I am currently designing a gainclone + preamp.
From datasheets I found out that tantalum caps are faster than a medium quality elco of the same value. They are unreliable though and most people say that they're "not done" in high end.
My budget is quite tight so I haven't got enough money to buy those fancy black high-end caps.
Could I use medium quality 1000µF elco's in the gainclone powersupply with 47µF tantalum caps or doesn't the tantalum caps improve things?
I would also use 10µF tantalum + 100nF polyester to decouple the opamps in the preamp.
Is this a good idea?
regards, Svokke
From datasheets I found out that tantalum caps are faster than a medium quality elco of the same value. They are unreliable though and most people say that they're "not done" in high end.
My budget is quite tight so I haven't got enough money to buy those fancy black high-end caps.
Could I use medium quality 1000µF elco's in the gainclone powersupply with 47µF tantalum caps or doesn't the tantalum caps improve things?
I would also use 10µF tantalum + 100nF polyester to decouple the opamps in the preamp.
Is this a good idea?
regards, Svokke
Tantalum caps have nothing to do in audio. They sound like crap and normal electrolytics are better sonically ( and more reliable ).
When budget is tight I can advise BCcomponents ( ex-Philips ) electrolytics series 37 and 38. Cheap and very good. Avoid the LL, low ESR types like series 135 as the normal ones sound better.
This doesn't mean that all low ESR types are bad, only this series 135 is not OK !
When you don't have a very tight ( but tight nevertheless ) budget you can try Panasonic FC which are very fine and reasonably priced ( about 3 Euro for a 2200 uF 50 V ). In fact they are so good that their price/quality ratio is far better than BG.
If you insist on extra decoupling I would use MKP or MKC to parallel electrolytics but quality of nowadays produced electrolytics like Panasonic FC really is better than we are used to and sometimes bypassing is not necessary at all. Bypassing Black gate for instance gives worse results than without bypassing.
My advise is to solder a small 0.1 uF MKP or a 0.1 uF 50 V BG NX HiQ ( 3 Euro ) directly from the + to - pins of your preamps opamps. To keep wires short it is best to solder them on top of the chips. Do this at every opamp. Please realise that replacing the opamps for something decent has to be done first.
When you can't find MKP/MKC, use MKT but realise these are not the best in film caps.
When budget is tight I can advise BCcomponents ( ex-Philips ) electrolytics series 37 and 38. Cheap and very good. Avoid the LL, low ESR types like series 135 as the normal ones sound better.
This doesn't mean that all low ESR types are bad, only this series 135 is not OK !
When you don't have a very tight ( but tight nevertheless ) budget you can try Panasonic FC which are very fine and reasonably priced ( about 3 Euro for a 2200 uF 50 V ). In fact they are so good that their price/quality ratio is far better than BG.
If you insist on extra decoupling I would use MKP or MKC to parallel electrolytics but quality of nowadays produced electrolytics like Panasonic FC really is better than we are used to and sometimes bypassing is not necessary at all. Bypassing Black gate for instance gives worse results than without bypassing.
My advise is to solder a small 0.1 uF MKP or a 0.1 uF 50 V BG NX HiQ ( 3 Euro ) directly from the + to - pins of your preamps opamps. To keep wires short it is best to solder them on top of the chips. Do this at every opamp. Please realise that replacing the opamps for something decent has to be done first.
When you can't find MKP/MKC, use MKT but realise these are not the best in film caps.
warning!Socalled "dry" (anode) tantalum electrolytic caps are very sensitive against sources with low internal- resistance!!
The neccesary minimum source resistance often is specified with about 3Ohm/V or so (check the datasheets) and therefore this kind of cap is not suitable for power supplies in audio power amps where low internal resistance is desired.
Also there are sensitive against reversed polarity (AC-ripple currents). Another reason against using them in ("high") power supplies.
Actually I can see tantalum-caps in quite a lot of preamps and CD-Players of the "high-end" audio class.
They could be used in preamps for rail decoupling purposes when they are protected against low source-resistance by some preresistance as it is often done for decoupling many opamp stages against each other when they are fed from one supply.
If it is a good idea to use them in such applications instead of modern high-quality electrolytic caps I don`t know but would trust on what said "jean-paul" about this as he seems to have a lot of experience with this.
BTW: Tantalum caps are often be found in military equipment as they can have high lifetime (when used in suitable places).
Jean-paul
With all due respect, I find the audio performance of the tantalums better than that of the cheap electrolytics. I unfortunately have no access to any quality caps - Rubycon is manufactured under licence, Nichicon and Panasonic are simply not imported - everything here is cheap no-name Chinese brands such as 'KVM', 'SEL', 'NCK' etc. Even film caps are very poor spec green chinese stuff - nobody imports high-grade components.
Edit: Did not complete my post.
I have tried normal electrolytics of various types and sourced from all over the local electronics market. Unfortunately none of them came close to the sonic performance of the Tantalums. Each tantalum costs about 20 cents. For that price, one can actually buy about 20 electrolytics of 10 uF, or a couple of 220 uF, etc. nobody stocks high-grade because they will not sell.
Jean I admire your experience and knowledge, and if you say that normal electros sound better, maybe you are right but since I do not have access to even moderate quality components I will have to be content with my lot. Philips electros are simply not available anymore, and the quality caps available are only in 47 uF onwards, if at all. Sometimes we come across computer-grade caps. You say they are not good for audio, but they are better than these cheap chinese stuff one has to live with.
And this is after A/B comparing a 100 uF electro head-to-head with a 1.5 uF Tantalum in an inverting STK 4191 (it was non-inverting, but has now become inverting due my repeated listening and deciding that the inverting was better) I will continue testing with various values, till the track on the board peels off, to find as good a cap as possible but I doubt I will be able to come near the tantalum.
Maybe I will be a long way away from maximising the sonic potential of the amp, but I will have to live with it. There's only so much one can do in a third-world country, and we just have to lump it. Digikey is not an option for the kind of quantities I will be looking at, and there're not enough of us audio nuts going around in the country to justify group buys.
With all due respect, I find the audio performance of the tantalums better than that of the cheap electrolytics. I unfortunately have no access to any quality caps - Rubycon is manufactured under licence, Nichicon and Panasonic are simply not imported - everything here is cheap no-name Chinese brands such as 'KVM', 'SEL', 'NCK' etc. Even film caps are very poor spec green chinese stuff - nobody imports high-grade components.
Edit: Did not complete my post.
I have tried normal electrolytics of various types and sourced from all over the local electronics market. Unfortunately none of them came close to the sonic performance of the Tantalums. Each tantalum costs about 20 cents. For that price, one can actually buy about 20 electrolytics of 10 uF, or a couple of 220 uF, etc. nobody stocks high-grade because they will not sell.
Jean I admire your experience and knowledge, and if you say that normal electros sound better, maybe you are right but since I do not have access to even moderate quality components I will have to be content with my lot. Philips electros are simply not available anymore, and the quality caps available are only in 47 uF onwards, if at all. Sometimes we come across computer-grade caps. You say they are not good for audio, but they are better than these cheap chinese stuff one has to live with.
And this is after A/B comparing a 100 uF electro head-to-head with a 1.5 uF Tantalum in an inverting STK 4191 (it was non-inverting, but has now become inverting due my repeated listening and deciding that the inverting was better) I will continue testing with various values, till the track on the board peels off, to find as good a cap as possible but I doubt I will be able to come near the tantalum.
Maybe I will be a long way away from maximising the sonic potential of the amp, but I will have to live with it. There's only so much one can do in a third-world country, and we just have to lump it. Digikey is not an option for the kind of quantities I will be looking at, and there're not enough of us audio nuts going around in the country to justify group buys.
Mmm, a lot of experience is too much said. I have experience with them and certainly the drop types are worthless in audio despite their low ESR and other good news. They simply don't sound good compared to other types of ( good ) electrolytics.
Still have sealed Long Life tantalums ( even wet types ! ) in stock which are better but I never use them. Try changing them in old Quad gear and you'll notice the difference. And where Svokke comes from there is some choice in components. At least BCcomponents can be found in Belgium, Holland or Germany. Cerafine, Silmic and the like are harder to find.
OK, Sangram, when nothing better can be found I understand you're using them. But trust me, brand electrolytics produced nowadays are OK. Besides that prices are low compared to tantalum over here
Believe it or not, I just returned from a bazaar where I found RS components 104-411 10,000 uF 63 v dated 2001 for 2 Euro per piece !
It wouldn't be much of a problem to ask members if they can send you some caps isn't it ? What values do you need ?
BTW I never said computer grade are not OK, I said BC series 135 is not OK. I gave the example of Pana FC which are outstanding ( and LL, low ESR, 105 degrees etc. etc. )
PS Do you know Punsumi caps from India ?
Still have sealed Long Life tantalums ( even wet types ! ) in stock which are better but I never use them. Try changing them in old Quad gear and you'll notice the difference. And where Svokke comes from there is some choice in components. At least BCcomponents can be found in Belgium, Holland or Germany. Cerafine, Silmic and the like are harder to find.
OK, Sangram, when nothing better can be found I understand you're using them. But trust me, brand electrolytics produced nowadays are OK. Besides that prices are low compared to tantalum over here
Believe it or not, I just returned from a bazaar where I found RS components 104-411 10,000 uF 63 v dated 2001 for 2 Euro per piece !
It wouldn't be much of a problem to ask members if they can send you some caps isn't it ? What values do you need ?
BTW I never said computer grade are not OK, I said BC series 135 is not OK. I gave the example of Pana FC which are outstanding ( and LL, low ESR, 105 degrees etc. etc. )
PS Do you know Punsumi caps from India ?
Well, Jean Thanks for the words, I will try out some more. I have tried out listening till 3 am, since your suggestion that electrolytics sound better did invoke some doubt in my mind.
I have finally settled on a tantalum pearl of 1.5 uF paralleled with a fillm cap of 2.2 uF. The film cap alone was also good, but too smooth and warm to my taste - the Tantalum added a bit of shimmer around the instruments and vocals, which i sort of like. The Tantalum alone was not passing enough bass. I will try the film alone tonight again, and let's see what happens over an extended listening session. In any case all this is somewhat a waste since the amps in question are going to become part of an active three-way system, but I want them to be able to reproduce the full spectrum well first, and then cut their bandwidth down...
All the elctros I tried were coming up short. I am today going to try some more combinations and electrolytics today - I find that small values do not work very well, the larger ones above 47 uF or so have better performance, specially if paralleled with film and/or tantalums. I will report again tomorrow, after trying out some computer grade capacitors and some more 'Rubycon' electrolytics. Unfortunately they are in large sizes - 47/100 uF, but will be of a good idea.
How do you know of Punsumi?
I have finally settled on a tantalum pearl of 1.5 uF paralleled with a fillm cap of 2.2 uF. The film cap alone was also good, but too smooth and warm to my taste - the Tantalum added a bit of shimmer around the instruments and vocals, which i sort of like. The Tantalum alone was not passing enough bass. I will try the film alone tonight again, and let's see what happens over an extended listening session. In any case all this is somewhat a waste since the amps in question are going to become part of an active three-way system, but I want them to be able to reproduce the full spectrum well first, and then cut their bandwidth down...
All the elctros I tried were coming up short. I am today going to try some more combinations and electrolytics today - I find that small values do not work very well, the larger ones above 47 uF or so have better performance, specially if paralleled with film and/or tantalums. I will report again tomorrow, after trying out some computer grade capacitors and some more 'Rubycon' electrolytics. Unfortunately they are in large sizes - 47/100 uF, but will be of a good idea.
How do you know of Punsumi?
I have finally arrived at the conclusion that the STK4191 just does not sound good. It is better than my commercial amps, but it is nowhere close to my reference amp. Also of no help is the fact that I have had to use a PCB, whereas the reference is built with P2P. The STK is too complex to build P2P, even the bias for the current stage is externally set, as are the muting flow resistors, etc.
Also since the STKs have two amps in the same package there is a loss of separation - not that much of a problem as eventually they will be operating mid+high of one channel.
In short, no capacitor I have tried has brought the STK close to the reference. The closest I got was the tantalum and film parallel, I would like to know of further possibilities. I have also tried smaller dimensioned and more expensive electrolytics - to no avail. I will just have to complete the assembly and concentrate on the XO.
Also since the STKs have two amps in the same package there is a loss of separation - not that much of a problem as eventually they will be operating mid+high of one channel.
In short, no capacitor I have tried has brought the STK close to the reference. The closest I got was the tantalum and film parallel, I would like to know of further possibilities. I have also tried smaller dimensioned and more expensive electrolytics - to no avail. I will just have to complete the assembly and concentrate on the XO.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.