DIY Audio Projects Forum

Hi all,

I'm planning another build of the venerable headphone amplifier. It seems like a very popular project for beginners so I'd like to take another whack at it.

First, I'm going to power it with an 18v transformer. The unloaded voltage should be in the neighborhood of 26v. To force unbelievably low ripple, I'm going to have a CRC filter follow the rectifier with 8800uF on each side and 2 ohms of resistance in the middle.

The tube filament will be the current controlling resistor for the current source in each side. At 6.3v and 0.15A each, the equivalent resistance is 42 ohms. So...a 6.8v zener diode controlling a TIP31 or similar current sink will do the trick.

The triode plate will be connected to the MOSFET gate through a capacitor. I'm not sure what type or value yet, but it need not be large. I want to dial the gate voltage in with a pot and make the MOSFET bias independent of the tube stage.

By running this circuit at 24v, I'm hoping it will open up the performance of some tubes that didn't work well before. With the JJ 12AU7, I recall it distorted the sound at low volume settings. Older tubes didn't do that.

I'm also wondering if I'll have to put in RF chokes to block out switching noise the diodes might create. Due to the high filter capacitance, it's possible the current waveform into the amplifier might be short, intense spikes.

Ed AE7TE

_________________

You can't have too much heat sink.

Hi Ed. Note that the heater current draw at 6.3V is 300 mA. At 12.6V it is 150 mA. Do you have a schematic drawn up.

Cheers

_________________
[ DIY Mains AC Power Cable Cord ] - [ Gobo LM1875 Amp Kit ] - [ Tang Band D4-1 Horn Speaker Kit ] - [ Monoblock Push-Pull KT88 Tube Amp Kit ]

I have one drawn up, but it's too large to upload. I might be able to redo it in a smaller file format but I can't right now, as I'm traveling.

The 12A_7 tubes have a filament that's center-tapped. You can choose 6.3v operation at 300 mA or 12.6v operation at 150mA. Each side of the filament wants 150 mA at 6v and you get to choose how you give that out. With tubes like 6DJ8, 6N2P, etc, you don't because the two are hard-wired in parallel from pins 4-5.

So, I could run each MOSFET at 12v/150 mA, with the remaining 12v dropped by a current source and half of the tube filament. The current source transistor maintains a constant 6v on the filament, and therefore a constant 150 mA through the IRF510. Everybody is happy. The only problems this causes is a 2-watt dissipation in the output device that forces me to be aggressive on the heat sinking. Not a huge deal. The other thing is that it costs a whole 300 mA from the power supply, not counting the small amount that I'm going to spend on the tube plates (which might be 0.5-1 mA each plate) and other things like LED indicators and whatnot.

A more efficient route would be to run the tube filament separately, as a 12.6v, 150 mA string, and then bias each MOSFET to 50 mA. Total current draw would be 250 mA for the power and filament stage.

My thinking is this - if I were too worried about efficiency, I'd be designing an amplifier that wasn't class-A. Then again, for a headphone amplifier, how crazy do we really want to get?

Ed AE7TE

_________________

You can't have too much heat sink.

Try this:
https://www.xnview.com/en/mobile/resizeme/

_________________
-= Gregg =-
* Ratings are for transistors - tubes have guidelines*
Home: GeeK ZonE
Work: Classic Valve Design

0.2Amps of bias from my exprience, is more than sufficent to drive any 8 ohm headphones at Class A
This results in about 1.28W of class a into 8.
Where calculation P=I^2R
1.28W=0.4*0.4*8
By deffiniton of class a signle end, maximum current draw is half of idle current.

If you want more sound quality from the circuit, I would higly recomend that you try capacitor coupping the tube gain stage to a SIMPLE OCL(push pull) output stage. This means your headpohnes are directly connected to the transistor, care should be taken as circuit fault can burn headphones, unless you have dc offset protection at output.

Capacitor output from se has quite a bit of disttortion, the bass would be sightly reduced and phase shift occurs.

Bear in mind the 12AU7 (ECC82) / IRF510 Headphone Amp has its output inverted, fixing this by a new schematic this may result in better sound quality.
Further more with the orignal 470uF cap you would not be driving any 8 ohms, because you can calculate the cut off freq at 8 ohms, this is way above 20hz. You can use a online calculator to do this. With higher impedance headphones, the Class A output power would be much more, Thefore your bias current needs not any more than 0.2Amps.
If you want to make sure your full class a, do a practicall test. Wire a amp meter in series with the power supply, monitor and increace volume untill the maximum current through meter is doubble the idle. If this volume is loud enough for you then your good.

_________________
I've quit gaming during meal breaks, that allows me to be more constructive

You can also email the schematic to the email at the top of the page and I can post it.

_________________
[ DIY Mains AC Power Cable Cord ] - [ Gobo LM1875 Amp Kit ] - [ Tang Band D4-1 Horn Speaker Kit ] - [ Monoblock Push-Pull KT88 Tube Amp Kit ]