Where do I start? (re: need a DIY amp project)

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I'm looking to start on a DIY desktop headphone amp project ASAP. It will be used to power my humble Grado SR60i (for now, upgrading in the future) with their 32 ohm impedance.

I have solid soldering experience, but not with SMT. I do not want to build an O2 or CMOY. I would prefer something decent quality (Class A), but have a budget of $300, preferably less. I like tube sound, but solid state is good too. I want to build something that will last, while I upgrade to better headphones, even if it's overkill for my Grados now.

I'd love something that was upgradeable (ie. either tube rolling or op-amps). I like to tinker, especially if I can hear the effects of different upgrades. I am willing to spend more than necessary for a good volume pot, for example. This stuff matters to me.

I need something with PCB(s) readily available now. I am not opposed to collecting pieces either from local electronics stores or over the web.

Any recommendations?
 
As you like to tinker, I recommend you start by winding a couple of ferrite transformers. That's because in my experience, typical headphone amps lack in dynamics due to inadequate power supplies for driving low impedances. If you wind a trafo with a 5:1 turns ratio (say) that will make your 32ohm headphones look like 800ohms to an amp - the dynamics will be noticeably better on practically any amp, assuming it still has enough voltage swing.

Mouser stocks some suitable cores now : 6695272521 Fair-Rite | Mouser

Since the sonic benefits of classA (IME) relate to having much lower load-induced noise on the supply, such a trafo will have an AB amp sounding much closer to classA.
 
Last edited:
I do not want to build an O2 or CMOY

You shouldn't lump the O2 headamp and a CMOY together, either. :) Not the same. A CMOY typically has just one chip that provides voltage gain AND output current drive. A single op amp chip likely cannot push even half the 100mA peak jcx calculated:

then calc I,Vpk for 120 dB SPL: 160 mWrms, +/- 100 mApk, 3.2 Vpk

The O2 uses a separate chip for gain and two paralleled high current chips for output that can source/sink up to 120mA or so of current and hit jcx's number. A single battery CMOY with virtual ground may not even be able to hit jcx's 3.2Vpk swing number either because you are dividing the 9V battery in half, minus chip overhead. The O2 uses two batteries giving around 7V peak swing. A CMOY with two batteries (they exist) still likely can't push even half of 100mA if using just one chip.

I say "likely" because the Grado RA1 headamp is essentially a single chip CMOY that uses one of the two output chips paralleled in the O2, so it can hit 60-70mA. Most CMOYs use chips that are only good for 20-30mA or so of output current.
 
Last edited:
DIY Projects | HeadWize is a good resource - if you look gilmore's amps are built point-point on perfboard

There is also a single layer pcb layout for the Gilmore that isn't too bad. You'll have to adapt it anyway since the input dual jfet are gone :(

The CKIII looks quite perfect to me for you. It isn't very fashionable but it's a well proven design in class A, the pcb is available, the project is well documented and it easily fits your budget. Add a dc protection board and you're good to go.
 
The Millet Minimax is perfect for you. The kit is 230$ well under your budget Beezar.com, it's a hybrid so you can rolling some tubes and tweakable. It's highly bias in class A and sound very good with grados and other headphones and you dont' need to do the case work which is the hardest part of Diy headphone amp.
 
Thank you everybody for the replies so far... I do not want to pollute this thread with a bunch of targeted replies, so I will just let you know I am considering my options presented.

I will need to do my research as I am not familiar with any of the kits/projects (Altronics, CKIII, and Millet Minimax) that have been suggested.

Sorry if I gave the impression I was more "advanced" at electronics than I actually am, but I'm not quite ready to wind my own transformers and at this point, I'm leaving the math out Side story, I started but did not finish electrical engineering in university; this might've been useful at this point :)
 
I strongly recommend the sapphire amp kit from RJM. The kit is easy to assemble, sounds really great and will come in well below your budget. Richard is great to deal with too and helpful with support should you need it.

Thread here:

http://www.diyaudio.com/forums/head...dio-sapphire-desktop-headphone-amplifier.html

Apart from being able to solder through hole components, the thing you need to be most careful with is to use twisted pair for the inputs and outputs.


Fran
 
Op-amps are your friends! They are the closest thing to an ideal amplifier that you'll find. The LM1875 to LM3886 series power amps have found extremely wide acclaim in the audio community, both in terms of measurements and listening tests, and are essentially little more than high-power op-amps. Most op amps, on the other hand, are made for line-level use, and usually expect a load higher than about 600 ohms. Some op-amps can push a lot more, and make great output stages. The O2 is a good example, will leave you with $200 in your pocket to spend on music, and in terms of sound quality it surpasses the limits of the human ear (based on what data we have available). Your Grados will love it: http://www.diyaudio.com/forums/headphone-systems/193977-objective2-o2-headphone-amp-diy-project.html Of course, if you want to go a more raw DIY route, there's the tried-and-true combination of a good op-amp for voltage gain and a good buffer for current gain.
As you like to tinker, I recommend you start by winding a couple of ferrite transformers. That's because in my experience, typical headphone amps lack in dynamics due to inadequate power supplies for driving low impedances. If you wind a trafo with a 5:1 turns ratio (say) that will make your 32ohm headphones look like 800ohms to an amp - the dynamics will be noticeably better on practically any amp, assuming it still has enough voltage swing.

Mouser stocks some suitable cores now : 6695272521 Fair-Rite | Mouser

Since the sonic benefits of classA (IME) relate to having much lower load-induced noise on the supply, such a trafo will have an AB amp sounding much closer to classA.

That's dubious, and a lot more trouble than it's worth. It's also really, really easy to make an adequate power supply. Seriously, you won't need anything more than an 500 milliamps at most for those Grados. Also, negative feedback will bring the output impedance of the amp to almost nil even if the power supply has an output impedance of a few ohms. Even the best audio transformers will contribute much more distortion than all the rest of a sensibly designed headphone amp. If you think 32 ohms is a problem, try talking to those speaker people with their 4- and 8-ohm loads! Also, do you have any evidence that hooking up a load to a class A gives and messes up the power supply less than with class AB? I've never heard this before.
I strongly recommend the sapphire amp kit from RJM. The kit is easy to assemble, sounds really great and will come in well below your budget. Richard is great to deal with too and helpful with support should you need it.

Thread here:

http://www.diyaudio.com/forums/head...dio-sapphire-desktop-headphone-amplifier.html

Apart from being able to solder through hole components, the thing you need to be most careful with is to use twisted pair for the inputs and outputs.


Fran
There seems to be some woo-woo going on in that design... Honestly, I can't recommend paying for woo-woo. Money is valuable but woo-woo isn't. :(
 
Last edited:
That's dubious, and a lot more trouble than it's worth.

Do please go on to explain why, I'm all ears.

It's also really, really easy to make an adequate power supply.
Sure, but is 'adequate' what the OP wants? - ISTM he's concerned about getting great sound quality. That doesn't happen in my experience with merely 'adequate' supplies - they deliver merely adequate SQ.

Even the best audio transformers will contribute much more distortion than all the rest of a sensibly designed headphone amp.
I suspect the evidence here isn't on your side - the Audio Precision test equipment uses one of the best audio transformers and I haven't heard them saying that's limited by transformer distortion.

Also, do you have any evidence that hooking up a load to a class A gives and messes up the power supply less than with class AB?
Does the evidence of ears count or are you a measurements only kind of guy? I should point out that merely having a classA amp does not guarantee that it won't mess with the PSU. Only a subset of classA amps leave the power supply invariant.

I've never heard this before.

So are you here to learn something new?
 
Last edited:
Marine boat has a point though... it's much easier (and cheaper) to build a very good amplifier for 32r loads than to wind audio transformers of a similar quality. There's about zero chance for a beginner to wind even a decent one.

One could buy a pair of sowter 8665 though. Doesn't come cheap.
 
He has a point for building an 'adequate' amplifier. If that's what the OP wants then I'll let him have the floor for the rest of this thread :)

I'm a beginner myself at winding trafos - they're very easy to wind the way I do them, which is a totally naive way probably far from the optimum because I've only been making them a couple of months having started from scratch. But I was so impressed with the results I decided to share my findings. Quite possibly the real trafo experts (BudP springs to mind) would class my trafos as very far from 'decent' yet they've sounded wonderful so far on headphones (not at all so wonderful on speakers though which reason I've yet to discover).
 
Do please go on to explain why, I'm all ears.

Sure, but is 'adequate' what the OP wants? - ISTM he's concerned about getting great sound quality. That doesn't happen in my experience with merely 'adequate' supplies - they deliver merely adequate SQ.
I was being modest.
I suspect the evidence here isn't on your side - the Audio Precision test equipment uses one of the best audio transformers and I haven't heard them saying that's limited by transformer distortion.
The people at AP may be magicians, but I don't think they can get their DC coupling with a transformer.
Does the evidence of ears count
I'd accept a blind or double-blind test gladly. Unless you can hear with your eyes I don't think that will be a problem!

or are you a measurements only kind of guy? I should point out that merely having a classA amp does not guarantee that it won't mess with the PSU. Only a subset of classA amps leave the power supply invariant.
The neat thing about audio is that, before getting to the transducer, the signal is electrical. The same electrical signal moves the same speaker (or earphone) in the same way and makes the same sound under the same conditions. We can measure electrical signals so tiny and precisely that we can measure the signal long before the speaker (or BA driver, etc.) makes any detectable movement, and long, long before the threshold of human hearing is breached. But no. I'm not a measurements-only kind of guy. I just wanted to say all that. :joker:
He has a point for building an 'adequate' amplifier. If that's what the OP wants then I'll let him have the floor for the rest of this thread :)

I'm a beginner myself at winding trafos - they're very easy to wind the way I do them, which is a totally naive way probably far from the optimum because I've only been making them a couple of months having started from scratch. But I was so impressed with the results I decided to share my findings. Quite possibly the real trafo experts (BudP springs to mind) would class my trafos as very far from 'decent' yet they've sounded wonderful so far on headphones (not at all so wonderful on speakers though which reason I've yet to discover).
I'll let you have the floor if the OP wants what's in the bold type!
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.