OpAmps and specs vs preference

[Peter Galbavy]

Been staring at the Caiman mods and wanting to play opamp rolling I started looking at the various chips mentioned here in the forum and on Farnell's site via their search tool. It got me to wondering if anyone has taken a pseudo-scientific approach to tabulating opamps specs and somehow tried to rate their audio performance (which is subjective).

This is from the perspective of looking at slew rates, gain, freq, voltage ranges etc. to see if any patterns emerge.

If no one has, I might try to have a go.

[Peter Galbavy]

Just ordered these from Farnell, initially to meet the minimum £20 credit card order, but most were very cheap (£1-£2 each). Apart from the common ones already mentioned in the various threads I have no idea if the others are safe to use. I will read up, but comments welcome. I wonder what the Burr-Brown ones are like given that they have not been mentioned much here.

ANALOG DEVICES AD826ANZ
TEXAS INSTRUMENTS OPA2132PA
NATIONAL SEMICONDUCTOR LM6132BIN
BURR-BROWN OPA2228PA
BURR-BROWN OPA2244PA
BURR-BROWN OPA2137PA
ON SEMICONDUCTOR SA5532NG
TEXAS INSTRUMENTS LT1013CPG4
TEXAS INSTRUMENTS TL082IPG4

Will build a spreadsheet at some point soon.

[leo]

Should be fun and look forward to your results.

You've got quite a mixture there, not checked them all but some are fet input and others are bipolar , those types alone tend to sound quite different.
We won't bother going into all boring stuff, as long as their stable in the circuit it'll be interesting to see what you find;)
Things can measure the same but don't always sound the same :eyebrows:

[sponge]

This is from the perspective of looking at slew rates, gain, freq, voltage ranges etc. to see if any patterns emerge.

If no one has, I might try to have a go.

As Leo says, you will have fun playing... But specs aren't everything - The humble NE5532 can sound very good and is well down the spec sheet. In my experience of opamp rolling the LM4562 (plastic or metal) will take a lot of beating - It was after all designed for audio, unlike some chips (eg. the THS4032) which are intended for video applications.

Just watch out for stability issues as Leo mentions - the THS4032 is particularly bad tempered and can cause some frustration.

I do like the AD826 and have stuck with it in my TC-7520 (with DAC upgrade) - Analog Devices chips do seem to have their own kind of sound - Quite unlike the Burr-Brown's.

In the end it's what YOUR ears prefer - Just have fun and give each chip a few minutes to warm up settle.

Ken

[Peter Galbavy]

I agree that specs are not everything but I am curious if there are any commonalities...

[sponge]

I agree that specs are not everything but I am curious if there are any commonalities...

Good luck on that - I'll look forward to your findings.

Ken

[Alex Nikitin]

NE5532 from TI is at the moment my favourite dual opamp in a purely analogue applications. I've found that it sounds more natural than LM4562 and the like. For a single opamp there is a better one - NE5534.

Alex

[Peter Galbavy]

Well, first package with 8 pairs of opamps turned up yesterday and I have a ship notice for the last pair also - I guess they were in a backroom box somewhere :)

I have printed out the datasheets for material to digest and will slowly put together a table of specs. In terms of the electronics side I am a neophyte, so time to read up on everything from scratch again (last time was maybe 25 years ago).

I hope I get time to play this weekend but with a friends BBQ on the casrd this may be difficult.

[Peter Galbavy]

While attempting to study the basics I came across this book online:

http://www.analog.com/library/analogDialogue/archives/39-05/op_amp_applications_handbook.html

It's the PDFs by chapter of Analog Devices publication described as: "Op Amp Applications Handbook, edited by Walt Jung (Newnes, 2006), comprises Op Amp Basics, Specialty Amplifiers, Using Op Amps with Data Converters, Sensor Signal Conditioning, Analog Filters, Signal Amplifiers, Hardware and Housekeeping Techniques, and Op Amp History. As the name implies, it covers the application of op amps, but does so on a broader scope. It also covers a host of closely related design topics, making it a formidable toolkit for the analog designer."

Chapter 1 - so far - is very readable and the chapter on op amp history is also very good and appears to be where all the pics and text on the Wikipedia article comes from.

[Codifus]

I agree that specs are not everything but I am curious if there are any commonalities...

The OPA2132 got alot of play on quite a few 7520s here. By the way, the TI OPA2132 *IS* the Burr Brown OPA2132. TI and Burr Brown merged (or TI bought Burr Brown) in 2000.


When i had OPA2132 in my setup, i found it to be laid back with musical timing, really really good with vocals, seducively so even, and a slightly monophonic soundstage, compared to the LM4562. I came away with the impression that the OPA2132 was the one for voice, but the LM4562 was better with instruments. The OP2132 was a very relaxing chip:)

CD

[UV101]

just ordered samples on TI OPA1641 (single) and 1642 (dual).

They are are new "sound PLUS" hi performance JFET offerings. They are literally just out and are on back order due end of the month for the duals and July for the singles. I'm looking forward to giving them a try as I think they may be pretty good.

I'm hoping they are detailed but not so forward as the LME479xxHA I guess I'll have to wait and see!!:rolleyes:

[Peter Galbavy]

Other thing is that on TIs website (Farnell's link to the 2132 datasheet is wrong so I had to go looking) they suggest the 2134 for audio applications. I have added a couple of those to my basket along with a pair of AD828s (again, simply based on researching IC makers own "alternative products" links). I will add a few more random pairs of opamps to build another £20 minimum while I prepare my brain for understanding the most basic electronics.

Still not had a chance to even listen to music with the out-of-the-box opamps and build a sample playlist. This is now looking like a next-weekend-at-the-earliest project :(

Oh, just saw UV101's post - the 1642's I saw but they are only offered as SOIC packages. I am not equiped for soldering and the like - anyone up for producing SOIC-to-DIP adaptors that don't require solder ? :) And Farnell have US stock of the duals now - but for a premium http://uk.farnell.com/jsp/search/productdetail.jsp?id=1786252&Ntt=opa1642

[StanleyB]

I will add a few more random pairs of opamps to build another £20 minimum while I prepare my brain for understanding the most basic electronics.
Have you considered a group buy on the AD826? The 10 piece break price is about half of the one off price. I am willing to take 4 or even 6 pieces of the AD826 if you do a group buy. I'll or course add the postal cost to the repayment.

[Peter Galbavy]

Happy to buy more. You're right of course, it's £5.65 each of £2.61 if you buy 10.

I already have a pair, but happy to buy 10 more to get the £20 min charge thing - I am not great at posting stuff to others, but I will drop you 6 or so 'round one evening next week - and I might take a speaker swicth off you too :)

Anyone else want anything ? Note, me posting stuff out - as above - may be a lazy process. Sorry.

[StanleyB]

I will drop you 6 or so 'round one evening next week - and I might take a speaker swicth off you too :).
I'll be more than happy with that :).

[sponge]

Peter, Take a look at:

http://cimarrontechnology.com/pre-mountedcomponents.aspx

They have some of the more exotic opamps (single and double SOIC) pre-mounted on adapters. Some are not exactly cheap, but I think they offer good value. I have used them several times and found them very quick and reliable.

Ken

[UV101]

you can get those empty browndog adapters here in the uk! :-)

[sponge]

you can get those empty browndog adapters here in the uk! :-)

Yeah, but he said he was not equipped for soldering...:(

BTW, who sells Browndog adapters in the UK?

Ken

[UV101]

Hi ken,

audioupgrades.co.uk stock them but I have also previously ordred from the states without issue.

Ian

[Mr. C]

Perhaps Leo could adapt an Twisted Pair I/V stage or deisgn a fully discret unit to use instead of omp amps, the difference is worth the effort.
I have not used op-amps in any of our designs for close to 4 years now.
Though space I feel in this case may be an issue anyway good luck.

[leo]

The only other niggle is the dac uses just a +15v single rail on the output stage supplies so its a case of finding something suitable with low noise and distortion on such a limited supply, of course an external supply could be added as an option but it then adds more cost
The dacs a voltage out type so something simple can be tried out without any I/V. Something similar to the Pass D1 maybe a laugh , could try both small bipolar or jfets

[StanleyB]

I have tried discrete output stages, including the Burson ones, against a variety of opamps. I am yet to be convinced that a discrete stage would give more bangs for bucks and allow for anywhere near the same level of flexibility as opamps do.

[leo]

I have tried discrete output stages, including the Burson ones, against a variety of opamps. I am yet to be convinced that a discrete stage would give more bangs for bucks and allow for anywhere near the same level of flexibility as opamps do.

With this dac I'd agree! most op-amps work better on the available low voltage supply found in the 7220/Caiman too

[StanleyB]

With this dac I'd agree! most op-amps work better on the available low voltage supply found in the 7220/Caiman too
There is lies the secret of the TC-7520 and Caiman. They are the first serious DACs that allow for the use of some very high-end opamps and regulators without having to worry about where the power feed to these alternative components would have to come from. That sort of flexibility normally cost a fortune to implement.

The AD826 (thanks Leo :) that I have been testing is noticeably superior to the LM4562NA, and about the same as the LM4562HA in terms of bass depth and controlled tops. Only the THS4032CD can do better in the bass, but it is far more expensive, near to unobtainable in perfect working condition, and won't work without an adapter PCB.

[Gazjam]

+1 for the AD826.

I've tried all the "usual suspects" and found this one gets the closest to stripping the hifi away and just lets you enjoy the music.

Didnt get on with the 4032..found it too harsh in the treble.

[Peter Galbavy]

I have ten more AD826's in as of yesterday. Between 6 and 8 are for Stan. Anyone else want the other 2 or 4 ? (I have two for myself from a previous order).

I also ordered samples from TI for the 2107 and 2111 - 4 of each - and I got a shipping notice, so hopefully will be able to test those *and* have saved between £40 - £80 (both those list as about £20+VAT on Farnell's web site).

I now really really need a free weekend to play. I want to build a playlist of about 3-4 tracks, 15 minutes maybe, and get used to it before switching opamps and comparing. But my musical tastes keep changing with my moods. Oh for a scientific mind...

[StanleyB]

I also ordered samples from TI for the 2107 and 2111 -
What's the exact product code? I can't find any info on the 2107 and 2111. There are a couple of letters missing at the beginning I suspect.

[Peter Galbavy]

I meant to OPA2107 and OPA2111 - the Burr Brown things I think. Hang on...

The OPA2107 (which has been mentioned by others):

http://focus.ti.com/docs/prod/folders/print/opa2107.html

the OPA2111:

http://focus.ti.com/docs/prod/folders/print/opa2111.html

I didn't really check spec on the OPA2111 to see if it was suitable, I just added it to the free samples basket while I was there.

[Butuz]

I have ten more AD826's in as of yesterday. Between 6 and 8 are for Stan. Anyone else want the other 2 or 4 ? (I have two for myself from a previous order).

Hi Pete.

I'll take 2 off yoru hands of that is OK! Let me know how much.

Are the AD826 a straight swap into a 7520? I.e no other work needs to be done?

Cheers

[StanleyB]

I didn't really check spec on the OPA2111 to see if it was suitable, I just added it to the free samples basket while I was there.
Let us know how you get on with them. I had a quick check of the specs. The noise figure is a bit too high for me, and the lowish slew rate is also not my cup of tea for use in the Caiman. Should be fine in the TC-7520 though.

[chrism]

Thanks Peter.

Anyone had a go with the AD826 yet. Be interested in any comparison to the LME49720HA as that's what I have currently in the Caiman.

Regards

Chris

[StanleyB]

I have just discovered a hidden gem from the same AD826 stable. It's the LMH6655. The only snags are:
1. It is a surface mount type, so needs to be fitted to an adapter. Tirna Electronics assembled a few for me and their workmanship is fantastic.
2. The max voltage is 12V. So it can only be used in the TC-7520 or in the line output of the Caiman.

Spec wise it is a close match to the AD826, but has an even lower noise voltage. Price wise it is far cheaper by a mile.

Initial listening test compared to the AD826:
Deeper bass.
Less metallic artefacts in the tops.

It sounds better than the THS4032, at about 20% of the price.

Well worth adding to any opamp collection, especially if you are a bass head. Scores highly on reggae.

[Peter Galbavy]

Not one I saw, but then I pretty much ignored all surface mount only opamps in my initial sweep.

Back to the spec vs preference stuff, I have not got my complete stable after FedEx finally delivered the samples that TI sent me weeks ago (they are crap, FedEx, aren't they ?). I will be building a basic spreadsheets with a column for each of the published specs and then some further "subjective" comments columns and I will see where I get over the coming weeks. I have an initial playlist and I will try to work through it a number of times before moving on the the next opamp. I will happily share my results here later for others to update / edit with their own.

[Peter Galbavy]

Attached is my first pass of all the opamps I have "collected" and some basic specs. Upon first inspection - and the point of this exercise is to hypothesise over which specs are useful - the OPA2228 is actually the closest to the shipped LM4562. Similarly, if the AD826 is your drug of choise then the AD828 is "better" is almost every published spec.

Please jump in and tell me what other data to collect (as a mechanical exercise) including trying to normalise some of the specs to particular arbitrary values, like 15V supply and/or 1KHz frequencies. I intend to add columns for subjective comments on performance - opinions very much welcomed.

I have not yet even opened my Caiman to try these. That's the plan for this afternoon :)

[StanleyB]

Great work Peter. But careful with the predicted thunderstorm later today. If it does materialize, best keep the Caiman closed. Stormy weather can create a high static electricity content in the air.

[Alex Nikitin]

Similarly, if the AD826 is your drug of choise then the AD828 is "better" is almost every published spec.

Peter,

AD826 and AD828 are essentially the same chip, however AD826 is unity gain stable and AD828 is only stable for gains +2/-1 or higher. For that reason AD828 can not be used as a universal drop-in replacement for unity gain stable opamps.

Cheers

Alex

[Peter Galbavy]

Thanks for that - now which bit of the spec tell me that so I can express that on the spread sheet ?

[Alex Nikitin]

Thanks for that - now which bit of the spec tell me that so I can express that on the spread sheet ?

For AD828 it is in the general description section of the datasheet. Sometimes it is in the AC parameters table as a "minimum stable gain" or something similar.

Cheers

Alex

[Peter Galbavy]

Thanks!