OpAmps for DIY Bass & Guitar Pedals

...and how to do so...
User avatar
khanyz
KVRAF
1556 posts since 16 Jul, 2004 from Deepest Yorkshire

Post Thu Feb 16, 2012 1:46 am

pj geerlings wrote:They are $4.90 each so I want to practice my surface mount soldering skills before tacking them down to the DIP-8 adapters that I (also) bought
You can get adapter clips so that you don't have to solder. I picked up some off ebay, but haven't tried them yet. I've stuck with DIP as that's what I'll use in the final build.

Image

Cheers.
I miss MindPrint. My TRIO needs a big brother.

xtp
KVRian
841 posts since 4 Jun, 2006 from The West Coast of New Zealand

Post Fri Feb 17, 2012 12:12 am

khanyz wrote: ...but I've gone through so many that I've really lost track.
Did you notice much difference soundwise with all the opamps you tried.

With the stuff I am doing I have prototyped with TL072 but I have purchased OPA2134 to go on the pcb versions... based upon nothing more scientific than the flip of a coin.

User avatar
khanyz
KVRAF
1556 posts since 16 Jul, 2004 from Deepest Yorkshire

Post Fri Feb 17, 2012 12:55 am

xtp wrote:Did you notice much difference soundwise with all the opamps you tried.
Yes, for clean signals, the noise/distortion levels did vary. However, I found if you put a discrete, fixed gain amp/buffer stage before them, this was reduced quite a bit. I use a simple FET design so that I can set the input impedance too. The J. Tillman one is a good starting point if you're not familiar with them. It's based on a standard design, tweaked for preference.

Mine's a little difference as I have +ve and -ve rails, as well as ground. It uses a Voltage Multiplier to accept a 9V input and give me the supply levels I want. There's also a bypass to use 18-36V input via a voltage regulator.

Rather obviously (IMO), the voltage supply to the opamps, particularly rail-to-rail, is important to overall performance. I spent quite a bit of time sorting out a supply circuit I was happy with, or approaching happy with (TBC). You can get some weird effects if you mess with it though, like sending signals, or LFOs though them.

Tonally, I couldn't really tell much difference. The frequency responce were similar, as was phase. The surrounding components had a much bigger effect on them.

For distorted/clipping signals, I did noticed difference in tone and how well the opamp reacted to high levels. I think this is where it gets subjective as better/worse isn't as quantifiable, as with noise.

Overall, there was a general bang/buck correlation for clean signals, but none really if you wanted to drive the opamp for character. I guess this is in line with the general opinion.

Of the 2 you mentioned, I prefer the OPA2134 for clean signals as it seems less prone to noise/disortion. If I remember correctly, it also accepted distortion well, but I wasn't particularly taken with the clipping tone.

The TL072 is a good workhorse type, which does well for either.

Hope it helps.
I miss MindPrint. My TRIO needs a big brother.

User avatar
Uncle E
KVRAF
10941 posts since 22 Nov, 2000 from Southern California

Post Fri Feb 17, 2012 10:17 am

xtp wrote:Did you notice much difference soundwise with all the opamps you tried.
I replaced a TL072 with a Burr Brown 2604 in a pedal I'm manufacturing (OCD-type circuit) and the sound instantly had considerably more highs, more lows, and less noise. It's such a significant change that I'm releasing the pedal with the opamp socketed so that people can make up their own minds over which they prefer.

xtp
KVRian
841 posts since 4 Jun, 2006 from The West Coast of New Zealand

Post Fri Feb 17, 2012 2:13 pm

Hi, thanks for the replies.

In my case I am building a BandPass Filter and a EQ specific for my playing style. Both are running 12v dc.

The OpAmps are used solely as input/output buffers and for makeup gain as both units have passive circuits.

I did wonder about tonal varience between Linear and Jfet, but not enough to prompt me to buy comparable parts. I certainly had no issues with the TL072 and really purchased the higher spec OPA2134 opamps out of curiosity.

:)

imcrazedandconfused
KVRer
9 posts since 18 Apr, 2012 from Germany

Post Sat Apr 21, 2012 1:16 pm

Hi guys,
from building and upgrading studio stuff, my conclusions are the following:

4558 - sounds very gritty. Most circuits have been using NE5532 and when cost calculation took place, these are oftenly used as a cheap alternative.
NE5532 are the first upgrade and are mostly a direct drop in that drops noise very significantly, good enough for stompboxes, if noise is what you want to lower.
MC33078 - a very clean and fast opamp as a replacement for the above ones for a very good price.
The BurrBrown are more on the beefy side of sound.
You can replace the ones mentioned with these, the other way round does not(!) always work.
TL072 - cheap alrounder, good enough for alot of circuits, depending on implementation.
OPA2134 - a considerable step up, the pre-decessor OPA2132 was a bit more stable and one of the most used upgrade opamps-
OPA2604 - very beefy and musical sounding opamp, at a very reasonable price.
For stompboxes, i wouldn't upgrade any higher than this.
OPA2227 - less slew rate, precision opamp, very quiet and quite pricey, may not fit in every position, but if it does, it can sound wonderful
The OPA404 has been mentioned (quad opamp). It is still available in PDIP, a very clinic/precise opamp, but very pricey.
OPA627 - for a long time this one is the flagship single opamp. Very pricey, but can also sound very good
AD797 - the direct competitor to the OPA627, very little noise, more neutral in sound

There are a couple others, that might be a good bet to experiment with, but the abive ones have been used in many audio circuits as direct replacements without the need to change the environmental parts for additional stabilization.
Just take care to upgrade your power supplies, additional power consumption adds up with every used chip. Take the consumption from the datasheets, multiply by the number of chips used add a safety margin and see if your psu is on the safe side.

Kind regards,
Martin

User avatar
khanyz
KVRAF
1556 posts since 16 Jul, 2004 from Deepest Yorkshire

Post Sat Apr 21, 2012 11:17 pm

@imcrazedandconfused: Have you tried the OP27GPZ? I really like it for boosts/mild overdrive. It is quiet and neutral, but you can really add character to it with the right feedback circuit. I'm getting to be a bit of a germanium diode nut.

Also, it seems TI have acquired National. I got an email saying they are merging my accounts.
I miss MindPrint. My TRIO needs a big brother.

imcrazedandconfused
KVRer
9 posts since 18 Apr, 2012 from Germany

Post Sun Apr 22, 2012 9:57 am

Hi,
no, i didn't test it myself, but from the datasheet it might be one of those chips that might sound really good in some applications, similar to the OPA227 range, a low slew rate, high precision opamp.
I didn't know about the aquiration of National, thanks for the head up.
I also didn't mention all the good chips from AD because i simply don't have them at hand, but have quite some stock of TI chips at home.
Also, i mentioned only those that are available in PDIP format still.
For effect pedals I would not go so far to go thru the hassle of using SMD chips on adapters, but if one is not afraid of that, there are lots of chips available, that are suitable for audio as well.
And in case one isn't afraid of trying to tame down the really fast more modern opamps e.g. video opamps, then there is an incredible range of opamps one my experiment.
But if this is really what someone is after - go ahead everyone...

Best regards,
Martin

imcrazedandconfused
KVRer
9 posts since 18 Apr, 2012 from Germany

Post Sun Apr 22, 2012 3:19 pm

OK, some rules of thumb for selection of opamps for audio:
Nearly all circuits are based on NE5532/TL072 type of opamps, these are most commonly used to build prototype circuits in the development stage of products.
Look out for the following specs in the datasheet;

Bandwidth
8MHz is the ballpark for commonly used audio opamps.
You don't really want to go slower than that. I would especially never go any lower than 6MHz, but without re-design of the circuit you may go up to maybe 20MHz without stabilization problems. If you go higher than that, e.g. with more modern audio and video opamps, you most probably have to tame them down with additional parts, or they may oscillate unpleasantly.

Slew Rate:
Normally, you want a high slew rate. A high slew rate preserves transients and you most obviously want this. It adds considerably clearness of the signal. Rule of thumb: You do not want to go below 12µV/s. Higher is better. This makes the MC33078 a very good opamp: despite it's low pricing, it is a very good allround audio upgrade opamp. If I remember it right, it has a slew rate as high as around 20µV/s or better.
OPA227 and OP27 have really low slew rate, but are very low noise and precise within their specs, and simply sound good (i checked, and i really did a mic pre once with OP27 and it sounded good, i just forgot it, but have looked up).

You also want preferably low noise opamps.
You may also want headroom, i.e. the possibility to feed them with more than +-15V to reach a good output swing.
You might want to buy unity gain stable opamps. They are simply more flexible for experimentation.

You may also want to check power consumption of the circuit after any upgrade.
Some of the better opamps are really hungry for power. They might not be suitable for battery-powered devices. Check it out before testing. Cross check with the originally used opamps and check if your PSU is capable to drive the circuit. No opamp can sound really good if the power supply is out-of-spec, i.e. not beefy enough or not quite enough (filtered). On a bench supply it may sound great, but remember, a bench supply has lots of power and is extremely stable. Go for a powerful and stable PSU for your stompboxes. Batteries are very quiet supplys, but with high consumption circuits, they may be exhausted in a view hours - and you don't ever want this to happen in an on-stage situation. Personally I would place my bet on a good powerful multi-device supply and have one spare for live and studio situations. Maybe even built one that is failure safe and switches to the spare trafo and circuit in case of failure with a relay. You may build something like that from scratch, using beefy trafos, some fast rectifier diodes (UF400x series are fast and good rectifier diodes and a step up from commonly used bridge rectifiers and 1N400x series) and some good elyts for filtering,use an adjustable voltage regulator, they have much better specs and can be setup for exact voltages, LM78xx and LM79xx (1A) are not really what you want, you may easily have voltage diffs up to 20%, LM317/337(max.1.5A) or LM350T (up to 3A!)are much better, voila a rock stable power supply.

Thanks for reading and best regards,
Martin

User avatar
khanyz
KVRAF
1556 posts since 16 Jul, 2004 from Deepest Yorkshire

Post Mon Apr 23, 2012 12:33 am

Some very useful info, thanks.

The only things I would add is that Bandwidth is usually expressed as GBW (or Gain Bandwidth). The bandwidth does vary with gain and the GBW usually refers to Unity Gain (if a value of G/Gain isn't given). You can play with this for effects, just be careful of the output levels.

Also, You might want to start with +-9V (18V supply). You can then step it up to +-18V for mains adapters, if you need the headroom (obviously checking the components can handle it). For pedals/batteries, you can either use 2x9v, a single 9v with voltage multiplier or drop down to 9V supply. These will have an effect on the sound though.

Most of the stuff I've been doing has been with 18V. I did try a Bass Preamp (Trace Elliot GP7ish) and it had to run at 36V to get the best out of it. 9V is convenient, with only minor reduction in sound quality (IMO).

Cheers,
Nigel
I miss MindPrint. My TRIO needs a big brother.

obiwanjacobi
KVRian
519 posts since 19 Jul, 2007 from Netherlands

Post Thu May 31, 2012 3:57 am

imcrazedandconfused wrote:[..]

Slew Rate:
[..] Rule of thumb: You do not want to go below 12µV/s. [..]
My guess is that you mean 12V/µS ..?

:hihi:
Grtx, Marc Jacobi.
Blog | VST.NET | GMPI.NET | MIDI.NET

imcrazedandconfused
KVRer
9 posts since 18 Apr, 2012 from Germany

Post Fri Jun 01, 2012 11:17 pm

SURE! Thanks, Marc, for clearing this up.
I like to stay corrected! Sometimes my fingers are too fast...and erroneous.
Unfortunately this is also the case when playing bass and guitar ;-)

Best regards,
Martin

PS: There ARE occasions, where I choose opamps that have low slew rate - because there are at least *some* that sound good and musical nevertheless!
One example is OPA 2227 (2.3 V/µs), the unity gain stable version of OPA2228 (11 V/µs). Both sound excellent in many places, where a standard audio opamp fits.
OPA2228 is ideal at higher gain, but tends to oscillate at low gain, there I use the OPA2227 at near unity gain in buffer stages etc.
Not a cheap opamp but well worth trying.

xtp
KVRian
841 posts since 4 Jun, 2006 from The West Coast of New Zealand

Post Tue Jun 26, 2012 7:10 pm

khanyz wrote: However, I found if you put a discrete, fixed gain amp/buffer stage before them, this was reduced quite a bit. I use a simple FET design so that I can set the input impedance too. The J. Tillman one is a good starting point if you're not familiar with them. It's based on a standard design, tweaked for preference.
Hi, eventually after about 3 years of experimenting I ended up going down this road. I have removed every op-amp & transistor, and replaced it with a JFet stage.

It took me a long time to get there but triode emulators, mu-amps and Jfet buffers are far more pleasing to my ears than any design a Op-amp has to offer.

All discrete here now.

User avatar
khanyz
KVRAF
1556 posts since 16 Jul, 2004 from Deepest Yorkshire

Post Tue Jun 26, 2012 11:44 pm

^ It's usually the only entirely discrete gain stage I have. Also, I've started taking a parallel out from it, adding an OEP transformer for a quick balanced out. I'm still testing it with some simple filters on it too.
I miss MindPrint. My TRIO needs a big brother.

User avatar
khanyz
KVRAF
1556 posts since 16 Jul, 2004 from Deepest Yorkshire

Post Thu Aug 09, 2012 3:11 am

FYI
New Japan Radio has delivered "PSpice Macro Model" as a privilege of
registration members on our site.
We will make it generally available starting from 9th of this August,
and we will terminate the membership registration service at the
same time.

There after all visitors can download the macro model without
registration.

We will delete the membership registration information rapidly.

We also provides "Application Simulator", a web based circuit design
tool to select proper external components for your application
requirements.

After the WebSIM(R) technology user registration in
the Transim website,you can use the simulator.

Thank you.
I miss MindPrint. My TRIO needs a big brother.

Return to “DIY: Build it and they will come”