Preamp based on TI THS6012

has anybody built something like this? I am in need of both a preamp and a headphone amp and read the article on and am initially impressed and optimistic. Digikey lists the part for 7 dollars a piece. Expensive, but not in the scheme of things. The board looks like the only catch, so I might wait a month or two to see if Mr. Elliott creates a project in that time.

The problem with the THS6012 is that it is a surface mount device and therefore not easy to use. There are many other high speed current feedback op-amps available that will give a similar performance but that are easier to handle. Two that come to mind are the Burr-Brown OPA603 (DIP8)and the Elantec EL2099C (TO220). The EL2099C will drive 440mA into 25ohm at +/-11V, has a slew rate of 1000V/uS and is a 50MHz device. The TO220 case ensures no problems regarding heatsinking. There are some useful application notes on high speed current feedback op-amps at the Elantec website.
There are a few more high BW, high speed, high current devices in mDIP packages:
Analog Devices AD811
National LM6171, LM7171

I've used the AD811 in a number of audio apps; it's a great part, but you MUST stick a heat sink on it for rails above +/- 12V. It idles high (I can see the Class A lovers starting to drool:)

There's also the AD816 (I think) dual in a higher power dissiption package.

I think Linear Technology also has some CFA parts that would be worth experimenting with, but ....
a bunch of the newer devices have rail voltage "limitations" (max +/-6v)

Surface Mounted Devices

Geoff, although I'll confess I haven't constructed any projects from new using SM ICs, I used to routinely replace defective ones. I feel that some people may be unnecessarily afraid of using them. Provided the board was designed for them in the first instance, I really don’t think they are any more difficult to install than conventional DIL packages. You will certainly need a good quality, fine tipped soldering iron, however I’m assuming everyone is using one of those anyway :)

Removing any surface mounted device, on the other hand, is a pain in the neck, even with a professional vacuum desoldering station. Unfortunately the solder tends to wick under the relatively broad surface area. For ICs I found the best solution was simply to cut the leads off, and remove the remaining leads individually.




2001-06-01 4:53 pm

Lots of headphone amp projects at At least one uses the LM6171 that mlloyd1 mentioned. Also see some of the discussions at for high-bandwidth devices.
Me, I wouldn't touch them. My amplifiers always oscillate, I'd hate to see one of them suckers take off. Actually, I probably wouldn't see scope only goes up to 100 MHz. SMD packages are probably a good idea.
Thanks for the response. I am somewhat new to this and the DIP packages are definitely more appealing since I dont have the knowledge or equipment to make a good surface mount board.

My next question would then be for the headphone amp, what kind of life would I get using +/-9V rails (2 9V batteries)

Are most of these high speed CFB opamps going to require pretty much the same layout considerations? I.E. Special attention to the inverting input, decoupled rails in proximity to the opamp, short leads on the parts etc. How close is close? millimeters? .5"? 1"?

You will probably do best by downloading the appropriate data sheets from the manufacturer and studying carefully. These days there is a wealth of information published by them which will go a long way to answering many of your questions. In addition look at other semiconductor manufacturers and the information they have available. Analog Devices is one such company that comes to mind that provides a great deal of useful information on their site.

Sometimes there's nothing like getting the information "from the hourse's mouth".



I agree that SMD devices can be soldered in the normal way, but I would not recommend this for a beginner with limited soldering experience (nor for those of us whose near focus is deteriorating due to advancing years - I have enough difficulty seeing the separate strips on a veroboard even with reading glases). Also, the THS6012 has a copper insert in the back of the case which must be soldered to a defined area of pcb for heat rejection purposes. This is not possible using manual methods.


Battery life will depend on which devices you eventually chose. Some are more battery friendly than others due to a lower quiescent current. The previously mentioned Headwize and Head-Fi sites offer some guidance on this topic

Some useful information on layout considerations etc. for high speed op-amps is given in the Elantec application notes Nos 9, 19, 22 and 23. Close for high speed devices is 1 to 3mm (less if possible :).

I suddenly remembered how Mr. Elliott ran the tests on this IC. I looked further and found the evaluation board. I am going to try to get a friend who works at TI to get me one, but if that doesn't pan out, it is listed at 50 dollars( was actually really surprised since my other experiences with eval boards have been $300+). I think this might be well worth the headache's I'll encounter when trying to do the layout and construction on a high speed board myself. Who said DIY was cheap :)