage and dissipation rating in less than one year where a carbon composition resistor (God forbid) will last for years. My theory is that to make anything over a few hundred Kohm in metal requires such a thin film that it is susceptible to erosion and internal arcing. A nice rod of high resistance carbon is not. I think it is not attending to detail to off-handedly use $5.00 resistors. In fact, it simply justifies the higher prices of products which disturbs Jack English in his R.I.P. piece. Too many of todays high-priced products are priced at the total of a grocery list of the "hottest parts". Is this "Engineering"?

I am glad you find tubes seductive, I do also. Furthermore, I find them as reliable and friendly as transistors when well applied. I have had extended talks with the tube engineers at Sylvania and General Electric and learned the following which I have confirmed in my own experience:

1. Receiving tubes, the kind we use in audio, were designed by GE and Sylvania for 10,000-hour life and 10,000 turn-on cycles. Their application was in radios and Color TVs where the turn-on transients were not unlike a modern amplifier. In my RM-9 I have not had one cathode or filament failure in well-made tubes (early Chinese tubes excepted due to poor workmanship which has of recent become quite good). I do not leave my amplifier on (except by mistake) and cycle it 3 to 6 times a day. Since tubes are good for as many hours as turn-ons any usage over an hour puts the eventual failure into the hours camp.

2. All this business about delaying the B+ (high voltage) is applicable to transmitter not home audio. In a transmitter, where the B+ is over 10 KV, you do run a danger of cathode damage. In an audio amplifier with a B+ of 500V it is not a problem.

3. The only application in which leaving tubes on seems to make a difference is in phono front ends where the tubes get noisy for reasons that even the manufacturers I talked with could not explain. There I have found it almost a toss-up between cycling and not.

I would not award the smoke test of the YBA to Murphy. I have designed a few transistor amps and know that it is not easy for transistors to tolerate full level output swings at high frequencies. It is often called "sticking" when the output transistors cannot turn off fast enough. The effect of both transistors conducting right across the power supply is shorted transistors and smoke. It has little to do with the load resistance other than a low value causes harder turn on which results in slower turn off. Most of us learned how to fix that in the late ‘70's. I would say that the amplifier indeed was at fault and that these conditions can occur in the wild. Where are his rail fuses? They are certainly sonically less objectionable than an output fuse that is after the feedback point. Is not this an important aspect of the amplifier to be brought to the attention of the public and the manufacturer?

Indeed there "was not a lot [you] could do." That is the nature of transistor equipment. I think most vacuum devices would have excused your test. At least they would have warned you by glowing red. Had you missed that warning replacement takes only a moment with a hot mitt.

Lastly, the output inductor that most reviewers disdain is probably the most important part in an amplifier and should not be omitted. It is virtually impossible to make an amplifier stable into all capacitive loads without one. Besides, why damn a few turns of heavy wire when it is a feature of many “audiophile” cables these days.

In closing, I think it is great that there are heaps of us out there making amplifiers and heaps of you reviewing them. We will probably never make the perfect sounding amplifier, nor the perfect ice cream flavor, or anything else where taste is involved. However, I think we can all do a bit better in making equipment more stress free in both initial and long-term costs.

Sincerely,

Roger A. Modjeski