|
|
Paul, I bet you are wondering which circuit is best. Unfortunately, you will have to do the work to find out. It is doubly unfortunate that the manufacturer hadn't done in the first place, as what works best is to build all the circuits at once, which of course requires many chassis and parts. Then give all the line stages to one listener for a few days. (By the way, only one set of tubes would be given so that the tube variable could be held constant.) Next, the many line stages would go to another home for three days. After about five such trips, a tally of preferred units is taken. The winner (or winners) is then duplicated. So now there are two identical line stages. The next step is to review the criticisms of the winning unit and modifications are made to only one copy in an attempt to improve it in those areas where it found complaints. A second round of traveling to different homes begins, but this time for only one day. If the modification did the trick, the unmodified unit is then given the same modification. (If it failed, the modification is removed and another put in its place.) And a second round of modification is made to one unit for evaluation. And so on… It is much like mountain climbing: you do not move up until you have secured your present position. Even this just described elaborate method is truncated, as it should be applied to all the proposed circuits first and then a final shootout can be made.
The alternative method is the one most commonly used: only one unit undergoes many modifications and some progress is other times it is lost and other times a giant step sideways is taken. The longer the modifications took to implement, the poorer the comparison's reliability. Additionally, or rather subtractively, since all of the modifications were evaluated on just one system, the results are seldom universal. In other words, the unit was optimized to compensate for the fallings of just one system. This is good news if the system belongs to you, but bad news for others with different systems.
One quick test I recommend is that you wire a high wattage 20k resistor across the first power supply filter capacitor. This will test if the power transformer to see if it can sustain any added current draw. If the B+ voltage falls drastically or excess hum from the speakers or buzz from the power transformer is heard or the power transformer becomes excessively hot, it cannot handle the increased load. Adding a separate transformer inside the chassis for just the heaters will help unload the potted transformer.
|
|
|
|
Details not covered: the power supply diodes should be replaced with high-speed Hexfreds and the power supply dropping resistors values must be changed if a higher current draw is used. The heater power supply must be somehow reference to the main power supply and should not be left to float completely.
Bytheway, thanks for the kind words, but careful with the comparisons with demised, or rather transmuted, Glass Audio, as I have already been accused of contributing to its demise by one reader. Paul, you have got some work ahead of you. Good luck and keep us updated with your progress.
Subject: Headphone amp
First of all, thank you for publishing such very useful information for free. I really enjoy studying your web magazine, although I don't understand everything. This leads me to my question:
I consider building a headphone amp for my Sennheiser HD580. Your circuit in http://www.tubecad.com/july99/ is really interesting, but I don't know how to calculate the gain of the amp. For my HD580 I need a max, output voltage of 7.7V (300 Ohms and 200 mW max.). Can I calculate the input-stage gain with the same formula as used for the grounded cathode amplifier (with Rp=20K and Rk = 930 Ohms)? Can you give some me further explanation regarding this input stage?
I've also seen the circuit of the month in http://www.glass-ware.com/tubecircuits/Tube_Headphone_Amplifier.html. But I don't like the low input impedance. So, what do you think about removing the 150k resistor and instead grounding the grid with a 1 Meg? Thank you in advance for your help.
Daniel
Germany
I also own the HD580 and I have measured (on a scope) the music signal delivered to the headphones. Even at excruciating sound levels the signal remained under 2 volts. (I must reveal that I listen to music at a much lower volume than any other audiophile I have ever met.) In other words, I am not convinced that we need 200 mW. Of course, spare power is reassuring. Either of these two circuits will provide a healthy output into HD580s. Both circuits will work with these headphones. The gain of the first circuit is set by the roughly the same formula as the grounded
|
|
