Is this the sound of op-amps?

I was at my friend Berrys' tonight and we fired up his DACT CT-101 pre-amp module which he is using with a power supply I built for him based on the BOSOZ derated to 24volts.

I was initially quite amazed at the sound quality, highs are nice with perhaps a bit of tiz but not aggressive, midrange is killer and the bass goes quite low although it seemed a bit uncontrolled at times. Large soundstage with decent depth and image placement. To put it another way the music sounded nice.

But, something seemed to be missing and for a while I could not put my finger on it untill I replaced the DACT with my BOSOZ and man, the music came alive, just got sucked into the sound and gear considerations out the window.

I came to the conclusion that the Op-amp based CT-101 was giving a similie of the music and my poor ugly BrideOSOZ was conveying the emotive content of the music although I might be in looneysville here.

What I am wondering is, do op-amps have a charecteristic sound? And are mosfets inherintly superior?

Anyone care to comment?

Tony D.
Well, that's a very difficult question, as there are many ways to use an opamp, and many varieties of them... FET/bipolar/hybrid, modern/ancient, voltage-feedback/current-feedback etc. Likewise, there are many different ways to use MOSFETs. So, I think it is a very bad idea to group all opamps or MOSFET circuits together and label each group with a characteristic sound. I tend to look poorly upon opinions which do not encourange an open mind, as this stifles the creative aspect of DIY, and closes doors to the possibility of exploring all the options.

Opamps in particular, i have found can have a very wide ranging sound, from that harsh "solid-state" sound to a low resolution, fuzzy and warm sound, to a very clear and transparent sound, and also they can produce a clean but lifeless sound... it's all in the particular device used, and what circuitry is around it. Similarly, i've built discrete MOSFET circuits which exhibit full, rich midrange and a wonderfully liquid sound, with a nice (if a bit soft) treble, through to very muffled and un-dynamic (is that a word?) sound. I think in general, it is much harder to choose a good opamp, and then to figure out how to extract the best performance out of it. So, an many DIYers will likely get better results out of discrete MOSFET circuits.

Inerestingly, the best preamp circuit I've ever built (IMHO) is a hybrid opamp/discrete MOSFET design. The circuit is really wonderful in transparency and preservation of the details, but also in the conveyance of that musical "soul". But, I'm not really here to plug a particular circuit...

I guess there's also your personal preference for what you want to hear. Lynn Olson authored an article entitled "Illusion Engines" (I's available on the net somewhere), which presents some valuable insight into what we are trying to acheive with audio reproduction. Personally, the more I try to compare real-life sound with what I hear from my stereo, the more I come to prefer a cleaner, more transparent sound that preserves the low-level details and dynamics. The greatest challenge for me recently has been to find a circuit that accomplishes this while also preserving (creating?) that emotional connection. I've built several circuits which have a very detailed, transparent sound, but are flat and lifeless.

I'd wouldn't mind taking a look at the DACT preamp schematics. Is there a webpage somewhere which has these?

[Edited by hifiZen on 10-04-2001 at 12:29 AM]
hifiZen, thanks for replying and your comments. It was not my intention to be closed minded or to lump all circuits using these devises into a group with "one sound" but I have read elsewhere on the net that op-amps tend to have a charecteristic sound signature and for my learning I polled for opinions here because of the diverse opinion pool.

I was surprised that even though the CT-101 sounded good and all the detail, soundstage etc. was in place the sound had a lifeless, unexciting quality.

You can check out the CT-101 at although I am not sure there is a schematic there and on the unit itself the op-amps have been painted over so it is not possible to say which ones they are.

Tony D.
Op-amps certainly have a definite sound of their own, sometimes sounding artificial or non-interesting.

It also has to do with the parts you use for the supply and for decoupling.

But I think it's mostly the chip type. I am not sure which dual chip DACT uses (could never find out which it is), but that might be a place to put a socket in and try several others.

Something like AD825, AD826 and LM6182 might be worth trying.

Local quality decoupling is essential too, and on the DACT they seem to be tantalum. I wonder why electronic engineers love tantalum, as they sound awful for bypass or DC blocking!

oops. i wasn't referring to you, carlmart - just rambling in general...

I don't like tantalums either, and i'm not sure what the attraction is... low RF impedence? dunno, but i stay away from them.

Of all the preamps i've built, i find that opamps will give that lifeless sound if they're left to drive the outputs directly. So, what i usually do now is add a discrete buffer inside the opamp's feedback loop, and this handles all the output current while the IC can relax and concentrate on doing it's job. The preamp I'm listening to uses a pair of power MOSFETs as outputs (IRF510/IRF9510), in a constant current SE class-A configuration. To improve the operation, made the output device a CFP with one bipolar transistor (2N3904 if memory serves me). The outputs are biased at about 70mA idle current, so they get quite warm. Now, the opamp is loaded very lightly (and the opamp's output stage is also pulled into SE class-A too), and doesn't have to drive any difficult loads (it doesn't take much to bog down an opamp) but still controls the output directly.

My current favorite opamps are the OPA627 and OPA134 from Burr-Brown. These seem to have rather more "life" than the high-speed opamps like AD811 and LM6182, but in all fairness, I havn't yet tried some high-speed opamps with a big external buffer (other than the BUF-634, which i've had little success with).

You made it sound as if I am the guardian against tantalum caps! Nothing of the sort, please!

BTW, there are some comments on another mail here comparing different caps that's worth reading.

What really sounds very interesting is how you mix op-amps with discrete transistors. Any design we can have more details of?

There was an article in TNT some time ago on how to pep-up an op-amp using discrete parts. Is it something like that?

On the subject of tantalum caps. I stay away from these critters as far as I can. I believe many of the more modern caps can out perform them. My experience with them is that they are high failure rate items and are primarily used by designers of digital circuits. Perhaps due to habit..

I agree the Burr Brown OPA627 is a fantastic opamp. We use eight of these in parallel in each channel of our moving coil preamps. These preamps have very accurate reproduction and a very low noise floor. The cost is rather high and gets even higher if you hand pick only the best of a lot.

My second choice is the OPA134 or its dual version. It has excellent performance and low cost. It replaced the OPA604 / OPA2604 in all my designs. Eight parralled OPA2134 opamps are used in each channel of my RIAA equalized preamps.

I have some pictures on my web site under projects showing the circuit board layout of these units. The site can be found at There is also a Class a MOSFET power amplifier which is another good example of the use of opamps in high quality audio equipment.

Modern opamps can provide excellent sound quality that is in my opinion indistinguishable from perhaps all but the very best discrete units. The important thing to remember is that opamps must be used in a properly designed circuit in which the impedances are kept as low as possible in order for them to deliver maximum performance. This is likely to be a good rule with no matter what type of device you use.

Where size is important it is hard to beat the compact size of a device like the OPA627. Even with all its associated external parts the entire circuit will still have a small footprint. As a example the size of the circuit board for one of my medium level discrete opamps with all supporting parts is roughly 4 by 5 inches. It would be quite difficult to parallel eight of these fairly large modules in order to lower the overall noise floor to the maximum practical level. The lenght of the interconnecting cables would likely cause serious problems with hum and noise pickup.

We should all remember that there are a great many opamps built into the equipment that recorded our favorite records and CD's. This very likely encludes recordings that are used for listening tests. Many of these opamps are likely to be of older technology and don’t live up to the current higher standards. The quality of these recording may be jugged to be excellent. If opamps are as bad as some has lead us to believe how can this be?

Judging differences between equipment:

When we start comparing equipment to see how it sounds we should also seriously consider the input load provided by the equipment, the output drive capability of the equipment and the cables connected to the equipment we are testing. Listening tests are nice but a good technical evaluation utilizing test equipment should also be performed to insure that problems are properly identified.

John Fassotte
Alaskan Audio

[Edited by alaskanaudio on 10-21-2001 at 01:55 AM]

I do have a number of these heat sinks left over. If you let me know exactly how many you need I can dig them out and verify what I have left over. I also have some that are not so massive that I used on a prior version of the amplifier you mention. These are also flat on one side but have shorter fins. The back plate on these is thinner.

I have found that heat sinks at can be rather difficult to find and I usually end be buying these in larger quantities than I need to get a overall better deal. If you desire to obtain some from me I can send you some pictures of the two types. I can be reached at [email protected] to work out any details.

John Fassotte
Alaskan Audio
I just posted this a day or two ago in a thread on biamping:

...At this point someone invariably comes up with the old chestnut about the sound having been through so many opamps in the recording studio that it won't hurt to use another one (or two, or three...). Muddy thinking. Use of an opamp isn't an all-or-nothing proposition; use of a single opamp, even an antique 741, will not kill the sound, it simply degrades it. Every additional opamp (or discrete device for that matter) degrades the sound a little more, like adding a little more grime to a window. You can still see outside, but the view deteriorates a bit with each speck of dirt. The objective is to keep the cumulative degradation to as low a level as possible. The studio used opamps? Well, rats. But don't punish the signal any's not its fault. Treat it kindly and it will reward you.

Note that I'm not saying that newer opamps haven't improved over the older ones, nor am I denying their obvious convenience, etc. etc. etc. I use them myself for some things.
However, to the extent that it can be claimed that discrete circuitry (whether tube, FET, or bipolar) has any performance edge over opamps, the argument that "the sound's already been through some opamps and it's still pleasant, so it won't hurt to put it through a few more..." sounds pretty empty to me. By that same logic, if you were to find someone who'd been beaten within an inch of his life, it'd be okay to hit him a few more times. Hey, why not? He's already been hit before and he's still alive. Obviously (at least I'd *hope* it's obvious) that line of reasoning is not going to sit well with a fellow who by all rights should be on his way to a hospital. The signal can't heal (unlike our hypothetical beating victim), so it behooves us to damage it as little as possible.


sorry for the lengthy absence...

anyway, i have an opamp + discrete output design up on my temp website at As I note in the webpage, it's a preliminary design, and needs a little work still, but the basic idea is to relieve the opamp of any stress due to the task of driving a load. Mainly, this is thermal in nature, but also we have the issue of rather high open-loop output impedence (which I've rambled on about in a previous thread somewhere). An additional benefit of an external discrete output stage is that we can pull the opamp's output stage into class-A, while providing the capability to deliver much more output power to whatever load is connected. All in all, the concept is working extremely well, and sometime in the future I should sit down and do some more thorough tests.