SE VT Headphone Amplifier

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Attached are a couple of pictures of the first prototype of a headphone amplifier using inexpensive vertical oscillator/amplifier dual triodes. I am going for a parafeed amp first, using some small power toroids for output XFMRs, as that's what I have on hand. I'l be scouting out options for legit parafeed transformers, but the troids should be enough to make a noise at least.This amp could also be run as a simple SE amp with gapped output XFMRs.
 

Attachments

  • Img_2191.jpg
    Img_2191.jpg
    169.7 KB · Views: 1,080
  • Img_2192.jpg
    Img_2192.jpg
    225.8 KB · Views: 1,061
As I described in my first post here, I'm using a pair of small toroidal power transformers for parafeed output transformers for now - you can see them in the picture.
The amp is touchy as to input level with the feedback loop open because of the excess gain, but the sound is surprisingly nice driving a pair of Grado SR225E headphones.
There is no lack of bass, and the control is reasonably tight, even for an open loop SE amp. The fun will begin when I close the feedback loop, as I have no real idea what the cheap output transformers will do when included inside the global loop. It's nice to have a gain-phase analyzer at work...
 

Attachments

  • Img_2259.jpg
    Img_2259.jpg
    311.1 KB · Views: 846
Well, I ran gain-phase plots on the system with loop closed and no compensation cap, and wonder of wonders, the amp did not oscillate. The phase margin at ~30 degrees leaves a lot of something to be desired, but I was expecting a lot worse. It looks like a 10uF output coupling cap is more appropriate for the headphone impedances and transformers I'll be dealing with.

Attached are gain - phase plots for the amp with toroidal power output transformers, 3uF coupling cap, 51 ohm resistive load, and no compensation. Attached is one more plot with a 10uF output coupling cap instead of 3 uF.

Instead of rolling off the HF gain to get better phase margin with the crappy output transformers, I'll be looking for a better transformer.
 

Attachments

  • VTHP_3u_out_G-P_50_ohm_A.jpg
    VTHP_3u_out_G-P_50_ohm_A.jpg
    175 KB · Views: 792
  • VTHP_3u_out_G-P_50_ohm_B.jpg
    VTHP_3u_out_G-P_50_ohm_B.jpg
    174 KB · Views: 745
  • VTHP_10u_out_G-P_50_ohm_A.jpg
    VTHP_10u_out_G-P_50_ohm_A.jpg
    179.1 KB · Views: 123
Transcendar has agreed to do a gapless version of their headphone output transformer that should work ok in parafeed mode. I'll be ordering those and some gapped transformers for yet another VT headphone design, and the normal SE version of this amp. I forsee interesting times ahead...
 
My transformers arrived today, one gapped pair, and one ungapped pair. The gappped pair is Transcendar's standard TT-106-OT intended for headphone use. The other pair is an ungapped version of the same XFMR. I'll be putting them to work this weekend in the parafeed amp described here, and an alternate board stuffing that configures the amp as a standard (albeit tiny) 2-stage single-ended amp. I'm interested to see the effect a properly made transformer will have on the gain-phase characteristics.
 
OK - here's the closed loop gain-phase using the Transcendar parafeed transformer with no compensation cap. The response is more, shall we say, interesting than the toroidal power transformers used as outputs, and there is reasonable phase margin at crossover. The peaking is a bit troubling, and will probably require a visit from Mr. Zobel - more on that later.

I'm eager to try the standard gapped single ended transformers with the alternate version of this amp. It may be a winner.
 

Attachments

  • Trans_Parafeed_G-P.jpg
    Trans_Parafeed_G-P.jpg
    162.6 KB · Views: 423
Several tries with a Zobel network on the primary side did not really damp out the 100 kHz resonance and phase dip. I will try a reasonably heavy network on the secondary side to see if it helps. after all, that's where I'm picking up my feedback. The other thing I want to fix is the amplitude rise near 1MHz, as there is not a whole lot of gain margin, and one could envision a situation where the amplitude would cross back over 0dB again - conditional stability, anyone?
 
Last edited:
This weekend, I tried a secondary Zobel network on the outputs of both my parafeed headphone amp and and its simple single-ended alternate stuffing. I ended up using 100nF + 47 ohms and some mild compensation of 3.3n + 330 ohms across the feedback resistor. Results are shown below. I think this is about as good as I can get using the inexpensive Transcendar output transformers, unless I really want to cut down the amplifier bandwidth. I may talk to Electra-print to see what they can do for me later on. Right now, though I'll declare a conditional victory and move on to other projects.

The first G-P plot shown is for the parafeed design, the second is for its single ended sibling. They look pretty similar, as they both use essentially the same transformer, the parafeed transformer being simply a gapless version of the standard Transcendar single-ended design.
 

Attachments

  • para082419_100n_47R_zobel_3n3_330R_comp.jpg
    para082419_100n_47R_zobel_3n3_330R_comp.jpg
    180.8 KB · Views: 359
  • sse082419_100n_47R_zobel_3n3_330R_comp.jpg
    sse082419_100n_47R_zobel_3n3_330R_comp.jpg
    181.8 KB · Views: 348
Last edited:
The single-ended version of this amplifier (vs. the parafeed) got boxed and took a trip to Burning Amp 2019. Several people auditioned it and a solid state class A headphone amp I also brought (see the thread Simple Single - Ended Headphone amplifier in this section). Listeners were divided ~50-50 in preference between the two amplifiers.
 
Here's a schematic for 1/2 of the single ended VT headphone amp that made its way to Burning Amp last Sunday. It's all Rolls-Royce with current source loaded first stage and LED-biased 2nd stage. You could do a knocked-down version of this with resistive and RC biasing, but it probably wouldn't sound the same. I plan to give it some better output iron for more bandwidth and (hopefully) better sound.
 

Attachments

  • SE_HP.gif
    SE_HP.gif
    26.5 KB · Views: 360
Here's a schematic for 1/2 of the single ended VT headphone amp that made its way to Burning Amp last Sunday. It's all Rolls-Royce with current source loaded first stage and LED-biased 2nd stage. You could do a knocked-down version of this with resistive and RC biasing, but it probably wouldn't sound the same. I plan to give it some better output iron for more bandwidth and (hopefully) better sound.

WOW! Congratulations on a Rolls-Royce amplifier with current source loaded first stage and a blah...blah 2nd stage.
What exactly is the point you're trying to make for this "VT headphone amp that made its way to Burning Amp last Sunday but needs some better output iron for more bandwidth and sound"??:rolleyes:

Why don't you share a design that'll simply "blow peoples socks off" or just post nothing and close these threads?
Nobody cares about mediocre sounding headphone amplifier design since the market is already flooded with them.
 
Mercedes -

The amp auditioned well at BA - not mediocre at all. My point is that it could be even better with better output iron.

As evidenced by the gain-phase plots shown, I was able to damp out most of the warts of the Transcendar output iron with compensation sleight-of-hand and achieve decent gain and phase margin. I did the best I could with what I had at hand, in the time I had available. The gain-phase analyzer I have at my work place is a valuable tool, showing things that wouldn't be necessarily visible from just a square wave response. A good square wave response is a necessary, but not sufficient indicator of decent behavior in the frequency domain.

I'll be talking to Cinemag about output iron possibilities, as they have some capabilities I didn't suspect until recently. I may also be talking to Electra-Print. The object will be extended bandwidth, but still with adequate gain and phase margin.

Obviously you don't care about what happens in this thread, so feel free to ignore it while I press on and develop this design.
 
Mercedes -

Obviously you don't care about what happens in this thread, so feel free to ignore it while I press on and develop this design.

Yes, you're exactly right. I do not care about what happens in your ongoing headphone amplifier design thread that when finalized will probably still NOT compare with many commercial designs such as the great sounding and inexpensive Magni 3.

And yes I'll feel free to ignore it as many others should here should do as well.

Perhaps someone here will gain some knowledge and decide to build this so-called "Rolls-Royce" amplifier.:)
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.