solderdude
Administrator
measureutternutter
Posts: 4,881
|
Post by solderdude on Mar 23, 2017 10:42:31 GMT
The RB3 series is bipolar and has clear BP markings on it and no stripe indicating the - But as you mentioned not above 10V rated for 47uF because these are miniature sized (5mm height) The larger R2B series does have 35V rated 47uF caps but these are bigger in diameter.
A bipolar cap will measure the same capacitance in both polarities. They do differ in leakage current when the reverse voltage increases which will give 2nd harmonic distortion.
When you use to of them like + - - + then it becomes bipolar as the leakage current is blocked by the one in the proper direction but capacitance halves.
So in order to get back to 47uF you need 2 of these series strings in parallel.
so:
+ - - +
- + + -
this will give a low distortion bipolar 47uF made from 4 x 47uF polar caps. No need to look for other caps when you have room to fit them. There is quite some room available in the HD681 cups.
the 10uF ones in the link above are bipolar but you will need 5 of these in parallel.
|
|
|
Post by ihack13 on Mar 23, 2017 11:22:49 GMT
Alright. So that makes 8 caps in total for left and right. The thing is though, I can see that some of them are not shiny on the surface. is it possible that the electrolyt has leaked?
Well. I just checked those and other caps with the DMM. Everything that you said made perfectly sense but now I am confused because the caps give me readings no matter how I pole them.
Does the DMM do some fancy things here and switch polarity internally? Or why do I get a reading both ways.
|
|
solderdude
Administrator
measureutternutter
Posts: 4,881
|
Post by solderdude on Mar 23, 2017 11:58:19 GMT
A polar cap is always the same capacitance, regardless of polarity. Also the test voltage of a meter is very small.
The polarity is only important for the DC voltage across it. In AC signals from an amp to the headphone there is NO DC voltage present. Wet electrolytic caps can handle a low voltage in reverse polarity without exploding. When a higher DC voltage is applied in reverse polarity the wet stuff inside will start to evaporate and the pressure inside the cap increases. Right up to the point where either a deliberate 'weak' point in the casing ruptures and gas escapes or the bottom rubber cap simply gives way launching the aluminium enclosure up in the air with a bang.
For the capacitance to be stable and leakage current to remain low a polar cap should have a DC voltage across it bigger than the AC that can possibly be there.
When used without that 'bias' DC voltage the cap will show more leakage (a certain resistance) when a half of a sinewave is present in reverse voltage. When 2 of these caps are in anti-series there is ALWAYS one of the caps in the correct polarity. In that direction the DC resistance is high. A high DC resistance = low current. The other one is in reverse and has a lower resistance. As they are in series a high + low resistance means the total resistance is always higher than the highest one. A high resistance = no leakage current so the wrong polarity cap cannot form gas (cause no current, no gas) and thus it works.
The shinyness of the finish has no relation to leakage of electrolyte. That last part only happens after many many years or when it has had a sizeable reverse voltage on it for a long time.
Don't worry... all these caps are fine and will measure the same capacitance in both directions as the capacitance is NOT polarity dependent, only the needed bias voltage is polarity sensitive.
|
|
|
Post by ihack13 on Mar 23, 2017 12:24:30 GMT
Thanks very much for the detailed explanation. Wanted to do the testboard with the variable resistors first. But seeing that the circuit takes 8/10 caps I'll just go for the direct soldering with the fixed resistors into the headphones. Hope it will be fine. Gotta pick the components up at my local stores when my order arrives. Hopefully 1206 hand soldering isnt too hard, without soldering on a board. They are called Surface Mount for a reason
|
|
solderdude
Administrator
measureutternutter
Posts: 4,881
|
Post by solderdude on Mar 23, 2017 14:08:34 GMT
Soldering 1206 onto a PCB is not hard. Using them as small resistors with little wires attached to it, is nearly impossible. The contact areas break-off from the component when just a little stress is applied. You really need through-hole components if you want to do 'spider' connections.
|
|
|
Post by ihack13 on Mar 23, 2017 17:07:57 GMT
alright. I think I will cut off a small piece of PCB then. Like this one.
|
|
|
Post by ihack13 on Mar 28, 2017 17:28:28 GMT
By the way I picked up my first delivery of parts today. The inductors weigh 20 grams. That feels a lot in hand. Are there lighter inductors? The SMD ones dont meet the Amp specs , most often. Gonna see what I can find. www.buerklin.com/en/suppressor-choke/p/74d3679
|
|
solderdude
Administrator
measureutternutter
Posts: 4,881
|
Post by solderdude on Mar 28, 2017 18:14:38 GMT
|
|
|
Post by ihack13 on Mar 29, 2017 13:24:22 GMT
0.07A ? really? Can you give me a source / youtube link to understand what the inductance does in this case? So I have a better understanding which values are of importance. I guess first of all the low resistance value of the inductor is important. Because in series we dont want to add another 12 ohm in series where we have resistors in 15 to 50 ohm range. That one I understand. But there is also the tolerance value and a lot of inductor types. Suppression, RF, data circuit chokes etc etc. And when I look up the same axial RF inductor value I get two differenz frequency values . One manufacturer Fasttron says 20khz and the other TDK says 1.2Mhz
|
|
solderdude
Administrator
measureutternutter
Posts: 4,881
|
Post by solderdude on Mar 29, 2017 14:28:25 GMT
0.07A = 70mA = constant DC current rating. Then there is the issue of core saturation which is primarily why the DC current is given. Above that DC current the inductance becomes less than specified. 70mA (rms) in a 32 Ohm load = 160mW ! 160mW = 120dB average level... you would be deaf if you needed those average levels Peaks it can handle even more so... you're safe with this inductor. Indeed the resistance must not become too high. 3 Ohm is not that high considering the driver is 32 Ohm already. The resonance thing is not of importance here as it is not used as a choke to filter out RF signals. This is the high frequency value you mentioned, usually between 100kHz and a few MHz depending on values. The lower frequency you mention 10kHz, 20kHz or 30kHz usually is the frequency at which the inductance is measuredNor does it matter where it resonates because of the low impedances and the large capacitance in parallel that tune the circuit far.. far ... far below its own resonance frequency. These numbers are of importance when filtering RF or when used in switched circuits etc. The tolerance isn't really important here as well because a 10% tolerance in inductance isn't worse than most caps (which are in parallel) and because the frequency changes with the SQR of L*C a small change difference in value won't change the resonance frequency that much. Because of the bandwidth of the filter itself it doesn't matter if the centerfreq = 8kHz or 9kHz That capacitance + inductance now is the (desired) resonance frequency around 8kHz. On top of that the resonance filter is quite well damped by the resistor in parallel which determines the bandwidth of the circuit.
|
|
natna
valued member
Posts: 4
|
Post by natna on Aug 27, 2017 12:55:58 GMT
A resurrection post haha
So glad I've seen your revisited circuit with measurments this time.
I got the HD681B + superlux velour pads. I have set up the recomended HD681F filter for my HD681B (low impedance output), since I want that -a bit- extra bass, and the B variant has a steep drop bellow 50Hz.
None of my sources gave very warm sound. Smartphone, PC, portable minidiscs, CD deck headphones out.
Now the highs are so well. The phony sibilant details are gone. Overly "Ss" are gone. I like the vocals - male and female- especially in good old jazz recordings. Piano, brass and guitars are very good. The stereo image is a bit shortened, but it was artifial anyway.
I only regret that I got big 650mA coils and they cannot fit in the caps.
I think I prefer your design over sonarworks HD681 profile or even a carefully chosen EqualizerAPO set.
Your circuit is also a great help for many sibilant earphones or IEMS. Many of them in the sub 200euro range have the same treble issues...
|
|
solderdude
Administrator
measureutternutter
Posts: 4,881
|
Post by solderdude on Aug 28, 2017 16:22:25 GMT
The filter is designed for 32 Ohm headphones and made to specifically adress the exact frequency, bandwidth and amplitude for these specific drivers. Other impedance headphones will require different values. For impedances above 120 Ohm and well below 16 Ohm passive filters are more difficult to create in practice. Also when the frequencies are a lot lower filters are hard to make passively.
For these headphones the filter works quite well.
|
|
natna
valued member
Posts: 4
|
Post by natna on Aug 29, 2017 8:50:16 GMT
Thanks for noting that Solderdude. I'm not ever gonna buy 250 or even 600 Ohm monsters, anyway If anyone has the Xiaomi Pro HD IEMs and has this circuit (without the bass R,C part) for his Superlux, do give it a try. It also has 32 Ohm impedance and crazy sibilance. You'll be surprised.
|
|
|
Post by ruslan on Dec 20, 2017 10:27:41 GMT
Dear Solderdude. Can you, please, provide information on tweaking passive filter similar to what you sent once to PinkFloyd (http://rockgrotto.proboards.com/post/61131/thread), especially I would like to know how to replace resistors R1 and R2 if I change R3 and R4. In Multisim changing parameters of R1 and R2 has no effect on the Bode plotter curve. Thanks.
|
|
solderdude
Administrator
measureutternutter
Posts: 4,881
|
Post by solderdude on Dec 20, 2017 11:30:59 GMT
When you want to see the effect of R1 and R2 in multisim you need to add an output resistance of say 100 Ohm or so in series with the generator. This simulates a higher out R of the amplifier. Only then you can see the effect of those parts. R1, L1, C1 and R2, L2, C2 are only needed when the headphone is connected to a higher output resistance aplifier. When used from low output R amps exclusively these parts can be omitted.
|
|