Auto-Bias board on Roy's website?

pichacker wrote:How robust is the auto bias board to tube flashovers?

If a tube goes short, or draws high current for any reason, then there is a possibility of a high voltage being presented to the cathode path. On a stock version this fuses the 10 ohm resistor. On the auto bias board this would of course make its way to the sensing circuit that would normally only be looking for 0.6V. With a defective tube we would have little control using the grid bias voltage.

mazeeff wrote:The problem of what happens when the auto-bias circuit encounters a tube with a plate to cathode short can be easily tested by Roy and/or Pavel. With the power off, remove a output tube, and short the plate to cathode, with a large jumper. This would emulate a common problem associated with redplating, where the excess heat of redplating results in metal fatigue, and quite often a short. This type of failure would put 400-500 volts on the cathode, and would normally vaporize the 10 ohm resistor. With eye protection, and a fire suit, flip on the power and see what happens. If the silicon based current limiter and/or the 10 ohm resistor fails, we have a problem. If the circuit works as Roy described in post #79, and limits the current, then we are fine. The alternative, is for Roy or Pavel to disclose part numbers associated with the current limiting circuit. With that information, we could quickly determine if the circuit can withstand a 500v short. In the past 10 years, I have had three tubes fail with a internal short like this, and two times it took out the 10 ohm resistor. Based on reading the forum, a number of folks have blown a 10 ohm resistor, like myself. It is not a rare event. I have wondered, if Roy purposely designed the VTA with the 10 ohm resistor external, because it is easier to replace when it is soldered to a socket and ground. Without an answer to this, we will just have to wait until one of the early adopters encounters a tube failure resulting in a short!

PS: Please do not take my posts as a criticism, in any way. I am a old retired EE, who was taught the value of being a "devil's advocate" early on. We used to hold VERY tough design reviews, where my EE peers were encouraged to find faults in my designs. Saved my Bacon more than once!

mazeeff wrote:Pavel. Thank you for a solid answer. Four 10 ohm failures, seems to justify our concerns. Back to Corndog71's idea of a fuse. If the fuse has a consistent 1.24 ohms, why not use this value as the bias resistance? You could easily mount PCB based fuse holders to make replacement easier. If I understand it, the value of 10 ohms was chosen to simply make the math easier! Would 1.24 ohms work just as well? Of course, these fuses would need to be 1.24 +/- 1% or so (same tolerance as the resistor). Pavel's experiences seem to run counter to Roy's post #79.

jfine wrote:Anyone got some detail pics of this installed in the M125 yet?

tubes4hifi wrote:J FIne,
almost all hi-end modern tube amplifiers are now using auto-bias.  Since we don't have the schematics for either the Pavel design we (Tubes4HiFi) are using, and we don't have the schematics for the PrimaLuna amps,
there's no way to compare for sure.   This version could certainly be called "adaptive", because if you substitute one tube for another then it's still going to automatically adjust bias to the preset value.
This is still a "fixed bias" method, which is the best for highest power and lowest distortion.  It is NOT cathode bias.
We are still using the 10 ohm resistor to measure and compare bias, same as before.   A "cathode bias" method would use an appropriate resistor for the tube type (say maybe 200 ohm 10 watts)
for an EL34, along with a large 220-1000uF capacitor in parallel with it.  An amp like that has lower power and higher distortion.

jfine wrote:Roy/Pavel,

What happens with the M125 when a tube begins to fail, I understand the other 3 tubes would still run with correct bias,  but is there some audible difference, like maybe less volume, etc., that would alert you that the tube failed? It would seem you could run with 3 tubes and maybe not notice?