So I built up a little tubed unit to replace the stereo channels of a class D 3116 2.1 that I use for desktop dac/computer usage...retaining the sub amp which sounds ok.
The tube amp is a tubecad cascaded 6dj8 (b+ 245v, gain of abt 50)) Hybrid Amplifiers & MOSFETs in Series driving a trioded sweep tube (pl504) cathode follower biased with LM337 in SEPP topology (245 b+ biased abt 15w, apparently with 2x gain and 2x current boost)...another Broskie nugget Circlotron & Book Review. Parafeed into a 500r opt. I wasn't expecting much.
My question is this. How the heck is this thing puting out 24v peaks into the speaker load??? as evidenced by my meter and ears lol. The 100w class D sub amp driving a BIG efficient sub just barely keeps up.
How can the LM337 which only gets abt 32v from the sweep tubes cathode manage to get 24v music peaks through the cap and OPT which divides the ouput by abt 6?? And do it cleanly???
Broskie states the 337 topology should double that 32v to 64v which would deliver way less than 24v. Why isn't the 337 smoking? I hate to think the 337 is on the verge of smoke knowing that if it goes I'd lose the big sweep preety quickly. A slip of the probe fried one regulator already and the pentode started red plating within abt 20 seconds. Luckily I was right there to power down
I'm really happy with this amp as it fits into an ATX PSU case, has great sound even for movies with 1.6r output impedance measured at the speaker terminals, super low noise floor,...better than the 3116 ( I can hear decent cymbals again) and seems almost as loud!
Has anyone else built up Mr B's SEPP 337 arrangement?
View attachment 576580
I was thinking of adding some more turns to my mains tranny to boost voltage a little and increasing bias current (this tube seems to do OK at 25w) a little but it's not needed...not lacking any clean volume for my purposes. I like that I can roll in 6n1p's, 6n6p, 6n23p aand other ecc88's into two different positions...lots of fun with oven mitts on lol.
TM
The tube amp is a tubecad cascaded 6dj8 (b+ 245v, gain of abt 50)) Hybrid Amplifiers & MOSFETs in Series driving a trioded sweep tube (pl504) cathode follower biased with LM337 in SEPP topology (245 b+ biased abt 15w, apparently with 2x gain and 2x current boost)...another Broskie nugget Circlotron & Book Review. Parafeed into a 500r opt. I wasn't expecting much.
My question is this. How the heck is this thing puting out 24v peaks into the speaker load??? as evidenced by my meter and ears lol. The 100w class D sub amp driving a BIG efficient sub just barely keeps up.
How can the LM337 which only gets abt 32v from the sweep tubes cathode manage to get 24v music peaks through the cap and OPT which divides the ouput by abt 6?? And do it cleanly???
Broskie states the 337 topology should double that 32v to 64v which would deliver way less than 24v. Why isn't the 337 smoking? I hate to think the 337 is on the verge of smoke knowing that if it goes I'd lose the big sweep preety quickly. A slip of the probe fried one regulator already and the pentode started red plating within abt 20 seconds. Luckily I was right there to power down
I'm really happy with this amp as it fits into an ATX PSU case, has great sound even for movies with 1.6r output impedance measured at the speaker terminals, super low noise floor,...better than the 3116 ( I can hear decent cymbals again) and seems almost as loud!
Has anyone else built up Mr B's SEPP 337 arrangement?
View attachment 576580
I was thinking of adding some more turns to my mains tranny to boost voltage a little and increasing bias current (this tube seems to do OK at 25w) a little but it's not needed...not lacking any clean volume for my purposes. I like that I can roll in 6n1p's, 6n6p, 6n23p aand other ecc88's into two different positions...lots of fun with oven mitts on lol.
TM
> cathode follower biased with LM337 in SEPP ... ...Parafeed into a 500r opt.
There's dozens of pictures on that page and none show classic inductor parafeed. Show us what you actually built. (Your "Mr B's SEPP" attach seems invalid?)
> How the heck is this thing puting out 24v peaks into the speaker load???
Many voltmeters over-shoot on audio. You want a steady test signal or an oscilloscope to know what you are measuring.
"24V peaks" is meaningless without an impedance. It's only about a half-Watt if you saw that at the 500 Ohm side. It is the expected 7 Watts in 40 Ohms. Since common 4-16r speakers will be 40 Ohms at bass resonance, a low-damping rig might kick 24V peaks.
Of course if you are happy, enjoy, don't ask "How?"
There's dozens of pictures on that page and none show classic inductor parafeed. Show us what you actually built. (Your "Mr B's SEPP" attach seems invalid?)
> How the heck is this thing puting out 24v peaks into the speaker load???
Many voltmeters over-shoot on audio. You want a steady test signal or an oscilloscope to know what you are measuring.
"24V peaks" is meaningless without an impedance. It's only about a half-Watt if you saw that at the 500 Ohm side. It is the expected 7 Watts in 40 Ohms. Since common 4-16r speakers will be 40 Ohms at bass resonance, a low-damping rig might kick 24V peaks.
Of course if you are happy, enjoy, don't ask "How?"
It is happening in exactly the same way that I observed 70 volt peaks across my Yamaha NS-10M's coming from a SE KT-88 amp. The woofer has a resonant peak in its impedance around 80 Hz where it isn't 8 ohms, it's more like 25 ohms. Add in some extra EMF produced by the voice coil moving in a magnetic field, and you can have peaks that do not appear clipped that would exceed the expected clipping point of the amp, by a LOT!
Thank you t's actually a line transformer. I tried wiring as an inductor by series connecting the primary and secs out of phase but it was too much dc for the opt so it ended up being used parafeed. I made a little scribble of what I built...
Hey thanks for responding. That certainly makes sense. I'm still thinking of adding a shutdown mechanism to save the op valves in case the regulator poops out.
I am amazed how good the cathode follower output works, and the cheap opt 'sounds' more expensive.
Is it fair to say that it's easier to make a 1k opt put out square waves as apposed to a 10k?
TM
Interesting way to reduce gain from the second stage. However, the output impedance from that stage will be raised as well.
Also, where is the PL504's screen grid connected? Straight to B+?
Why is the grid leak resistor on the output tube only 47k? Won't that load down the second stage 6DJ8?
--
Also, where is the PL504's screen grid connected? Straight to B+?
Why is the grid leak resistor on the output tube only 47k? Won't that load down the second stage 6DJ8?
--
Last edited:
The screen grid is currently fed by 200r from b+ and locally decoupled with 10uf to ground. Its not my circuit so I cant tell you why B has 47k but Im using 470k....Ill try 47k at some point.
I should have mentioned that Ive got 2.2k and 270r on the second tubes cathode to bring up the gain and now have 2x 6.8r on the LM377s which is overdissing the pentode a little.
I also switched the input cathode r to a red led which I dont really like as well so will likely switch back to 180r. Havent tried ANY other feedback but may try a little at some point; prob dont need it.
With a simple PSU - cap multiplied with a hv NPN, two small resistors and 2x more smallish smoothing caps per channel its dang quiet! AC filter circuit (on a tiny board) from a broken microwave with x and y caps and a small torroid got rid of the hiss I was hearing. Triodes are dc (MC34063A step down easy circuit AND regulated = good for tube rolling) heated and the pentodes are 27vac from a modified torroid.
Every other tube circuit Ive made uses a dc step up (MC34063A +*mosfet) for B+. This one uses a nice big torroid but I cant tell the diff.. I think Ill keep using the step up in future projects when I run out of the box of torroids I just got.
This is the highest volt amp Ive worked with....self imposed limit...lol...by this time next year Ill hopefully have graduated to 400.
TM
Just measure the maximum output voltage before clipping into an 8 ohm resistor, from this you can calculate your maximum power. I suspect you are closer to 2W output, maybe even more.
Anyway you are having fun, learning and have something that sounds good to you.. All good as they say.
Anyway you are having fun, learning and have something that sounds good to you.. All good as they say.
It can definitely overdrive my Mark Audio chr 70s which dont get used since they hurt my ears (headache)...its the only amp I can tolerate with these speakers! With 93db speakers itès plenty powerful for my office room. Total cost was well under a hundred bux.
TM
TM
Last edited:
- Status
- Not open for further replies.