OPA 1622 Integrated Headamp Project

Here is a project I completed several months ago that really worked out well. A friend over in France had this design in his head and shared over at DIY Audio, TI was coming out with specific op amps designed for Headphone amplifiers! So with this really niche market I was very curious to how these new headphone targeted op amps would work in a headphone amplifier…

Usually with these new IC’s you wait for months before OEM vendors offer a product with the new technology, Schitt and THX use some of these op amps in this series of new ic’s in some of there new offerings…so to get ahead of this I decided to build this friends design…(back then I had NO idea what Schit or THX was up to).

A neat thing is one of TI’s design engineers that designed this specific op amp joined in the discussion several times watching us noobs using his “baby”…

Chris’s “Ampli Casque Integre” Schematic:

The op amp in particular is the TI OPA1622. At the time I built “Chris’s” design there was not a commercial product with this chip in it…and the size of the chip was very, very small compared to existing DIY parts we are used to. On other DIY projects like my AGDR ODA I got to solder really small 805 parts. This amp has even smaller parts and they can not be soldered with a conventional soldering iron. Their are “leadless” QFN parts.

Whats a QFN: Quad Flat No-leads package, they have no leads or pins like previous op amps, they are flat and you have to use a different method of soldering them to a pcb. Commercially they use solder pots, re-flow ovens etc to get these devices to attach properly. There is a thermal profile as well that you need to be aware of, too much heat for too long can destroy these small parts and for a DIY’r they can be "expensive’…

QFN Op Amp:

So I had to learn to solder all over again with a hot air gun instead on an iron, using flux pens, and solder paste instead of normal or regular solder. With some experimentation I learned how to do this, after a few test runs with how much flux, how much solder paste, I watched as the tiny little chips floated and wiggled into position as I heated them up…the solder turns molten and the surface tension helps move an postion magically to the correct pad location…its a really neat thing to see in this tiny macro world…

I had to get a set up flip up binocular magnifiers to actually see well to do this…so easy to short leads out underneath the parts…

Heres in the pcb with the IC attached as well as a power reg . Notice the small IC upper left of the large tall standing part, this is a BIG 8 pin SOIC part thats smaller than the op amps you normally plug into a DIP socket…and then notice the other three smaller parts…they get very, very small especially the one in the middle right that is surrounded by a circle of pads…!!

My workstation where I do some of my magic…

Getting there!

Testing…It works!!! Phew!! (Back when the ODAC was in favor)…

Finally all packaged up!

Ok for now, this was a lesson on how old farts can learn new things…after 40 years of soldering old school I can now lay down parts many time smaller and dense that I ever thought would be possible in the past!! From Vacuun tubes to advances solid state…

Take a look inside the THX AAA amps or the new Schitt Magnis and you will see these type ic’s being used in some of the latest headpone stuff and especially in phones and portable stuff…

How does it sound??

Thats next.

Link to DIY poject file: https://www.diyaudio.com/forums/headphone-systems/304684-ampcasq-opa1622-integrated-headamp-project.html

I still have a pcb if anyone is interested in trying this one!!! :>)


I built this amp, for two reasons, to learn how to deal with new very small ics in DIY headphone projects and secondly to see how the worked in our headphone amp world…

TI designing chips with extremely low distortion numbers came out over the past few years…and this was a chance to “hear” them…or not!

Chris (friend over in France, can meet him at diyaudio.com) put this simple design together as a testbed for his future projects. He had some simple design paramters:

  • It should stand in a small aluminum box, in order to sit on desk
  • It should be powered by an AC adapter I had, which provided 17VDC
  • It should have a volume control
  • It should have RCA and jack 3.5 inputs
  • It should have jack 3.5 & 6.35 outputs
  • It should allow me to experiment crossfeed
  • Final price should stay under control

This says nothing on distortion levels etc…I took this as is, whatever TI stated I had to accept and using this in a “reference” type circuit was ok with me at the time. Circuit layout in a design is VERY important.

We lucked out with absolutley no audible noise or hiss or oscillations with all kinds of loads, ie cable lenghts of headphones, impedance etc.
DEAD SILENT at all volume levels…zip, zero, nada. Made me smile.
Even with some of todays stuff you hear some noise at extreme levels with higher gains.

Here is my first impression back then:

I spent a few hours this morning listening subjectively to Chris’s amp I just completed and several other amps I have build and bought. I listened to AGDR’s Desktop version of the O2 amp, the AGDR ODA, the O2 with Booster Board (another AGDR mod), Bottlehead Crack with Speedball mod, and a Schitt Vahalla 2 all tube amp.

Using several of my demo recordings from friends in the industry I listened over and over etc to the same recordings all with my Beyer T1 2nd gen headphones. 600 ohm. But these cans impedance vary all over the place.

Many state that a higher voltage amp is needed to drive the T1’s best.

With the T1’s in Chris’s amp I cant listen to the T1’s over about half way on the volume control. The amp will play these louder but it gets painful and its definitely in hearing damage area IMO. So to me this little amp plays the T1’s rather nicely…and I compared with an amp that has many times more power at 600 ohms.

All the amps used sound really good, a statement of the present state of design, good components and good engineering and technology. Some of the amps dont play well with lower impedance cans.

To me a personal preference is that lower impedance cans seem to sound better to me that the T1’s with this amp. I have no idea why, its probably my subjective bias, knowing this amp is a lower voltage amp than some others?

That said the T1’s with this amp do very well…this is a dead silent amp and I am using a 16vdc IBM laptop power adapter at present and dont really know if this supply and is introducing any audible issues. Compared to the other amps I would think not at this time. I will try some good old DC batteries and have a regulated DC supply on its way as well.

The x-feed circuitry to me is not really doing anything great that I can hear with most of the recordings I listen to. I tried a wide varierty of recordings and thought I would wear out that switch!! But…never could really hear any really audible difference to my ears. With one channel only there is a drastic noticeable difference as things do indeed “move” in space. Its just with two channels hooked up its harder to tell the difference…IMO. I found some older tunes, ie Blood Sweat and Tears, that have songs starting out mostly in one channel for several seconds and if I flip the X-feed Switch down, you can hear the shift of things for sure so its working.

The sound is squeaky clean and no audible distortion of any kind is heard. In its present state it holds its own with these other amps. The only comment on a real difference between all these amps is that the soundstage is the one thing that seems to change somewhat. Powering down sometime seem to produce a small thumping or discharge type of sound, but turning the volume down makes this go away? But there are times when shutting down there is not sound.

The Bottlehead amp is like your back in row 10 listening at the symphony. while the Schitt V2 is like front row center, Chris’s amp seems like its very much like the Schitt V2, right there up front IMO. This is much influenced by the quality of the recording as well.

All in all this good example where TI and its new chips have brought us…they allow for a very small package to be built that has very good specs etc

Doing some calculating with a headphone power calculator:
These were the cans I had back then.

Using the headphone power calculator of Bob Robinette give the following results with the cans you mentioned :
• DT1350 (80R, 109dB/mW), OPA1622 with +/-5V supply should be ok, it even can reach the max SPL of the can : 129dB
T1 (600R, 102dB/mW), max SPL with OPA1622 with +/-5V (clipping at 4.1V) is 113dB. Certainly too short, it should be better with +/- 10 V supply, allowing 120dB SPL max.
• T90 (250R, 102dB/mW), OPA1622 with +/-5V supply should be ok, max reachable SPL : 117dB
The max SPL section of the spreadsheet (‘How Loud Will an Amp Drive Your Headphones’) is based on the max output voltage of the OPA1622 which is appx Rail-0.9V with most of loads.

The only thing I did not like with this amp was the crossfeed circuitry. It didnt really work as I thought and I spoke to Chris about this and he agreed it was well…not needed.

The one thing I did find annoying was with lower gain cans the volume know didnt have to be cranked up very high to get to loud listening levels, so we removed the crossfeed stuff and re-engineered its switch to have three slelectable gains…voila problem solved!

I have had this neat little OPA1622 amp since 2017 and I still have it and still use it!

Hope you have enjoyed this early journey into “exotic QFN” headamp building!!