Gainclone Transformer

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hi, Daniel said that in January 2010, maybe actually it's not the same? Today it's Vishay: BC / Vishay 0.1uF Capacitor | Chipamp Electronics But for Cz, what do you think of Vishay Ero MKP 1837 or Wima MKP 10?
Phil.
The little blue BC/Vishay caps are not of sufficient quality to use anywhere in an amplifier enclosure. So, I suppose that a small portion of the trouble is from the bad caps shipped with the kit, again. Those aren't helpful.

AVX T* or a Small red Wima should work acceptably, and if they happen to be too efficient for your RC, all you'd have to do is change the resistor value, which may be slightly inconvenient. These work perfectly for your power hf bypass needs, but possibly not the best choice for output RC (zobel), although it is doable.

It could be much more convenient to use the expected part for the application, which is a polyester/mylar dip ("blob," not box) cap, in which case the datasheet's published resistor value is probably correct. You can find a two pack of the little green Tracon polyester dip caps at the Radio Shack, and those will work perfectly for output RC use.

There are two missing parts, which are the output inductor with paralleled resistor (not included with starter kit); and they are, the other half of the output filter parts (see official LM3886 datasheet from TI and/or NatSemi). Fine tuning the output filter for better stability (that also means better tone) would be a lot easier if you had all four parts to it so that it could work appropriately.

P.S.
Main problems with tone on the chipamp.com kit is:
1). The power decoupling caps are far too small at 100u (natsemi specifies a minimum of 470u 50v)--the suggested range for that application is 680u~1500u (smaller for more forward tone usable with TV, or bigger for more laid back tone usable with music), preferably panasonic caps.
ONE (1) Cornell Dublier's Mallory SEK 4.7uF 250v, specifically that one (source Allied, albeit both SK and SEK will work as *might* others) can be placed on each amplifier board, from V+ to V- where it will basically turn some useless tonal distortion into useful DC. That is only mildly helpful, but when you've got such a big challenge as this kit, I suggest doing it; being that we're so very far away from a home run that coming any closer is of excellent benefit.
2). The 47u inverting input signal coupling cap "Ci" is far too small to allow sensible bass--the suggested range for that application is 220u~680u, and lower voltage caps (16v~63v range) can be used at Ci, as needed to fit the board. If you'd like some treble "air" consider adding bypass such as 4n7~10n polyester and/or 0.47u~1u bipolar electrolytic in parallel with the main signal cap.
The build guide and the Internet advises to omit (short) Ci, but you should know what that does--going without that cap gives you less amplifier harmonic signature at the cost of even less dynamics and a shortfall of speaker protection, so I have to say this: Try to get the amplifier's harmonic distortion (tone) under control by other means so that you can use Ci and experience the full range of dynamics for lifelike music replay (and a modicum of speaker protection as well as a more durable amplifier). If you want dynamics, then I'm sure you DO want Ci, but that will be SO much more pleasant if it is the right size.
If necessary, you can recoup the tone after installing Ci (of right capacitance) by lightening the input load, from 22k to 27k (for use in cases when there is also a potentiometer, in which case, even higher values can be used so long as the sum is not significantly higher than the feedback resistor value). That'll compensate the tone. And you still get to hear your dynamics for lifelike music.
3). The kit is built for dual-mono or monobloc, involving a transformer per each channel (so two transformers if two channels); however, the grounding scheme used by the kit is extremely confusing if a stereo build involving only one transformer for two channels.
Here's a fix for stereo builds:
If a stereo build with only one transformer, then jumper PG+ to PG- at the power board, with a thick solid wire to create a 0v tap, and then run one (1) ground cable from the power board's new 0v tap to the amplifier board's "CG" (instead of the amplifier board PG+ and PG- which are not used).
In this case, speaker ground cables go to 0v tap (that you installed) on the power board.
At your option, this may be further embellished by putting 6a05 or 10a1 or similar diodes, at the amplifier boards, into the V+ and V- terminals, in series to the dc umbilical cable, so as to form virtual dual mono, a little more stability and of course that means a slightly better tone. That will work on almost all audio amplifiers.

P.P.S.
Each mod mentioned is only slightly effective apiece, but in combination, may work almost well enough for a pleasant experience from the chipamp.com kit. The remainder may depend on speakers, and if yours aren't dull, then you may like padding resistors to both midrange and tweeter (or a BSC if 2-way speakers).
In my own experience, the mods helped, but there was still a slight shortfall as for practical useful tone; but, when I also added a BSC to the speaker, the chipamp.com kit DID succeed. This seems to indicate that success will eventually rely on the speaker.
I have other news: An equalizer most certainly will NOT help against tonal distortion; however, the mods (and bsc) do.

P.P.P.S.
We would like to hope that a kit will be more convenient than that. And if you've got somewhat dull speakers and your usage is TV, then your experience could be very convenient indeed. But for other cases, I hope that my notes have helped you. If this plan was a shortfall or bust, then try the MyRef instead. Or if you needed a LOT more power because party, then try my inexpensive Parallel TDA7293 (with the published mods, of course).
 
Last edited:
Hi Daniel,
Thank you for your accurate and long answer, I appreciate.
But I need some time to understand all.
My speakers are FE126eN (two per side in OB, back to back, with BSC, Xo 190 hz)..
Phil.

Well, if you're collecting parts:
You might appreciate four panasonic FC 50v caps (tall skinny version) anywhere in the range of 680u to 1500u, to use for amplifier board power decoupling caps (because 100u is too small).
Choose the 680u for TV (more forward vocals) or the 1500u for music (more laid back tone). That particular 1500u should be exactly the same brand and model number as used by the audiosector kits (that research already done--we don't need to work hard for it).
And you might like a pair of good quality caps, approximately 16v~25v in the range of 220u~680u to use for Ci (because 47u is too small).

Just a bit of quality control there.

P.S.
Also a bit of Gel Flux (applied to vias before soldering), because those gold plated vias are almost nonstick.
 
Hi Daniel,
Thank's again.
I'm talking about this recently amp :

An externally hosted image should be here but it was not working when we last tested it.


Pot is 10 Ko, I remove R1 and input cap, R3 is 1 Ko and Rf is 27 Ko, add 220 pF between pin 9 and 10, dual mono power suply (27/0/27) is snuberized with 10000 uF caps, dual safety loop breaker.
I think your recommendations are also valid for snuberrized PS..

It's works well in my case cause I used fullrange drivers crossed at 190 hz and I mostly listen jazz music or acoustical music..

I still has another kit of pcb from Chipamp, and for next amp I try yours advices.
Phil.
 
I think 47uf is enough for Ci. It gives about -0.2dB for 20Hz. The speaker will have minimum -10dB drop there at least, so 0,2dB compated to that is nothing.
I'm not sure bigger than 100u is required for TV cap. I use only 100n there as there are reservoir caps. Some setups use 1500uF near the chip as there is no reservoir cap at all.
Not sure why should it be the exact same TV as suggested by Audio sector.
Audio is about preferences. Because someone likes X or Y sound, doesn't mean other likes it too. So if you only have different kind available you shouldn't just buy one for much more just because it was told. If you want to test, then yes buy different kind, but not a must.
 
Hi,
I think that I described most of the elementary fine tuning possibilities, which turned out to be unnecessary because for "190hz and up treble amp" driving full range speakers, the easiest fine tuning, is a simple padding resistor.

Unnecessary ? I don't think so, contrary it was very interresting..
Two other things, remove input cap changes sound, crosstalk and clarity are better. Remove resistor input (R1 1 Ko), changes again sound, more "natural"..
Remove both, and you have no componants at the input signal path.
But this resistor is useful for DC, if I remember well ?

About gain : I listened gain 22, 28 and 33. For my taste 28 was better, a good compromize..
Phil.
 
Hi,Unnecessary? I don't think so, contrary it was very interresting. Two other things, remove input cap changes sound, crosstalk and clarity are better. Remove resistor input (R1 1 Ko), changes again sound, more "natural". Remove both, and you have no componants at the input signal path.
But this resistor is useful for DC, if I remember well?
To answer your question, the cap is useful for getting rid of DC, and that resistor didn't help much.

Instead of omission of useful parts, it is generally better to correct/adjust the value of the part. For example if that 1k +in base stopper is too much, try reducing it to 330R or 390R (which is a "general purpose" fairly neutral value, typically evenly balanced neither overly inductive nor overly capacitive but rather centered, for typical 1/4w resistors, averaged greatly), or even try using a different type of resistor material, such as carbon composite or cermet, or silver oxide, etc. . . because in the presence of gain, any differences are more noticeable.

Even really offbeat resistor values could work, as long as they aren't too high. To make a wild guess on that +in series resistor value, I'd have to say that it shouldn't be bigger than your feedback-shunt resistor, 680R, and it IS remotely plausible to use one as small as gain factor less than 680R, thus resulting in a smaller range of values to explore: 27R~680R, so yes the 1k is too high, but I do not know by how much.
Most of the parts that come with the kit are the wrong size, so the unhelpful 1k isn't a surprise.

Removing coupling caps gives you less amplifier sonic signature at the cost of shortchanged dynamics, less durable amplifier and higher risk of speaker fires, meaning that removing the caps is a sort of wrong way fix. If it was a good idea to leave them out, then we'd never have put them in. And, like that comment about the resistor, caps also come in different values and certainly different qualities, and in the presence of gain, aspects are all amplified.
Prerequisite: Actually, squashing the amplifier's tonal distortion is usually done by adjusting the amplifier board's power decoupling cap size (100u is just begging for a shouty result--caps are too small and may need replaced), and then after that task is complete, your signal coupling efforts will begin to work at high quality.

Firstly, I would suggest to find reasonable component values, and then after finding/using the most appropriate/practical figures, then later it is time to try different qualities/materials/brands, etc. . . (right values come first, because it would not be useful to try all different brands of a wrong size part).
 
Last edited:
lly, squashing the amplifier's tonal distortion is usually done by adjusting the amplifier board's power decoupling cap size (100u is just begging for a shouty result--caps are too small and may need replaced), and then after that task is complete, your signal coupling efforts will begin to work at high quality..

After some more critical listening at higher volume levels I did find the sound slightly "harsh" in the upper frequencies and with denser types of music like Bombay bicycle club. It seems to confirm what you are saying. Could this be in part because of the tiny decoupling ci on the Amp board ?
 
Last edited:
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.