Ok, the thread just developed a life of its own 
I try to get one thing after another.
Value of Ci.
The Cin/Rin combination gives me a cutoff of ~0.00072Hz with 4.7uF/47k
So I should probably move to the more widely accepted: ~0.00154Hz with 4.7uF/22k
Therfore I declare my Ci to be ~ 0.000769Hz
Since I would like to keep Ci as small as possible (sizewise) I choose Ci to be 33uF (as enzoR suggested)
So the formula gives me an Ri of: 62.7k
The cap will have a diammeter of 12mm which will be a huge pain arranging on the board.
Also Ci will change as soon as I have succesfully rearanged the parts to allow for Ci.
Additionally enzoR volunteered to measure my design (with some fancy audio analyzer) against a p2p lm3886, so since most of the design-changes about the ground are more or less surgical, I'd like to wait for the test results, to for one get a better understanding where negative influences come from and on the other hand to see how bad my grounding scheme really is >.<
cheers!
P.s.
However I think suggestions can be put in until monday.
I try to get one thing after another.
Value of Ci.
The Cin/Rin combination gives me a cutoff of ~0.00072Hz with 4.7uF/47k
So I should probably move to the more widely accepted: ~0.00154Hz with 4.7uF/22k
Therfore I declare my Ci to be ~ 0.000769Hz
Since I would like to keep Ci as small as possible (sizewise) I choose Ci to be 33uF (as enzoR suggested)
So the formula gives me an Ri of: 62.7k
The cap will have a diammeter of 12mm which will be a huge pain arranging on the board.
I guess an additional parallel 50pF will do?AndrewT said:
It would, but it would also elevete the chip by approx 2.5mm and I think it might be better to keep the track-length down.DaveR said:
If I understand you correctly you say the zobel should tie in closer to the chip 0V?infinia said:
Also Ci will change as soon as I have succesfully rearanged the parts to allow for Ci.
Additionally enzoR volunteered to measure my design (with some fancy audio analyzer) against a p2p lm3886, so since most of the design-changes about the ground are more or less surgical, I'd like to wait for the test results, to for one get a better understanding where negative influences come from and on the other hand to see how bad my grounding scheme really is >.<
cheers!
P.s.
However I think suggestions can be put in until monday.
Oh yeah, i think michaels comment about the jumper on the chip gnd line is a good one. Rather get rid of it, maybe make a small narrowing in the trace instead.
And rf attenuation across Rg can be done in parallel underneath if required i guess. Not much board space anymore! lol. Amazes me how much u can fit on 40x30mm!
And rf attenuation across Rg can be done in parallel underneath if required i guess. Not much board space anymore! lol. Amazes me how much u can fit on 40x30mm!
Last edited:
Ah btw, forgot to mention, I modified my laminator today, and it works like a charm, 0.07mm pitch no problem (no typo ) and 0.12mm as track width worked also very well.
For those interested in the details:
Samsung ML-1865 Printer,
reichelt-catalog Paper (good distributor in germany, has moderately glossy paper)
HAMA A40 DIN A4 Laminator ~17€
cheers!
P.s jumper is fixed, new PCB-update tomorrow.
) and 0.12mm as track width worked also very well.For those interested in the details:
Samsung ML-1865 Printer,
reichelt-catalog Paper (good distributor in germany, has moderately glossy paper)
HAMA A40 DIN A4 Laminator ~17€
cheers!
P.s jumper is fixed, new PCB-update tomorrow.
Last edited:
So toner transfer method? Thats amazingly good.
Im lazy! I just print on normal white paper, then place it toner side against the photoresist and shove it in my uv scanner. Learnt from a friend, i never believed him until i saw with my own eyes how good the results were!
But still, even though you can achieve such fine pitch, its better to space them as much as possible apart (leakages, dirt ect..)
Im lazy! I just print on normal white paper, then place it toner side against the photoresist and shove it in my uv scanner. Learnt from a friend, i never believed him until i saw with my own eyes how good the results were!
But still, even though you can achieve such fine pitch, its better to space them as much as possible apart (leakages, dirt ect..)
Hmm a scanner, good to know! I never wanted to buy two chemicals to do the job ;-)
New components/updates:
Now with 100% more Ci and at least 70% more standing resistors.
Also renamed Rg to Rin, seemed more sensible.
Hmm so much free space
AMP: v5.7
cheers!
New components/updates:
Code:
Ci: 33uF T+F bipolar d12mm
Ri: 63.4k
Cd: 50pF
Rin: 22kNow with 100% more Ci and at least 70% more standing resistors.
Also renamed Rg to Rin, seemed more sensible.
Hmm so much free space
AMP: v5.7
An externally hosted image should be here but it was not working when we last tested it.
cheers!
Hi,
1/4inch spade terminal pads can be inserted at Vn & Vp, if placed top to bottom.
Similarly 1/4 spade terminal pads for Power Ground between Vp & Vn but placed left to right. This is where you will bring in the twisted triplet from the PSU.
Add a pair of pads for a 10r resistor in the SIG GND to Gnd trace.
This is not needed if you build only monoblocks. If you build a chassis with two or more of these PCBs inside then the resistor is used to attenuate the current in the ground loop so formed.
I would move RnL (R//L) off board.
1/4inch spade terminal pads can be inserted at Vn & Vp, if placed top to bottom.
Similarly 1/4 spade terminal pads for Power Ground between Vp & Vn but placed left to right. This is where you will bring in the twisted triplet from the PSU.
Add a pair of pads for a 10r resistor in the SIG GND to Gnd trace.
This is not needed if you build only monoblocks. If you build a chassis with two or more of these PCBs inside then the resistor is used to attenuate the current in the ground loop so formed.
I would move RnL (R//L) off board.
Noted, but I like the screw-down terminals better. Rearranged them anyways, now looks a little bit cleaner.
And for now I'd like RnL to stay on the board, will wait for the test-result
No mono-blocks, so added resistor, new part:
Rb2: 10R
AMP: v5.8
An externally hosted image should be here but it was not working when we last tested it.
is Ci that huge?
MULTICOMP|NP100V336M13X20|CAPACITOR, N/P, 33UF, 100V | Farnell Nederland
i thought it would fit in the original space...
MULTICOMP|NP100V336M13X20|CAPACITOR, N/P, 33UF, 100V | Farnell Nederland
i thought it would fit in the original space...
So this is the final, testing version of the AMP,
AMP: v5.9:
cheers!
AMP: v5.9:
An externally hosted image should be here but it was not working when we last tested it.
cheers!
Now that you have a final version... 
What point is there in going "smallest" on the amp PCB when you're just offloading capacitance to the power board? IMO, it would be better with both on the same PCB, or reduction in power board size by moving at least a couple of those capacitors on the amp board.
I know how you did it is how most people do, but it seems to have been a strange evolution in the chipamp crowd, originating from the basic circuit in National's datasheet, and from the idea of modularity with an amp board or a PSU which started out with the PSU having only "enough" capacitance. Then even more ironically people started suggesting use of lots of low(est) ESR capacitors on a remote capacitance *island* PCB instead of putting more on the amp board.
Also, forgive if I have overlooked this detail in the topic thread, but your pads and holes for the power input and amp output seem far too small for the wire (or robust sized connectors' pins), and also too small on the power PCB.
What point is there in going "smallest" on the amp PCB when you're just offloading capacitance to the power board? IMO, it would be better with both on the same PCB, or reduction in power board size by moving at least a couple of those capacitors on the amp board.
I know how you did it is how most people do, but it seems to have been a strange evolution in the chipamp crowd, originating from the basic circuit in National's datasheet, and from the idea of modularity with an amp board or a PSU which started out with the PSU having only "enough" capacitance. Then even more ironically people started suggesting use of lots of low(est) ESR capacitors on a remote capacitance *island* PCB instead of putting more on the amp board.
Also, forgive if I have overlooked this detail in the topic thread, but your pads and holes for the power input and amp output seem far too small for the wire (or robust sized connectors' pins), and also too small on the power PCB.
Last edited:
I've thought about this at some point but decided for the modular version (makes it easier to adjust the PSU to ones special needs). If the AMP works I might create one version with everything on one PCB..
The pad-size will hopefully be enaugh, and we will see, once its built. You could view this version as closed beta. The holes on the other hand are only there to point the drill, actual drill-sizes should be individually choosen.
cheers!
The pad-size will hopefully be enaugh, and we will see, once its built. You could view this version as closed beta
. The holes on the other hand are only there to point the drill, actual drill-sizes should be individually choosen.cheers!
I see.
I've been thinking of "snubberizing" my PSU after having read parts of this thread (http://www.diyaudio.com/forums/chip-amps/43423-high-cap-unregulated-psu-chipamps.html) Since I'm not too keen on reading 50+ pages, is there some other (simple) source for this? Or could the design on page 1 still be considered up-to-date?
cheers!
I've been thinking of "snubberizing" my PSU after having read parts of this thread (http://www.diyaudio.com/forums/chip-amps/43423-high-cap-unregulated-psu-chipamps.html) Since I'm not too keen on reading 50+ pages, is there some other (simple) source for this? Or could the design on page 1 still be considered up-to-date?
cheers!
^ The design as a whole is subject to variation, basically it's a typical unregulated (though you could regulate it if you wanted to add that) linear PSU with a 0.1uF capacitor between each power rail and ground, and (at least 1/2W) 1 ohm resistor + 0.1uF capacitor in series between each power rail and ground.
Yes the original is a variation of that, but it was later decided that the resistor value could be much lower than 5W.
Now a word about layout, placement on that design. You'd have noise coming from the diodes or from the chipamp chip. For best results the circuit and parallel capacitor will be closer to one or the other or both, not right before the output pins on a separate PSU board. Here is where the audio voodoo comes in. Specific diodes were also used, because one person felt one brand of those sounded best. Then they switched sources, and other people have different tastes... generally speaking, soft recovery type diodes can make as much difference as the snubber IMO, or doing the "put a 0.01 to 0.1uF capacitor in parallel with every diode in the rectifier bridge" modification.
When you start talking about 100 variations of a basic design, there are at least thousands of different possible results. Most don't sound different enough to notice, but I do like adding some 0.1uF capacitors here and there in circuits, and they are inexpensive and don't take up much space. It's best to build the basic PSU with spots on it to add addt'l components then listen to it, then add the components and decide for yourself if it matters as there are so many variations possible.
Here is yet another variation that sounded good to me powering LM3886 based gainclone, though snubber resistor, capacitance, and pass transistor values are larger than they needed to be simply because I had those parts lying around and wanted to use a spare transformer I had that was a little too high in voltage for the amp without some type of regulation... so I'm suggesting, alter a design to suit your exact situation and needs.
Yes the original is a variation of that, but it was later decided that the resistor value could be much lower than 5W.
Now a word about layout, placement on that design. You'd have noise coming from the diodes or from the chipamp chip. For best results the circuit and parallel capacitor will be closer to one or the other or both, not right before the output pins on a separate PSU board. Here is where the audio voodoo comes in. Specific diodes were also used, because one person felt one brand of those sounded best. Then they switched sources, and other people have different tastes... generally speaking, soft recovery type diodes can make as much difference as the snubber IMO, or doing the "put a 0.01 to 0.1uF capacitor in parallel with every diode in the rectifier bridge" modification.
When you start talking about 100 variations of a basic design, there are at least thousands of different possible results. Most don't sound different enough to notice, but I do like adding some 0.1uF capacitors here and there in circuits, and they are inexpensive and don't take up much space. It's best to build the basic PSU with spots on it to add addt'l components then listen to it, then add the components and decide for yourself if it matters as there are so many variations possible.
Here is yet another variation that sounded good to me powering LM3886 based gainclone, though snubber resistor, capacitance, and pass transistor values are larger than they needed to be simply because I had those parts lying around and wanted to use a spare transformer I had that was a little too high in voltage for the amp without some type of regulation... so I'm suggesting, alter a design to suit your exact situation and needs.
An externally hosted image should be here but it was not working when we last tested it.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.