=20
The Poor Serbian Man's Optical Attenuator
Prodanovic Aleksandar and Slobodan Pavleski
,=2023rd October 2010
A gadget made by Spavleski & Zen Mod
Article scribed by Zen=20
Mod
=20
You=92re a Macho Man. You have all the gadgets which real Macho Guys =
absolutely=20
must have: A CD player, or even a transport with a separate DAC, and =
recently=20
you made a dedicated media PC, which is dead silent and as capable as =
HAL 9000!=20
You have nice and potent speakers too. Maybe you also have an ancient =
Black Disc=20
Spinner, but that's another issue. Or is it? It gives you even more=20
opportunities to later make another gadget, to amplify those tiny =
signals=20
randomly collected from black grooves...And you just finished making the =
best=20
amps money can buy, just because you=92re lucky, you=92re also potent, =
same as your=20
speakers, and you=92re hanging with right Macho Bunch, at the right =
place,=20
diyAudio, sharing ideas and the newest tips on how to reach Audio=20
Nirvana.As written above, you=92re macho, and you like listening = to your=20 music loud, and you also like to watch your fave Football or MotoGP, and = actually listening to them, not just staring at a moving picture. There = is=20 nothing better than to hear bones cracking (Football) or the metal to = asphalt=20 scream (MotoGP)! Wait a minute, you have a Wife, or even a Mom? And they = aren=92t=20 always in the mood for your torture of loud music every day? So, you = really are=20 not so macho and you need a Volume Control. =93gotcha!=94 =96 JB = .
What is important in an Optical Volume Control ?
First, = its name=20 must be written with Capital Letters. Second, it must be lossless, = unlike those=20 MP3=92s, which you hate even more than a Volume Control! Third, it must = maintain=20 peaceful coexistence/cooperation with the preceding and following stages = at the=20 same time. OK, except for the first, capital letters, the following two = aren=92t=20 easy tasks to achieve, so now we are going to have to get little more=20 technical...
Everyone here knows what a resistive divider is, so = I have=20 no intention of teaching you that just to entertain you. In fact I=92m = pretty=20 lazy, so I=92m typing all this just buying time until I really must = paste some=20 schematics and pictures in here, and start to explain a few things. OK, = now=20 here=92s our first pic, a voltage divider:
Pic.=20 1
As you can see =96 that=92s pic of a plain ole = potentiometer. The=20 Optical Volume Control is however, not made physically of one resistor, = with a=20 contact track exposed to a wiper, but from two variable resistors = pictured as=20 this:
Pic.=20 2
These two variable pots are ganged in a way that, while = the=20 series one is decreasing in value, the shunt one is increasing in value. = To=20 complicate/or simplify things further, imagine that the pots are = shaftless=20 (funny term, isn=92t it?) and that these two resistors are = light-dependent, or=20 better defined, their resistance is inversely proportional to amount of = light to=20 which they are exposed. So the pic can look something like = this:
Pic.=20 3
Assuming that we are still following the same law of=20 decrease/increase of series/shunt resistance, it looks that we have = gained a=20 couple of things.
1. We have contactless variable resistance, = which means=20 that we have succeeded in the first factor needed for losslessness, = hehe...=20
2. We need a proper circuit to make these LEDs behave in the way = we want=20 them to exactly. But number two actually has a hidden benefit. We must = make some=20 unnecessarily complicated circuit blocks just to preserve the impression = that we=20 are still Macho, doing things which our Chicks really aren=92t able to = understand.=20 The fact that they aren=92t even slightly interested in understanding = our toyz =91n=92=20 gamz isn=92t part of the prospective of this clever = document.
Luckily,=20 thanks to the kind people from the Pro Oriented Industry, who are pretty = much=20 unaware of the existence of a bunch of wakoo people obsessed with using = real=20 parts for something that is intended to make some funny noises called = Music, we=20 have those nice and precious parts: LDRs, or Light Dependent Resistors. = For=20 further info on other projects with these funny devices, including = GeorgeHiFi=92s=20 thread on the Lightspeed Attenuator, see the Links to the side of this = article.=20 That boy started the entire Frenzy on Big DiyAudio.
But we boyz = from=20 Serbian DiyA, had an idea how to ensure the good behavior of our tiny = LDRs with=20 current sources, which must be of course, both adjustable and variable. = For the=20 schematic look at =93PCB01 Schematic=94 at the end of this = document.
As you=20 can see, we decided to use the old fashioned approach, without the use = of fancy=20 add-ons such as microprocessors etc. A pot is still the main adjusting = part, but=20 this time we use the cheapest 2x10K linear one, and the audio signal is = nowhere=20 near it. Its role is to handle variable voltages, which are steering = adjustable=20 Current Sources.
For your convenience, the dedicated power supply = schematic is in upper left of schematic sheet. Of course both the PSU = and the=20 steering circuit are on the same pcb, accidentally named as PCB01. The = supply=20 schematic is pretty self explanatory. Center tapped or non CT = transformer, a few=20 diodes, an LM317 as a 70mA CCS, and a few Zener diodes working as a = reference=20 for the shunt Darlington on the output. Exact output voltage is the sum = of Zener=20 voltages plus 2x the Vbe of Darlington (that=92s 1V35). 15V is drawn on = the=20 schematic, but who cares if we have 1V35 more...
So, now go back = to the=20 exact LDR circuit schematic. If you=92re looking at these pages on your = =91puter=20 screen, now is the time to either print these two schematics, or use two = screens. Do whatever you wish if you are not ready to jump back and = forth,=20 reading text and looking at schematics. Here is what you will=20 discover:
-Q6, along with LED1 and VR1 form a reference voltage = string,=20 with 3mA of current flowing through. We have available there two fine = voltage=20 points, 2V5 on top of VR1, and ~ 2V more on top of LED1 .
-We are = using a=20 reference voltage of 2V5 to inject it into our main Governor (Hasta La = Vista -=20 ring a bell ?) Pot itself, so we have nice clean and steady DC voltage = to play=20 with. Q3 is any small N type Darlington (or even two small signal diodes = connected in series if you=92re a cheapskate ) and is nothing other than = a=20 convenient voltage shifter. Using it there is to try to null TempCo of = the same=20 type Darlingtons used as control elements in two main Current Sources = (look=20 down) .
-So, we have clean and steady voltage (2V5 less two Vbe) = on top=20 of Pot1. Now look at the vertical string starting from 15V, made of LEDS = in 4=20 LDRs (OK7,5,3,1), closed jumper JMP01, and dedicated Current Source, CS = made of=20 Q1 steered with voltage from POT01, cleverly cascoded (by Mighty ZM) = with Q4.=20 That one, Q4, is biased from the previously mentioned voltage tap on top = of=20 LED1. That vertical string is dedicated to series LDRs.
-I=92m = clever, so=20 I=92ll paste the above paragraph here for the right vertical string, = just changing=20 the appropriate part labels. So now look at the vertical string starting = from=20 15V, made of LEDS in 4 LDRs (OK8,6,4,2), closed jumper JMP02, and = dedicated=20 Current Source, CS made of Q2 (steered with voltage from POT02), = cleverly=20 cascoded (by Mighty ZM) with Q5 . That one, Q5, is biased from the = previously=20 mentioned voltage tap on top of LED1. That vertical string is dedicated = to shunt=20 LDRs.
Idea is, with the main pot at zero (7 O=92clock) position, = minimum=20 current is flowing through the left series LDR LED string (on top of Q1) = so the=20 mating LD Resistors have max resistance. At the same time, max current = is=20 flowing through the right shunt LDR LED string (on top of Q2 ) so the = mating LD=20 Resistors have minimum resistance. The result is that our poor tone = signal is=20 attenuated maximally.
That crossed behavior is achieved with a = crossed=20 connection of reference voltage to a two section of main pot, while the=20 reference voltage is connected to the =93start=94 end of POT02, and = it=92s connected=20 to =93end=94 end of POT01. Incidentally, at the 12 O=92Clock position of = the main pot=20 these two voltages coming on both wipers must be same. But we are = clever, so we=20 introduce a =93fake log law=94 inducing resistor in each section, R200 = and R300.=20 That=92s an old trick, to use a linear pot because the section matching = of them is=20 better than with cheapskate class logs. Put a few resistors from wiper = to ground=20 and Voila!, we have an Audiophool pot for pocket money, or we hope we do = at=20 least .... where=92s da :devil: smiley when a man needs it ?
OK, = now look=20 again at Pic.3.
You see where the signal is coming in, where ground = is, and=20 where the signal is coming out. Now, look at the Schematic of PSM LS. We = have 4=20 (or just two, for stereo unbalanced, if you wish so) sections. The first = one is=20 made of OK1 & OK2 (R-), the second one is made of OK3 & OK4 = (R+), the=20 third one is made of OK5 & OK6 (L-), and the fourth one is made of = OK7 &=20 OK8 (L+).
If you obtained your LDRs from me, use Series X and = Shunt X as=20 a pair, where X is 1 to 4. So, Series 1 + Shunt 1, Series 2 + Shunt 2, = Series 3=20 + Shunt 3, Series 4 + Shunt 4. They=92re matched with that in mind. =
I=20 think that everything is clear regarding the points where the signal is = coming=20 in, where Ground is, and where the signal is coming out. Just one note:=20 cleverly, on both the schematic and pcb itself, the audio ground isn=92t = connected=20 to any point of the steering circuit. My Fullrange Bro Spavleski, = responsible=20 for those funny patchwork pcbs we made, even uses the top pcb layer as = ground=20 for the steering circuit, without any ill effects.
The = electrolytic caps=20 and block caps, connected in parallel with LDR LEDs, are there for = cleaning up=20 any possible residual noise in the PSU. Use them or not, your choice, = they=20 aren=92t critical at all.
The P4-P11, 2M2 trimpots are there just = in case=20 that the LDRs aren=92t tightly matched. If you have tightly matched LDRs = you don=92t=20 need them at all. Later in this text I=92ll explain a better cure, in = case you=20 feel or measure that some section is slightly weaker than = another.
We=20 will leave the Optical Volume circuit now, and look at the buffer = schematic;=20 That=92s the =93PCB02 Schematic=94, also on end of this document = .
First =96 look=20 at the the buffer PSU schematic on the top left of the page, when = oriented as=20 you=92re used to looking at a page and still able to see letters
.Transformer=20 =96 its wisest to use common one for both Optical Volume and buffer = circuits.=20 Whatever. While you=92re free to use center tapped or not center tapped=20 transformer for Optical Volume circuit, here you must use a center = tapped one,=20 because we need a two-leg supply, where we have both a positive and a = negative=20 leg. Of course now, because the power grounds on two circuits/pcbs = aren=92t=20 connected through any other possible path =96 you can wire it for PCB01 = in any of=20 previously mentioned ways, and for PCB02 you must connect the center tap = where=20 it belongs .
The Reg itself, is pretty similar to the already = explained=20 reg for Optical Volume. So it=92s pretty straightforward; transformer, = snubber=20 cap, Graetz diode bridge, electrolytics + solid caps filter for each = leg, then=20 LM317 (100mA CCS for pos. leg), LM337 (100mA CCS for neg. leg), then an=20 interesting thing, a sort of common (for both legs) Zener reference = string, made=20 of D5,D6,D7,Q1,D8,D9 and D10. The first three are a voltage reference = for the=20 shunt darlington Q2, while the last three are voltage reference for the = shunt=20 darlington Q3. Q1 is a little CCS for the Zener string, serving both = halves,=20 upper and lower, and not being connected to ground.
OK, then we = have a=20 string made of Q2 and Q3, and their emitters are connected to output = ground,=20 while each Zener string half is finding their potential path/point via = the=20 corresponding Darlington. Anyway, to cut the story short, and my pains = in=20 explaining things on non-native, you=92ll have a pretty low noise reg, = with=20 5V1+0V65+5V1+1V3=3D13V15 on each output. Twice 5V1 is obviously from the = two=20 Zeners, 0V65 is from conducting connected uni diode (1N4148) and 1V3 is = 2Vbe of=20 the Darlington.
Stability vs. Time isn=92t of much importance = here, just=20 because once all parts reach their temperature equilibrium, the output = voltage=20 is pretty stable. Conducting connected 1N4148, which have opposite = TempCo to=20 Zeners is purely from habit here.
C5 & C6 are noise = suppressors for=20 the Zener string. CX and CY are possible oscillation killers and tiniest = noise=20 removers. Now I need rofl smilie, but I don=92t have it handy
.=20
C7,C8,C9,C10 - noise killers, what else ?Now look lower on the=20 schematic. Pick the one you want of the 4 framed schematic parts, = they=92re=20 identical. Assuming that you=92re looking at the farmost left, that=92s = nothing=20 other than a JC/Borbely (or Borbely/JC=96who knows?) buffer stage, made = of two=20 complementary Jfets. Unfortunately they=92re the scarce Toshiba ones, = but you=92re=20 probably so eager to spend a few Greenies making the ultimate Macho = Volume=20 Gizmo, that you already have them.
R1A: Usual=20 anti-oscillation/helper/niceplaceformylatestbestbrandofshinyresistor. = R2A=20 defines the input impedance of the buffer, and at the same time = conveniently=20 taking the gates of both Jfets to ground. Q1A & Q2A, they=92re the = workers=20 here. Ever see two Woodcutters working with a crosscut saw? They=92re = doing same=20 thing here, pushing and pulling (Got ya! a crosscut saw can be pulled = only
C4A & R4A , C3A & R5A are snubbers. = C1A=20 &C2A are local decoupling, along with snubbers. R3A is there to = clamp=20 outputs to ground, even when the outputs have no load. Well, one buffer=20 explained. You must believe me when I say that the other 3 are the same, = just=20 differing in parts nomenclature and with the suffix B,C, or D instead of = A.
Before taking this writing (for me), and reading(for you) = adventure=20 further, lets sum up a few things ...
The entire gadget is made = of 4=20 volume blocks on PCB01, called PSM Optical Volume, and 4 buffer blocks = on PCB02,=20 called PSM Buffer. We can use/combine them in several ways, and only you = can=20 decide the best way to do it. Certainly the simplest way of combining = them is a=20 pure balanced attenuator, buffered on the output, meaning that each of 4 = volume=20 blocks/channels is buffered on the output.
OK, shall we power up = the=20 bloody thing(s)? Naah, first we must be sure that you didn=92t mess up a = few=20 things regarding LDR orientation. Lets look at some nice pics, brought = to you=20 courtesy of spavleski. See here:
Be=20 careful! The short (LED) legs are fragile. Even if you have a dot on the = LDR=20 body, re-check the dot position with the diode tester on your DVM. The = rest of=20 the picture is self=96explanatory .
LDR=92s are in opposite = position in right=20 and left columns. Nota Bene: LDR=92s are drawn as photo-transistors on = our=20 schematic (now you=92re saying before I can: =93on end of document, = PCB01=20 Schematic") But don=92t worry, the LED=92s are drawn the right = way.
Its now=20 time to ask you =96 Did you wire that funny pot in this way?
Now=20 is also the time, assuming that you checked twice how the LDR=92s are = positioned=20 and soldered, and that you already soldered all the other parts = (including those=20 funny PC-like jumpers JMP01 and JMP02). There is no need to mount any = heatsink=20 now on IC1 and Q7, just check their temperature later with a finger. = Warm is OK,=20 if hot then mount some small heatsink of any sort you have in your parts = drawer.
Ready for powering up PCB01? Do not connect the = transformer leads=20 to PCB02. At this time we need it just for PCB01.
- remove JMP01 and=20 JMP02.
- turn main potentiometer to 12 clock.
- turn mains ON and = check=20 PSU voltage. A convenient place for the black probe is any GND point = (upper=20 ground plane - LSP1 point ), and for your red probe the upper pin of = either=20 JMP01 and JMP02. You=92re measuring PSU voltage through the LDR=92s LED = chain in=20 fact, but that's convenient to check LED=92s orientation in same time. = Expect=20 voltage of around 16V6. If you=92re serious about programmed voltage, = change one=20 of the 5V1 Zeners to a 3V9 one, and you=92ll be in the 15V ballpark. If = that=92s the=20 case check few more things:
-Is LED1 shining? if yes, OK (it=92s = supposed=20 to shine )
-Check the voltage (ref to ground of course) on the = middle leg of=20 Q3. It must be around 2V5.
Next, power down and connect a mA = meter=20 between the pins of JMP01 with the main potentiometer still in a 12 = O=92clock=20 position. Then power up. The current is probably few mA. Turn the main = pot=20 slowly clockwise, taking care that the current does not rise above 15mA. = When=20 you are on max. position with the main pot (5 Clock), set the current to = 15mA=20 with P1.
Now power down, remove the mA meter clips from JMP01, = place a=20 jumper on it, place the mA clips on the JMP02 pins, and turn the main = pot to 12=20 O=92clock. Then power up and check the current. It must be in range of = few mA.=20 Turn the main pot slowly anti-clockwise, taking care that the current = does not=20 rise above 15mA. When you are on minimum position with the main pot (7 = O=92clock),=20 set the current to 15mA with P1. Power down, remove the mA meter clips = and place=20 the bridge on JMP02. Bingo, you=92re done with PCB01 for = now.
Shall we=20 progress to PCB02 powering up now?
If you=92re finished with = that first=20 beer, don=92t use another one .....
=20
If you=92re in for another mug of coffee, now is the time.OK, = you can=20 leave the transformer wires connected to PCB01, and you must connect = them to=20 PCB02 now. You didn=92t forgot to solder one tiny jumper, situated = between the two=20 right most 100uF caps and right most 5V1 Zener did you? That jumper is = for=20 connecting all ground-planes to actual ground .
There is one = slight=20 omission in PCB02 ; you must solder a wire bridge between two bottom = pads =96 look=20 at the picture =96 drawn as a yellow line. Its on the rev.B pcb, and we = are going=20 to solve that mistake in later pcb issues. Then this note won=92t be = needed, along=20 with this pic:
You=20 don=92t need heatsinks on LM=92s and Darlingtons to start. Check later, = as I=20 mentioned before. Warm is OK, but if hot mount any small heatsink you = have. Pads=20 for buffer Jfets are oriented to ensure a face to pace position of = little=20 critters. Feel free to put a small blob of thermal goo between them and = tie them=20 with plastic wire tie or even small piece of heatshrink.
Power up = da=20 thingie. Check positive and negative PSU rails with the black probe on = ground=20 (mid point between two 2200uF caps is convenient place), and red probe = to the=20 middle pin of BDX53C for the positive leg, and then red probe to middle = leg of=20 BDX54C for the negative leg. If these two voltages are in the ballpark = of 12V be=20 happy. Next check DC voltages across every 50R output trimpot. Its fine = if you=20 see around 250-300 mV That=92s the sign that both Jfets in each buffer = are=20 conducting. While you are there, if everything is OK after 10min or so, = put=20 black probe on ground, red probe on each ( one at a time) buffer output = and=20 adjust DC offset to zero with the trimpot.
Now, you can power off = your=20 new Semi-Macho gadget and think of what you really want to make. It=92s = sorta=20 complicated to describe the several possible variants, so I made few = sketches=20 and I hope that things are clearer that way.
First =96 full blast = .........=20 nope, just in case, I=92ll start with simplest = configuration:
It=92s = pretty obvious that this configuration is for plain stereo attenuation = =96 two=20 attenuator channels, two buffers. That means that each channel output is = buffered. It=92s handy to have spare pair of outputs so you can add an = additional=20 pair of RCA=92s and feed them ditto from attenuator outputs, before the=20 buffers.
Next one is somewhat more complicated :
As=20 you can see, two attenuator sections and four buffers. That way each of = the=20 stereo unbalanced channels is buffered on both the input and output. = Same=20 thinking applicable as in the previous configuration, you can mount an=20 additional pair of RCAs and feed them pre-output buffers . You can also = mount a=20 small 2-deck toggle switch on the front panel and bypass the input = buffers=20 whenever you feel that=92s wise in case that you have one source with = enough=20 muscles to drive the LS directly, and other sources that are weaker in = driving=20 capability. A toggle switch can be used to control a small relay, which = will=20 usually have better contacts than the switch itself, so it=92s = beneficial to use a=20 relay as a bypass switch. You can steal a few dozens of mA for the relay = from=20 the raw PSU before the PCB01 reg. Be creative and find these points, or = for a=20 wiser/purist solution, make a small pcb with a tiny Graetz diode bridge = and a=20 small electrolytic cap, and put an adequate resistor in line with the = relay=20 coil. That way you=92ll not pollute any part of the DC grid in the=20 gadget.
OK, the next one is the (in fact =96 almost) full blast=20 configuration. A fully balanced attenuator , buffered on output. Look at = this:=20
The=20 picture is pretty self- explanatory, even without the text below
Now,=20 just one final teaser ...... This is what I call full = blast!:
You= =20 would need two PCB02 boards for that final configuration, and I firmly = doubt=20 that anyone needs that, or has the time and will to make = it.
Well, while=20 I=92m in mood for writing, maybe I can write few words about some final = tweaks=20 needed not so much for proper functioning of our PSM LS, but for peace = of mind.=20 You=92ll need a plain DVM (a =91scope is even better) and any sort of = signal=20 generator. It can be a proper siggene, it can be a tube oldie, it can be = a test=20 CD, or a simple PC sine wave proggie output from sound card of crappiest = PC you=20 have at home...
Now I=92ll throw in some handy pictures here = (courtesy of=20 spavleski) of several most convenient ways of wiring Da Thingie. First = the most=20 complicated:
nice=20 place for two pictures. The first one is the crosscut saw in action .... = Wood=20 Diy=92ers understand this..
if = these two are like Jfet buffer then these must be analogous to a = cascoded Jfets=20 buffer :
Of = course =96 you can make a mix of XLR/RCA connectors on both ends, input = and=20 output, just like in this picture. And I presume that you must know how = to wire=20 and mix these two if you=92re willing to make gadget like this = one:
Be=20 warned, this circuit isn=92t any sort of balanced/unbalanced or=20 unbalanced/balanced converter. You can use it balanced in/unbalanced out = (or=20 vice versa) but you must tie pins 1&3 on the empty XLR while using = the=20 adjacent RCA. And yes, I just recently remembered the XLR pin outs, just = in case=20 you recently forgot:
Pin 1 =96 Ground
Pin 2 =96 Positive = Leg
Pin 3 =96=20 Negative Leg
Now I can presume that you have decided which=20 arrangement is for you, that you read both the thread and this funny = paper, that=20 you wired all needed wires between the transformer, both PCBs, input = selector=20 (no, I can=92t teach anyone how to wire the selector, hehe), and all = input and=20 output connectors, so we are ready for the final procedure.
We = need any=20 sort of signal generator, set for ~ 500Hz and a DVM on AC voltage or a = =91scope.=20 I=92ll show block schematic for balanced stereo. If you=92re using = unbalanced=20 stereo, just cover half of the schematic with blank paper. (if you are = reading=20 from a screen, tape paper on the screen hehe).
This=20 procedure is brought to all of us by Mighty spavleski
.=20
OK =96 whazzat ? Nothing other than a signal generator connected to = ALL the=20 hot inputs in parallel. Obviously, all input grounds are tied together. = That way=20 all 4 channels/sections are fed with the same signal. On right side, = connect=20 your AC meter, (either the DVM on AC or a =91scope) to ground and to one = output=20 (HOT) at a time .
Step 1 =96 position main volume pot on 9 clock; = measure=20 Out 1 & write down, measure Out 2 & write down, measure Out 3 = &=20 write down, measure Out 4 & write down.
Step 2 - position main = volume pot=20 on 12 clock; measure Out 1 & write down, measure Out 2 & write = down,=20 measure Out 3 & write down, measure Out 4 & write down.
Step = 3 -=20 position main volume pot on 3 clock; measure Out 1 & write down, = measure Out=20 2 & write down, measure Out 3 & write down, measure Out 4 & = write=20 down.
Step 4 - position main volume pot on 5 clock; measure Out 1 = & write=20 down, measure Out 2 & write down, measure Out 3 & write down, = measure=20 Out 4 & write down.
If you=92re lucky, all values will be in = 5% or even=20 smaller bracket. If not, you have one beautiful trick as a weapon tnx to = my=20 clever engineer friend spavleski.
Lets say that 3 channels are = pretty=20 equal in attenuation and that one is weaker. Lets say that the weaker = channel is=20 channel No.2, intended for R channel, negative leg. If you look on the = PCB01=20 schematic, you=92ll see that the LDR responsible for that section is OK1 = as series=20 LDR and OK2 as shunt LDR. That channel is weaker than the others IF = shunt LDR=20 (OK2) consistently has a smaller resistance than needed. The way to cure = that is=20 to steal some current flowing through it. Take a look at this:
I=20 already wrote previously that we are not going to mount these 8 trim = pots on=20 PCB01. If you bought them in the meantime, you can curse me, or you can = keep=20 them for something else funny, and curse me again.
So now we have = a few=20 empty holes in that PCB, and they are a perfectly adequate place for = soldering=20 two parallel 3mm red LEDs in parallel with our nasty behaving OK2. That = way we=20 are stealing some current from it in the entire range of operation, not = just in=20 some narrow area around the desired (listening volume) range, as is case = with=20 parallel connected trimpots. This solution is certainly better. Of = course you=20 can try for a start, curing it with just one red LED in parallel to OK2, = and=20 repeat the procedure as needed. This time you need to just compare the = =93faulty=94=20 channel with the nearest one. If needed, add another LED in parallel to = first=20 one.
Now, if you=92re using a =91scope for this procedure, = power up once=20 again that signal generator, position the main pot between 7 and 8 = clock,=20 connect the =91scope to any output, and fiddle with P3. The goal is to = see just a=20 slight signal wobble on the screen, meaning that we choose a natural = position=20 where we can just hear music singing through our new gadget. 7 is=20 zer/nada/zilch/mute, and around 8 is minimum. If you have just a DVM, = perform=20 that part of procedure with a signal source connected to input and an = amp &=20 speaker on the output when you finally place the PSM LS in your system.=20
Final notes =96 to achieve the same thermal conditions for all = LDRs, we=20 prepared one nice 3mm hole ditto in center of LDR area. Look at these = small pics=20 :
So=20 =96 I think that the technical part of this text is almost or completely = finished.=20 I=92ll add a BOM, in fact a list of parts, without a list of places = where it is=20 both cheap and clever to buy them. I=92m anyway from a different part of = world=20 from most of you, and all the parts are probably easy enough for you to = get at=20 the nearest Tobacco Store .....
Just few more = facts:
Lowest input=20 impedance of LS itself is in range of 350 Ohms, at somewhere around the = 9=20 O=92clock position of the pot, while max is up to 35K near both extremes = of the=20 pot. The input impedance of the buffer stage is pretty much the same as = the=20 input gate(s) resistor, 220K, while output impedance is well in the = range of 50=20 ohms.
PCB01=20 - PDF Download
PCB02=20 - PDF=20 Download
-------------------------------------------------= -----------------
=20
thread;
=20
=20
=20
=20
kudos=20 for at least triple grammar and spell check work ; you boyz are the best =
P.S.=20 please - do not edit that announcement