Hi,
I am after a few opinions on the decoupling requirements for the OPA627, LT1115 and AD8610.
The opamps are for use in a line stage pre amp with either the 1115 or 8610 in the front end and the 627's as the output stage due to their output current capability.
- Rail to rail, rail to ground? Value and types of Caps etc.
All thoughts welcome!
Regards
Sheriff
I am after a few opinions on the decoupling requirements for the OPA627, LT1115 and AD8610.
The opamps are for use in a line stage pre amp with either the 1115 or 8610 in the front end and the 627's as the output stage due to their output current capability.
- Rail to rail, rail to ground? Value and types of Caps etc.
All thoughts welcome!
Regards
Sheriff
Start with 100nF polyester(63V) or ceramic close to supply pins. Connect them down to groundplane if possible, otherwise good signal ground.
Because your load is related to ground (I suppose) the decoupling should also be connected to ground.
You could also use 10 nF + 10 uF alu or tantal.
Take a look at my SMD headphone amp where I have used AD8610/OPA627 decoupled with SMD 100 nF directly down to groundplane.
If you are thinking of a line stage I recommend the AD8610, FET inputs and offset voltage of 10-20uV! My headphone and has DC-offset at the output of circa 80 uV. Besides that it is a really high performance opamp.
Because your load is related to ground (I suppose) the decoupling should also be connected to ground.
You could also use 10 nF + 10 uF alu or tantal.
Take a look at my SMD headphone amp where I have used AD8610/OPA627 decoupled with SMD 100 nF directly down to groundplane.
If you are thinking of a line stage I recommend the AD8610, FET inputs and offset voltage of 10-20uV! My headphone and has DC-offset at the output of circa 80 uV. Besides that it is a really high performance opamp.
The sonic differences from the different decoupling methods?
Hi Per,
Thanks for your quick reply.
I have noticed that alot of the data sheets seem to point toward 100n, yet elsewhere in this forum people swear by 10uF tants.
The 627 datasheet recomends trying 1uF tants as decoupling rail to gnd, as they may improve the dynamic performance of the thing....
Your headphone amp looks really interesting BTW. Thanks for sharing this with us.
You have obviously done a lot of work on this subject, so could you tell me know the sonic differences in the two decoupling methods in an application such as mine.
P.S Can i assume that the 1115 is ot in the same performance league as the 8610 and 627??
Thanks again for your help
Best Regards
Sheriff
Hi Per,
Thanks for your quick reply.
I have noticed that alot of the data sheets seem to point toward 100n, yet elsewhere in this forum people swear by 10uF tants.
The 627 datasheet recomends trying 1uF tants as decoupling rail to gnd, as they may improve the dynamic performance of the thing....
Your headphone amp looks really interesting BTW. Thanks for sharing this with us.
You have obviously done a lot of work on this subject, so could you tell me know the sonic differences in the two decoupling methods in an application such as mine.
P.S Can i assume that the 1115 is ot in the same performance league as the 8610 and 627??
Thanks again for your help
Best Regards
Sheriff
Re: The sonic differences from the different decoupling methods?
Sheriff, I would suggest you to use the small caps like Peranders says, but also with 100uf electrolythics.
The small caps bypass the big ones.
All these caps near the op-amp supply pins (from + to ground and from - to ground).
They like it, the sound gets much better.
For (fast and good) op-amps, it's much more important to have caps near the supply pins than a super-regulated and with big caps supply with the big caps far away and only small caps nearby.
Sheriff said:
Sheriff, I would suggest you to use the small caps like Peranders says, but also with 100uf electrolythics.
The small caps bypass the big ones.
All these caps near the op-amp supply pins (from + to ground and from - to ground).
They like it, the sound gets much better.
For (fast and good) op-amps, it's much more important to have caps near the supply pins than a super-regulated and with big caps supply with the big caps far away and only small caps nearby.
Re: Re: The sonic differences from the different decoupling methods?
The advices seen in datasheets are for "professional" use. The opamps needs some capacitance in order to work as intended. I suspect that 1-10 nF is an absolute minimum and 100 nF is some sort of "average" value. If the voltage regular is nearby the need for extra decoupling decreases. Tantal is not good to use because tantal is a rare metal. Think of the resources on earth.carlosfm said:
op-amps like food nearby
I once changed 2.2uf tantalum caps (they were very near the op-amps pins) with 100uf electrolythics bypassed with 0.1uf ceramics.
The deam preamp started to sing.
Before that, my friend was not convinced.
We had tryed OPA2604, OPA2132, LM6172.
My friend (a high-end salesman, with very good ear to detect subtle sounds) was still prefering the original NE5532s. (!)
Yes, the NEs happen to work well even on bad circuits.
Manufacturers use them because they're cheap, and they work on cheap circuits.
Then I gave the circuit my "special treatment".
It was a Micromega integrated amp.
After my cap "cure" the guy didn't whant to believe what he was hearing.
We tested all the op-amps again, and he picked the OPA2132s.
Yes, they sound better.
But well bypassed.
And this was done without expensive or fancy caps.
The guy is impressed with the sound of his amp now.
As I said, to get the best out of good and fast op-amps (some people call them fancy) you'll have to do it properly.
Or else the worst op-amp may play better.
This is my practical experience.
I once changed 2.2uf tantalum caps (they were very near the op-amps pins) with 100uf electrolythics bypassed with 0.1uf ceramics.
The deam preamp started to sing.
Before that, my friend was not convinced.
We had tryed OPA2604, OPA2132, LM6172.
My friend (a high-end salesman, with very good ear to detect subtle sounds) was still prefering the original NE5532s. (!)
Yes, the NEs happen to work well even on bad circuits.
Manufacturers use them because they're cheap, and they work on cheap circuits.
Then I gave the circuit my "special treatment".
It was a Micromega integrated amp.
After my cap "cure" the guy didn't whant to believe what he was hearing.
We tested all the op-amps again, and he picked the OPA2132s.
Yes, they sound better.
But well bypassed.
And this was done without expensive or fancy caps.
The guy is impressed with the sound of his amp now.
As I said, to get the best out of good and fast op-amps (some people call them fancy) you'll have to do it properly.
Or else the worst op-amp may play better.
This is my practical experience.
Hi!
I´ve just bought a couple of 4,7 uF "dry elyts" to use together
with 100nF (ceramics) for bypassing AD797, OPA627 and NE5532:s.
The catalogue from the swedish distributor (ELFA) claims that
the dry elyt:s have the same characteristics as tantalums, but
lacking the negative side of tants (spontaneous short-circuiting).
The "dry elyts" sure looks like tants, just looking at them,
but I would like to know if there is anyone here that has tried
them out and what your opinion on these devices are?
/ eplpwr
I´ve just bought a couple of 4,7 uF "dry elyts" to use together
with 100nF (ceramics) for bypassing AD797, OPA627 and NE5532:s.
The catalogue from the swedish distributor (ELFA) claims that
the dry elyt:s have the same characteristics as tantalums, but
lacking the negative side of tants (spontaneous short-circuiting).
The "dry elyts" sure looks like tants, just looking at them,
but I would like to know if there is anyone here that has tried
them out and what your opinion on these devices are?
/ eplpwr
We now pause for this commercial message
"The advices seen in datasheets are for "professional" use. The opamps needs some capacitance in order to work as intended. I suspect that 1-10 nF is an absolute minimum and 100 nF is some sort of "average" value. If the voltage regular is nearby the need for extra decoupling decreases. Tantal is not good to use because tantal is a rare metal. Think of the resources on earth."
I didn't know that the IC knew it was in a non- professional application . Good audio high end design pratice usually even goes beyond the recomendations in the data sheet. The distance from the reglator is not really not the issue since these high frequency decoupling caps. and the regulator usually has a fairly large value electrolytic and is inductive at RF frequencies. IMO use of Tantalums on a few hundred preamps is not going to have an effect on the tantalum mining. I don't use them because they sound bad. Some of the polymerized organic semiconductor types like the POSCAP series would be a better sounding cap with simular ESR and low indutance to the solid tantalum types.
http://www.semicom.co.uk/manufacturer/sanyo/sanyo_oscon/oscon_construction.html
Peranders! When are you going to stop thread jacking and abusing the forum to sell your products? I have seen you do it here of on other audio forums several times. I have complained to the moderators and now I am complaining to you in public. I and some others I have spoken to are finding this Spam to be very obnoxious.
Its bad enough to have a link to what appears to be mainly a commercial web site for you products. The design appears to be an inferior copy of the highly reguarded Meta 42 design.
http://tangentsoft.net/audio/meta42/
BTW you might want to point out that the AD 8016s are designed for a maxim supply voltages of +/- 13 V and not 15 Volts.
Back to the topic:
Ceramic and polyester don't sound good for bypass caps and I would use good polypropylenes like EROs or Panasononic from DikiKey. With the right electrolytic cap you don't need additional small caps. See gratuitous picture with post.
We now return to our noncomercial ( at least it use to be) forum.
"The advices seen in datasheets are for "professional" use. The opamps needs some capacitance in order to work as intended. I suspect that 1-10 nF is an absolute minimum and 100 nF is some sort of "average" value. If the voltage regular is nearby the need for extra decoupling decreases. Tantal is not good to use because tantal is a rare metal. Think of the resources on earth."
I didn't know that the IC knew it was in a non- professional application . Good audio high end design pratice usually even goes beyond the recomendations in the data sheet. The distance from the reglator is not really not the issue since these high frequency decoupling caps. and the regulator usually has a fairly large value electrolytic and is inductive at RF frequencies. IMO use of Tantalums on a few hundred preamps is not going to have an effect on the tantalum mining. I don't use them because they sound bad. Some of the polymerized organic semiconductor types like the POSCAP series would be a better sounding cap with simular ESR and low indutance to the solid tantalum types.
http://www.semicom.co.uk/manufacturer/sanyo/sanyo_oscon/oscon_construction.html
Peranders! When are you going to stop thread jacking and abusing the forum to sell your products? I have seen you do it here of on other audio forums several times. I have complained to the moderators and now I am complaining to you in public. I and some others I have spoken to are finding this Spam to be very obnoxious.
Its bad enough to have a link to what appears to be mainly a commercial web site for you products. The design appears to be an inferior copy of the highly reguarded Meta 42 design.
http://tangentsoft.net/audio/meta42/
BTW you might want to point out that the AD 8016s are designed for a maxim supply voltages of +/- 13 V and not 15 Volts.
Back to the topic:
Ceramic and polyester don't sound good for bypass caps and I would use good polypropylenes like EROs or Panasononic from DikiKey. With the right electrolytic cap you don't need additional small caps. See gratuitous picture with post.
We now return to our noncomercial ( at least it use to be) forum.
Attachments
Re: We now pause for this commercial message
META 42 is a nice design, among the better ones (with nice looking pcb's, important!) by also you have different tastes and you have never seen or listened to my stuff. Don't you guys recommend listening first before make judgements? I do my stuff and I don't have look at what the DIY community think it's right.
Fred, I don't mind if you show us your DIY stuff (can't remember I haven't seen any).
Fred, I have not yet written any building directions or circuit descriptions. Of cource I will point out very clearly why you should use +-13 V at the most and what happens if you use +-15 V. BTW: My amp lives still with +-15 V and AD8610, will probably do so if not broken yet. AND Fred AD8610 + BUF634 is very very very good, according to my ears despite the fact that you like the META42 better.
Best regards from Per-Anders Sjöström
Boy, you are in the mood today. You are not complaining about this thread http://www.diyaudio.com/forums/showthread.php?s=&threadid=12186 and Chipcomp?...and AKSA....Tangent. and ... more? I don't mind if people at hobby level or slightly above try to promote their things nicely.Fred Dieckmann said:
META 42 is a nice design, among the better ones (with nice looking pcb's, important!) by also you have different tastes and you have never seen or listened to my stuff. Don't you guys recommend listening first before make judgements? I do my stuff and I don't have look at what the DIY community think it's right.
Fred, I don't mind if you show us your DIY stuff (can't remember I haven't seen any).
Fred, I have not yet written any building directions or circuit descriptions. Of cource I will point out very clearly why you should use +-13 V at the most and what happens if you use +-15 V. BTW: My amp lives still with +-15 V and AD8610, will probably do so if not broken yet. AND Fred AD8610 + BUF634 is very very very good, according to my ears despite the fact that you like the META42 better.
Best regards from Per-Anders Sjöström
Re: We now pause for this commercial message
Fred, you know what I mean! If you want a professional and commercial advice how would you design an amp with help from an opamp it's hard motivate the costs and the benefits if I would do as you suggest. It's quite a difference if you make it for fun or want to make money. If I (at work) want to use some chip I want practical and sensible advices which have some connection to reality.Fred Dieckmann said:
I'm no authority on the subject by a long shot, but 100µf seems overkill, and certainly if you have a ground reference nearby you should use two caps...one from '+' to gnd, and one from '-' to gnd.
The large electrolytics will not bypass the higher frequencies well. Normally, I use a 4.7µf electrolytic and a .18µf non-inductive stacked metal film cap parallelled on each rail to ground. In tight spots, I have simply used one single .18µf film cap from '+' to '-' (not referenced to ground) and had no problems. But the first method seems to be preferable.
The large electrolytics will not bypass the higher frequencies well. Normally, I use a 4.7µf electrolytic and a .18µf non-inductive stacked metal film cap parallelled on each rail to ground. In tight spots, I have simply used one single .18µf film cap from '+' to '-' (not referenced to ground) and had no problems. But the first method seems to be preferable.
I think we can agree on that this capacitor is first of all there for stability reasons....only! Slow opamps like 741 don't need much but a opamp with speed over 10 V/us need something and over 100-1000 V/us it's crucial, very very important! The goal with this capacitor is to create sufficiently low impedance of the power supply voltage. This is the main reason.
10-100 nF is required, 10-100uF normal, >100 uF seems to be a little overkill but doesn't hurt. The main thing is to create low impedance over 1 MHz.
10-100 nF is required, 10-100uF normal, >100 uF seems to be a little overkill but doesn't hurt. The main thing is to create low impedance over 1 MHz.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.