Despre surse, rejectii, stabilizari, etc in audio

[Greierasul]

Eram sigur ca nu ai masurat bine sj ai interpretat gresit ceea ce trebuia. Nu aia este tensiunea continuua despre care vorbim noi aici. Ufffff.....

[Victor]

Ba da .

Se face un filtru trece jos ( care diferentiaza componenta AC de DC ) si dupa aia se masoara componenta DC .

Eu am facut filtrul trece jos din droselul si condesatorul unui neon de 40 W ( model vechi ) ,

Tu ai evidentiat cu osciloscopul doar THD-ul retelei , exagerat de grav deteriorat , situatie datorata probabil mai multor consumatori casnici avand cos fi mai mic de 0,85 .

Pentru a remedia acest lucru ai nevoie de un POWER CONDITIONER , chiar si varianta pasiva prezentata de mine pe acest forum este foarte multumitoare in cazul THD-ului tau .

[Greierasul]

Leco, daca tensiunea continuua de pe retea ar fi de 10Vcc cat ai gasit tu, pai atunci transformatoarele ar lua foc pur si simplu. Iti dai seama daca rezistenta primarului e de 1ohm , cam cat e la un traf ceva mai puternic, ca vom avea doar pentru partea reala a impedantei un curent de 10A ? Iar asta s-ar traduce intr-o putere de 100W disipata in primar ? Asta la modul simplist neimaginandu-ne curentul de magnetizare care ar apare in atare situatie...

Revenind la explicatiile teoretice, tensiunea continuua apare pe retea in urma debalansarii tensiunii alternative. Practic tensiunea alternativa se muta cu totul mai sus sau mai jos cu tensiunea continuua respectiva. Deci tensiunea continua apare la axa nu la varfuri. Tot acolo se si masoara. Considerand ca impedanta retelei nu este zero, daca tragi curent numai dintr-o alternanta vei debalansa reteaua. Asta este unul dintre motive pentru care nu este recomandata redresarea monoalternanta.

Daca vrei sa masori tensiunea CC de pe retea, asa cum ti-am spus si eu si Victor, ai nevoie de un filtru integrator, practic un filtru trece-jos cu frecventa de taiere setata mult sub 1Hz. Cel mai simplu este sa inseriezi o rezistenta de 100k ohm cu o capacitate bipolara de 500uF. Acest circuit serie RC se va conecta la bornele retelei de 230Vac iar tensiunea continuua se va masura pe condensator cu un simplu multimetru setat pe tensiuni DC. Masuratoarea o faci dupa circa 5 minute de la conectare la priza, nu imediat. Capacitatea bipolara o obtii din 2 electrolitici polarizati de 1.000uF legati intre ei la ornele de + sau de - , dupa preferinta. Nu conteaza tensiunea condensatoarelor, poti sa folosesti din cele de 6,3V fara probleme(fiind mai ieftine).

[Greierasul]

La fel de curios sunt si eu sa vad intr-un bloc ce tensiune continuua exista in medie pe retea. Vezi dimineata si seara. Apoi vezi in weekend-uri cum sta treaba. Trebuie subliniat ca cele mai afectate de tensiunea continuua sunt toroidalele de putere care sunt de calitate, cele care au caderea de tensiune mica sub sarcina , de 5%. Adica pe scurt sunt mai afectate aparatele audiofile. Aici nu e nevoie de volti de DC, practic cateva zeci de mV incep sa dea peste cap transformatorul toroidal. Mi s-a intamplat de cateva ori mizeria asta, tocmai de aceea cand te duci in alta casa sunetul se poate schimba dramatic, iar apoi un filtru care are un DC-trap sa faca asa-zise minuni.

[Greierasul]

2,2Vcc pe tesniunea de retea ????????? Ce faceti frate in blocul ala de ai o tensiune in halul ala ? Am mai auzit si eu de 0,2Vcc dar de 2,2vcc....in cazul asta nu te ajuta nici DC-trap-ul. Nici nu vreau sa stiu cum functioneaza toroidalele in aceste conditii. Daca ai vreo scula audiofila e musai sa treci pe trafuri E+I. Nu ma pricep la SMPS-uri, dar nu sunt mai potrivite in astfel de situatii ?

Dupa cum vad eu ca e valoarea tensiunii alternative, de doar 224Vac, devine clar ca impedanta retelei este exagerat de mare.La mine tensiunea variaza intre 230Vac cand e incarcare mare pe retea si 238Vac seara-noaptea. Ma indoiesc ca ai in bloc consumatori asimetrici foarte mari, ci mai degraba cred ca impedanta retelei este prea mare. Probabil ca aveti pe scara sau la intrarea in bloc un contact imperfect. Sau poate chiar Enel-ul sa aibe un traf coborator zonal subdimensionat, cine stie.

Chiar acum o sa refac si eu testul dar de data asta cu condensatorii in paralel cum i-ai pus tu, nu inseriati cum i-am avut eu.Ma indoiesc ca vor diferi rezultatele dar sa totusi sa fie cat de cat apropiate testele de masura.

[Greierasul]

update; reluata masuratoarea in spiritul "Leco style" . La mine DC pe retea era de circa 14mV. Firma unde am facut masuratoarea avea pornite in acest timp 144 de neoane. Puterea instalata 10kW trifazat.

">https://www.youtube.com/watch?v=P3kRNompntA&feature=youtu.be

[Shambala]

Nici unul nu ati avut curiozitatea sa cititi linkul catre DIYAudio despre dc-trap pe care l-am pus.

Circuitele dc-trap se pot face cu oricate diode in serie, nici vb de vreo limitare la 2 diode ! Daca sunt 2,2V.. se pun pur si simplu cate 4 diode in serie. Merita sa rasfoiti macar ...

[Greierasul]

Sigur ? Diodele reprezinta doar un accesoriu si nu piesa principala in ecuatie. Condesatorul este regina. Daca te apuci sa limitezi caderea de tensiune pe condensator la mai mult de 1,4V cu ajutorul diodelor esti obligat sa nu mai folosesti electrolitici ceea ce duce catre inchiderea proiectului. Mai departe, daca totusi vrei sa folosesti electrolitici trebuie sa-i pui cap in cap ceea ce face sa-ti creasca reactanta capacitiva la frecventa de interes de 4x. Deja incepe sa nu mai fie audiofil. Daca totusi insisti si vrei sa te duci catre tensiuni si mai mari, de 10V sa zicem, ramane valabila solutia cu electroliticii cap in cap dar va trebui sa-i alegi la tensiuni foarte mari asa incat sa nu explodeze la curenti mari. Nu-i de joaca atunci cand trebuie sa manipulezi curenti mari cat si tensiuni mari in acelasi timp in condensatori.

Sa fim seriosi, pana la 1V DC-trap-ul este viabil, aproape transperent , ieftin si compact. De la 1V in sus este mai degraba o problema majora a retelei de alimentare si trebuie remediata altfel. Inseamna ca impedanta retelei este prea mare si chiar daca scapi de tensiunea continuua de pe retea aceasta incepe sa danseze dupa ritmul muzicii din cauza impedantei mari

[Greierasul]

Vezi ce impedanta are reteaua la tine acasa. Masoara tensiunea in gol apoi baga un resou in priza de 1kw. Masoara din nou tensiunea si vezi cu cati volti cade aceasta sub sarcina si imparte la curentul resoului pe care-l stii. Rezulta impedanta. Daca ai o cadere de tensiune mare atunci problemele sunt la tine. Daca e rezonabila atunci e de la ei. Ma indoiesc ca lucrarile la metrou se fac doar cu grupuri. In proportie de 90% problemele vin de la constructia metroului,acum e clara treaba.

[Greierasul]

Impedanta la tine in casa e de 0.1ohm. O valoare foarte buna, apropiata de idealul practic. In schimb impedanta retelei externe este mare, de circa 0.7ohm. Probabil cablarea blocului subdimensionata, sau vreun contact imperfect, sau traf coborator mic. In manualul de utilizare al unei centrale termice electrice se specifica sa nu de depaseasca valoarea de 0,1ohm la impedanta retelei. Fa audiofilie in conditiile astea daca mai poti...

[Victor]

Deja imi pun problema daca merita sa continui cu acest subiect ... sper ca cel in cauza sa se abtina in viitor de la aberatii tehnice in subiectele altora .

Sa trecem la putina teorie despre DC Trap :

Va prezint un filtru de retea fabricat de ceainizi impotriva RFI dar si separator galvanic 1:1 pentru consumatori mai " delicati " :

Va prezint o varianta DIY pentru DC Trap care foloseste diode Zenner de 5W / 6,8V , aviz celor ce au tensiuni DC foaaarte mari suprapuse peste reteaua de AC 230 V , sper sa remarcati ca nu s-a facut niciun compromis la electrosecuritate :

Ca fapt divers , varianta DC trap din simularea B functioneaza perfect cu condesatori de 6.800 micro / 16 V , intr-un aparat cotat ca fiind audiofil .

Nu doresc sa-i mai fac reclama ...

[Victor]

[Victor]

Acum sa trecem la lucruri cu adevarat serioase :

Aici aveti schema , in varianta pentru Europa , 230 V AC

[Greierasul]

Toate schemele de DC-trap pe care le-ai postat astazi Victor, intr-adevar pot lucra la tensiuni continuue oricat de mari (limitate de tensiunea condensatoarelor) dar au si o parte negativa. Acest tip de DC-trap face ca intregul curent de pornire (inrush current) sa fie incasat de catre condensator cu consecinte dezastruoase pentru fiabilitate. Deci este obligatorie folosirea unui circuit softstart care sa limiteze curentul la o anumita valoare tolerata de condensator. Impedanta DC-trapului este de 4x mai mare fata de cazul in care condensatoarele ar fi fost in paralel. In plus sunt necesari condensatori de valori foarte mari ,cu ESR-ul cat mai mic si curenti de lucru cat mai mari daca se doreste ca aceste circuite sa aiba o viata cat mai lunga. Pe scurt sunt necesare condensatoare foarte scumpe si de dimesniuni mari.

Al-2-lea circuit reduce tensiunea pe condensator cu 0,7V din cauza diodelor puse in antiparalel fata de primul circuit. O idee buna la prima vedere dar ne alegem cu 2 diode care vor functiona in continuu ceea ce implica zgomot de comutatie suplimentar.

Al-3-lea circuit DCtrap simulat in locul diodelor pune o rezistenta. Cu cat creste curentul de lucru prin ea cu atat va scade tensiunea pe condensator si il vom proteja pe acesta din urma. Ideea nu este rea dar asta presupune sa fim de acord cu cresterea impedantei retelei cu 1ohm. Suplimentar la un curent de 10A aceasta rezistenta trebuie sa disipe 100W.

Al-4-lea circuit reduce valoarea rezistentei si o inlocuieste cu o reactanta inductiva. Buna ideea , se reduce disipatia pe rezistenta dar ramane impedanta mare a retelei iar suplimentar avem o defazare U-I si posibilitatea de oscilatii si influenta campului elmg astfel creat.

Tinand cont de faptul ca tensiunea continuua de pe retea afecteaza transformatoarele in special la puteri mici, cu cat crescand puterea de lucru influenta negativa diminueaza, prefer DC-trap-ul clasic foarte bine si rational calculat. Daca tensiunea continuua de pe retea este in limite normale, sub 0,1Vcc, cel mai bine este fara.

[Victor]

Nu conteaza ce preferi tu , oricare dintre variante este superioara recomandarii initale de a nu folosi DC Trap .

Desigur , inteleg retinerea ta datorita pericolului potential de electrocutare a unor useri nefamiliarizati cu tensiuni periculoase , dar asta nu-i impiedica pe unii sa foloseasca acasa amplificatoare cu lampi , care la randul lor folosesc tensiuni mult mai periculoase decat un amarat de DC Trap .

Ca sa revenim la subiect : ceainizii au lat o carcasa foarte frumoasa , au montat in ea trafo separator galvanic , filtre antiparaziti , cateva prize pentru distributia tensiunii 230 V si cam atat , pentru uz audiofil , studiouri de inregistrari etc .

Remarca ta ca este necesar un softstart pentru limitarea curentului initial al condesatoarelor din DC trap este fara sens logic , mai gandeste-te inca o data .

Cat despre " calculul unui DC Trap clasic foarte bine si rational calculat " nu pot decat sa te felicit , sunt ferm convins ca e de preferat sa mergi doar pe cai batatorite .

Eu voi merge pe o alta solutie , mai exotica , inca nepopularizata pe plaiurile mioritice .

Momentan o pastrez doar pentru mine , dupa o testare serioasa voi vedea daca se justifica sau nu o alta solutie decat cea clasica .

Pe alte forumuri , se discuta foarte intens aceste probleme legate de variatiile de tensiune , puritatea spectrala a retelei de 230 V AC si compensarea factorului de putere , de tensiunea continua aleatoare suprapusa peste cea alternativa , paraziti RF .

Unii propun scheme minimale gen : Softstart , DC Trap , GND Lift , RFI Filter si uneori separare galvanica 1:1 , considerand ca peste acest prag costurile cresc astronomic iar variatiile retelei sunt cu adevarat importante doar in zonele cu consumatori industriali .

" Noroc " ca la noi nu mai e cazul , consumatorii industriali au disparut .

Uite o propunere decenta a unor audiofili constructori de pe alte forumuri , montaj care de altfel care functioneaza excelent fiind evaluat pozitiv de un mare club audiofil ( rusesc ) , asa ca nu e cazul sa insisti cu ineficienta sau pericolul electrocutarii , deoarece nu exista nimic antiprosti :

[Victor]

Punerea in evidenta a mizeriilor produse de sursele in comutatie etc si alti paraziti suprapusi pe reteaua de 230V care ne alimenteaza aparatele noastre trebuie rejectate cu eficienta .

Va prezint un asemenea aparat care se poate adapta la aproape orice filtru RFI la care avem acces la bobina din intrare , aparatul pune in evidenta nivelul parazitilor iar filtrul netezeste varfurile din retea si merge combinat perfect cu schema de mai sus .

editat: altadata ori "iti bati capul" ori te abtii de la alte comentarii.

"""

How it works

The filter section begins with six VDRs, which are Intended to remove the damaging effects of high energy tran-sients on the mains.

To some extent they will reduce Impulsive Interference effects loo but will not eliminate them.

The filter section will remove RF Interference from the power lines.

The current balanced inductors in com-bination with the Y-capacltors (08, 9, 12,13) serve to clean up common mode Interference, while the X-caps (C1-6, 10, 11) do the same for differential mode noise.

The current balancing in the toroids prevents the cores from saturating under the effects of the current drawn by the load.

The pick off coll from the first toroid detects any imbalance caused by interference currents flowing to ground via via (he Y-capacitors.

The signal is amplified by IC1a and passed to the detector circuit consisting of Q1 and 2 and associated components.

This detector responds to the peak value and to the duration of the signal, so a short, high voltage pulse wfll give the same reading as a sustained, low amplitude burst.

IC1b feeds the detected voltage to IC2, which is a common or garden bar-graph drive 1C.

The LEDs are fed with current pulses from D5 to reduce the overall current consumption of the circuit and dissipation in IC2.

The 1C is switched to dot mode twice each cycle of the mains (via pin 9) to reduce the current requirements still further.

If you look closely at the display, you might just discern a difference in brightness between the highest dot and the rest of the bar but the overall effect is of a continuous bar display.

The power for the low voltage circuit is derived from the mains via C14.

R3 prevents damaging inrush currents If the mains happens to be close to its peak value at the time the circuit is switched on.

R2 provides a discharge path for C14 when the conditioner is disconnected from the mains or if the fuse should blow.

This kind of power supply does not isolate the low voltage circuit from the mains and is only suitable for use in completely self-contained pieces of equipment like the conditioner.

The supply capacitor will be large but nowhere near as bulky or heavy as a mains transformer for circuits requiring small currents (up to 100mA or so).

A capacitor used In this way should be X-rated since it is effectlvely connected across the mains.

Mains conditioner photo

Year by year the pollution of the mains supply grows steadily worse.

In addition to the usual industrial effluents from rotating machinery, waste products from switch mode power supplies, sewage from drills, washing machines, vacuum cleaners and oven thermostats, there are now plans afoot to pollute the mains deliberately.

I hardly need to mention the consequences — streaky TV pictures, popping and crackling radios, mushy hi-fi sound.

Greenpeace — where are you when we need you?

Mains borne interference is not a thing to be taken lightly.

Spikes of 1 kv and above are a common (in some areas frequent) occurrence and this can and does damage unprotected equipment.

A simple voltage dependent resistor (VDR) connected between live and neutral of the mains plug will usually forestall damage to the equipment but it doesn't prevent the annoying interference effects.

Apart from spikes and impulsive interference, there is a constant background of more regular interference which gets steadily worse as time goes on.

RF interference has become more of an annoyance since the CB boom and the increasing use of switch mode power supplies adds its own contribution.

The latter are supposed to be suppressed at source but this only serves to reduce the interference and doesn't eliminate it.

Another development has been the increasing use of the mains for signalling purposes.

At its lowest level this can be equipment such as cordless intercoms but the problems associated with sending digital signals through the mains are rapidly being overcome.

Some years ago National Semiconductors introduced the Bi-Line system, the front end of which was an IC (the LM1893) which puts data through the mains by means of an FSK modulation system.

It was, by its nature, for localised use but this and similar systems — even the home computer add-ons for through the mains control — are all adding to mains borne interference.

A system to eliminate gas and electricity meter readers has now reached the stage of field trials.

The idea is that meter readings are sent via the mains as far as the nearest sub-station, from where they will be transferred to the telephone lines by means of a modem.

This long distance use of mains signalling obviously can't be suppressed, so a band has already been set aside for it.

One can envisage a time when the 'mains waves' will be just as strictly regulated (and just as crowded) as the air waves.

The effects on hi-fi and audio equipment have yet to be seen.

In addition to all this man-made interference, there is another source which will always be beyond any kind of legal regulation and control — the weather.

Electric storms and even lightning strikes make their presence felt through the mains.

The only way to be sure of an unpolluted power supply for your audio equipment, TV or computer is to clean it up yourself.

The power conditioner is the tool you need for the job.

Inside the conditioner the mains supply is purified, transients are cleared and RF interference is blocked.

The clean supply is then fed to a socket or multi-way outlet which can supply power to all your sensitive equipment.

If you find it hard to believe that the mains is really as polluted as I say, this project will certainly convince you.

A unique feature is its bar graph display which actually lets you see how much interference it is removing.

As you watch the LEDs move and occasionally flick way up towards the top of the scale, you'll be in no doubt that the power conditioner is working for its living.

The correct way to avoid any problems with mains connections the gospel goes is to plate all your plugs with gold.

The reasoning behind this was explained to me by the proprietor of Hi-Price Audio to be something like this:

The gold plating on the plug acts very much like the uniform of the doorman at the Dorchester Hotel.

Nice, well-bred sine waves know that they will be welcome inside, whereas interference is overawed By the golden splendour of the doorman's uniform and embarrassed by its own scruffy appearance.

It knows that it will feel out of place in such magnificent equipment and wanders on in search of the electronic equivalent of a Yummy Eater fast food bar.

"Besides," he said. "if punters ; fink they can hear a difference, am I going to argue?"

I was impressed by his logic and bought a dozen.

Of course, back in the real world we have a mains filter which works to consider.

Mains conditioner circuit

The filter begins with six VDRs.

This is partly a concession to the fringe hi-fi community who believe that if one is good, six must be six times as good.

For a given spike, the clamping voltage will be reduced by an infinitesimal amount by having a number of VDRs In parallel, due to the highly non-linear voltage to current relationship of these devices.

It's rather like hoping to reduce the forward voltage drop of a diode by wiring half a dozen in parallel.

It will be reduced very slightly but not so's you'd notice the difference.

For more rational beings, there is another reason for having half a dozen VDRs.

A VDR will only absorb a certain amount of energy from a spike before becoming stressed beyond its limits.

If these limits are exceeded, it can result in the VDR breaking open and scattering zinc oxide far and wide.

After that, your equipment is no longer protected.

One of the essential figures on a VDR data sheet is the maximum energy it can absorb in a short period of time.

Figures of 5 to 20 Joules in 10ns are common for small components.

To increase the energy you have the choice of buying a larger VDR or using several in parallel.

Parallels

The parallel option has the advantage that you can choose how much protection you want to give (an upgradable mains filter!) and that the average absorption over a longer period of time will be greater than for a single large VDR.

It could be that because of an electric storm you get just the conditions to pop a large VDR (and your equipment) but which would allow the parallel combination to continue giving protection.

Speaking as one whose new TV set has just been zapped by a thunderstorm which exploded the plug VDR too, the more protection you can give, the better.

For those of you who are not familiar with the characteristics of VDRs, they are rather like AC versions of the zener diode, although the voltage clamping is not so sharp.

Below their rated voltage they are virtually an open circuit.

A little above this they begin to conduct until at about twice the rated voltage they have virtually no resistance at all.

It may seem that a sharper cut-off would be an advantage but too quick a conduction would lead to blown VDRs every time there was a long term surge in the mains voltage.

They are, in fact, very well suited to their job.

The clamping voltage is usually measured at 100A and will be somewhere between 600V and 800V for a device rated for 240V mains operation (which will begin to conduct at about 350V — just above the mains peak).

The peak current for even a small VDR will be many hundreds of Amps but this can only be sustained for a few microseconds.

High peak currents for a very short time is exactly what impulsive interference will give.

Capacitor, heal thyself

The main section of the filter consists of a pair of current balanced inductors and banks of capacitors to remove RF interference.

A number of capacitors in parallel are used in preference to a single large capacitor to take advantage of the much higher self resonant frequency of the smaller caps and also because they are generally able to withstand short term thermal and voltage overload better than their larger brothers.

The value of the capacitors to earth is limited by the need to comply with earth leakage regulations — they are the maximum allowable values, taking into account their tolerance and should not be increased under any circumstances.

Connecting capacitors across the mains puts them under enormous stress and components not designed for the job can easily catch fire, short circuit, or at best just quietly fail — even if the voltage rating is high enough.

Construction of a metalised film capacitor

The basic construction of a metalised film capacitor.

(a) The two strips of metalised film are rolled together.

(b)The ends of the roll are coated with metal and the leads welded on.

© Finally the whole assembly is dipped or encapsulated to give the capacitor you buy from your component supplier.

Capacitors which have been designed to withstand the stresses and to comply with the appropriate standards are divided into three main categories: Class X1 These are for connection between live and neutral in situations where pulses of over 1.2kV can be expected.

Class X2 These are for connection between live and neutral where transients will not exceed 1.2kV.

Class Y These are made to the highest standard of all and are used for connection between a power line and earth or any other situation where failure might expose someone to a lethal shock.

Most capacitors for mains use have the rather magical sounding property of self-healing.

This is a consequence of the metallised film construction, the essentials of which are shown in Fig. 2.

The dielectric material is coated with a very thin layer of aluminium — around 300 Angstroms (3 x 10^8 metres) thick.

Two dielectric strips will be coated — one with a margin on the left hand side and one with a margin on the right.

The two will then be wound together so that the metal film of one 'plate' extends to one side of the roll and the other to the opposite side.

To make the connections, the two sides of the roll are sprayed with metal from a flame or arc gun and the leads attached.

You can see this kind of construction in the 'naked' metallised polyester capacitors — the block shaped ones with metal at either end and leads that fall off at the slightest provocation.

These caps are layered in long strips and then sawn up into individual capacitors rather than being individually wound, but the principle is the same.

The difference between class X and Y capacitors and the cheap 'n' cheerful metallised types is mainly in the standard of construction.

The mains capacitors may be interleaved with paper (sounds an odd material but it has some excellent properties), be vacuum impregnated with epoxy to remove air pockets where ionisatlon may take place, be series wound to reduce electrical stresses, have several layers of bonding metal, be encapsulated in fire retardant material and so on. Construction varies from manufacturer to manufacturer.

If the dielectric is punctured by a high voltage spike, instead of short circuiting through the carbonised mess left behind when the dielectric burns, the very thin metallisation is vapourised away from the area and the capacitor carries on as if nothing had happened! Strictly speaking, the metallisation is oxidised, the oxygen being supplied by the decomposition of the dielectric. The oxide doesn't conduct, so the damaged area is sealed off. It's not quite self-healing but almost as good!

Construction

Mains conditioner layout of components

The component overlay for the project is shown above. Some of the components are mounted vertically to save space — the leads should be bent carefully and not too close to the body of the component to avoid stressing the bonding.

The best way is to hold the lead just above the component body in a pair of pliers, then to bend the lead in a smooth curve with finger and thumb.

The VDR positions have two holes for the 'live' connections, allowing components with either a 0.2in or 0.3in lead pitch to be mounted.

Similarly, the capacitor which supplies the low voltage circuit has two pads for one of its connections to allow two popular sizes of capacitor to be mounted.

The remaining hole is left unused.

Each coil on the two toroids has 15 turns of 1mm diameter enamelled or the circuit will not work properly.

In addition to the power windings, T1 has a further pick-off coil of 15 turns of 0.25mm diameter wire over the centre of the coil in the neutral line.

This connects to points A and B on the circuit board.

The direction of this winding is not important.

The 1 mm diameter wire is firm enough to support the toroids on its own (in fact, you'll need quite strong fingers to wind it into a neat coil) but holes have been provided on the PCB for strapping them down with cable ties, just to be sure.

Figure 4 shows details of the inlet and outlet cables and connections.

A 2BA bolt and solder tag is used to earth the metal chassis of the case and to provide a connection point for all the earth wires.

Strain relief grommets must be used on the panel cable holes to clamp the leads firmly in place.

The front panel is drilled with a line of holes at 0.2in intervals for the LEDs.

I used 3mm round red LEDs in the prototype but there is no reason why you should not use other shapes or colours if you wish.

The usual black mounting clips can be used but they will have to be pared slightly with a sharp knife to fit the 0.2in spacing of the holes.

Otherwise, you may prefer the appearance of the LEDs without clips.

Whether or not the clips are used, the LEDs should be stuck in place with epoxy resin so there is no possibility of the leads touching the panel or slipping through and becoming exposed.

The low voltage section of the circuit is not isolated from the mains, so for safety purposes must be thought of as being live.

When the LEDs and the inlet and outlet cables have been attached to their respective panels, you can solder the power connections to the PCB.

The LEDs are best left unconnected until the case has been assembled, otherwise you won't know how short to trim the leads.

Screw the chassis together, with the PCB resting on the bottom flanges of the side pieces. Turn the whole assembly over and check that there is enough clearance between the metal flanges and the pads and tracks of the PCB.

Check also for solder blobs, untrimmed leads or any swarf on the flanges that might cause a short between the metal and the PCB tracks.

When you are sure that all is well, fit the chassis into the bottom section of the case and screw the PCB to the support pillars.

The LED leads can now be trimmed to size and soldered to the header pins on the PCB.

All that remains is to put in the fuse, screw down the lid of the case, press in the rubber feet and your Power Conditioner is complete!

Front and rear panels and connection to PCB

The front and rear panels and connection to PCB.

Testing

There is very little that could be wrong with the filter section of the circuit except for open or short circuits (you did check the PCS carefully, didn't you?)

Before plugging in, it's best to do a quick resistance check.

Set your multimeter to a high resistance range and check the resistance between ground and live on the inlet lead, then between ground and neutral.

Both should appear as an open circuit.

If there is any movement of the meter whatsoever, don't attempt to use the conditioner.

Check the PCB again, check your input lead connections and if both of these seem OK, take out each Y-capacitor in turn and check its resistance.

The fault can only be in one or other of these places, so you won't have far to look.

A resistance measurement between live and neutral on the inlet or outlet lead should show up a resistance of about 220k — the discharge resistor.

If it is much below this (say, below about 180k, which could just be the result of resistor tolerance and meter Inaccuracies) take out the fuse to the low voltage circuit and see if this makes any difference.

If not, check the PCB carefully and as a last resort check the resistance of each of the X-capacitors.

A final possibility — if you've damaged the coating of the copper wire on the toroid coils and allowed the two coils to touch (I hope not!) this will also cause problems (to say the least!) If all is well so far, check the continuity of the live, neutral and particularly the earth connections.

(Check the resistance between the input earth and output earth and make sure it's zero and so on).

After making sure that there is a suitable fuse in the plug, apply power to the conditioner but don't plug anything into the output socket yet.

You should see the LED display flick upwards as you turn on the power, then the LEDs will go out one by one until they are all extinguished. If you keep watching the display for a while, you'll probably see it flick upwards every now and again as the conditioner catches some interference. Even with nothing connected to the output, it still removes pollution and gives an indication of how much there is around.

If all the LEDs light up and remain lit, don't instantly conclude that there's something wrong.

Take a look around and see if you can find anything that might be causing a lot of interference.

When I first tested the prototype in the ETI lab, all the LEDs lit up and I spent several minutes puzzling what could be wrong — everything seemed OK.

Then the photocopier in the next room stopped printing...

Now is the time to find out how good a job you've made of winding the coils.

Plug your hi-fi, TV set or whatever into the outlet socket and take another look at the LED display.

The sensing circuit will always pick up a certain amount of 50Hz signal from slight imbalances in the inductor and from slight differences in the Y-capacitor values, but it should not be enough to swamp the display.

If most or all of the LEDs remain lit ten seconds or so after plugging something into the output socket, there is a good chance that you have one turn too many or too few on one of the coils.

If one or two LEDs remain constantly lit, you can improve matters by adjusting the coils (or ' re-winding them if they're untidy!) or as an absolute desperation measure the value of R5 can be reduced to bring the display into line. The heavier the load, the more apparent any imbalances will be — an electric fire makes a good test load.

If the display section does not seem to be working properly, don't attempt to test it with its capacitor power supply.

Remove all connections from the mains, set your bench power supply to about 16V, connect the negative lead to the negative lead of C15 and the positive lead to the junction of C14 and R3.

Connect the negative lead of the multimeter to the negative terminal of your power supply.

Check the voltages on pins 9 and 3 of IC2.

Both should be 12V (or within 1V either way).

If both are higher, ZD1 is probably faulty.

If only one is higher, check D5 or D6.

If either or both are low, disconnect the power and check all the diodes (in particular, check they are the right way around).

Also check C15 and C18 and the PCB for shorts.

If the readings are OK so far, check the voltage at pins 6 and 7 of IC2 and pins 1,2 and 3 of IC1 They should all be the same at about 6V.

Touching a finger to pin 2 of IC1a should cause all the LEDs to light.

Remove the finger and they should turn off one by one.

If this works but the display doesn't seem to pick up anything from the mains, check R4 and the connections to the pick-off coil.

If nothing happens at all, measure the voltage at the positive plate of C17 and see if it rises when you touch the 1C pin.

If not, check for a short in C17 (or a solder blob across its pads!) and the connections of 01, 02 and C16.

If the voltage across C17 rises, but the LEDs don't light, check the voltage at pin 5 of IC2.

This should also rise.

If not. IC1 is faulty.

If it does rise but the LEDs don't light, check all the connections around IC2 and replace it if necessary.

If the voltage across C17 remains high at all times (without the finger), suspect 01, 02 or C16.

Using The Conditioner

In the form presented so far, the Power Conditioner can be used with loads of up to 1.5kW.

It will, in fact, cope with loads of 2kW intermittently — I tested the prototype by running it for an hour with a 2kW electric fire as a load.

It didn't come to any harm but it did get rather hot!

Most domestic equipment will have a label or tag on it somewhere to say how much power it consumes.

If you are using a multi-way output socket, don't forget to add the loading of all the equipment you have plugged into it.

As a very rough guide, a TV set consumes 100 to 150W, a 100W per channel hi-fi will consume about 300W with the volume turned up to full blast, a home computer may be anywhere between 10W and 250W depending on whether it has its own screen, disc drives, or whatever.

It is also important to use mains cable that is suited to the load.

To be on the safe side, you could wire the conditioner up immediately with 13A cable but it's wasted if you're only running small, sensitive devices.

The normal 0.5mm2 mains flex will cope with loads of up to 750W total.

The thicker 0.75mm2 cable will be OK up to 1.5kW, so this is probably the best compromise.

Unless you intend to load it to the limit, a 5A fuse in the inlet plug is advisable.

If you are in doubt about any of this. your local electric shop will probably have an electrician who can advise you.

The conditioner will cope with all likely loads as it is (you don't really want to decontaminate the power to your electric fire, do you?) However, there are always one or two big-number enthusiasts who want to upgrade to the limit.

The way to do it is simply to use thicker wire to wind the toroids.

You'll be faced with the option of using fewer turns (which is OK as long as all the coils have the same number, although lower frequency performance will be impaired) or of overlapping the turns slightly.

I wish you luck!

If you do have an application for the higher current version, it would be advisable to solder some thick copper wire along the main current carrying tracks (the wide ones) on the PCB.

Unless you can find a way of winding the coils evenly, or are willing to accept fewer turns, you will probably find the bar graph registering 50Hz pick up.

Reducing the value of R5 will prevent it from swamping the display, which will then be less sensitive but should still give a •good indication of the suppression.

There is no lower limit to the value of R5 — it's up to you to choose a suitable compromise between rejection of unwanted pick up and display sensitivity.

In areas of high RF interference, it is a good idea to keep all leads after the conditioner as short as possible.

Use the inlet lead to give you the reach you need, then keep the outlet leads trimmed short.

Most of the time this will not be critical but it's worth bearing in mind if you live next door to a CB enthusiast Twelve-ten till we do it again, good buddies.

RESISTORS

R1, 2 220k 1/2W

R3 68 1W

R4 4k7

R5 100k

R6 47k

R7 10k

R8 27k

CAPACITORS

C1. 3, 10 10n class X2

C2, 4-6, 11 33n class X2

C7 100n class X2

C8, 9, 12 13 2n2 class Y

C14 330n class X2

C15 2,200ufd 16V radial electrolytic

C16 10n ceramic

C17 2ufd2 16V electrolytic

C18 2pfd2 tant or 10ufd electrolytic, 16V

SEMICONDUCTORS

IC1 LM358

1C2 IM3915

Q1,2 FS40

D1-6 1N4001

VDR1-6 V250LA2, Mullard 593/4 series, or equivalent

LE01-10 3mm red LED

ZD1 12V 1.3W zener

MISCELLANEOUS

T1, 2 FX4054 coated toroid cores wound with 1mm and 0.25mm enamelled wire as per the text

FS1 PCB mounting fuse clips and 50mA fuse

PCB;

case;

20-way right angle PCS header;

mains plug;

mains socket or multi-way connector;

0.75mm2 mains cable: strain relief grommets;

LED clips;

nuts and bolts """

[Greierasul]

Revenind la subiect, cum saptamana asta n-am de lucru la AudiophileStarter iar proiectul SARA l-am tras pe partea dreapta din cauza lipsei de interes din partea clientilor, m-am gandit ca ar fi interesant sa incep a taia puternic din costuri la SARA. Pur si simplu vreau sa vad ce ramane daca latura economica primeaza. Ca atare, am castrat alimentarea, zona din care se pot obtine economii buna de bani. Tocmai de aceea am si postat aici acest experiment, fiind legat de surse si rejectii.

Deci, am umblat la transformatoare. Daca in proiectul original erau 2 trafuri toroidale de cate 350VA fiecare, adica un total de 700VA, am castrat unul dintre ele, pastrand un singur transformator de 350VA pentru o putere de 2x125W in 8ohm. Nu m-am oprit aici si am taiat drastic si din capacitatea de filtrare. Initial , pentru fiecare canal in parte existau circa 60.000uF sub forma CRC. Acum am luat pur si simplu modulul de alimentare/redresare de la AudiophileStarter pe care l-am pus in locul bateriei originale. De la 60.000uF am scazut la 20.000uF(2x10.000uF) asezati simplu, C.

Am pastrat configuratia monobloc de la puntile de redresare in aval deoarece nu era niciun castig semnificativ sa inlatur vreo punte.Dupa o recablare necesara noii configuratii am trecut la masuratori. Puterea de test a fost 50w in 8ohm.

Surpriza a fost foarte placuta pana acum. Diferentele la masuratori nu sunt mari, aproape ca nici nu merita a fi de luat in seama. Din nou singura buba pe care nu reusesc s-o rezolv elegant si ieftin tine de fierul carcasei care preia campul magnetic al trafului si induce in circuite un semnal perturbator de 50hz. Valoarea sa nu este mare , ba din contra, la -93dbV este chiar mica, dar este loc de mai bine, mi-as dori sa fie in ton cu restul zgomotului , sub -110dbV.

Inca nu am audiat noua configuratie, daca masuratorile sunt foarte bune si nu indica o degradare de evidentiat a performantelor electrice, totusi uechea are si ea importanta ei, nu de putine ori poate scoate in fata aspecte pe care nu te-ai gandit initial sa le masori. Daca termin astazi, bag o ureche sa vad ce si cum.

[Greierasul]

Surprinzator THD-ul la 20khz se tine foarte bine si nu vrea deloc sa se lase mai prejos fata de varianta originala. Nici IMD-ul nu renunta la performanta inalta si arata ca se pot obtine distorsiuni demne de cele mai bune preamplificatoare sau DAC-uri dar in conditiile de putere mare la borne.

Banuiam pe undeva ca teoretic daca lucrezi atent amplificatorul si alimentarea se obtin rezultate frumoase si cu varianta clasica a alimentarii, dar trebuie sa recunosc ca doza de surprindere n-a fost chiar mica. Realitatea respecta teoria cata vreme stii ce si cum.

[Mishu]

Să înţeleg că vrei să "reînvii" proiectul Sara?

[Greierasul]

Da, dar la costuri minime asa incat sa devina accesibil. Daca originalul era 2500 de euro, varianta noua nu trebuie sa sara de 1500 euro dar performantele sa fie aproape identice. Rezultatele noi sunt superbe, thd-ul la 1khz e de 0.0003% iar la 20khz e de 0.0009%. IMD-ul e de 0.0015%. Sunetul ramane superb dar este ceva mai dependent de temperatura condensatoarelor din alimentare. In schimb are un sunet mai coerent fata de original. Cea mai mare provocare ramane cablarea.

[Victor]

Ramasesem dator cu schema unui SofTStart , prin intermediul caruia sa se alimenteze toate etajele prezentate anterior :

[Greierasul]

Da, inteligent...alimentarea direct din retea, simplu, ieftin si eficace. Singura parte cu minus ar fi ca butonul de pornire vede tensiunea de 230Vac.

[Victor]

Nu .

Releul de pornire On / Off ale carui contacte alimenteaza / taie tensiunea de 230V AC va fi alimentat doar din tensiunea special generata de preamplificator , de obicei acea tensiune se numeste " Trigger On " .

Preamplificatorul , pe durata functionarii , pe langa semnalul audio , genereaza si cateva tensiuni " Trigger On " pentru comanda amplificatoarelor din lantul audio .

Evident , preamplificatorul ar trebui sa aiba si telecomanda IR , pentru comoditatea utilizatorului . In acest fel , o data cu preamplificatorul porneste si lantul de amplificatoare , fara a mai fi necesara actionarea vreunui buton separat de pornit / oprit la fiecare amplificator de putere din lantul audio .

[Greierasul]

Aha, deci este vorba despre o schema logica mai mare. Eu m-am limitat doar la ceea ce am vazut in poza ta. Da, pare un concept interesant si totodata complex.

[Victor]

Desigur , va stau la dispozitie cu orice lamuriri suplimentare , daca va fi cazul .