What output Z for headphone amp? - Page 2 - diyAudio
Go Back   Home > Forums > Amplifiers > Headphone Systems

Headphone Systems Everything to do with Headphones

Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Reply
 
Thread Tools Search this Thread
Old 6th August 2003, 11:13 PM   #11
PRR is offline PRR  United States
diyAudio Member
 
PRR's Avatar
 
Join Date: Jun 2003
Location: USA
> the maximum amount of relatively undistorted voltage swing is produced when the anodes are loaded by the proper impedance. If 50Vrms is available at the primary, for 100mW total output a current of 2mA will flow (Z=25k). This turns into ~1.79Vrms and 100mA when transformed to a 32R impedance. The maximum undistorted signal at the secondary, then, is limited to 1.79Vrms. As far as I know, connecting 300R cans to the 32R winding will not automatically enable the output tubes to produce a higher voltage swing at a lower current.

Where to start?

You get the most power at the optimum load. Into higher loads, you get higher voltages (but less power).

Pentodes are funny and IMHO probably not a good idea for a hi-fi low-power amp that must face varying loads.

A Triode is roughly a constant resistance source. You would think (from basic power engineering) that best power transfer will happen when Rl=Rp. In fact the best power is more like Rl=2Rp (it takes a couple pages to explain why).

When Rl=2Rp, the peak voltage across the load is about Rl/(Rl+Rp)= 2/3rds of the supply voltage (really more like B+ minus a few dozen volts for origin-bend).

If you use a higher load resistance, the output is still Rl/(Rl+Rp) but with the high Rl the voltage approaches the supply voltage (minus origin effect).

Also, in a triode, distortion drops as load impedance increases, and power does not drop fast. Rl=5Rp is a nice value if THD is more important than raw power.

In headphone amps, quite small tubes can make more power than headphones require. We are rarely so desperate we need to match for max power.

> Z=25k ... 50Vrms ... 2mA

??? That is an awful high impedance for a transformer, especially a small one. It needs super-fine wire, which raises cost per pound and tends to break in the winding machine. The many-many turns needed to get 25KΩ in the bass have high self-capacitance and leakage inductance and will ring near the top of the audio band. Remember the total impedance is 2 or 4 times higher for push-pull.

Anyway, what tube will even notice 2mA? Maybe 12AX7 would have trouble, but you probably want something fatter like 12AU7 or one of the TV tuner tubes. These do over 10mA with ease. That allows a much lower transformer impedance and saves money and reduces the cost of making the transformer response nice and flat.

It is probably better to pick a transformer first, because stock trannies are VERY much cheaper than customs. One very common configuration is 5K:16Ω. Put a 32Ω load on it, the primary is 10K. Find a tube with a plate resistance about half of that: 6BQ7 is not my first choice but it will do and similar types are dirt-cheap. Bias it at 100Vp, 5mA, 50V 5mA peak swing. The RMS power is 35V^2/10K or around 122mW: you may want to raise your plate voltage and current (there's room in the specs). But that is single-ended (too hot here to do a push-pull workup). The plate resistance runs about 6K in this area. Reflected through a "5K:16" transformer the source impedance is about 19Ω.

The "5K:16" transformer has a voltage ratio of 17.7:1. The 50V on the plate reflect as 50V/18= 2.8V peak, 2V RMS at the secondary, or 125 mW in 32 ohms.

With a high-Z load the plate swing will run more like 79V peak. The unloaded output voltage is then 79/18= 4.5V peak, 3.1V RMS.

We can guess (or calculate) that for 100 to 600Ω phones we will get between 2V and 3V out. 2.5V RMS in 100 ohms is 62mW, not too shabby. However 2.8V RMS in 600Ω is 13mW, and we know that 3V in some 600Ω phones is not always enough.

If you follow through, a standard 5K:16 tranny with 180V or so plate supply will make ample power into almost any headphone.

Push-pull is almost over-kill for headphones: you have the extra cost of a phase-splitter, and your plate-to-plate impedance tends to wind up higher than your transformer winder wants to go. An old TV V-Sweep tube with hi-Mu and lo-Mu sections will give more than plenty single-ended power plus excess gain for feedback (which you are going to need if your amp is barely powerful enough and must have low distortion and wide response).
  Reply With Quote
Old 6th August 2003, 11:15 PM   #12
diyAudio Member
 
jeffreyj's Avatar
 
Join Date: May 2003
Location: Tampa, FL
Quote:
Originally posted by dhaen
Jeffreyj,
Not how I see it, at least for triodes.
The maximum undistorted voltage swing will occur when the load is near infinite.
Pardon my oversight: I forgot to include the clause, "for a specific output power," in my previous post.

(However, I think it was entirely obvious what I meant since I specified the output power and load impedance in the next sentence.)

edited because I just saw PRR's post...
  Reply With Quote
Old 6th August 2003, 11:31 PM   #13
diyAudio Member
 
jeffreyj's Avatar
 
Join Date: May 2003
Location: Tampa, FL
PRR, thanks for taking the time to post such a thorough answer!

Quote:
Originally posted by PRR

Also, in a triode, distortion drops as load impedance increases, and power does not drop fast. Rl=5Rp is a nice value if THD is more important than raw power.


Check.


Quote:

With a high-Z load the plate swing will run more like 79V peak. The unloaded output voltage is then 79/18= 4.5V peak, 3.1V RMS.


Is this assuming no negative feedback? With NFB, won't the signal swing at the anode remain 50Vrms regardless of the loading (as long as it is not overloaded, of course)?
  Reply With Quote
Old 7th August 2003, 12:13 AM   #14
diyAudio Member
 
jeffreyj's Avatar
 
Join Date: May 2003
Location: Tampa, FL
Quote:
Originally posted by PRR

> Z=25k ... 50Vrms ... 2mA

??? That is an awful high impedance for a transformer, especially a small one.
Sorry, that was just an example. It isn't the actual operating condition of the output stage. That, I believe, will have a B+ of 100-120V and a standing current of 12mA per channel. The output tube selected at this point is a 6DJ8.

I can well imagine that a 25k impedance would require impossibly fine wire! As a matter of fact, I am building a winding machine right now to make a run of tiny pot core transformers which have secondaries of 42awg - the wire is thinner than a hair, but less strong!
  Reply With Quote
Old 7th August 2003, 03:00 PM   #15
G is offline G  United States
diyAudio Member
 
G's Avatar
 
Join Date: Apr 2002
Location: Belleville, IL.
Hi JeffreyJ,

A while back I drew up a push pull 6SN7 headphone amp. I believe the transformers I was looking at where Hammond 1609 transformers. Off the top of my head I think they were 10K PP/8 ohms. They are rather large for the application but they will work for a tube with a plate resistance of 3K or so. Also those little PP transformers that Harman Kardon used in their amps once upon a time will work also. They are smaller than the 1609s. Unless you just insist on winding your transformers, for the sake of doing it, ready made transformers can be adapted to need.

G
__________________
Gavin
  Reply With Quote
Old 7th August 2003, 03:47 PM   #16
diyAudio Member
 
jeffreyj's Avatar
 
Join Date: May 2003
Location: Tampa, FL
Hello G,

I foolishly thought I could wind the OPTs myself, but after perusing a few resources on this somewhat arcane art, I've come to my senses and will be more than happy to just specify them!

But my original question remains: what happens when, e.g., a 300R headphone is connected to an 8R secondary. Does the primary voltage rise to partially offset the lower load or does NFB keep it more or less constant. If the latter occurs, then 300R cans will not be driven with near enough power!
  Reply With Quote
Old 7th August 2003, 06:00 PM   #17
EC8010 is offline EC8010  United Kingdom
diyAudio Moderator Emeritus
 
EC8010's Avatar
 
Join Date: Jan 2003
Location: Near London. UK
(Assuming triodes.) While it's true that output power doesn't fall as fast as one might expect when RL is raised beyond 2ra, it certainly doesn't rise fast enough to compensate for such a large change as going from 8R to 300R.

The issue is about the difference between load matching and output resistance. Valves need to have their loads matched to their internal resistance to obtain maximum power. Negative feedback can reduce output resistance as seen by the load, but it doesn't change the amount of power that can be extracted from the valve. The best example of this problem is an OTL. Adding 40dB of global feedback reduces the output resistance to the point where it can damp a moving coil loudspeaker, but the valves' internal resistance hasn't disappeared, it's just that most of the power is developed across it, rather than the load, and that's why OTLs are so inefficient.

It doesn't take much negative feedback to hold the load voltage very nearly constant when going from an ideally matched 8R load to an almost open-circuit 300R load.

If one had power to spare, one could simply use a 4W amplifier matched to 8R, and there would be sufficient voltage for 300R headphones as well. If the amplifier is smaller than that, it has to be load-matched.

All you need to do is to match to your chosen headphones and be done with it. After all, nobody uses more than one set of headphones, so it's not as if the amplifier needs to be universal!

There are some things best left to the experts. Filming sharks underwater is one, and winding output transformers is another.
__________________
The loudspeaker: The only commercial Hi-Fi item where a disproportionate part of the budget isn't spent on the box. And the one where it would make a difference...
  Reply With Quote
Old 7th August 2003, 07:36 PM   #18
PRR is offline PRR  United States
diyAudio Member
 
PRR's Avatar
 
Join Date: Jun 2003
Location: USA
> With NFB, won't the signal swing at the anode remain 50Vrms regardless of the loading (as long as it is not overloaded, of course)?

I thought we were discussing overload point.

Feedback does not change Power. (Feedback may move the THD at a given power, but the amp still overloads at about the same point, just more suddenly.)

If it swings 80V in a hi-Z load before overload, then you load it in a matching impedance, the overload point moves to about 40V.

Sure, with heavy feedback and an output well below overload, the output voltage will not change. That may not be what you want: lo-Z cans need much less voltage for a given volume. However even phones of the same Z have very different sensitivities (over 20dB different) so you gotta have a VOLume knob and let the user set the gain.

> a B+ of 100-120V and a standing current of 12mA per channel. The output tube selected at this point is a 6DJ8.

I thought we might come aroud to the DJ.

Modjeski says that in commercial use, he had reliability trouble at 1.1W, and prefers 0.6W. But he was building voltage amps, not power amps. (Note also his comments about "fake" 6DJ8s- actually some of these will work well in push-pull where they did not in single-ended..)

Looking at the 6DJ8 zero-grid and plate dissipation lines, 100-120V is too small for best triode power. Also the load resistance for best power is comparable to plate resistance (this one defies the rules because you and the DJ-specs have restricted the B+), which means higher THD and poor damping. However you can get over 1/2Watt this way, far more than you need, so you can fudge.

The real problem is practical transformer impedance. The 6DJ8 Rp runs about 3K. That suggests 6K loading, or 12K-24K for p-p class A. That's a high impedance for a hi-fi power winding.

Because the plate to plate impedance of two 6DJ8 is about 6K, you can base the transformer inductance on that, not on the reflected load impedance. For low core distortion in a push-pull transformer, you want the inductance to give response far below 20Hz (6Hz is a typical figure for good hi-fi). That says a 6K source needs over 100H primary winding, a lot.

Considering Modjeski's comments, and also improved voltage swing, I would idle at 5mA or 6mA per device (which may be about the same as your "12mA per channel"?). At high impedance, it would swing about 100V peak per plate. At very low impedance it would swing 25V 15mA per plate (which is class AB). At maximum pure-A it would swing 60V 6mA peak per plate.

60V 6mA per plate is 180 milliWatts, a healthy headphone driver. It is 10K load per plate, good for THD but an awkward transformer design. I think you have to design for inductance and turns ratio, and see where other parameters fall.

> 100mW into 32R is the design spec

1.8VRMS, 2.5V peak.

The plate swings 60V peak at max class A power loading. 60/2.5= 23.8:1 ratio from one plate to secondary.

Think ahead: no matter how much pencil you burn, you have to breadboard it with off-the-shelf parts to see if it comes close to goals and is worth getting costly good parts. 23.8:1 == 20:1 within tube accuracy. A 120/240VAC:6/12VAC power transformer gives that ratio, can handle over 100V swings on the primary, won't suck bad at 60Hz and will usually be flat well into KHz. And is cheap. And the math is easier. You can certainly test for power and distortion at 200Hz in varying loads. With dual primary you can check the effect of different ratios in different impedances. With a trace of feedback you can listen to music and know if the plan will NOT fly.

Eyeballing the plate curves, you will get 2.5V peak in 32 ohms, 3.8V peak in 120 ohms, 5V peak in 500 ohms. This is 100mW, 59mW, 24mW. In my opinion, there are several hi-Z phones that do want more than 3 or so volts RMS.

I say 7VRMS 10V peak in hi-Z is nice. The plates swing 100V in hi-Z. Let's try a 10:1 ratio from one plate to the load. (That's why I said to get a 120/240:6/12 transformer.)

With two tubes and 32 ohms, each plate can swing about 60V and 10mA. The 32Ω power is almost 300mW, however it is in class AB above about 100mW. At 120Ω the plate will swing about 70V peak, 7V pk in load is about 200mW. In 500 ohms the plate will swing about 75V pk, 55mW.

Either of those would be a fine design. 20:1 would have lower distortion but could feel weak in some hi-Z cans. 10:1 isn't high-THD and will really sock 32Ω cans (and you won't hear it slip from class A into class AB).

> use a single 4pdt switch

Way-overkill. If you must provide taps, wind the secondary all as one tapped winding. Center-tapped will work fine. Ground one end. Mount two headphone jacks. One is fed from the top of the winding and labeled "hi", the other from the center-tap and marked "lo". You can try to explain the difference to users, but they will be fine if they just try both jacks and use the one that sounds best.
  Reply With Quote
Old 7th August 2003, 10:03 PM   #19
diyAudio Member
 
jeffreyj's Avatar
 
Join Date: May 2003
Location: Tampa, FL
PRR:

Quote:
If it swings 80V in a hi-Z load before overload, then you load it in a matching impedance, the overload point moves to about 40V.


So the voltage swing on the plate does increase with reduced loading of the secondary. Whither NFB? If the plate has not been overloaded, then why does the voltage swing change at all. The input stage will compare a fraction of the output voltage, see it is incorrect, and do something about it. Or at least that's how it works in power supplies, transistor amplifiers, servomechanisms, etc.

Interesting article, btw. Makes me all the more wary of the tubes I have. I better take them apart to make sure the grid pitch is constant.

EC8010: the part about filming sharks was priceless!
  Reply With Quote
Old 8th August 2003, 03:09 AM   #20
G is offline G  United States
diyAudio Member
 
G's Avatar
 
Join Date: Apr 2002
Location: Belleville, IL.
Hi JeffreyJ,

Look at it this way. If you have a transformer that will present a 3K load to an amplifier with a 8 ohm load connected to the secondary then if you connect a 300 ohm load to the 8 ohm tap the primary of the transformer will now present a (300/8) x 3K = 112.5K load to the amplifier. Once you know the impedance ratio that you want it is easier to find an appropriate output transformer to suit your needs. I would recommend that you choose a tube with a low plate resistance. For say a tube with a 2K plate resistance I would use a transformer that will present a 20K load to the amplifier with a 300 ohm load on it's secondary. You can get away with that much damping in a headphone amp. The transformer I described would probably have to be custom made though. It's going to be easier to adapt a standard output transformer to a 32 ohm load. Then you will only have to worry about putting enough current into the load. How about push pull 6AS7?
__________________
Gavin
  Reply With Quote

Reply


Hide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
4 Pin Amp output to 3 Pin headphone jack? StrayS2k Headphone Systems 3 17th February 2009 04:46 PM
Aikidos Headphone with Output Transformers? wonsup Headphone Systems 4 15th April 2008 05:34 PM
Add headphone output to preamp whitelabrat Headphone Systems 2 24th January 2008 07:36 PM
How to connect 62 ohm headphone to 8 ohm output Sunsun22 Headphone Systems 15 16th June 2007 01:29 PM
Headphone Output to GC amp StalfoS Headphone Systems 10 29th March 2005 12:08 PM


New To Site? Need Help?

All times are GMT. The time now is 05:15 PM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio

Content Relevant URLs by vBSEO 3.3.2