• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

Ultra-Linear Midget Amp ?

You don't have to use UL mode "finals", if you regulate g2 B+. Full pentode or UL, some form of loop NFB is (IMO) indicated. Even horn speakers need some damping factor help.

FWIW, I'd avoid the small signal tubes and go with the 6BM8. About 2 WPC is what you get. Better to have power that's not used than have a speaker wrecking amp that's clipping most of the time. I've taken the liberty of uploading a 6BM8 design that's based on something Mullard did.

A "weird" option would be 3X 3V4s running full pentode PSE. If you regulate g2 B+ at about 70 V., cranking the anode B+ voltage up to 110 V. should be safe enough, provided you watch the anode dissipation carefully. Each 3V4 requires its own "floating" filament supply of 2X alkaline "D" cells, which should last a reasonable amount of time. Also, "fixed" bias, with individual trim pots., is needed or "current hogging" will be a problem.
 

Attachments

  • dcsedx2_fixed.gif
    dcsedx2_fixed.gif
    44.4 KB · Views: 471
  • dcsepartsdw1.png
    dcsepartsdw1.png
    221.1 KB · Views: 463
For the users of horns or other high efficiency speakers who build amplifiers with 700 or 900 miliWatts maximum power; would a small signal pentode running Ultra-Linear make a musical amplifier?

Why not just use a triode connected 6E5P ? That would hit your power requirements . Use a 3k-5k output transformer , 330 ohm grid stopper , 220k grid leak and 3 x schottky diodes in the cathode circuit for the entire power stage . I found the 6E5P to make a better sounding spud amp than the 6S45P , trouble is a slight lack of gain with my 97dB co-axes but should be fine with horns

316a
 
Hi Eli! It is good to hear from you again.
 
I see several of those caps are electrolytic. Are any of them in a position where a change to film & foil, or metalized plastic would improve the sound?
 
I understand little of how Ultra-Linear operation improves the sound. When you say:
"You don't have to use UL mode "finals", if you regulate g2 B+. Full pentode or UL, some form of loop NFB is (IMO) indicated. Even horn speakers need some damping factor help."
Is the idea that U-L increases damping factor over triode operation?
Does voltage regulation on grid2 B+ increase this damping factor?

Brian
 
Is the idea that U-L increases damping factor over triode operation?

No, UL operation increases damping factor over pentode operation, but is still not high enough to be used without negative feedback in most cases.

Of the triode, UL or pentode choices, triode offers the highest damping factor before negative feedback ('open loop'), but lowest power output.

A triode output tube may have high enough damping factor to allow its use without negative feedback, and it can sound nice that way (although adding negative feedback will lower distortion and increase damping factor).


Does voltage regulation on grid2 B+ increase this damping factor?

Not that I'm aware of. Regulating the screen grid supply improves distortion and overall performance of a pentode, but doesn't reduce its high plate resistance.
--
 
Last edited:
"Better sound" is a subjective term. This forum is the only place on earth where I've encountered folks with sufficient maturity and restraint to avoid discussions of subjective phenomena.

Objectively, UL operation yields a sort of middle ground between pentode and triode operation of the same tube. It was developed at a time when 'powerdrive' wasn't practical in consumer audio equipment. The primary advantage of UL is that it yields nearly as much power output as pentode operation with conventional drive circuits, but with reduced distortion figures. My opinion is that triode operation with powerdrive often makes more sense now that it's easy to accomplish.

My advice is to go ahead and build something. Don't wait. Listen, study, and experiment. Don't waste money on boutique components -- just buy stuff that's adequate and put in the effort to make mistakes and learn from them. High-end components are useless without good circuit design, and sometimes even more useless WITH good circuit design. I'll get down off my soapbox now. :D
 

Shoog

Member
2002-08-15 10:16 pm
Eire
Built a nice little UL ECL82 amp with about 1 1/2W of output. Took feedback from the plate of the pentode to the cathode of the triode. Worked out really great and sounded fantastic considering it was using some salvaged console transformers for the outputs.

So yes it can be done and produce excellent results.

Shoog
 
Last edited by a moderator:
I regret having mentioned 417 & EF86. I thought it would be helpful to give an example of what I was thinking of with this project. I did not mean to give the impression that I had a preference for those two tubes in this particular application.

Thank you, Bas Horneman and Shoog for the replies.
 
Shoog, Your report of building a low power U-L amp and having it "sound fantastic" is most encouraging.
 
Brian
 
Last edited by a moderator:
I'm intrigued by plate-grid local feedback around the output tube.

I'd like to try a single-ended amp that yields about 5W output into 4 ohms. I was looking at the RH84 recently, threw it into LTspice. It seems the 12AT7 is run at a really strange place, with only a 22k plate load resistor and 240R cathode load resistor (unbypassed). The 12AT7 thus runs very hot (5mA) and with a lot of distortion. The RH Universal with 12AX7 is even worse, running the 12AX7 with a tiny 360 ohm cathode resistor (again unbypassed), therefore with only about 0.7V grid-cathode voltage (grid current?). It doesn't look like a competently designed circuit, but I may very well be missing subtleties that are beyond my understanding.

Is there a good, recommendable, but simple, SE amp design using plate-grid feedback? Is there an improved version of the RH84 extant? Or should I mimic George's (Tubelab) SSE with its cathode feedback?

thanks, and sorry if this is off-topic.
--
 
I'm talking about the original RH84 (cause that is the one I built and listened to). The cathode resistor is 680ohm.

That's actually an "RH88." Did you build that with a KT88 output? That would have much higher B+, maybe 450V, which would certainly help the 12AT7 driver work better.

Oh jeez. I think I confused ECC81 with ECC83, and wasted my time in spice last night. As Emily Liltella (Saturday Night Live) used to say... "Never mind."

Now, how to scale this down to a +320V supply without draining all the oomph out of the input driver. In the RH84 for EL84 output, the 22k plate load is too heavy a load on the 12AT7. Maybe something like a 6J9P in triode would work better. But 'Schade' plate-grid feedback works against the rp of the driver tube, and 6J9P-triode will have much lower rp than 12AT7/ECC81. Perhaps switch to 12AX7/ECC83 with 100k plate load resistor? or better yet, a 6AU6A pentode?

--
 

Yeah, it doesn't help that I put in the wrong driver tube!
:headbash:

I didn't mean to bash the RH design. I've seen the controversy, but there are too many satisfied DIYers with RH amps out there to automatically dismiss it.

I've never had an SE amp, and I don't fully understand 'Schade' feedback. I figured I could kill two birds with one stone by deciding on an SE design with Schade feedback. If I hate it, I could always switch to the time-honored SE with DHT output.

--
 
Last edited:
That's pretty cool! The problem for me would be finding and buying the correct transformers to use as the OPTs. I'd also have to source the output tubes, since I don't have any of the ones you used. I have some 6AU6A in my stock.

For the OPT, is this the kind of thing I'd be looking for?
50VA 50W Toroid Transformer Primary 0 115V 230V Secondary 6V 0 6V | eBay

I'd like to start a new thread, about creating an 'RH Slayer' with more common tubes. Would you be OK with that?
--

PS - Actually, that transformer has a 6-0-6 V secondary, which is probably different from the one you used... maybe.
 
Last edited: