C
chessrock
Banned
Robert D said:
Isn't Discreet Opamp sort of an oxymoron? Kinda' like Jumbo shrimp? Or Assistant Supervisor? Or Freezer Burn? How 'bout Student Teacher?
.
Isn't Discreet Opamp sort of an oxymoron? Kinda' like Jumbo shrimp? Or Assistant Supervisor? Or Freezer Burn? How 'bout Student Teacher?
.
D
dementedchord
Psychotic State alumni ch
chessrock said:Great explanation.
So would it be safe then to assume that, if a given unit wasn't utilizing a "more expensive topology" ... didn't have a transformer anywhere in the path, and used essentially lower-grade caps and other supporting circuitry ... would it be difficult, then, to justify a significantly higher cost for said unit?
Or, more accurately ... could you at least see where one might raise that question or suspicion, without being totally out of line? Do you think it would be a valid concern, without even having listened to the unit or having direct experience with it? Hypothetically speaking, of course.
.
not sure i'm comfortable being painted into this corner for (suspected)purpose of advancing your agenda/argument on someother thread... but if one assumes that they also took similar approaches to all areas of design... your concerns would be justified it seems to me...
tkingen
Djembes Rock
LeeRosario said:specify what chip?
Thats like asking, "if all cars had a V8, would they all run the same?"
Sorry, not trying to be a smart ass.
Through the discussion on this thread I've learned that the chip is not the primary determination of the sound of the preamp.
G
Gilliland
New member
I'm not sure I'd accept that as an accurate statement. How about this as a substitute: "The chip is not the only determination of the sound of the preamp."? In truth, the chip probably IS the "primary" determination. The real question here is whether the "primary" provides 60% of the preamp's sound or 95% of the preamp's sound. And the answer is probably: Both. 60% in some cases, 95% in others. (Obviously, these numbers are hypothetical.)tkingen said:
For example, we know that the M-Audio Octane, the Studio Projects SP828, and the Grace M801 are all eight channel mic pres based on the same basic Burr Brown chip. But the Grace uses discrete resistors in its gain control, avoids electrolytic capacitors in its signal path, and has a hefty power supply. I'm sure it has other circuit virtues as well. As a result, it costs more and presumably (I haven't actually heard it) sounds better. The Octane and SP828, on the other hand, probably aren't that far apart. For those units, the chip may well be 95% of the sound. For the Grace, other factors come into play, so the 60% figure may be closer to the mark. (Again, all numbers are hypothetical.)
M
MCI2424
New member
Gilliland said:
Read my post above. In a minimal design (where all the weight is thrown on the IC) there is not much flexability and, hence, the chip will be about 95% of the sound. If the designer used his noggin and designed a great frontend and set the ICs parameters to compliment his goal. the IC would probably be in the 10-20% of the sound. These ICs are only amplifiers with added circuitry to allow much more of a configuration than say a standard Op amp.
In reality, there are instrumentation amplifier ICs that cost over $100 each and go into the gigahertz range. I don't think these would be very good for audio, but they have their applications.
Bottom line: It is up to the designer if he uses good judgement, good parts in a great design as to how these ICs will sound. Discreet component design offers even more flexability as well as the ability to match critical components for lower noise, crosstalk, crossover distortion etc.
These ICs can have a night and day difference in sound depending on who uses it.
R
Robert D
New member
chessrock said:
No, not really. Op-Amp is short for operational amplifier, and really just means it's a collection of circuitry designed for signal amplification, condensed into a module with an output and a differential input, for which the gain can be easily set by use of negative feedback. Hell, there were vacum tube op-amps before the transistor was invented, let alone ICs.
Oh, and you forgot the oxymoron of the day, military inteligence.
"The Army reserve....... money for college, in case you come back."
M
MCI2424
New member
chessrock said:
A discreet component means a seperate singular OR a set of components.
A discreet op-amp is an integrated circuit but is a seperate packaged component in itself. You can make a instrument amplifier with a discreet op-amp and its supporting components OR integrate the whole thing into another integrated circuit (like what we are talking about here).
The integrated instrument amplifier has its supporting caps, resisors etc. etched permanently into the silicon and depending on how many "hooks" the designers provide you with to allow access to their circuit, the flexability will be nill to vast. That is why high-end designers like discreet component designs. They can manipulate every single parameter (except to get into the discreet active components themselves) to achieve the desired results as well as hand match complimentary components.
A IC based design can be as good depending on HOW flexable the IC is to begin with.
Neither can be definitively dismissed without looking at the design and the quality of components/build and most importantly, sound.
As a side note, specifications are much ridiculed by some here. Specs are the #1 thing that designers look for before actually designing any circuit because of the obvious fact that there are hundreds on components out there in every catagory you can think of. How do I design anything when I am told to start with an op-amp? There are op-amps with bandwidths form a few hundred hertz to well over 1 gigahertz. Where do I start?
Specifications. Data sheets. These data sheets give all parameters of the component in question from teperature limits to noise, bandwith and many more test parameters. Save designers tons of time. Sorting through a few op-amps for an ideal design is better than having no specs, having to test every single parameter and sorting through a few hundred.
This goes for every component in any system. So, specs on anything is only a starting point to help you figure out what set of things MIGHT do the job. It is up to the designer to actually test his/her design for optimum performance.
I know, waaaay too long winded.
G
Gilliland
New member
Yes, I've read your post. It sounds like you and I are in pretty close agreement, even though we've used different numbers to make our points. That's fine, the difference between mic pres is a pretty subjective area anyway. The numbers/percentages that we're throwing around are all pretty meaningless. In reality, if two mic pres sound only 20% (your number) the same, or only 60% (my number) the same, one of them is broken. The actual difference between mic pres is MUCH smaller than that - no matter what the design.MCI2424 said:
The important point is that we agree that some designs are going to be very generic in their use of a given chip, while others are going to show considerable differences, all depending on the choice of support circuitry.
evm1024
New member
A bit of history
Not an oxymoron.
Opamps is (as has been noted in this thread) operational amplifier. Opamps were originally designed as a differential amplifier ( + and - inputs) that allowed the creation of different mathmatical operations in analog computers. By changing the topology and values you could make them sum, multiply, divide, integrate, and so on. And they were very fast (realtime in many cases) and much faster than the digital computers of the day. A bitch to program too.
As it turns out having a differential input was very useful for all kinds of applications outside of analog computers and thus they worked their way into audio.
By my direct experience there is a significant difference in sound of opamps. But it is of the last 5 percent and in many cases the last 0.1% type of difference. And that difference between opamps only really shows up when you keep the same topology and only change the opamp. Differences in topology make large differences, as do differences in the passive components.
A simple example is the change of the output coupling electrolytic cap to a non-elecrolytic of high quality (and perhaps huge size) will make a significant change in the amps "color".
Monolithic opamps have wide variation in componet values onthe die. Some devices have laser trimming to better specs and thus are closer to the designed spec. Descrete opamps allow the opamp maker to hand select the individual componets in the opamp. (I've got 8 Jensen 990 that I built on the bench waiting testing, Most of the parts were hand selected) Thus the DOA represents a hand crafted optimized opamp. IF DOA were built like monolithic opamps we would use 20% parts and make many and then bin out the out of spec parts.
The designer of a preamp gets to select opamp, topology, component specs in order to achieve a specific set of design goals. We can alter the sound of the amp by substituting other components. The nuance here is that all of the mix of board layout, components, topology has a sweet spot where they all work in better harmony or perhaps not so harmony. A great design is in that sweet spot. (more zen like design left for your imagination)
Time to go.
chessrock said:
Not an oxymoron.
Opamps is (as has been noted in this thread) operational amplifier. Opamps were originally designed as a differential amplifier ( + and - inputs) that allowed the creation of different mathmatical operations in analog computers. By changing the topology and values you could make them sum, multiply, divide, integrate, and so on. And they were very fast (realtime in many cases) and much faster than the digital computers of the day. A bitch to program too.
As it turns out having a differential input was very useful for all kinds of applications outside of analog computers and thus they worked their way into audio.
By my direct experience there is a significant difference in sound of opamps. But it is of the last 5 percent and in many cases the last 0.1% type of difference. And that difference between opamps only really shows up when you keep the same topology and only change the opamp. Differences in topology make large differences, as do differences in the passive components.
A simple example is the change of the output coupling electrolytic cap to a non-elecrolytic of high quality (and perhaps huge size) will make a significant change in the amps "color".
Monolithic opamps have wide variation in componet values onthe die. Some devices have laser trimming to better specs and thus are closer to the designed spec. Descrete opamps allow the opamp maker to hand select the individual componets in the opamp. (I've got 8 Jensen 990 that I built on the bench waiting testing, Most of the parts were hand selected) Thus the DOA represents a hand crafted optimized opamp. IF DOA were built like monolithic opamps we would use 20% parts and make many and then bin out the out of spec parts.
The designer of a preamp gets to select opamp, topology, component specs in order to achieve a specific set of design goals. We can alter the sound of the amp by substituting other components. The nuance here is that all of the mix of board layout, components, topology has a sweet spot where they all work in better harmony or perhaps not so harmony. A great design is in that sweet spot. (more zen like design left for your imagination)
Time to go.
evm1024
New member
MCI2424 said:
A discrete opamp is not an integrated circuit. Link to dictionary...
http://www.computeruser.com/resources/dictionary/definition.html?lookup=7647
S
sdelsolray
Active member
ADK released a preamp (last year I think) that allows you to swap opamps and transformers. Obviously, the remaining circuitry does not change between swaps.
Dan Richards over at the Listening Sessions has posted comparison samples for those that are interested.
Dan Richards over at the Listening Sessions has posted comparison samples for those that are interested.
M
MCI2424
New member
evm1024 said:
In our industry it is acceptable to define a discreet op-amp as a op-amp with all of its parameter hooks available to the user. The pins would allow single resistors, capacitors etc. to be selected and connected to get the desired functions. Some op-amps only have a +, - and power supply connections (like comparators). Like everything, the current definition is alittle behind the times as "discreet" was coined in like 1960s (Technically, the formal definition (used to be) of "discreet component" was only applied to passive components. By this definition, a transistor or a diode is not a discreet component, therfeore, it would be impossible to build any amplifier with only "discreet" components.)
where ICs were totally new. In today's electronics where a single op-amp is in IC form, it is considered "discreet" in relation to an Intel Pentium 5.
The two are vastly different in that a Pentium 5 is a total system and a single op-amp can be used for almost any imaginable purpose.
But, it is a matter or semantics to a degree. John Hardy believes he designs a "discreet" pre-amp despite the fact that he has a non-discrete op-amp in his designs.
EDIT: Just a sample of what I am talking about from a paper written about test systems:
"...signal conditioning circuitry This subsystem filters, amplifies, attenuates or formats signals sent between
the plant and the software execution computer. This principally consists of discreet components
(resistors, capacitors, op-amps, etc.)...."
Last edited:
evm1024
New member
Techno Babble
Well Andy,
In the late 60's and early 70's I worked in the wafer fab at Tektronix and then at Intel. We were building monolithic and Hybrid opamps with 2 GHz unity gain opamps for Tek scopes at that time. I do not recall that there was any confusion between a potted discrete opamp and a hybrid or monolithic opamp.
In addition I recall no distinction between passive components and active components in the defination of discrete vs integrated.
So to me, based on 40 years experience in our industry your assertion is false.
You can make opamps out of tubes, discrete transistors and passives, a hybrid solution or a monolithic integrated circuit. (or any combinnation of the above) but you cannot make an integrated circuit out of discrete active or passive components. IC are by definition on a single die.
Of course, All those folks at Tek and Intel might be wrong.....
-E
I assume that you will not take my word for it. I do recall tube computers and my first transistor. I did work with an analog computer in school and more than a few discrete transistor based machines. My first IC based machine was a PDP-8I, Guess what the I stands for. (Well that and a 370)
Here is an excerpt from this URL
http://www.answers.com/topic/operational-amplifier
History
The operational amplifier was originally designed to perform mathematical operations by using voltage as an analogue of another quantity. This is the basis of the analog computer where op-amps were used to model the basic mathematical operations (addition, subtraction, integration, differentiation, and so on). However, an ideal operational amplifier is an extremely versatile circuit element, with a great many applications beyond mathematical operations. Practical op-amps, based on transistors, tubes, or other amplifying components and implemented as discrete or integrated circuits, are good approximations to the ideal.
Op-amps were originally developed in the vacuum tube era, where they were used in analog computers. Op-amps are now normally implemented as integrated circuits (ICs), though versions with discrete components are used when performance beyond that attainable with ICs is required.
PS
Tutor's tip: It would not be "discreet" (to show tact or reserve) to try to join "discrete" (consisting of unconnected or distinct parts) and competitive departments.
MCI2424 said:
Well Andy,
In the late 60's and early 70's I worked in the wafer fab at Tektronix and then at Intel. We were building monolithic and Hybrid opamps with 2 GHz unity gain opamps for Tek scopes at that time. I do not recall that there was any confusion between a potted discrete opamp and a hybrid or monolithic opamp.
In addition I recall no distinction between passive components and active components in the defination of discrete vs integrated.
So to me, based on 40 years experience in our industry your assertion is false.
You can make opamps out of tubes, discrete transistors and passives, a hybrid solution or a monolithic integrated circuit. (or any combinnation of the above) but you cannot make an integrated circuit out of discrete active or passive components. IC are by definition on a single die.
Of course, All those folks at Tek and Intel might be wrong.....
-E
I assume that you will not take my word for it. I do recall tube computers and my first transistor. I did work with an analog computer in school and more than a few discrete transistor based machines. My first IC based machine was a PDP-8I, Guess what the I stands for. (Well that and a 370)
Here is an excerpt from this URL
http://www.answers.com/topic/operational-amplifier
History
The operational amplifier was originally designed to perform mathematical operations by using voltage as an analogue of another quantity. This is the basis of the analog computer where op-amps were used to model the basic mathematical operations (addition, subtraction, integration, differentiation, and so on). However, an ideal operational amplifier is an extremely versatile circuit element, with a great many applications beyond mathematical operations. Practical op-amps, based on transistors, tubes, or other amplifying components and implemented as discrete or integrated circuits, are good approximations to the ideal.
Op-amps were originally developed in the vacuum tube era, where they were used in analog computers. Op-amps are now normally implemented as integrated circuits (ICs), though versions with discrete components are used when performance beyond that attainable with ICs is required.
PS
Tutor's tip: It would not be "discreet" (to show tact or reserve) to try to join "discrete" (consisting of unconnected or distinct parts) and competitive departments.
Last edited:
R
Robert D
New member
Yep. I've been working in Silicon Valley since '77, and I've never seen transistors and diodes not referred to as discrete components. They are active vs passive discreet components - caps, resistors, etc. belonging to the latter group. And Integrated circuit has always meant you put the whole circuit on a wafer.
M
MCI2424
New member
evm1024 said:
OK, but, there is plenty of high-end stuff floating around that uses op-amps (IC) and are still billed as being totally discreet designs, so, right wrong or indifferent, you won't know until you open the box.
This includes Neve also (arguably the best there is). That is a fact.
Semantics aside. An Op amp is considered by to be discreet by M.I.T's standards anyway. (re-read my post above, I edited in a quote from a M.I.T paper). Since the current IC based Op-amp is the only package you can get (there are no tube op-amps available at this time) and it is such a staple in today's electronics, it stands on it's own as a single discrete device.
And, no, you are not wrong and they were not wrong back in the 60s.
Time marches on and today we have ASICS, Mixed signal ICs with up to several million devices etched in a single die. Discreet means something different in 2007.
BTW: You are right, my spelling of discrete is wrong. Oh well.
Go here for a heart attack
:http://www.bitpipe.com/tlist/Discrete-Component-Interconnection.html
Last edited:
evm1024
New member
Language usage
Say, I thought that American English was your native language .... LOL
dis·crete Pronunciation (d-skrt)
adj.
1. Constituting a separate thing. See Synonyms at distinct.
2. Consisting of unconnected distinct parts.
3. Mathematics Defined for a finite or countable set of values; not continuous.
to recast: This principally consists of distinct components.
So Andy if we use the logic that you have shown us those resistors, capacitors and op-amps are all discrete devices?
You may have a number of discrete (distinct) IC on a board along with an ASIC and one discrete (made from individual components) op-amp.
Oh By the way. I read that app note (an1126) and no where in the app note is the word discrete (in any spelling) used.... That app note is from Cypress Semi, I did a little consultinig work with one of their design groups in Tigard in 1999 or perhaps 2000.
English is a fun language where a word can have many meanings. "Discrete opamp" has a specific meaning. ASIC designers would not know one if it was sitting on their desk. And marketing people have less of a clue.
-E
MCI2424 said:
Say, I thought that American English was your native language .... LOL
dis·crete Pronunciation (d-skrt)
adj.
1. Constituting a separate thing. See Synonyms at distinct.
2. Consisting of unconnected distinct parts.
3. Mathematics Defined for a finite or countable set of values; not continuous.
to recast: This principally consists of distinct components.
So Andy if we use the logic that you have shown us those resistors, capacitors and op-amps are all discrete devices?
You may have a number of discrete (distinct) IC on a board along with an ASIC and one discrete (made from individual components) op-amp.
Oh By the way. I read that app note (an1126) and no where in the app note is the word discrete (in any spelling) used.... That app note is from Cypress Semi, I did a little consultinig work with one of their design groups in Tigard in 1999 or perhaps 2000.
English is a fun language where a word can have many meanings. "Discrete opamp" has a specific meaning. ASIC designers would not know one if it was sitting on their desk. And marketing people have less of a clue.
-E
D
dementedchord
Psychotic State alumni ch
he does have a bit of a point as to the eveolution of the language unfortunately.... i recently jammed with some guys and wass asking about day jobs durring a break... the guitarist claimed to be a tech... i said i had been one myself and was comfortable working at a component level.... he says he is too... if it needs a new drive or vid board he's ok with that... sic...
M
MCI2424
New member
evm1024 said:
I do agree. For fun I asked 6 electrical engineers about discrete op-amps and all of them but 1 picked the IC based Op-amp as it. 1 guy (about 63years old) had the opposite opinion.
Things are changing and definitions are changing with the staggering amount of absolutely huge IC based chips out there. Analog as a industry is becoming rare in the industry and more "chip based" designers are pumped out of colleges.
Anyway, the original question asked in the thread has been answered long ago.
dgatwood
is out. Leave a message.
MCI2424 said:
Just because 5/6 don't know what discrete means, that doesn't mean it's okay to call an IC a discrete device.
A discrete op amp is, by definition, one composed of individual transistors and other parts. Discrete and integrated are antonyms, therefore, by definition, nothing integrated can be discrete, and to use the word discrete in that way is a blatant abuse of proper terminology.
See also:
http://en.wikipedia.org/wiki/Integrated_circuit
IC said:
Note the usage of "ICs over discrete circuits" indicates that no member of the former set of devices can be a member of the latter set.
http://en.wikipedia.org/wiki/Discrete
Discrete said:
Now, in theory, you could blur the lines and have an op amp on a chip that might qualify as somewhat discrete. You'd have to build it with a series of individual components wired together inside the chip with microscopic wires like the ones that connect the circuit inputs to the pins, then sealed into a single unit with epoxy.... That said, I can't imagine any chip manufacturer in their right minds building such a device.
M
MCI2424
New member
dgatwood said:Just because 5/6 don't know what discrete means, that doesn't mean it's okay to call an IC a discrete device.
A discrete op amp is, by definition, one composed of individual transistors and other parts. Discrete and integrated are antonyms, therefore, by definition, nothing integrated can be discrete, and to use the word discrete in that way is a blatant abuse of proper terminology.
See also:
http://en.wikipedia.org/wiki/Integrated_circuit
Note the usage of "ICs over discrete circuits" indicates that no member of the former set of devices can be a member of the latter set.
http://en.wikipedia.org/wiki/Discrete
Now, in theory, you could blur the lines and have an op amp on a chip that might qualify as somewhat discrete. You'd have to build it with a series of individual components wired together inside the chip with microscopic wires like the ones that connect the circuit inputs to the pins, then sealed into a single unit with epoxy.... That said, I can't imagine any chip manufacturer in their right minds building such a device.
They did that all the time in the 70s.