mwhouston wrote:
Matt: What cap brands and types do you favour?
Actually, I tend to be rather pedestrian in my choice of most components; sticking with what has generally worked for me in the past without issues. I tend to focus on technology for the given application and then go with a reputable quality manufacturer.
Case in point, for signal pass elements at audio frequencies (like interstage coupling caps) I like quality polypropylene film capacitors. These are generally very stable, have good tolerances, good aging characteristics, high voltage tolerance, are in many cases self healing, and exhibit graceful degradation at end of life. So this is what I'd use.
As for manufactures, I really don't want to start a fight. Everyone has their favorite and prices are all over the map. In general, the wilder the marketing claims, the less likely I am to buy. I know there are people out there paying big bucks for special paper and aluminum foil capacitors rolled on the thighs of Japanese virgins by the light of the full moon... but I'm not one of them. I also try to stay away from commodity manufactures in central and east Asia because I can't be guaranteed of their performance. I've had very good luck sticking with companies like IC or Sprague (Vishay). (Like I said... pedestrian). A tremendous number of the supposedly "Audio" capacitors are either "manufacturer of the month" with good marketing, or rebranded major brands at 1000% to 100000% markup. Beware of anyone that want's more than a few dollars for a signal capacitor. You're likely paying for marketing rather than quality.
Here's a good tip. Any quality reputable parts manufacturer will have a technical support line you can call and actually talk to an Engineer. A parts Engineer will talk your ear off concerning the minute details of part construction, materials, processes, etc. If you get a "that's proprietary" response every time you ask a question, that's a red flag that they don't actually make their own parts. Buyer Beware.
I know I haven't really answered your question, but it's about the best I can do. I do have to say that I am always amazed that the same people who will extol the virtues of a $100 capacitor will, in general, pull a random schematic off the internet and begin changing things without a thorough analysis. I have always believed that proper design and analysis comes first and (sensible and prudent) component selection second.
Well, now I've done it. Let the firefight begin!
mwhouston wrote:
While playing the amp for about 10 hours the other day I notices a slight resonance or distortion on some notes. It ocured mainly on low notes of the female voice. The easiest change for me to make was to change the preamp tube. At the time I think it may have been a 12AT or Y7 in the amp (markings are gone). I replaced it with a NOS Toshiba 12AU7. I found the 12AU7 sounded one of the better tubes in both the 4S MKIII and 4S MKIV Universal. Sure enough play the same tracks again resonance has gone.
I was think about your topology and it may be that you are seeing mild bias excursions that tended to resonate at a particular frequency.
The basic amp circuit first stage gain is about 17.6 and can handle an input voltage of approximately 3.5 volts peak before going into grid conduction. The power stage is however biased at -34v so it will go into grid conduction any time the input voltage to the first stage goes beyond about 1.9v peak.
With a 12AU7 in the "universal" topology, the gain is about 20dBv (unbypassed). With the volume control at half volume it's abut a -20dBv loss so the result is unity gain. With a 12AT7 for example, the gain is about 26dBv so with the same volume setting you have a gain of ~6dBv (2.0 v/v). So dependent on your input voltage and dynamic range, you could be pushing the power stage into grid conduction and bias excursions. Of course, all of this assumes that the power stage is running at full power (~2.1W).
If you were not at near full power, then it could have also been due to the preamp stage going into grid conduction. With the 12AT7 this would occur at about 2.7v peak input, but the 12AU7 would require about 3.2v peak input. This may not seem like much of a difference, but if you were operating on the edge (with just some dynamic peaks popping over the limit), you could see it.
Or it could also just be that the 12AU7 has only about 30% of the distortion of the 12AT7 in the universal design.
Or it could be that, as you say, the 12AU7 was just a better tube.
This is what makes tube rolling so addictive.

B.T.W. watch the gain on that 12AX7. With a gain of 29dBv (~28v/v) it will be really easy to overdrive the 6EM7 input stage. (~0.07v peak minus the volume setting. e.g. 0.7v peak with the volume control at 1/2 volume.)
In any event, it's good to hear that you're enjoying the amp. I'm very glad that the design worked out so well. Have fun.