| Radian |
Hi,
I put this post in the chip amp forum because Carlos and some others mentioned using this OpAmp.
Please ignore any wrong grammar and spelling, this is my second language.
I have a very limited understanding of electronics, so forgive me if my questions seem to be silly.
I want to upgade my Marchand XM1 active crossover with the OPA2132.
As Carlos mentioned a wile back this is one of his favorite OpAms if it is bypassed with 47-100uf cap and a smal ceramic cap.
As I understant, this bypass is to be done on the negative and the positive rails of the OpAmp.
Each crossover bord uses three OpAms: one for the buffer and two for filtering. This makes all together 12 OpAmps for a three way stereo unit.
Since all the OpAmps are fed by one psu, do I need to use a cap after the psu with 12 times the requierd capacitence for one chip, or is it better to still have 12 induvidual caps close to the chips. The other question would be, if I already used plenty of capacitence for filtering at the psu, does that spoil the induvitual bypassing.
I appreciate any help,
Klaus |
|
|
| peranders |
No worries about your english. It's good!
Good engineering pratice is to have 100 nF/63 V polyester or ceramic close to each supply pin of the opamps and down to ground. You can also have for instance 10 nF + 10 uF electrolytic capacitors but this OPA2132 or OPA2134 aren't so sensitive. It becomes more sensitive the faster the opamp gets.
Your last question: Strive alsways to have decouplings caps as close as possible to the "current consumer".
Doesn't your filter already have sufficient decoupling? |
|
|
| carlosfm |
| quote: | Originally posted by peranders
You can also have for instance 10 nF + 10 uF electrolytic capacitors but this OPA2132 or OPA2134 aren't so sensitive. It becomes more sensitive the faster the opamp gets.
....
Doesn't your filter already have sufficient decoupling? |
Who told you so?
Just by looking at the datasheet?:clown:
Since I've registered here that I've been insisting on how these op-amps must be used.
The way you are "advicing" this member to use them they sound very bad.
Thanks for that.
This is also one of the major flaws of your "SMD Gainclone".
The way you use op-amps I should advice that the best results you can have is with the NE5532, they are very happy this way and when used this way they sound much better than modern and much superior op-amps.
Keep up the good work.
Oh, and please use the AD8620, a much inferior op-amp than the OPA2132, but sounds better the incorrect way you use it in audio applications, of course. |
|
|
| macboy |
| quote: | Who told you so?
Just by looking at the datasheet?
Since I've registered here that I've been insisting on how these op-amps must be used.
The way you are "advicing" this member to use them they sound very bad.
Thanks for that.
This is also one of the major flaws of your "SMD Gainclone".
... |
Carlosfm,
Your post could have been more constructive. Instead of just telling others that you think they are doing something wrong, you could tell them how you think they are doing it wrong. That is, tell them what you think is the right way. You could have saved yourself a lot of typing by just saying, "In my experience, the OPA2132 sounds better to me when I used 100 uF bypass caps instead of much smaller ones.". That would be a constructive addition to the thread. |
|
|
| moving_electron |
I'm trying some different op amps but keeping the same bypass caps before changing the cap values again.
OPA2132 sounds rather good to me and is the current winner with 47uF bypass caps per power pin. Has 22uf between rails though at a single point on the board and 22uF from each power rail to ground where power comes on board. Each has a 100nF ceramic as well. 100nf ceramic between each op-amp supply pin.
I may leave this bypassing alone for a while though as I am quite happy with it for now. Sounds good in the preamp and the phono preamp. Crisp, clean but not harsh or "clinical".
Other opamps to try though. |
|
|
| Radian |
Hello Carlos,
Please tell me how you would go about in my case.
Since I have 12 OpAmps on one power supply do I still need to put 47uf close to each pin of the OpAmps or can I use 12X 47uf at the PSU and just 100nf close to the chips?
I actually don't have the 2132 but the 2134.
Cheers,
Klaus |
|
|
| carlosfm |
| quote: | Originally posted by Radian
Hello Carlos,
Please tell me how you would go about in my case.
Since I have 12 OpAmps on one power supply do I still need to put 47uf close to each pin of the OpAmps... |
Yes.
| quote: | Originally posted by Radian
... or can I use 12X 47uf at the PSU and just 100nf close to the chips? |
No!:bawling:
| quote: | Originally posted by Radian
I actually don't have the 2132 but the 2134. |
The same.
Enjoy.:cool: |
|
|
| carlosfm |
| quote: | Originally posted by moving_electron
OPA2132 sounds rather good to me and is the current winner with 47uF bypass caps per power pin. |
You are getting there.
You are on the right path.
Take those 47uf caps and put just 100nf caps.
You will be shocked.
You will say: **** op-amp!:bawling: |
|
|
| moving_electron |
| quote: | Originally posted by carlosfm
Take those 47uf caps and put just 100nf caps.
You will be shocked.
You will say: **** op-amp!:bawling: |
Are you saying to take off the 47uF and just use the 100nf caps?
My experience with the OPA2604 was that things got better as I added capacitance on the power pins although at that point I only had 10uF on them.
The OPA2132 sounded better to me than the OPA2604 when I tried them at 47uF. I made a number of changes to the system as I went from 10uF to 47uF.
Looking forward to being shocked! Mentally, not physically.
:) |
|
|
| carlosfm |
| quote: | Originally posted by moving_electron
Are you saying to take off the 47uF and just use the 100nf caps? |
No!
I was telling you to try because it will sound very bad!
The word was C R A P, but it was censured.:confused:
| quote: | Originally posted by moving_electron
My experience with the OPA2604 was that things got better as I added capacitance on the power pins although at that point I only had 10uF on them.
The OPA2132 sounded better to me than the OPA2604 when I tried them at 47uF. I made a number of changes to the system as I went from 10uF to 47uF. |
The OPA2132 is better than the OPA2604, and as I said lots of times here, I recommend between 22 and 100uf capacitance on each supply pin to ground.
You can also bypass the electrolythics with small 100nf multi-layer ceramics, but this is not so critical. Note: the LM6171/2 needs them.
Also a 100~330nf poly cap directly from V+ to V- PSU pins on the op-amp. Usually you can do that under the circuit.
:att'n: This gives me very good results everywhere, not only with the OPA2132.:att'n:
For me, it's a rule.
After so many tests I made years ago, this is what works best.
Always.
Also, if you don't do it this way, forget the OPA627, it will NOT sound good.;) |
|
|
| digi01 |
another way is BP caps,a couple of small bp caps will do.
I use 2X6.8uf BP caps(Solen,WIMA-MKP10 film caps)at the regulator power supply part and 0.1uf close to each chips.
the sound of opa2604 is sweet,but I like 2134 more. |
|
|
| moving_electron |
| quote: | Originally posted by carlosfm
No!
I was telling you to try because it will sound very bad!
|
Ah, makes sense now. I would have been shocked if it sounded better without the 47uF caps. I have two nearly identical phono preamps, one without much capacitance on the power supply pins and one with. I definitely liked the one with the capacitance more. Same for my preamp.
| quote: | Originally posted by carlosfm
The OPA2132 is better than the OPA2604, and as I said lots of times here, I recommend between 22 and 100uf capacitance on each supply pin to ground.
You can also bypass the electrolythics with small 100nf multi-layer ceramics, but this is not so critical. Note: the LM6171/2 needs them.
Also a 100~330nf poly cap directly from V+ to V- PSU pins on the op-amp. Usually you can do that under the circuit.
:att'n: This gives me very good results everywhere, not only with the OPA2132.:att'n:
For me, it's a rule.
After so many tests I made years ago, this is what works best.
Always.
|
I am a pretty enthusiastic believer in this and have been using it since your original comments way back when with good results.
| quote: | Originally posted by carlosfm
Also, if you don't do it this way, forget the OPA627, it will NOT sound good.;) |
One of these days I will be listening to the OPA627 in the Regulated BIGC. Slow but steady progress... |
|
|
| KBK |
OK, carlos...why do you wnat to modulate the the PS rails against each other? I'm confused about the logic of the small poly cap between the two power rails, right at the op-amp. After all, these ARE global fedback devices, and the speed of that feedback IS Finite....
Maybe to control PS micro-modulation under an AC audio signal load? (feedback micro-timing delay issues?) Please supply an explaination, if you so desire. I'm very curious. |
|
|
| Stabist |
| quote: | Originally posted by carlosfm
The OPA2132 is better than the OPA2604, and as I said lots of times here, I recommend between 22 and 100uf capacitance on each supply pin to ground.
You can also bypass the electrolythics with small 100nf multi-layer ceramics, but this is not so critical. Note: the LM6171/2 needs them.
Also a 100~330nf poly cap directly from V+ to V- PSU pins on the op-amp. Usually you can do that under the circuit.
:att'n: This gives me very good results everywhere, not only with the OPA2132.:att'n:
For me, it's a rule.
After so many tests I made years ago, this is what works best.
Always.
|
So if I understand correctly - (btw: I'm trying to make a pretty decent MC chipbased preamp - so I'm opened for many good ideas) - those "between 22-100uF" capacitors are soldered as close as possible to the chip +/-legs and gnd and that 100-330nF poly cap are positioned between +/- legs ...
Btw - in my test phase I'm trying the circuitry with NE5532 and OPA2604 (because I had those at home) - but which chips are recomended for low noise MC preamp? Is opa2132 good choice? Or is it better to look elsewhere - eg LT1028.
Thanks |
|
|
| Gaucho |
| quote: | Originally posted by carlosfm
Also a 100~330nf poly cap directly from V+ to V- PSU pins on the op-amp. Usually you can do that under the circuit.
|
:D
I'd like to mention the fact that I've achieved very good sonic results trying electrolytics (usually a 10uF with a 0.1uF film) in this position. The results very with the op-amp being used. |
|
|
| peranders |
| quote: | Originally posted by carlosfm
The way you are "advicing" this member to use them they sound very bad.
This is also one of the major flaws of your "SMD Gainclone".
|
| quote: | Originally posted by carlosfm
Oh, and please use the AD8620, a much inferior op-amp than the OPA2132, but sounds better the incorrect way you use it in audio applications, of course. |
| quote: | Originally posted by moving_electron
IOPA2132 sounds rather good to me |
Carlos, don't forget that your opinion is not facts and certainly not in others ears, so if you change your statements to more like "I think", or "my opinion is.." I think we will have a better discussion climate. Besides even though my Gainclone has design flaws you have never heard it I'll suppose. We know one who was very unsatisfied. We have also heard that everyone else are _very_ satisfied. So what does this mean?
This means maybe that there are some exaggerations. "Really bad" isn't maybe so bad. If you say "sounds really bad", how bad is that? |
|
|
| Upupa Epops |
| Bypassing depend on concrete connection, design of PCB, type of opamps and several other conditions, but Carlos maybe like " end solutions" as " Holy Inkvisitor " - " No doubt, no problems " :cool: . |
|
|
| carlosfm |
| quote: | Originally posted by peranders
Carlos, don't forget that your opinion is not facts and certainly not in others ears, so if you change your statements to more like "I think", or "my opinion is.." |
Take it always as my oppinion, try it and listen.
If I say "I think it sounds better" is because I'm not sure.
As I'm sure it sounds better, I say so.
The negative "climate" may be generated by others, not me, as I'm just reporting my experiences.
Everyone is free to take it or leave it, try it or not, but please discuss with reasons.
Curiously, those who discuss are always those who don't bother to try what I'm suggesting...
Sometimes I ask myself why do I bother to post on audio forums...:dodgy: :confused: |
|
|
| NealG |
Because we like you Carlos! :D
I have tried the 100nF between + and – supplies in a DAC with the OPA2134 and can’t say I liked it. It pushed the midband forward highlighting detail at the expense of the treble which lost air and ambience.
Also regarding recommendations for opamps in phono preamps I always understood it was better to use a bipolar design as they generally have a much lower voltage noise than FET. I understood FET’s are better suited to MM due to a lower current noise. |
|
|
| carlosfm |
| quote: | Originally posted by NealG
Also regarding recommendations for opamps in phono preamps I always understood it was better to use a bipolar design as they generally have a much lower voltage noise than FET. I understood FET’s are better suited to MM due to a lower current noise. |
Absolutely.:angel: |
|
|
| peranders |
| .... but it means not so much in the real world, I mean if you compare to the noise from the vinyl itself. |
|
|
| carlosfm |
| quote: | Originally posted by peranders
.... but it means not so much in the real world, I mean if you compare to the noise from the vinyl itself. |
I have lower noise with the OPA637 than the AD797 on my MM phono pre.
Also, the OPA637 sounds much better, but that's another story.
It's not that you notice when listening to the music, you won't.
It's just that I'm very picky and always test with the volume to the max and the ear on the speaker.:D |
|
|
| Fixup |
hi guys,
This is my first post to this great forum. I make world's smallest headphone amps (SuperMacro etc).
For the reason of Carlos' V+ to V- bypassing, you can find it in TI (BB)'s datasheets, such as OPA690:
"In addition to the usual power-supply decoupling capacitors to
ground, a 0.1 µF capacitor is included between the two power-
supply pins. In practical PC board layouts, this optional-added
capacitor will typically improve the 2nd-harmonic distortion
performance by 3dB to 6dB."
3dB to 6dB - that's a huge difference!
I found this trick by my own and used it in my headphone amps, then I was very excited to find that Carlos and TI also discovered this!
www.fixup.net |
|
|
| lcleven |
| The schematic for the XM1 shows 10uf, 50V electrolitic caps at each pin 8. The XM1 uses LF 353 opamps. I have an XM44 which uses OPA2134 opamps and has 0.1uf monolithic axial ceramic caps at each pin 8. |
|
|
| carlosfm |
| quote: | Originally posted by lcleven
I have an XM44 which uses OPA2134 opamps and has 0.1uf monolithic axial ceramic caps at each pin 8. |
The Arcam CD7SE has the first cap at more than 5cm from one of the op-amps.
The other one is at some 10cm from the caps.
The caps are shared between the two op-amps.
Nothing shocks me anymore.:clown:
:bawling: |
|
|
| Gaucho |
| quote: | Originally posted by carlosfm
The Arcam CD7SE has the first cap at more than 5cm from one of the op-amps.
The other one is at some 10cm from the caps.
The caps are shared between the two op-amps.
Nothing shocks me anymore.:clown:
:bawling: |
That's why it sounds muddy! Excellent sleuthing. :D |
|
|
| pinkmouse |
| I totally agree with Carlos, in fact, I have even gone up to 470uf for chips like the AD811. |
|
|
| peranders |
| quote: | Originally posted by carlosfm
I have lower noise with the OPA637 than the AD797 on my MM phono pre.
Also, the OPA637 sounds much better, but that's another story.
It's not that you notice when listening to the music, you won't.
It's just that I'm very picky and always test with the volume to the max and the ear on the speaker.:D |
How much difference if we talk dB's? Is there any practical difference, I mean when you _not_ are have the volume at full throttle? |
|
|
| carlosfm |
| quote: | Originally posted by peranders
How much difference if we talk dB's? Is there any practical difference, I mean when you _not_ are have the volume at full throttle? |
I only start listening to the "hiss" at around 3/4 volume.
The difference in noise from the two op-amps in this application is quite evident. |
|
|
| DC Dave |
| With regards to the original question, is there a difference between bypassing with ceramic versus polyester film? In both situations using 100nF in combo with 100uF electrolytic. Or is it just a matter of personal preference? |
|
|
| theChris |
as with everything in audio it is science vs art.
scientifically, the ceramic will probably be the best bet, ceramics have better high frequency properties (the whole reason a single cap isn't good enough in the first place).
artistically, some argure that the film-cap will allow for better sound. |
|
|
| Fixup |
In this case:
* The cap is not in the signal path, no need to use a film one
* It works at frequency far above audio, ceramic is better for high freq.
* It acts as short-circuit, ceramic's non-linearty has no chance to show up (low freq. is short-circuited by a large electrolytic or tantanium cap).
Therefore, ceramic. |
|
|
| peranders |
| quote: | Originally posted by carlosfm
I only start listening to the "hiss" at around 3/4 volume.
The difference in noise from the two op-amps in this application is quite evident. |
Yes, but how loud is it then? Party volume? |
|
|
| Upupa Epops |
| P-A, let it be, it is lost fight. When somebody use boxes as headphones ;) , certainly will hear differences... |
|
|
| moving_electron |
| quote: | Originally posted by DC Dave
With regards to the original question, is there a difference between bypassing with ceramic versus polyester film? In both situations using 100nF in combo with 100uF electrolytic. Or is it just a matter of personal preference? |
This page at Anolog Devices has some interesting information related to cap types.
http://www.analog.com/library/analo...versary/21.html
The high inductance characteristics of polypropylene caps are interesting.
I thought there was a more detailed paper following the same flow as this page on the AD sight but I cannot find it at this point. |
|
|
| moving_electron |
In the link there is a comparison chart at the bottom. I am curious about the comment in the monolythic ceramic (multilayer) "disadvantages" box.
Does "poor stabilty" refer to stability of the capacitance value with tempature change? Or do they mean that the capacitance value is prone to shift over time?
CarlosFM. I think you have mentioned that you didn't notice any difference between Polypropylene or ceramic for cap bypassing at the pins (V+ to ground, V- to ground) but that you prefered Polypropylene for the V+ to V-.
Is the poly preference for V+ to V- poly just a personal preference for what you had on hand or did you really hear difference vs. mono (multilayer) ceramic? |
|
|
| carlosfm |
| quote: | Originally posted by peranders
Yes, but how loud is it then? Party volume? |
It's party volume already at 1/2 volume.
| quote: | Originally posted by Upupa Epops
P-A, let it be, it is lost fight. When somebody use boxes as headphones ;) , certainly will hear differences... |
Did someone ask you something?
Why do you keep threadjacking all around?:confused: |
|
|
| Upupa Epops |
| Big brother is watching you... :clown: |
|
|
| pinkmouse |
| quote: | Originally posted by Upupa Epops
Big brother is watching you... :clown: |
Yes, he is watching...;) |
|
|
| Sinbios |
Sorry to bring up a somewhat old thread, but what is the problem with sharing the electrolytic with all the IC's when I bypass with 0.1uF ceramics close to each IC? On OPA690 datasheet, it says:
[In addition to the 0.1uF caps,] Larger (2.2ƒÊF to 6.8ƒÊF) decoupling
capacitors, effective at lower frequencies, should also be
used on the main supply pins. These may be placed
somewhat farther from the device and may be shared
among several devices in the same area of the PC board.
So, is it OK to use two electrolytics for all the IC chips, or does each chip need its own electrolytics? |
|
|
| carlosfm |
| quote: | Originally posted by Sinbios
Sorry to bring up a somewhat old thread, but what is the problem with sharing the electrolytic with all the IC's when I bypass with 0.1uF ceramics close to each IC? On OPA690 datasheet, it says:
[In addition to the 0.1uF caps,] Larger (2.2ƒÊF to 6.8ƒÊF) decoupling
capacitors, effective at lower frequencies, should also be
used on the main supply pins. These may be placed
somewhat farther from the device and may be shared
among several devices in the same area of the PC board.
So, is it OK to use two electrolytics for all the IC chips, or does each chip need its own electrolytics? |
Sure, the datasheet says that.:bawling:
But your ears will say otherwise.
Try both and find out. |
|
|
| kimbo |
So Radian......
what did you end up doing....and how sucessful was it?
I'm about to upgrade my XM1 by replacing the stock opamps with OPA2134 as suggested by Phil Marchand. He also suggested that if I really wanted to I could "....replace the 2K trimpot with a fixed metal film resistor."
I figure that whilst i have the unit apart I'll make the suggested mods ....adding bypass caps......and am curious about how yours worked out in the end.
Anyone else out there upgraded their XM1 and woudl like to share their experience?
Kimbo |
|
|
| Radian |
So Kimbo....
actually I had only finished a mono XM1 because I decided to switch to fullrangers. But my advise is not to use the OPA2134 as the new LM4562 should be far better for the job. I like the TSH4032 best, but it is not unity gain stable so it is out of the picture for this job.
Google for the TeddyRegs to use it as PS these are fantastic.
If you have them close to the chip you can spare the 40-100uf cap.
But otherwise do exactly what CarlosFM suggested, he really knows his stuff.
Greets,
Klaus |
|
|
| kimbo |
Hi Radian,,
thanks for the reply.
ahhh......LM4562. Magentar also recommended this in my other similar thread in the Speakers forum.
http://www.diyaudio.com/forums/show...threadid=112894
Looks like I'll may have to go down this route. However I just spent an hour online trying to find a local Aussie supplier for them without any sucess. I think that as I already have the OPA2134 I'll go with them for the time being, and I'll install all the bypass caps as recommended. Once I have a bank balance again (ie. after the Christmas consuming frenzy is over) I'll grab some 4562s and pop them in.
Quick google of 'teddyregs' only threw up a few hits........gotta fly off to work now so will chase up more info tonight and see what these are......I'm not familiar with them
Many thanks. |
|
|
| endeeinn |
| quote: | Originally posted by carlosfm
No!
I was telling you to try because it will sound very bad!
The word was C R A P, but it was censured.:confused:
The OPA2132 is better than the OPA2604, and as I said lots of times here, I recommend between 22 and 100uf capacitance on each supply pin to ground.
You can also bypass the electrolythics with small 100nf multi-layer ceramics, but this is not so critical. Note: the LM6171/2 needs them.
Also a 100~330nf poly cap directly from V+ to V- PSU pins on the op-amp. Usually you can do that under the circuit.
:att'n: This gives me very good results everywhere, not only with the OPA2132.:att'n:
For me, it's a rule.
After so many tests I made years ago, this is what works best.
Always.
Also, if you don't do it this way, forget the OPA627, it will NOT sound good.;) |
Carlos,
I am new to this so this may be a stupid question but I read that opamp bypass caps should be ceramic COG or NPO types but 100uf is too big for those types. What type of 100uf bypass cap are you reccomending, electrolytics? If electrolytic would they be polarized or non-polarized? I understand the 100-330nf is a poly cap. Thanks for your help. |
|
|
| Nuuk |
| No, Carlos was suggesting 100 nF for bypassing the 100 uF electrolytics! :att'n: |
|
|
| endeeinn |
Is this what the circuit looks like? Since I have the 2134 is a dual, I assume each opamp will have two circuits that look like this .. correct?? |
|
|
| kimbo |
Although the OPA 2134 is a 'dual', the pair of op amps within the single DIL package share common +v and -v pins (pins 8 & 4), so your schematic applies for the pair rather then for each individual op amp.
Otherwise your schematic is how I interpreted the thread. I've now got the caps and will crank up the iron over christmas........sure is gonna get crowded under the board!! I note that somewhere in one of these posts it is suggested that they be soldered directly on the pins, but I figure this is too messy and probably not neccessary and soldering onto the pins of the DIL sockets under the board would be OK. Perhaps Carlos could confirm this and the schematic. |
|
|
| AndrewT |
Hi,
the small HF caps should be soldered in with the shortest possible route between cap and opamp. This means trimming the cap leads short and soldering them direct to the same pads as the opamp, preferably without a socket.
The electrolytics are less critical and may be soldered in upto 10mm away from the opamp pins. Well, you might get away with 20mm if space is a problem. |
|
|
| bichi |
carlosfm,
Fascinating stuff!
- any preference on electrolytic type? (standard, low-ESR, tantalum, solid-polymer, MLCC?)
- have been experimenting with large-value MLCC X7R types for both audio-coupling and power decoupling, posting results in a few weeks... |
|
|
| Stuey |
| Carlos doesn't post on this forum anymore. |
|
|
| SimontY |
| quote: | Originally posted by Stuey
Carlos doesn't post on this forum anymore. |
That's a pity, I always enjoyed his posts. Do you know where he does post?
Thanks,
Simon |
|
|
| KSTR |
| In another forum (with the same themes, basically -- some evident exceptions, though) where they call this forum only the "other" forum... |
|
|
|
