Headphone amps to build or buy

[Ant]

Ive kind of got into headphones recently. Circumstances mean i dont get to use the main system very much at all, and tidal on my phone and some nice headphones has been an enjoyable avenue.

The headphones i have available are as follows
B&w p5
B&w p7
Sennheiser momentum 1
Audio technica at706 electret which need a power amp to drive them
Sennheiser hdr180 wireless jobs.

Of these the p7 and the at ones are the best sounding, the p7s edging clear as the at electrets are sounding abit tired, and the at's dont block out extraneous noise half as well.
The p5s are the best at this but are not the best sounding.

The wired ones are all low impedance, all around 30r

Any thoughts from the panel?

Cheers ant

[Mike H]

Tubed or solid-state?

Ready-made or DIY?

Did a Google, there's loads. Some prices are barking (just IMO).

If DIY, lots of tube and SS kits on eBay (search 'headphone amplifier kit'). Most are cheap and might be fun to play around with.

[Ant]

I dont mind either way mike, buying is easier, building is more fun

I just dont know where to start

[ed]

if your looking for a bugdet solution you could do far worse than my latfet example. I've been happy with it for the last 2 years and although I winged about it being a tad bright, Rowuk was on the money suggesting different headphones have different presentations.

You could have had the spare boards if you were interested but I can only find 1 of the 5 I had, but I think 1 of my sons had a pair. If you're interested and he hasn't used them I'll get him to send them to you. For about £15 in parts plus a wall wart it's awesome value.

hpa se circuit.jpg (219.06 KiB) Viewed 14763 times

[Ant]

Yep id be interested in that ed, it'd be nice to do some soldering again
If he hasn't used them let me know
Cheers ant

[Ant]

Ed, dumb question, is it 12v dc or ac from a wall wart?

[Nick]

it will be DC. 1 amp should be enough 350ma or so a side looks like from the CCS's

[Ant]

Thought so. Yet another complete brain fart moment

[JamesD]

Just chiming in to say my P7s need 50mA for max drive and about 1-5 mA in practice to be loud!

Happy New Year!

James still lurking around

[IslandPink]

Happy New Year James !

[pre65]

James still lurking around

Glad to hear that James.

[ed]

Hi Ant

I dug out the spice project, it shows 300 ma with 12v per channel. If you use the relay shown in the hand drawn circuit then whichever relay you use would need the solenoid current factoring in......it's all over thinking so I would just say use a 12v 2a smps.

The boards are en route apparently....It was my first attempt at a pcb so there are a couple of features, namely: I forgot the mounting holes and I got the wrong layer for the text so it's missing. I will send a a bom that matches the part values to the board positions. Bias the fet at about 265ma , about 1.25v across the the resistor at the bottom of the 317, shown as R4 on the spice diagram.

hpa spice.jpg (108.07 KiB) Viewed 14416 times

[Nick]

My bad. The 350ma was because I saw R4 5ohm and then divided 1.25 by 4 instead of 5

[Tony Moore]

Hi Ant,

I have some pcbs left over from my DIY headphone build in 2021 which you can have free if you wanted to go the class-A PP route.

It's a design based on a circuit by Kevin Gilmore which I captured in KiCad and made my own layout.

Original site I used for reference:

http://www.pavouk.org/hw/dynalo/en_index.html

The only downside is it uses matched Toshiba 2SK170BL and 2SJ74BL JFETs for the input stage so a bit difficult to source these days. However, there is a seller in the US with genuine matched sets but at a bit of a cost. (punkydawgs on Ebay)

It's a fine sounding head amp though, absolutely quiet and very dynamic.



There are things I probably would've done slightly differently in a mk2 version of the pcb but you live and learn! You're welcome to the KiCad files too if you wanted to make updates on the pcb.

Cheers,
Tony

[Nick]

Bias the fet at about 265ma , about 1.25v across the the resistor at the bottom of the 317, shown as R4 on the spice diagram.

Forgive my ignorance, but as the 317 is wired to be a CCS and its ref voltage is 1.25v how can you bias the fet for any more than that current and 1.25v across the sense resistor. I would expect the voltage at the source to just start to rise but no more current to pass. I would have thought you would bias the fet to put the CCS in limit and then enough extra to give you whatever headroom you need on the output.