advantages of potted "discrete" op amps

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
G

Guest

Guest
i see these op amps used in various pre's and have always wondered what was the big deal in potting the things in epoxy or whatever they use. they look cool. but i am wonderiug can you really hear a 75$ op amp over a burr brown 604 or 5532? i am about to build a 16 in summing box for a protools studio, i was looking into the best op amps for high headroom plus coloration. i like trying different things(especially with other peoples money) so i figured i'd blow some money on a few big op amps just for the heck of it. anybody got a suggestion on which ones to get(i need 2)?

side note i need a beefy power supply already assembled on a board ready to go. i could get by with
+- 18, but +-24 or variable supply up to +-24 would be great

if anybody knows a place to get one right away please let me know

i am in la cali, so i need a board from in the us.

thanks
 
Some potting serves as a heat sink for output transistors and is a necessary part of the design. Potting also keeps jerks like me from stealing your design. If you can hear, you can positively hear a difference between one opamp and another. If you are asking if a $75 opamp sounds better, it depends what you want, but all these things most certainly sound different. If they didnt, all gear would sorta sound the same.

contact roll music, they can build a fully assembled supply for you, or at least assemble the board for you and all you have to add is the power transformer. Ive used their power supplies and they are very nice and justin is a really fair guy.

dave
 
thanks for the tips. i suppose my question was not too clear, but apart from hiding a design and heat issues, should there be a discernable 75$ difference in the sound? probably not. just like using a $45 transformer versus an ina134 or ssm2142, the differnce good or bad isn't radical in my opinion, at least as far as my experience with a few transformers.

i was just wondering if there was an increase in headroom or other functional benefit to the potted unit. in terms of sound quality i suppose that is always subjective and requires a listening test to determine.

i guess the next question should be can an op604 run at 24v +- without a heat sink?
 
i was just wondering if there was an increase in headroom or other functional benefit to the potted unit.

Headroom depends mostly on supply voltage. A opamp (discrete or IC) running @ +/-24 V ends somewhere @ 26 dBu. +/-18 V gives you 23 dBu. This is no breathtaking improvement IMO, but one can disagree here.

Headroom is actually not what we (OK, let's say I) want, we want dynamic range, i.e. very low noise and high headroom. A 604 will be around 10 dB (only a rough estimate) noisier than a JE-990 in your application, with the same headroom, so 10 dB less dynamic range here. Even a 5534 will give you slightly higher dynamic range (altough it has lower headroom), because it is much quieter.

Some discrete units have very good driving capabilities - a JE-990 will be very happy driving 300 Ohms, most ICs won't.

Is that what you wanted to hear?

BTW, the very last improvement in sound is always the most expensive one...

Samuel
 
If you ask me, and you have, you can buy an extremely good chip opamp for $7.50. It will sound very good when you use it in a circuit designed around it.
The $75 opamp will sound very good in a circuit designed for it as well.
The two opamps will not be interchangable as most opamps aren't, and don't let the Audiophiles phool you with their articles on the net using the same circuit to "objectively" evaluate two dozens of opamps.
These devices will have different properties. The chip opamp will have:
- More gain-bandwidth
- Less measurable distortion
- Very low DC offset
- Low power consumption
- Limited current drive when it comes to diving low impedance transformers

Most likely the discrete opamp will have:
- A little more headroom typically
- Less gain-bandwidth
- More distortion of sorts that people may prefer
- Some will have outright shitty DC offset properties in variable gain circuits
- High current consumption
- Excellent current drive for low impedance output transforners

Let us assume that the $7.50 opamp will sound 95% as good as the $75 discrete opamp.
Would you ask yourself: "Is that 5% worth to ME an extra 67 dollars?" For some people the answer is yes, for many it is no.

Sound is a matter of taste. I have had listening sessions where one subject preferred the AD797 to the JH990 at a 45dB gain setting because he though it was more accurate, but another person in the room liked the JH990 better because it sounded more alive and could not care how accurate it was.

Cheers,
Tamas
 
I have had listening sessions where one subject preferred the AD797 to the JH990 at a 45dB gain setting because he though it was more accurate, but another person in the room liked the JH990 better because it sounded more alive and could not care how accurate it was.

Let me throw in that the JE-990, as far as I'm informed, is not surpassed by any other opamp regarding noise, THD, overshoot and gain-bandwith (for unity-gain compensated units).

Many other discrete opamps will have worse specs, as Tamas pointed out.

Samuel
 
[quote author="Samuel Groner"]
Let me throw in that the JE-990, as far as I'm informed, is not surpassed by any other opamp regarding noise, THD, overshoot and gain-bandwith (for unity-gain compensated units).
Samuel[/quote]

Even the Forssell 993?
 
[quote="originator"}should there be a discernable 75$ difference in the sound? probably not. just like using a $45 transformer versus an ina134 or ssm2142, the differnce good or bad isn't radical in my opinion, at least as far as my experience with a few transformers.[/quote]

As with lots of stuff, when you compare two things side by side, the difference is often really subtle and one doesnt seem to make much of an impact over the other. When you compare the cumulative results, often the difference is quite large. If you've recorded all your drums, for instance with one type of opamp, it can show really different results than had you recorded everything in your drum buss with another type of opamp, same idea goes for transformers, etc. Sometimes that cumulative difference is obviously worth paying for it when the same opamps compared side by side just only seem to make a marginal difference.

dave
 
Even the Forssell 993?

I would be surprised if a JFET-design were as silent and as low-THDish. The datasheet says 1.5 nV/sqrt(Hz) for the 993, where the JE-990 states 1.13 nV/sqrt(Hz). Distortion specs seem to be lower as well, but no comparable data is available.

AD 797 and LT 1028/1128/1115 are the closest ICs I'm currently thinking of. Any others?

Samuel
 
I would be surprised if a JFET-design were as silent and as low-THDish. The datasheet says 1.5 nV/sqrt(Hz) for the 993, where the JE-990 states 1.13 nV/sqrt(Hz). Distortion specs seem to be lower as well, but no comparable data is available.

Specs and shmecks... :evil:

One more time, these units are not interchangeable in most applications. The JH990 is designed for low voltage noise input, but it has a large amount of current noise. JFET inputs like the JFET993 are just the opposite as they are designed for high impedance systems. Yes, the input FETs used most of the time are very low voltage noise, but they fare better with 2k input and up. Noise and distortion are not even 1/10th of the story.
Also, I have yet to see a discrete opamp sold these days that is unity gain stable. The JH990s I have are pretty close and they are very stable at x2 gain, but the JFET992s I have need at least an x3 gain to be stable.

Tamas
 
I would also like to say that theJH990 is a very low priced discreet opamp at just $45. So, the diference isn´t THAT great (like the mentioned $75)when comparing to modern monolytic opamps...

Look at the prices of AD797.

It´s more like $30 diference...
 
that is a lot of great info. the ad797 shows a max voltage of +-15. i was hoping to use only one supply of +-24 for the whole thing, although i have yet to find a bal line receiver with +-24 capability. the ina134 is
+-18 max. the opa604 shows +-24 as max, so i was going to buffer all the inputs with those chips, then try to sum with some with the same supply rails. maybe even use the 990C at 24v.

i could tap off the +24 and set up some +-18 regulators if i can't find a balance line receiver that will operate at 24, then just use the ina134's. i guess transformers are always an option too, it's only 16 channels.

thanks for the tips guys.

i love this forum!
 
although i have yet to find a bal line receiver with +-24 capability

INA 103? Get some free samples, this stuff is expensive to buy for 16 channels... Would be great as summing amp as well, if you go balanced.

JFET inputs like the JFET993 are just the opposite as they are designed for high impedance systems. Yes, the input FETs used most of the time are very low voltage noise, but they fare better with 2k input and up. Noise and distortion are not even 1/10th of the story.

You tell me that an opamp with 1.5 nV/sqrt(Hz) voltage noise is designed for high impedance systems..? IME FETs start to get quiet @ 10k and up, altough I think FF did a pretty good job with his 993 regarding noise (and believing the data he published).

Agree with you about unity gain stability.

Samuel
 
obviously i wrote that backwards. the bb line drivers (ina134) are showing on the datasheet as +-18, not the 103's which do run at +-24. sorry for the error.
 
Back
Top