| sangram |
Hi
In a frantic bid to eliminate all caps in the signal path (don't ask, the quality of all caps in India is way below average, I get horrible results with any kind of cap and good sonics with no caps), I have removed the output coupling caps on my CD Player as well as the input cap on my P2P-wired LM 4766 chipamp (currently with its legs in the air mounted on a heatsink and no other case).
I get about 20 mV offset in worst case (volume pot turned to zero) and about 1mV at best (fully open pot) and the offset varies with the travel of the wiper on the pot (due to the input impedance varying). Before removing the caps my offset was about 1 mV-2mV on both channels. Input offset (from CD-player) is about 3 mV.
I guess in theory I'm flirting with disaster. In practice, has anyone done this and what are the dangers of doing so? Currently I'm running a set of cheap woofers, so I'm not going to cry if they blow but I'm looking at a speaker upgrade real soon and I don't want to do one more immediately afterwards. |
|
|
| sagarverma |
| quote: | Originally posted by sangram
Hi
(don't ask, the quality of all caps in India is way below average, I get horrible results with any kind of cap and good sonics with no caps),
|
hi Sangram,
plz don say so in this forum.what image u r projecting of your own country?think,,,,,,,plz:angel:
the main prob lies in p2p(believe me),my first amp was p2p and horrible,transferred to pcb....still up and running great.
CAPS,use philips(only those can b gurranteed genuine bec,,purchase only from philips shop[there is only one official philips shop in lajpat rai market here in Delhi,i purchase all the req. yes all the req. from there only.]).they sell in bulk(for 100nf etc,500pcs min@50p),i don mind payin when i know thing is good and genuine.
u can make pcb very easily using the toner xfer process.(discovered recently,, 4 me;) ).
Sangram,rest assured,pcb will solve prob(noise,sonics etc)
regarding caps.::::::
i removed power decoupling caps from tda8563 amp,even removed the signal caps!yes,its true.the amp sound improved greatly.maybe ,the caps were degrading sound.issue can be::1)quality,bec they were local,Rs 2 for4700uf/25v!:::2)polarity could b wrong,wrongly inserted by me.......
it was a bore amp(to make),so i never bothered bout all that.but,one fine day i started the amp,and it didnt work.what could b the reason.its still dead.maybe lack of capacitors ate it away.(most likely,bec rest all modules like xformer,rectifier etc working else where sucessfully).there was no spark,noise sound etc,jus silence,amp dead! |
|
|
| I_Forgot |
Yes, you ARE flirting with disaster.
You may be upgrading your speakers sooner than you think.
Is speaker quality in India higher than capacitor quality?
I_F |
|
|
| sagarverma |
| quote: | Originally posted by I_Forgot
Yes, you ARE flirting with disaster.
You may be upgrading your speakers sooner than you think.
Is speaker quality in India higher than capacitor quality?
I_F |
In India,concept of diy is nacent,few do it.
my country got all,but in bulk only.u have to search and research hard to get good things loose.
its like sifting thru ATM for change. |
|
|
| sagarverma |
hi nordic,
that was bec of shorting(the bench i was working on had wires strewn all over).i dunno what shorted what,but i lost the setup. |
|
|
| sangram |
| quote: | Originally posted by sagarverma
hi Sangram,
plz don say so in this forum.what image u r projecting of your own country?think,,,,,,,plz:angel:
|
The truth? :) Seriously, it is true, you say later down the post that DIY is nascent. Availability of good components is very very poor.
| quote: | Originally posted by sagarverma
the main prob lies in p2p(believe me),my first amp was p2p and horrible,transferred to pcb....still up and running great. |
Oh make no mistake. Mine runs great too, and I've tried all sorts of cap combos. Paralleling films, only film caps, etc.
You're joking about P2P, right? The only time a PCB is better is when you can't actually solder correctly in P2P and wires fly fast and thick, come loose etc. The sound quality beats many commercial amps. But without the cap.
| quote: | | Originally posted by sagarverma CAPS,use philips(only those can b gurranteed genuine bec,,purchase only from philips shop[there is only one official philips shop in lajpat rai market here in Delhi,i purchase all the req. yes all the req. from there only.]).they sell in bulk(for 100nf etc,500pcs min@50p),i don mind payin when i know thing is good and genuine.[/B] |
Don't make me laugh dude. Those Philips caps stink for series application in line level apps. I know, I've tried.
They're just OK for supply decoupling and filtering, but for series audio, even the cheap chinese greencaps have lower ESR and sound better.
| quote: | Originally posted by I_Forgot
Is speaker quality in India higher than capacitor quality?
|
:D :D
I think no more need be said, I'll order a few BG-Ns from Digikey. |
|
|
| sagarverma |
hi Sangram,
u r a seasoned diy'er.i m new in this field(serious diy).but my experience to get components is OK OK(good not possible here,bec hard work req.).
it seems from your posts that u have had very bad experience.
| quote: | | u have to search and research hard to get good things loose. |
this is what i said:) .searching is needed to get the sources who will give u good components in small quantities.
easily available components r :mad: .good ones r available,but in bulk,now its upto u how u get them in quantiies u req.after about 2 yrs,its a breeze to get what i need(almost 90% r got).
currently i m looking for opa134 series.if i fail to get it in Delhi,i'll get it from b'lore.
| quote: | | The sound quality beats many commercial amps |
it does and that to by big margins.what commercially available is ****.i have creative 4.1 inspire 4 my comp.,but even tda1554 amp beats it in quality and sound by big margins(cheap Rs400 ,pair of speakers).
| quote: | | Don't make me laugh dude. Those Philips caps stink for series application in line level apps. |
:apathic: .not those with BC written on them.they r dupli.there r dupplies of every component flying around.i wonder what u got:confused:
| quote: | | even the cheap chinese greencaps have lower ESR and sound better. |
how u cal. ESR?(don't say it was printed on them) |
|
|
| sangram |
No not a very bad experience - until the point I had a system good enough to reveal the difference.
You can get a good idea of the quality of the cap and some of its basic specs. All parameters have an effect, and ESR (or at least what I understand as ESR) is not very different. Since I've been doing lots of tests in the last few weeks, I automatically correlate the sonic signature to ESR. Technically it may not be correct, but basically I sense a HF rolloff which is easily corrected by paralleling a film cap.
AFAIK Philips does not make passives any more, so I think what you're getting is all NOS. Which is not necessarily a bad thing. BC is now part of Vishay, I don't know if they still have a license for India where they use the Philips name for marketing - I doubt it.
Getting good actives is never the problem. Good passives, transformers, cases etc are the issue. But specially passives.
Anyway let's not sidetrack this discussion. There's a lot of India-specific threads with regard to components. Referring to those will be better, or posting there on this aspect.
As far as input caps go, I'll look at getting some quality caps from Digikey or RS India. To update results, at about half travel the offset is at a very respectable 2 mV on one channel and 1 mV on the other, and I do take care to switch on the amp after the player and switch off before, so that speakers are protected from transients, but I do believe that under some conditions there may be a DC surge in the CDP output, so I'll take some hit on SQ to put in caps. I need about 100 uF for a lower F cutoff of 1.6 Hz (1 Kohm load - inverting amp) which will be roughly where I want the cutoff to be. Which means exotic films are out. The lowest possible is 22 uF, for about 7.2 Hz, but we'll still get some audible band problems...
The other option is to use 10000 uF on the outputs and parallel the hell out of them (1000, 100, 10 and 1 and .1 uF) to get acceptable performance, but that is a distant second choice. |
|
|
| youyoung21147 |
I'm running my bridged LM4780 without any coupling cap.
I get 3mV offset when the pot is down, and 9mV when at maximum. I use a laptop as soundcard right now. Don't think it has any coupling cap, I opened it once and didn't see any.
Never tried to use a cap, but sound is so transparent without it that I don't wanna try ! :D |
|
|
| jackinnj |
| quote: | Originally posted by sangram
Hi
In a frantic bid to eliminate all caps in the signal path (don't ask, the quality of all caps in India is way below average, I get horrible results with any kind of cap and good sonics with no caps |
seems to me that you are describing an opportunity, not a problem! if the market is so scarcely subscribed, there is an opportunity for indigenous manufacture!
(I have made caps out of mylar film, scotch tape, teflon tape, vinylidiene chloride - with copper tape from 3M -- more out of a "proof of concept" than any usefulness).
but getting back to the topic -- the bypass caps in the National series of chipamps really required -- first supply issues (attaining low localized impedance) but secondly the protection circuitry can oscillate and this will put transients on the power rails. You should be able to get decent 100nF ceramics anywhere in the world that this forum is read.
if you don't want to put a Zobel on the output use a small choke wound on a 10 ohm 2 watt resistor --this will keep high power nasties away from your speakers.
Also, analyze where the offset stems from -- offset is usually caused by a mismatch in the input impedances of the opamp -- remember that a small amount of current flows from the input pins of an opamp -- I*R * K (where K is gain) is usually the amount of offset and can be measured. Unexplainable offset may arise from electromagnetic or radio frequency interference. "Unexplainable" can often be eliminated with good bypassing techniques. |
|
|
| sangram |
No not that desperate, only signal path.
I already have very good decoupling, I'm not sure there is really a possibility of running the amp without decoupling. I have 10uF and 100nF film cap in parallel. Zobel is in experimental stage, I am using 4.7 ohm and 100 nF because that's what I had lying on hand, but will look at pushing it down to 10nF - 15 nF over time...
My offset at the CDP output is about .1 mV. My problem is the volume pot. At min position it is basically driving the - input to ground, whereas the + input is at ~954K (1K+33K in parallel since the same values are in the feedback loop). This is generating about 20mV on the output.
As the pot is raised the offset comes down. At max travel the offset is a paltry .3 mV, but in a normal listening position (halfway) the offset is only about 2 mV on one channel and 1 mV on the other.
youyoung: The PC soundcard always has an output coupling cap as they operate off single supply. You don't even have the option of removing it unless you're digging deep in your laptop. I would assume 2.5VDC would be present if you removed it...
The pot is what creates the offset you are facing. Because it results in the opamp impedances difering on both inputs. But I assume you're using NI config, in which case your input impedance is much higher and the variance is also smaller. |
|
|
| youyoung21147 |
| quote: | | youyoung: The PC soundcard always has an output coupling cap as they operate off single supply. You don't even have the option of removing it unless you're digging deep in your laptop. I would assume 2.5VDC would be present if you removed it... |
Yes I'm stupid, there are two sides on a laptop PCB :D it must be on the other side !
But maybe also it could operate fully differential, then an output cap is made unuseful. Don't know if some soundcards use this concept.
Anyway, I love my LM4780 without cap. As it is bridged, it would need 47µf to have a 7Hz corner frequency !!! It would be a very expensive (if it exists) MKP cap :D
Concerning preamps, some use opamps that have external offset compensation (I think of AD797). It could be nice trimming offset with the preamp plugged + GC, so any offset could be removed easily :smash: |
|
|
|
