OP AMPS

From circuitbending to homebrew stompboxes & synths, keep the DIY spirit alive!

Moderators: luketeaford, Joe., lisa, Kent

User avatar
jimmyedgar
Ultra Violet Vibrator
Posts: 537
Joined: Wed Nov 02, 2011 11:08 pm

OP AMPS

Post by jimmyedgar » Mon Nov 07, 2011 10:57 am

I was hoping to get some good information on types of OP AMPS...

DO's and DO NOT's
Experience with different types and modifications
Sounds qualities, Noise qualities, specific usages

Why are there so many types of opamps?
Does it matter which one I use?
Can something bad happen if I use the wrong one?
What specifications are important in my application?

http://www.synthtech.com/tutorials.html

They were supposed to do a tutorial.. but maybe some of us could have some good information too...

User avatar
Dave Kendall
Set The Controls . . .
Posts: 1222
Joined: Thu Jul 16, 2009 7:20 am
Location: England

Post by Dave Kendall » Mon Nov 07, 2011 12:07 pm

Hi.

I'm no expert, but if there's one opamp to always have lying around for synth circuits, it's the TL071, or TL072 and TL074 2- and 4-channel versions.
It gets used all over the place, and if no opamp type is specified, it's a pretty safe bet.

The TL062 is a lower power version, which can be handy for battery powered circuits. The 082 is abolut the same price as the 072, but it's specs are not quite as good, IIRC......

cheers,
Dave
Want to Buy: KURZWEIL KSP8

"Everything in moderation, including moderation"

User avatar
Neutron7
Oldschool bleeper
Posts: 1950
Joined: Sun May 23, 2010 12:24 pm

Post by Neutron7 » Mon Nov 07, 2011 1:06 pm

I always though the rule of thumb was

The '07x has lower noise figure, so is better for AC signals.
The '08x has lower offset, so is better for DC signals

User avatar
Tombola
Super Deluxe Wiggler
Posts: 1878
Joined: Sun Jan 23, 2011 11:33 am
Location: London, England

Post by Tombola » Mon Nov 07, 2011 2:07 pm

I'd like an op amp with the same Pinout as a TL0*4 but with low distortion and high current output - to better drive a spring reverb.

I built it with a 74 and swapped it with a 84. - I think it sounds better, cleaner, lower distortion, but I don't trust my ears against placebo effect.

Only one quarter drives the spring, the other three are doing regular audio amplification and buffering

User avatar
Neutron7
Oldschool bleeper
Posts: 1950
Joined: Sun May 23, 2010 12:24 pm

Post by Neutron7 » Mon Nov 07, 2011 2:21 pm

i would use an op amp designed as a buffer or headphone amp, but they
do not come in quads.

http://www.ti.com/product/rc4560

http://www.ti.com/product/buf634

http://www.intersil.com/products/device ... pn=HA-5002

User avatar
Dave Kendall
Set The Controls . . .
Posts: 1222
Joined: Thu Jul 16, 2009 7:20 am
Location: England

Post by Dave Kendall » Mon Nov 07, 2011 2:23 pm

I always though the rule of thumb was
The '07x has lower noise figure, so is better for AC signals.
The '08x has lower offset, so is better for DC signals
Thanks - didn't know that...

cheers,
Dave
Want to Buy: KURZWEIL KSP8

"Everything in moderation, including moderation"

User avatar
daverj
Vintage Video Wiggler
Posts: 8514
Joined: Fri Jun 19, 2009 11:09 am

Post by daverj » Mon Nov 07, 2011 2:49 pm

According to TI "The TL07x series are the low-noise versions of the TL08x series amplifiers". But in looking at the specs, the noise figures are the same. The TL08x has higher input offset voltage and input currents for the standard version. The A and B versions have the same as the TL07x.

Input offset voltage (max):

TL08x = 15mv
TL07x = 10mv
TL06x = 15mv

TL08xA = 6mv
TL07xA = 6mv
TL06xA = 6mv

TL08xB = 3mv
TL07xB = 3mv
TL06xB = 3mv

Input offset current at 25C (max) (standard, A, B):

TL08x = 200pA, 100pA, 100pA
TL07x = 100pA, 100pA, 100pA
TL06x = 200pA, 100pA, 100pA

Input bias current at 25C (max) (standard, A, B):

TL08x = 400pA, 200pA, 200pA
TL07x = 200pA, 200pA, 200pA
TL06x = 400pA, 200pA, 200pA

Differential Amplification (typ at 25C):

TL08x = 200,000:1
TL07x = 200,000:1
TL06x = 6,000:1

Unity gain bandwidth:

TL08x = 3Mhz
TL07x = 3Mhz
TL06x = 1Mhz

CMMR (typ):

TL08x = 86db
TL07x = 100db
TL06x = 86db

Supply voltage rejection (typ):

TL08x = 86db
TL07x = 100db
TL06x = 95db

Supply current per amp (typ):

TL08x = 1.4ma
TL07x = 1.4ma
TL06x = 0.2ma

Slew Rate (typ):

TL08x = 13V/us
TL07x = 13V/us
TL06x = 3.5V/us

Equiv. Input noise:

TL08x = 18nv/sqrHz
TL07x = 18nv/sqrHz
TL06x = 42nv/sqrHz

THD:

TL08x = .003%
TL07x = .003%
TL06x = (not specified)


I do not see any electrical reason to select the TL08x. It's only advantage is that it's cheaper. The TL06x is lower power but slower and noisier.

User avatar
EATyourGUITAR
has no life
Posts: 5495
Joined: Tue Aug 31, 2010 12:24 am
Location: Providence, RI, USA

Post by EATyourGUITAR » Tue Nov 08, 2011 12:12 am

the LM386 is more for driving a speaker so maybe that would work on a spring reverb? I think the pinout is completely different though.
WWW.EATYOURGUITAR.COM <---- MY DIY STUFF

User avatar
gwaidan
Super Deluxe Wiggler
Posts: 1938
Joined: Sat Mar 07, 2009 7:26 pm
Location: Sinney, Orstraya
Contact:

Post by gwaidan » Tue Nov 08, 2011 12:48 am

LF412/442 is not much more than the TL0x2, but better specs-much better offset voltage performances and the internal frequency compensation seems to have more of a comfort zone for gains less than 1.

User avatar
fonik
Super Deluxe Wiggler
Posts: 3314
Joined: Fri Sep 19, 2008 5:54 am
Location: Germany
Contact:

Post by fonik » Tue Nov 08, 2011 2:06 am

LM324 swings to the rails.
Big Boss at fonitronik.de
Tech Buddy at Random*Source/Serge

User avatar
fonik
Super Deluxe Wiggler
Posts: 3314
Joined: Fri Sep 19, 2008 5:54 am
Location: Germany
Contact:

Post by fonik » Tue Nov 08, 2011 2:08 am

i always found this link helpful as a hint:
http://users.ece.gatech.edu/~lanterma/sdiy/datasheets/
Big Boss at fonitronik.de
Tech Buddy at Random*Source/Serge

User avatar
Paradigm X
Super Deluxe Wiggler
Posts: 2056
Joined: Tue Nov 02, 2010 5:43 am

Post by Paradigm X » Tue Nov 08, 2011 5:56 am

gwaidan wrote:LF412/442 is not much more than the TL0x2, but better specs-much better offset voltage performances and the internal frequency compensation seems to have more of a comfort zone for gains less than 1.
Thanks was just going to ask about this. In one of Thomas Henry's books he specs an LF444, and i wondered why (he seems to spec every hard to get/expensive component ever, bless him :lol:)

As a general rule why would you chain two opamps (again, TH, but a different circuit in the drum cookbook) rather than just increasing the gain of one? Is there a gain limit per opamp? Is this in the datasheets somewhere, ive looked at the 741 datasheet (op amp TH spec'd) and didnt find anythign which obviously states this.

Cheers

User avatar
fonik
Super Deluxe Wiggler
Posts: 3314
Joined: Fri Sep 19, 2008 5:54 am
Location: Germany
Contact:

Post by fonik » Tue Nov 08, 2011 6:49 am

Paradigm X wrote:As a general rule why would you chain two opamps (again, TH, but a different circuit in the drum cookbook) rather than just increasing the gain of one? Is there a gain limit per opamp?
i did not find such a circuit in the drum cookbook. maybe you took an OTA for an opamp?

anyways, no, there is no such a rule.
re the gain limit: depending on the opamp you would significantly increase the noise ratio when increasing the gain. IIRC not all opamps are suited to do this.
the reason to use two opamps in series is to be able to use these in the inverting configuration (which makes mixing signals easier).
Big Boss at fonitronik.de
Tech Buddy at Random*Source/Serge

User avatar
Paradigm X
Super Deluxe Wiggler
Posts: 2056
Joined: Tue Nov 02, 2010 5:43 am

Post by Paradigm X » Tue Nov 08, 2011 7:59 am

fonik wrote:
Paradigm X wrote:As a general rule why would you chain two opamps (again, TH, but a different circuit in the drum cookbook) rather than just increasing the gain of one? Is there a gain limit per opamp?
i did not find such a circuit in the drum cookbook. maybe you took an OTA for an opamp?

anyways, no, there is no such a rule.
re the gain limit: depending on the opamp you would significantly increase the noise ratio when increasing the gain. IIRC not all opamps are suited to do this.
the reason to use two opamps in series is to be able to use these in the inverting configuration (which makes mixing signals easier).
yeah, youre right sorry, its the white noise source; he says if its too quiet (which it is in my application) to add another opamp stage. its not drawn though.
:)

In this case, noise is not an issue!!! :D

User avatar
daverj
Vintage Video Wiggler
Posts: 8514
Joined: Fri Jun 19, 2009 11:09 am

Post by daverj » Tue Nov 08, 2011 1:14 pm

There are many issues with stability as gains get too high. Another issue is bandwidth. Op amps have a "gain bandwidth product" spec, which means that as the gain goes up, the bandwidth goes down. So, for example an opamp with a 1Mhz GBWP will have a 1Mhz bandwidth at 1x gain, 100Khz at 10x gain and 10Khz at 100 times gain. So if you need to amplify something (like a tiny noise signal) 200 times and want a bandwidth that goes above 20Khz, you need two stages or an amp with higher bandwidth. (but higher bandwidth amps are sometimes more noisy)

That's just one of many things to consider.

User avatar
Paradigm X
Super Deluxe Wiggler
Posts: 2056
Joined: Tue Nov 02, 2010 5:43 am

Post by Paradigm X » Tue Nov 08, 2011 2:16 pm

Cool, thanks.

Reading up a bit more the basic equation is 1+rf/rg; so i cant turn the thing i want to turn down (noise) to 0 anyway.

So i think in what i want to do, which is make the noise gen much louder, but 'turn off-able' i need to make a 2 stage (TL072) opamp, then hang a passive attenuator off the back of it.

Cheers, all, some helpful stuff in this thread :)

Babaluma

Post by Babaluma » Tue Nov 08, 2011 3:50 pm

don't forget discrete opamps, they are supposed to have a nicer sound, but i think they are far more expensive/large/power hungry.

wmonk
Common Wiggler
Posts: 241
Joined: Thu Jun 03, 2010 1:01 pm
Location: Utrecht, The Netherlands

Post by wmonk » Tue Nov 08, 2011 4:53 pm

Babaluma wrote:don't forget discrete opamps, they are supposed to have a nicer sound, but i think they are far more expensive/large/power hungry.
:eek:
discrete just means its not in an integrated circuit package. Of course they can be made to spec, but what makes em have a nicer sound? More nonlinearities?
An opamp in IC package is also just a bunch of transistors, resistors an capacitors.
My Website: kits in Europe
My Weblog

User avatar
fonik
Super Deluxe Wiggler
Posts: 3314
Joined: Fri Sep 19, 2008 5:54 am
Location: Germany
Contact:

Post by fonik » Tue Nov 08, 2011 5:38 pm

daverj wrote:There are many issues with stability as gains get too high. Another issue is bandwidth. Op amps have a "gain bandwidth product" spec, which means that as the gain goes up, the bandwidth goes down. So, for example an opamp with a 1Mhz GBWP will have a 1Mhz bandwidth at 1x gain, 100Khz at 10x gain and 10Khz at 100 times gain. So if you need to amplify something (like a tiny noise signal) 200 times and want a bandwidth that goes above 20Khz, you need two stages or an amp with higher bandwidth. (but higher bandwidth amps are sometimes more noisy)
i am learning a lot from your post, dave. thanks for sharing.

what does (in)stability exactly mean? what is the effect of instability? going to the rails? erratic behaviour?
Big Boss at fonitronik.de
Tech Buddy at Random*Source/Serge

User avatar
daverj
Vintage Video Wiggler
Posts: 8514
Joined: Fri Jun 19, 2009 11:09 am

Post by daverj » Tue Nov 08, 2011 8:20 pm

Instability is a blanket term that covers a lot of different things, primarily AC signal problems such as ringing, overshoot, oscillation, and distortion. Though it can also include other things like DC non-linearities and various forms of drift.

A real life opamp is different from the ideal theoretical opamp. An ideal opamp has infinite open loop gain and infinite bandwidth. Real life ones have neither. In fact they have two bandwidth limitations, small signal and large signal (slew rate). Plus they have input capacitances and resistances, and the physical circuit adds parasitic capacitances and inductances to the theoretically perfect circuit.

In theory an opamp feeds back the input error signals in a closed loop and is a stable amplifier. In reality any frequency limitation or phase shift that happens in the loop causes the signal being fed back to not cancel out the input error signal completely and the differences can cause the output to not be what you were expecting.

As the gain of a circuit goes up, the gain, bandwidth and phase characteristics of the opamp start deforming the ideal opamp model and increasing instabilities.

Here are a couple of application notes about stability, but they can get pretty deep into math pretty quickly.

http://www.ti.com.cn/cn/lit/an/sboa054/sboa054.pdf

http://www.ti.com/lit/an/sloa020a/sloa020a.pdf


If you really want to get deep into it, the guru of opamp stability is Tim Green (from Burr-Brown / TI). Here's a link to a series of articles he wrote. Note that the links on that page are out of order. What is shown as part 8 is actually part 1, part 7 is actually part 2, etc... Part 9 and 10 (the one at the top of the page) are correct.

He takes a different approach, using spice models to model the causes and effects of instability.

http://www.en-genius.net/site/zones/acq ... cqt_092407

User avatar
fonik
Super Deluxe Wiggler
Posts: 3314
Joined: Fri Sep 19, 2008 5:54 am
Location: Germany
Contact:

Post by fonik » Wed Nov 09, 2011 3:42 am

now that's a comprehensive answer, dave! thanks a million.

daverj wrote:In theory an opamp feeds back the input error signals in a closed loop and is a stable amplifier.
isn't this the golden opamp rule?
Big Boss at fonitronik.de
Tech Buddy at Random*Source/Serge

User avatar
daverj
Vintage Video Wiggler
Posts: 8514
Joined: Fri Jun 19, 2009 11:09 am

Post by daverj » Wed Nov 09, 2011 12:06 pm

fonik wrote:
daverj wrote:In theory an opamp feeds back the input error signals in a closed loop and is a stable amplifier.
isn't this the golden opamp rule?
Rules are made to be broken.

:goo:

For filters and distortion circuits you want the feedback to not match the input.

User avatar
fonik
Super Deluxe Wiggler
Posts: 3314
Joined: Fri Sep 19, 2008 5:54 am
Location: Germany
Contact:

Post by fonik » Wed Nov 09, 2011 2:38 pm

:hihi:
Big Boss at fonitronik.de
Tech Buddy at Random*Source/Serge

User avatar
Neutron7
Oldschool bleeper
Posts: 1950
Joined: Sun May 23, 2010 12:24 pm

Post by Neutron7 » Fri Nov 11, 2011 8:57 pm

daverj

Thanks for that great rundown, i was sure misled about the TL08x

I have a few things i used 08xs in instead of 07xBs! ill be swapping those out if i have to service or move them in the case. I got a huge batch of 074B and 072Bs a while ago for super cheap.

free upgrade that takes almost no time, but i probably wont hear :)

User avatar
Dr. Sketch-n-Etch
Super Deluxe Wiggler
Posts: 7842
Joined: Thu Jan 21, 2010 1:46 pm
Location: Vancouver, BC, Canada

Post by Dr. Sketch-n-Etch » Sat Nov 12, 2011 2:34 pm

Paradigm X wrote:
gwaidan wrote:LF412/442 is not much more than the TL0x2, but better specs-much better offset voltage performances and the internal frequency compensation seems to have more of a comfort zone for gains less than 1.
Thanks was just going to ask about this. In one of Thomas Henry's books he specs an LF444, and i wondered why (he seems to spec every hard to get/expensive component ever, bless him :lol:)
Hi. Dr. Sketch-n-Etch here, analog designer guy for Intellijel. I tend to use the TL07x for everything, as (regardless of what the "gurus" say) it tends to be pretty bullet-proof. However, there are times when I prefer an LF44x. The main reason is that it has superior INPUT BIAS CURRENT specs but isn't that much more expensive. This means that the input terminals admit less current than TL07x. The spec for TL07x is 65pA typical, maximum 200pA. The spec for LF44x is 10pA typical, maximum 100pA. (Most opamps sit near their typical values on most specs.) This means that the LF44x is better for things like sample and hold buffers (less current flow means less droop), so I tend to use them in my analog shift register circuit.
Paradigm X wrote:As a general rule why would you chain two opamps (again, TH, but a different circuit in the drum cookbook) rather than just increasing the gain of one? Is there a gain limit per opamp? Is this in the datasheets somewhere, ive looked at the 741 datasheet (op amp TH spec'd) and didnt find anythign which obviously states this.
The best reason to chain two opamps together is to reinvert after an inverting stage to get the signal back to the right way up, and the only good reason to do this is because you require an overall gain of less than one (unity) or are mixing multiple signals together (which you can only do properly with the inverting configuration). If you require a gain of one or more for just one signal, then a single non-inverting stage is best.
Neutron7 wrote:Thanks for that great rundown, i was sure misled about the TL08x

I have a few things i used 08xs in instead of 07xBs! ill be swapping those out if i have to service or move them in the case. I got a huge batch of 074B and 072Bs a while ago for super cheap.
I would be careful there. Most designers tend to use the TL08x for CV processing (presumably because of its slightly better input bias current specs...?), and the TL07x for audio signal processing (because of its lower noise specs). In fact, the noise specifications of the two opamp families are identical according to their datasheets (18nV/root Hz typical at 1kHz). If you pay attention to the "gurus", there are far better opamps for both types of applications, for example from Analog Devices (AD series) and Burr Brown (OPA series). However, these tend to be many times more expensive than the trusty TL series, and I believe that the advantages are dubious at best, particularly for synth-level applications. Danjel and I have used TL072 and TL074 in all of the latest Intellijel analog modules, and they are getting rave reviews. Plus, those who have heard my personal handbuilt modular rig have always commented on how "high fidelity" it sounds, and it's full of TL07x as well. I believe that what you do with an opamp is more important than the opamp itself, as long as you are aware of its limitations. For example, the TL07x comes in three different classes, and these relate to its input offset voltage specs (3, 6, or 10mV maximum). In fact, I tend to find that most of them fall around the 3mV range, and this is small enough for most applications. Some people get overly concerned about this spec for things like CV summers in VCOs, but it really doesn't matter all that much, and opamps with very low input offset voltages often have other problems (since they typically have bipolar input stages with high input bias currents). However, there are times when the input offset voltage can bite you in the ass (as, for example in the servo opamp of a linearized 2164 VCA, if the offset is of a certain polarity, which occurs on average about 50% of the time) in which case you just have to be aware of the problem and compensate for it. Offset voltages are easily trimmed out if they are a problem.
this night wounds time,

Post Reply

Return to “Music Tech DIY”