• 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.

help me decide,headphone amp - parafeed or normal tx ?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I am building and amp for friend who has large budget.

He wants a valve based headamp. To my mind that conjures a couple of good options

1) CSS or choke load parafeed with sowter transformer (Sowter Type 8650 Headphone Output transformer)

2) Single ended with sowter gapped transformer (SOWTER HEADPHONE TRANSFORMERS) - the 8968

3) CSS or choke load parafeed with a magnequest autoformer (TL-404)

Both look like they will work well with a myriad of tubes (d3a, 6h30p...the list goes on, i could even use DHTs).

My question is, given a very large budget, which would you build and why. I havent heard a parafeed amp, but sites such as epc.cc seem to like them very much. I also asked the question on head-fi a while ago, but its crunch time and i need to start building. He will be using them predominantly with HD650's, and very occasionally with grado's.

Has anyone done both with relatively highend iron, any comparisons?
 
Last edited:
My personal suggestion is one of Gary Pimm's self-bias CCS loads for a 5687 tube, with parafeed magnequest, electraprint, or lundahl magnetics. Something in the 4:1 or 5:1 ratio range. See the Espressivo Espressivo Headphone Amplifier but the modern self-bias CCS no longer needs a pentode at these power levels.

If you really have the money to burn, build exactly this http://www.nutshellhifi.com/Raven-MarkII.gif
 
Thanks gents, you have made my mind up.

I am going to go with lundahl magnetics, 6h30p.. led bias, dact attenuator, dn2450 cascode ccs.

Two chassis build to avoid hum, PS tube rectifier, clcl - probably lundal chokes, obbligato film caps.

or to simplify things, i may ask jack (electraprint) to do all the magnetics - people speak very highly of his transformers.
 
Last edited:
Jack's stuff is great, especially getting custom power trans for exactly what you need. He doesn't seem to be a big fan of choke-input supplies, and when I've asked him about using his chokes as choke-input, he usually seems wary. I've successfully used the large hammond enclosed chokes as choke-input as long as you run the current way under the rating. (~120ma in my 300ma choke). No mechanical hum in mine but the magnetic field is colossal, keep those OPT's far, far away from choke input supplies.

Or alternatively, hook up the supply w/ a dummy load and put a scope or FFT software onto the disconnected output transformers, you can then twist/move the transformers around possible mounting spots and find the dead zones where the field lines are all at right angles and the hum is minimum. I found I can mount my opt's about 12" in front of the choke, if they are perfectly in line vertically, and horizontally about 3" to either side, the hum is extremely low (measurable but not audible) and I didn't need a separate chassis.
 
Thanks gents, you have made my mind up.

Pitty. IMHO "tube sound" is as much of a "tranformer sound" as it is that of the tubes so the output - if buyer desires the "tube sound" - should incorporate both of these, and nothing else.

Parafeed = less OPT sound and more capacitor sound which probably isn't what the buyer is looking for.

Somebody wit large budget might also be interested in tube amplifier because of its "bling" value rather than actual sound so CCS (and other 3+ legged sandy components in general) won't be a big selling point for him.

My choice would therefore be a normal (SE) OPT in straightforward configuration suitable for speakers, driving higher Z load that is headphones. Alternatively a load-matched OPT could be used, but this will provide waaaay too much output (speaker OPT into high-Z headphone load = awesome damping factor and enough swing regardless as less power is required).

With generous budget you mentioned you might as well build 2 or 3 architectural alternatives and have him keep the one he likes the most ... who knows, it just might be the CCS load parafeed variant suggested above.
 
If it's just the ultimate sound quality you want, I'd use a 2a3 output - low anode resistance and wonderful sound. And for the ultimate in performance feed it direct coupled from a 26 in filament bias. I don't expect you to build this, but I'm going to!! I have AKG K-701s which are 50 ohms, and I'm scratching my head about what SE output transformer to use, or indeed if I should use parafeed. Parafeed maybe even with a toroidal mains transformer with the right step-down ratio. Still at the pondering stage but it wouldn't be hard to knock together - take a day I imagine since I already have power supplies and filament supplies in separate cases. Useful to be modular!

andy
 
Except that in a normal SE output you still have the capacitor, it's just the final power supply cap instead of the parafeed cap.

Unless you use some 3 legged sandy components and make a regulated supply.
But then you just swap the capacitor for the regulated supply - i.e. swapping a simple and improvable component for a complex circuit!
 
But then you just swap the capacitor for the regulated supply - i.e. swapping a simple and improvable component for a complex circuit!

That was kinda my point. Parafeed w/ an active load (or a really good plate choke I suppose) isolates the supply. Yes you have a capacitor but it is just one, and it's relatively small. SE connection requires either supply regulation or accepting that the final power capacitor (and other parts of the supply depending on impedances) is part of the audio path. I'm not saying one is guaranteed better than the other just that you have to consider the current paths. It is not trivial to isolate them completely.
 
swings and roundabouts it seems.

He is coming from a darkvoice OTL, which is coupled through an electroltyic, so either will (or should) be a step up from that.

it should be a fairly straight forward build. I have already built up some CCS on perf board, the rest is pretty basic P2P and the supply is simple. I'll see how AC behave on the filaments but may need to go DC regulated.

I'll probably draw the circuit tonight and post it up for critique
 
I've just built a simple SRPP, one lytic as the output cap (220uF), no transformer, and had excellent results from 6N6P run at about 18mA.

Even with 32ohm mp3 phones, earpiercing volume, sounds great.

Hexfreds- CRCRC psu, bypassed with 0.47uf PIO.
Allen Bradley resistors 150 ohm lower cathode resistor. upper cathode resistor split - output between 100 and 50 ohm resistor - Broskie claims it helps to reduce the impedance load from the phones.

lytic bypassed also with some small films.

That's about it.
 
I've just built a simple SRPP, one lytic as the output cap (220uF), no transformer, and had excellent results from 6N6P run at about 18mA.

Even with 32ohm mp3 phones, earpiercing volume, sounds great.

Hexfreds- CRCRC psu, bypassed with 0.47uf PIO.
Allen Bradley resistors 150 ohm lower cathode resistor. upper cathode resistor split - output between 100 and 50 ohm resistor - Broskie claims it helps to reduce the impedance load from the phones.

lytic bypassed also with some small films.

That's about it.

I have built similar prototypes. They sound good, but are ultimately at the mercy of the big cap at the end - this is where the transformer should make the difference. With the tx, i wont need the output impedance benfits of the SRPP, so keeping it simple a single triode (maybe parallel) will be used.

I may well be able to get away with a simple crcrc supply, the CSS should provide enough PSRR to get away with skimping on the supply with no noticeable degrading of the sound. Skimping is not in the design brief though.. so try and do optimal everything.
 
I am building a headamp similar but chose SE gapped transformer using Salas HV shunt. This basically gives you the advantage of parafeed but with practically no caps in the signal path. Its something new anyway.

Also just forget DHT unless you want to pay for the tent supplies.
 
I have built both para-feed and typical SET with transformer output.
They all sound good to me.
First I am with others here the headphones sit on your ears there is no place for hum or buzz. I use and enjoy the Tubecad.com Solo power supply for the B+ and regulated heater supply. Dead quiet is worth the money spent. Regulated power supplies are highly recommended.
Second the headphone amplifier should be designed to work with your headphones to be used. Anything else is a compromise. Match the output impedance to the needs of the headphones and preference of the user.
The parafeed can have large output capacitors in the chain. You can end up paying more for the capacitors than the transformers. However parafeed can get lower output impedance if that is a need.
I ended up preferring good old SET’s with custom output transformers. First I had Edcor make what they labeled GXSE3-300-5K. With 6BQ6GTB’s these sound very good to me. Then I had Electra-Print make a pair of 8000K : 300 ohm 3 watt center tapped custom transformers. The balanced output into HD300’s really rocks.
Much more work and hard to get 307A’s with the same transformers above are really cool.
The PCB heater power supplies for the DHT 307’s sold by Pete Millett really work well.
DT
Just for fun!
 
here is a first draft... anything wrong here?

ps caps will be film (probably obbligato oils), parafeed cap will be something very nice.

An externally hosted image should be here but it was not working when we last tested it.


What do you think to the operating point of the d3a, and the transformer spec (going on whats available from sowter, so if something else would be better please say).
 
Last edited:
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.