I’m now back in business. Building a new 4P1L PSE output stage so will reuse the 300B/4P1L Flexible PSU. I never managed to post an update on the troubleshooting I had to do to get this HT PSU to its optimal state.
The output voltage was lower than expected and the 50Hz component extremely high. Something was wrong. So I traced the issue down. I found a bad solder in one of the rectifier’s cathode. The supply was operating in half-wave mode.
The supply is choke-input with 6AU7 rectifiers (hybrid bridge with FRED rectifiers). The transformer is custom made and has multi-taps for 300-400 and 500V. The tuning capacitor for the choke input is 470μF, then choke is 2.5H into 50uF oil cap. The filtering stage per channel is 20H + 100μF Oil caps;
Here’s a test of the supply at 330V/60mA per channel. It’s very rewarding now to see no 50Hz component and that the ripple at 100Hz is just 4mV (ignore the mA typo on the image):
Hi,
i got inspired by your concept and am now building an amplifier along the lines of your 2p29l-4p1l PSE idea. I originally planned for a PSE with 6E9P`s, but the lure of DHT…
I do have a set of very fine OPT that i want to use, 4k:8 so i will chose a different bias probably, from simulations 300V 32mA looked fine. I am planning to use fixed bias for the 4p1l, i don`t want the dissipation of the filament bias and the distortion should also be better.
On the driver side 2P29L i want to test an idea i had some time ago: filament bias with a chain of LED, as the dissipation <1.5W easily permits some experimentation.
thanks for sharing your ideas, greetings Harry
Hi Harald,
Not much gain is needed to drive the 4P1L PSE output stage. So a 2P29L is ideal given the low mu. Yes, I’d run the 4P1L at 30mA/300V. They work fine at 300V.
I’m implementing filament bias now as will reuse existing supplies from my previous 814SE Amp. I’m also testing Rod’s new bias regulator before he releases to the market. Looks promising and probably something to consider for fixed bias.
Regarding filament bias vs LED, you need to consider the amount of current (150mA) as well as the 10V or so of voltage drop. Need high current LEDs as well as several of them. A nice 100R wire wound resistor can work just fine in filament bias, so worth comparing the options first.
Cheers
Ale
Yes, Rod told me about his bias regulator when i ordered the the filament regulators from him. Sounds interesting, i agree. I have started on a shunt-regulated version, but am generally interested in his ideas. Rod has many interesting insights that is shure.
I will keep you informed about my build and it would be nice to exchange experiences.
Greetings, Harry