I'm a bit late to this particular party, but love the case 
I saw one just like that on Taobao and was thinking of things to put into it....
Just one question - why are the reservoir cap boards so distant from the amp boards? Power supply impedance is where its at for best SQ (lower=better) so shorter wires are the way to go.
I saw one just like that on Taobao and was thinking of things to put into it....Just one question - why are the reservoir cap boards so distant from the amp boards? Power supply impedance is where its at for best SQ (lower=better) so shorter wires are the way to go.
Option: More friendly/smoother tone can be done by installing additional amplifier board power decoupling caps, exactly same model as those on the board, and the additional caps will probably fit instead of the green terminal blocks. In this case, the DC umbilical cable gets soldered at the bottom side of the board, directly to the new cap's pins (keep the wire at least 1.5mm off the board in case of obstructive traces). For example, if the caps on board are 470u, then an additional identical cap in parallel is 940u.
470u||470u amplifier board decoupling caps is like this: Same soundstage/imaging as 470u, tone like 680u, and bass slam like 940u (the actual value)
Therefore more friendly/smoother tone without hindering soundstage at all.
In the document AN1192.pdf by same manufacturer as the LM3886, multiple parallel 470u caps are suggested for amplifier board decoupling.
That's an excellent answer, I laud your honesty
Yes I notice that too, at least DIY amps. But I wonder - if you ask them 'Why do you do this?' they may well reply 'Because that's what I see in other amplifiers'. So it seems its nothing more than current fashion.
I don't notice many commercial amps being designed like this though, I looked at dozens of pictures of internals over at 6moons. Very instructive. One which got excellent listening results was by a company I'd not heard of before, called Bakoon. They put a whole lot of capacitors right next to the output devices (MOSFETs, not chipamps).
In my experiments, putting more capacitors close to the chipamp gives more bass impact (snappiness) and generally a cleaner sound overall. However that was with a regulated supply and you don't have regulators.
On the chip itself, the supply voltage is but micrometres from the input.....
Oh I'm so grateful that you mentioned that!
Mystery solved!!!! YAYZ!!!
Regulators, especially tracking-pre-regulator and similar multi-stage discrete regs, are the ideal series element for full bandwidth audio amplifier to have excellent imaging/soundstage performance.
However simpler means such as diode, resistor or fuse (either at edge of amplifier board) or custom dc umbilical cable are possibilities and more easily doable.
That is not advantageous for subwoofer's bass impact; however, if you wanted imaging/soundstage, then you'll be needing a series element, preferably at the edge of the amplifier board.
The regulator has the greatest voltage drop and is therefore more effective.
The diode has a lesser voltage drop but has a very small voltage variance so it works well over a wide range of current, which is a great and well performing convenience, and a bit easier to install.
Resistive means, such as fuse, resistor or customized dc umbilical cable work over a much smaller range of current but can perform well in cases that use only a small range of current, such as for when extremely efficient speakers are used and/or when the target current output is mostly un-changing and known; however, when the current range is vast, such as one audio amplifier to do sublimely at both quiet and loud, then simple resistive means is not optimal.
Actives, such as regulator, transistor and diode (see datasheet and choose one with a current versus voltage drop graph that is nearly straight up like a wall for minimum sag) will do better than resistive means (resistor/cable/fuse) for Bass Impact power.
See the bottom half of post#26 above for the visually hideous yet audibly charming and super-easy way to do the job without excessive current dumping and without complexity.
Compare:
The subwoofer amplifier will not need the series element that I'm talking about because the subwoofer amplifier has no need of imaging/soundstage features.
Contrast:
When the decoupling is forced to work well by using a series element with voltage drop but minimal voltage fluctuation, then the useful output power AND bass impact may increase considerably by as much as 1/3rd higher than a solid copper bar, because PI filter can both alter amplifier clipping pattern And maximize peak output power.
It can be surprising.
But, it is not applicable to a bass-only subwoofer amplifier without need of imaging/soundstage performance.
Last edited:
Hi Daniel,
Thank's.
Caps on board are 1000 uF, and I have only 1200 uF, no more other values..
I think it's ok ?
When you said "friendly/smoother tone", that means more "musical" with better dynamic ?
Phil.
Hi abraxalito,
Maybe this one :
http://audioaddiction.net.au/wp-content/uploads/2012/09/AMP11_5.jpg
Like Daniel said, thank's, this is the first time I hear this explanation !
Phil.
Toroidal core inductors are what I go for when I don't want the complexity of regulators.
Adding more caps without any series elements I'd not recommend. Presenting a lower impedance path to the charging pulses from the trafo means polluting the local ground. Series elements help to keep the current loops small but since most of the PSU noise is load-induced, it pays to have the lion's share of the capacitance as close to the point of consumption as possible.
Haha you nailed it in one! Note the use of many small paralleled caps to get as low an ESR as possible.
From what I recall that box is just one of two boxes, the top one. The lower one includes more caps, but very importantly a couple of relatively huge chokes which look just like trafos. So as Daniel implies, the series elements are very important. Don't just add caps without consideration for where currents are flowing!
<edit> Just checked, there are 4 huge chokes, not two. So looks like they implemented a CLCLC type supply, with the final C right on the OPS doorstep.
Last edited:
http://audioaddiction.net.au/wp-content/uploads/2012/08/AMP11_6.jpg
Very interresting.
Do you have an idea of the chokes value?
Phil.
That is apparently the correct value; however, larger value caps may be more difficult to select for quality.
Also, that is probably true of every non-inverting LM3886 build.
I wish that worked.
In theory, that is fantastic, but in practice, the typical 20% tolerance of capacitors should require ballast for each because despite identical decoration/appearance, the things may be up to 20% different, and you would very much like to avoid many 20% different copies of your signal because we'd have to call that a blur.
Yes, most such good intentions result in a blur.
In practice, 2 caps parallel is uncomplicated and works far better than many more caps parallel. If more than 2 paralleled, you'd have to get a lot more specific about the individual qualities of each capacitor and use better matching (or ballast, such as a series element). That could be laborious if there were many caps.
Practically speaking, the prospect does not work on the amplifier board (except for precision matched caps).
However! It does work on the power supply board with not more than 7 caps parallel. It is useful for up to 7x2200u or up to 7x3300u power supply boards.
The idea is not useful at small signal and therefore it doesn't go on the amplifier board.
Edit:
Practical: Not more than 2x caps parallel on the amplifier board. Not more than 7x caps parallel on the power supply board. Exceptions may be possible for specialty tight tolerance caps.
Last edited:
That post was long, but this was the good part.
. . . on the amplifier board, that is.
At this time, it was THE good way !
Need more listening to confirm, try other speakers, build new passive preamp with Tocos RV24 I found in UK..
Actually it's blue ALPS, not so good with midrange.
Close to prefect ? I hope so, and I hope that others build the same !
About price : it's around 400 euros, but I spend a lot of money for case, quality connectors and Caddock.
I think it's possible to build about same with only 300 euros.
In the coming months, maybe we can test it with Modulus86 and some Kaneda amps..
A case to follow !
But I am sure it was the winner regarding price..
Some build details :
all resistors are matched, I bought 10 (1%) and arrange by mesurement
solder is tin with 4% silver
wire output is CAT5 (or PTT298), 4 twisted wires
audio wires is silk insulate, 4 solid copper wires, 0,3 mm and 0,5 mm for hot point, 0,4 and 0,6 for cold point, like my diy interconnect cables.
Phil.
this is too high a temperature for general purpose soldering.
You really should use a much lower temperature eutectic.
63/37 183°C
62/37/1cu 183°C
62/36/2silver 179°C
Solder - Wikipedia, the free encyclopedia
Andrew, sorry for mistake, in fact solder (3% Ag) is :
Commandez Soudure Soudure 1mm, 0.5% Copper, 96.5% Tin, 3% Silver, 500g MBO SN96.5 AG3 CU0.5 10/10 RT15 - 500G sur fr.rs-online.com et soyez livrs ds le lendemain.
227 °, it does not change much.
What precautions to take with hight temp soldering?
The only problem that I see would be for caps, if time soldering is up to 2 secondes?
Some brands give max temp and time for their componant.
Phil.
Commandez Soudure Soudure 1mm, 0.5% Copper, 96.5% Tin, 3% Silver, 500g MBO SN96.5 AG3 CU0.5 10/10 RT15 - 500G sur fr.rs-online.com et soyez livrs ds le lendemain.
227 °, it does not change much.
What precautions to take with hight temp soldering?
The only problem that I see would be for caps, if time soldering is up to 2 secondes?
Some brands give max temp and time for their componant.
Phil.
from Wiki
Risks damage to any semiconductor.
Difficult to solder to planes.
There are no advantages to using a high temp solder, except to prevent a solder connection coming undone at very high operating temperatures. (I used High temp in slot car motors, they ran very hot!)
A leaded eutectic is far better for general purpose soldering.
still a bit too hot. 220 max.
Risks damage to any semiconductor.
Difficult to solder to planes.
There are no advantages to using a high temp solder, except to prevent a solder connection coming undone at very high operating temperatures. (I used High temp in slot car motors, they ran very hot!)
A leaded eutectic is far better for general purpose soldering.
No, I could only make a wild guess as I've never come across chokes like that in a catalogue or on a vendor's website. 22mH perhaps?
Yes but at sounding, is not so "clean" than tin/silver/copper..
For my taste, now, someone might say it's "snake oil".
Some picts :
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
Phil.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.