The first DC Power Scaling kits were the SB Super Budget kits: SB1 for fixed-bias amps and SB2 for cathode biased amps. These had a very economical circuit and were problem-free to install. Many techs and amp builders still use them and we supply PCBs for them.
The simple circuit required the use of a potentiometer with special characteristics, and the only way to attain that was to use a mil-spec pot. "Mil-spec" is a misunderstood term inasmuch as most people believe that the part is going to be better and/or last longer than an average part. That can be true but is not always the case. Mil-spec merely means that there is an official description of the part so that parts procurement can be uniform. In the case of the pot, its special characteristic is not contingent on it being mil-spec; rather, it just fell out this way as it was the only part available with the given parameter rating.
When I began using that pot style, I was getting an extremely good deal on the pricing. I ended up using a few different values and standardised the use of them in my early production amps. But then the prices went up precipitously. "Shock shock horror horror", as the song goes.
I designed out the expensive pots, which was fine for all the traditional control positions in an amp, but more complicated for Power Scaling. I needed an "amplified pot" and began on a path of designing many different solutions, each with its benefits and pitfalls, with each successive design aiming to improve on what came before. The final solution turned out to be great on paper but fragile in some installations. It eliminated the pot value as a variable in the control range and when I test the circuits here it always works for me. But I'm always lucky
A couple of very experienced Power Scale installers encountered issues with the latest circuits and I made a decision to go back to the SB approach. It is fail safe. It always works. No one has ever had a problem with it as long as they mount the mosfets in a location where air moves over that part of the chassis. Even while the latest-greatest was going out, I was researching the expensive pot options to see what the impact would be on kit pricing.
Frankly, I sometimes thought that I should have just raised the kit prices and kept on with the SB1 and SB2 as they were.
One of the issues with the SB approach was that the Power Scale pot did not physically match the Drive Compensation pot. The expensive pot had far fewer options than standard pots do, requiring different knobs, or ending up with the knobs sitting at different heights. The solution is to get a dual version of the expensive pot, which is even more expensive and larger, making finding places for the two controls more difficult on some chassis. The dual could be used for a single-pot Power Scale solution in some circuits.
Along with looking into the pot availability, I worked on the Automatic Drive Control issue which resulted in two kits to perform the drive compensation automatically. This means there only needs to be one panel control. There is a thread about that on this forum.
All of the PCBs for the new kits have been ordered.
Where the old-style SB module boards carry a single mosfet and its related components, the new SBX module has all of the high-voltage circuitry as one module, with any required interconnections already made. The ability to have different Va and Vs voltages is also accommodated. The Power Scale control ties to this, and there are pads to tie into the ADC and TREM controllers, as required. There is also a provision to change the effective taper of the Power Scale pot to tailor the sweep of loudness to the individual player or amp designer. One of the "discoveries" in the recent search was that the previous variety of tapers is no longer offered.
The circuit is still economical in complexity by being very simple and not needing too many parts, so circuit density is low. None of the parts are critical yet the end result is precise. The expensive pot has its one special characteristic but is otherwise as imprecise as every other kind of potentiometer, but the circuit designs makes that not matter.
SB1 is an SBX + TBS, and only requires the addition of RBX and ADC-L or ADC-H as needed.
SB2 is an SBX + VCK, and only requires the addition of TREM and ADC-L or ADC-H as needed.
I'm considering including the ADC in the SB1 and SB2 package.
I have an SV2-M36 in a cathode-biased 4×EL84 Rocket-style amp.
In this amp the Powerscaling delivers an exceptional Overdrive or distortion sound.
I would like to implement a footswitchable OD function which, when engaged, reduces the Power Scale setting by a proportional amount while simultaneously increasing a Lar-Mar PPIMV setting. It would also need to simultaneously reduce the reverb level.
I’ve read the switching section in TUT4, where preset PS levels are selected using separate Power Scale controls/units. On the SV2-M36, is it practical to switch an additional resistor/network around the existing 1M PS control to lower the output voltage, or would you recommend a different method because of the DC voltage present on the PS control and any concerns around instant changes of power?
I also have an ERK available. Can ERK/JFET switching safely be used anywhere in the SV2 control network, or should the PS function be switched with a relay/MOSFET arrangement instead?
Current SV2 setup: 1M PS pot, R5 calibrated to about 262k effective, PS3 fixed at 164.2k, full-power regulator drop about 6V.
Japan is often left out of the discussion when designing PTs.
Japan nominally uses 100Vac mains, with some places having 200Vac. The widely used 100Vac is 15-20% lower than the typical North American mains, where PTs to match may be rated at 115Vac, 117Vac or 120Vac.
London Power always specifies PT primary windings to be 120V each, as this provides some protection against high mains situations. On the other hand, it makes it worse for use in Japan.
My later PTs accommodate Japan by providing taps on the primaries. The usual way to do this is to have identical primaries rated thusly:
PRI-1: BLK (120) - RED (100) - WHT (0)
PRI-2: BRN (120) - GRY (100) - ORG (0)
The above windings can be wired in many ways matching to 100V, 120V, 200V, 220V and 240V. Typically the manufacturer of the amp would make the interconnections appropriate to the designated sales region. If the amp was resold, say, and the new owner is in an area with a different mains voltage, the interconnections would have to be changed. To flip between the options using a single switch would require 4-poles and 5-positions.
To keep things simple for myself, I only wanted to match to 100V, 120V and 240V. Using the above windings would require a 4P3T switch. But I got lucky... I saw a schematic for a brilliant solution to this problem that only requires a 2P3T slide switch. I have not seen the original since and cannot remember who it was.
In any case, the solution requires asymmetric windings as so:
PRI-1: BLK (120) - RED (20) - WHT (0)
PRI-2: BRN (120) - GRY (100) - ORG (0)
The wiring is fairly simple. The slide switch has 4-pins per side (per pole). Designate one end to the 240V position. These end pins are linked together.
On one side of the switch the next pin is tied to BRN, then the third to LINE and the fourth to GRY.
On the other side of the switch the next pin ties to WHT, the third to Neutral, and the fourth to RED
As with all dual-primary PTs, LINE (L) ties to BLK through a fuse, and NEUTRAL (N) ties to ORG through a second fuse. The fuses are the same value.
With the above 2P3T slide switch wiring, one end is 240V, middle is 120V and the other end is 100V.
The usual 2-way range switch for 240V and 120V switches the BRN and WHT ends of the windings so the windings are in series or parallel.
I would like to split the signal coming from guitar.
The original Best F/X Loop had a switch to accomodate line level vs guitar signal level.
For the loop that was, not for the input.
Is this not needed in the newer version, the one that is in the kit?
Is the input good for guitar? Or is it better after a preamp?
For what levels is the Sigma designed?
The input is suited for guitar as I read in the description.
Is there an update on the schematic, just like the BFX had?
So I am in the UK and shipping costs from Canada are quite expensive, so when I bought SV1 and RBX for an amp some time ago I also bought SV3 & VCK having a cathode biasdd amp planned.
I finally got around to building that cathode biased 4 x el84 amp and after getting it all working correctly, I fitted the SV2–M36 and VCK.
The amp is based on a trainwreck rocket which was based around an ac30, but I’ve put my own slant on it also adding a few extra features and enhanced cathode adjustment/separation.
I’ve got the SV2-M36 working , but I’m seeing a fairly large voltage loss at full Power Scale and wondered whether this version needs tailoring for the amp’s B+.
Raw B+ at M2 drain is about 336V. At PS fully clockwise, M2 source/output is only about 281–285V, so M2 is dropping around 55V. Gate-source is about +4.7V. PS1/PS2 at full CW are about 33.2/33.3V.
The scaling otherwise works properly, although the 1M pot sweep gets to minimum voltage at less than half of its rotation, so I inserted a 500k trim pot and now the rotation is perfect. SV2 will take the output stage down to about 1V (I haven't rechecked this since inserting the 500k trimmer, having set by ear instead), and VCK is holding the downstream B3 supply very close to its original voltage.
I’ve checked R1 and R9, replaced D1, repaired a poor R23 joint, and the FQU2N90S M1 tests normally. M2 is running around 80°C at full PS, versus about 50°C for M1.
Is ~55V drop at full PS normal for the SV2-M36 at this B+, or should the QP4/QP5 control network be adjusted? If it should be lower, which resistor values would you recommend changing for a raw B+ of roughly 335–340V?
Also I pulled all the power tubes and retested, still getting the large voltage drop so can confirm its not load related.
I found somewhere that there might be some component values that could be tweaked to overcome this but not sure which ones and what values to try, any help would be greatly appreciated.
Many Vox amps have bias-modulated tremolo which adds a wrinkle to Power Scaling such amps. Fender has their Vibrosonic amp, as well. In both cases the modulation signal is applied via the Depth control to the output stage grid-leaks either directly, or through a bias pot. We need a way to make this signal shrink as we dial Power Scale down to low-power.
The easy solution is to use a p-jfet since it requires positive gate voltage to turn it 'off', wired as a shunt element across the Depth control. A series resistance is needed for the jfet to work against. This is the basis of our TREM kit. We used the last type of p-channel jfet that was available as a through-hole device, i.e. a component with leads. Unfortunately it is obsolete as are so many devices suited to audio.
So what would be the basis for TREM of the future?
Ironically, the answer was another primitive device known as an LDR - light-dependent resistor. (The junction-field-effect transistor was one of the earliest semiconductors devised.)
To achieve a functioning identical to the jfet circuit, the output of the LDR would have to be a high resistance that becomes a low resistance at low Power Scale settings. The LDR output is naturally 'high', so the control circuit has to provide more current to the LED input side of the LDR as power reduces. This requires that the LED be powered all the time but "held back" for full power and "let loose" at low power. This simply requires the use of current-steering as we saw in TUT chapter-9. Amps with bias-modulated tremolo have B+ of 350V to 450V, so I made the control range adaptable for 250V to 550V using two trimpots. The PCB for this circuit is 1.2" x 1.3" (29.4 x 31.9 mm) with space for a mounting bracket.
An alternative approach adds variable resistance between the Depth control and the modulation point. Here, the default resistance is 'low' and attenuated resistance is 'high'. Although this seems contrary to how the LDR is inherently, the control circuitry becomes greatly simplified in the Power Scale situation. The board size is reduced to 0.9" square (22.1 mm square). Again, trimpots allow the stock supply voltage to be 250V to 550V. This board is TREM-L (not on the site yet).
Since the introduction of Power Scaling back in 1990, I have occasionally used automatic drive compensation in some amps I built, eliminating the need for the second control. At that time, I referred to the control as 'Limit' and the automated feature as 'Auto-limit'.
Drive compensation is a necessity in a Power Scaled amplifier, but having a separate control for it is a turn-off for some players, and physically difficult in many installations. Some circuits are amenable to the "single-pot solution" using a dual pot to perform both functions (Power Scaling and drive compensation). However, a purely electronic approach is desired by many players and installers. To that end, I designed two kits that perform this function.
Mickey Corrieri of Soultone amps asked me for help with an idea he had to automate the drive control using LDRs (light-dependent resistors). The LDRs are actually more than just a light-dependent resistor, as they have a light source built in and are a unitised opto-coupler. The light source could be neon lamps or LEDs (light emitting diodes). LDRs were used in Fender amps as the actual modulation element to couple the tremolo/vibrato oscillator to the audio path. Fender, Rivera, Mesa and many others used LDRs for channel and feature switches.
As a drive compensator, there are two ways to wire the LDR output: in series with the signal or in parallel with it. The two situations require that the LDR resistance default to be 'low' or to be 'high', respectively. This requires two different types of control circuit, hence two different resulting kits: ADC-L and ADC-H. The natural default for an unpowered LDR unit is to have a high-resistance output.
The Soultone application was specific inasmuch as the supply voltage is known and there was an auxiliary supply that could be used to power the LEDs of the LDRs. Standard LEDs can be current hogs, needing about 10mA to be lit to a level suitable as a front-panel or foot-switch box light. They can be driven with up to 25mA, some up to 50mA, for extra brightness. However, for the integrated LDR unit we only need a few milliamps for a useful range of output resistance. Mickey wanted to use the LDRs in series with the signal, for which the output resistance default had to be 'low'. This means for the full-power condition of the amp, the LEDs in the LDR have to be 'on'. Using the auxiliary supply, there was no problem supplying LED current. We used current-steering (explained in TUT) to control the LED and thus the resistance imposed in the signal path. This was optimised for the LDRs he chose and for the voltage environment in the amp.
The current-steering was actually "inverted" for his application and in retrospect could have been a current-mirror instead. This accommodates high-current LEDs in the most common LDRs.
For a general purpose form of the custom application, I chose a different LDR with a more sensitive input, and whose data sheet shows the relationship between LED current and LDR resistance. This keeps waste heat to a minimum considering we want to track high-voltage and only use B+ for any other power required. The circuit is able to work properly in voltage environments ranging from 250V to 550V. For the signal control approach used above, the ADC-L AutoDriveComp-Low-resistance default would be the up-to-date circuit choice. It is highly simplified because the application introduces an inversion of the drive. This board is 1" square (25.4mm square).
To shunt a standard MV requires that the default resistance be high. However, the tracked voltage moves in an opposite direction to the LDR resistance change we want, requiring that we go back to the current-steering approach and a larger board; 1.3" square (31.9mm square). The ADC-H board defaults to a high LDR resistance and has to work against a fixed resistor in the audio circuit to form a voltage divider. Again, a supply voltage between 250V and 550V is accommodated. Both kits have trimpots so the control range can be calibrated to the supply voltage present. Both kits also have two LDRs with floating output resistances that can be used however they need to be.
The kits are not on the site yet, but they are ordered and on their way.
In an audio system there has to be a point in the signal path where the signal can be made smaller - all the way to zero - allowing the listener to regulate how loud the sound is. A voltage divider made from fixed resistors with the tap points selected by a switch will accomplish this goal, but not very gracefully. We can add more resistors in the divider string, creating more tap points and add more positions to the switch, for smaller loudness increments. Integrated forms of this feature highly in many hifi offerings.
We can expand this idea to as many points as we wish and there are many hifi systems that use relays to select the corresponding taps in a stereo system. The simplest use the panel switch to control the relays directly, while others replace the switch with a vector control and digital interface. Integrated circuits are available as 'electronic volume controls' where the resistive elements are selected by CMOS switches during the signal zero-cross to reduce switching noise.
A variable resistance will do this much more easily and with the potential for infinite resolution, i.e. equivalent to an infinite number of resistive taps and switch positions, and thus infinite loudness levels.
The potentiometer has a resistive track with a total specified resistance from end to end. There will be a tolerance for the value, which is usually 20%. This means a 10,000-ohm pot can be 8,000-ohms up to 12,000-ohms and still be within specification. Since the pot is the voltage divider in its entirety, this sloppiness of production does not matter, at least until you want to change more than one thing simultaneously.
A mechanical moveable contact can be wiped from one end of the track to the other. This contact is creatively called the 'wiper'.
The rate of change of resistance versus the rotation of the wiper is called the 'taper'. If the resistance varies in direct proportion with the percentage of wiper sweep, the taper is 'linear', denoted by a raised-letter 'B' in the value. A linear 10k pot is denoted as either B10k, or as 10kB.
For the loudness change to feel linear to our ear, the power has to change logarithmically. For it to seem half as loud, power must go down by ten times. A simple reduction by half would only feel "a little bit quieter'. Therefore, we need pots with a log taper, meaning that the signal must change very quickly in amplitude over the sweep of the pot. Where a linear pot set half way will have half the input voltage at its wiper, a log taper pot will have one-tenth. Log taper is denoted by a raised-letter 'A' as A10k or 10kA, for a 10k pot.
It is very difficult to lay down a logarithmic resistive track, so instead, small linear sections are laid down and the hope is that the transition points are not too audible.
In a mono audio system, the issues of pot value tolerance and taper irregularities can be minimized, and are generally non-problematic.The taper of the pot is selected for each pot position in the circuit, based on the circuit itself, and how the sonic result should be over the sweep of the pot. The basic division by circuit is 'passive' or 'active'.
Passive controls work freely comprised of resistors, capacitors and maybe chokes, but are otherwise independent of gain elements. All of these pots will be log taper, for Volume, level, and tone controls.
When a pot is inside the feedback loop of a gain stage, we may use linear or reverse audio, taper 'C' to achieve a linear loudness or effect change.
In a stereo system, or one with yet more audio channels, ganged pots with multiple sections are used. Each section is specified as for a single-section pot, but now we have a new parameter called 'tracking'. If you feed the same signal into all of the pot sections, will all the wipers have the same output voltage? For conventional pots, the answer will be 'no. Even for expensive pots of this type tracking will be within 3dB, which is quite noticeable.
ALPS from Japan has their famous 27mm stepped pots which are 42-position switches with resistors between the positions. Using 1% resistors allows these pots to track very well, within 1dB which is the nominal limit for Human perception of a loudness change. despite the opportunity to calculate the steps very accurately, many of these pots have a wide overall tolerance but do divide the signal properly. For modern equipment made ever smaller, this pot option is discarded.
The physical size of the pot has a relationship to its expected life time, measured as 'cycles', end-to-end wiper sweeps. Tiny pots like 6mm and 9mm diameter, may be rated for 1,000 to 10,000 cycles, or have no rating. 16mm are usually 15,000, as are average 24mm pots. 28mm mil-spec pots might be rated from 15,000 to 25,000. This is for carbon tracks.
Pots can be made with conductive plastic tracks and these generally have 100,000-cycle life ratings, at least down to 16mm.
All of the above are panel pots, with shafts that accept fixing a knob to, that can be easily manipulated by the equipment user and hopefully which matches the aesthetic of the equipment. A single rotation from CCW to CW is generally 300-degrees of the circle.
Wire-wound pots are made using a wire coil of resistance wire with a wiper that sweeps across one side of the coil. These or most often designed to handle high power, although there are precision units made by Bourns and Spectrol used for fine-setting of machine and instrument parameters. These are typically 10-turn pots, where the pot shaft is rotated ten times in the same direction for the wiper to move fully from one end of the track to the other. Life cycle is usually 1,000,000, but keep in mind it takes 10 turns to go end-to-end. These pots are not something that you make broad adjustments with, so they are unlikely to be used by hobbyists except for a bench power supply.
The 10-turn pot is also a panel-mount and requires a Vernier dial to be useful and for the operator to know where along the track the wiper is situated. A vernier dial shows which revolution the pot is at and the infinite division of that rotation. A small plate is secured under the nut that secures the pot to the panel; the plate has a tab that lines up with a slot in the knob and a set screw secures the knob to the shaft. These controls are surprisingly small, typically about 17mm diameter, not including the solder tab connections.
In an audio amp or musical instrument amp, even one with tubes, the signal voltages across the pots are generally small, less than 10V or so. This is good, because the pots used in these devices are often limited to 50-250Vdc. Linear pots have higher voltage ratings than log pots, where, for example, 16mm linear is rated at 250V and 16mm log is only 50V. This rating is based on the physical construction of the pot and has nothing to do with the resistive value. Moving up to 24mm pots gives more leeway at 500Vdc for linear and 100V for log.
Voltage ratings on pots is something that nearly all amp designers are unaware of. Even for those who know about it, they often make poor choices or assumptions for specific applications, such as when adding Power Scaling in MI amplifiers. In the basic form the pot sustains high voltage and must be of sturdy design, explaining why London Power used a mil-spec pot rated at 1,100V. This rating applied even to log taper pots, which turned out to be better for the application than linear pots. Copyists skimped and used standard 24mm pots with dire results. The present kits use less expensive 16mm pots that fit into smaller spaces, but they see much reduced voltage and will last their normal life.
Power ratings for pots are often routinely ignored. Again, linear pots have higher ratings than same-size same-value log pots due to the segmentation of the log pot resistive track. In most applications the power rating is not a limiting factor provided the resistive value is high. In guitar amps the lowest values might be 10k for a bias pot, or 25k for a midrange frequency control, with all others 100k or higher; majority values are 250k to 1M.
A 16mm linear pot is typically rated for 125mW and log is 60mW. If this is a 250k pot, the voltage across the track is limited to 176V and 122V, respectively, so not a big concern in common tube amp circuits. Hifi amps may use down to 1k pots, with 10k being the most common these days (and nights). Voltage limited by power is now 35V and just under 8V, respectively for linear and log at 1k.
In tube amplifiers that are fixed-biased, the bias pots should be not too high in value. London Power uses 25kB as standard.
Capacitors are everywhere in audio circuits and their power supplies.
A capacitor is formed when two conductors are separated by an insulator. That insulation is called the dielectric, and can be air (not very useful, but the cause of 'parasitic' capacitance), paper, oil, plastic, ceramic, mica, electrolytic materials, polymers, or organics. The conductors are called 'plates' and are effectively large flat metal surfaces with leads attached. The entire assembly is covered in a suitable coating to keep the environment out.
Capacitance value is in Farads, which is a huge value for an audio circuit, or even for a power supply, although small "super caps" of a few Farads were used in lieu of a battery as backup power in VCRs and similar to support memory circuits. In audio, we see values from picoFarads (pF) up to milliFarads (mF). Note that microfarads should be denoted as uF. Ancient schematics will show uuF instead of pF.
Man-made capacitors are imperfect with the following flaws in various extents:
DA - dielectric absorption is the inability of the dielectric to release built up charge
DF - dissipation factor is the inability of the dielectric to follow fast changing charge/discharge cycling
ESR - equivalent-series-resistance is a fixed internal resistance not the impedance based on frequency
Remember: the power supply is the other half of the signal path so the caps used here make a difference that is audible.
Ceramic caps
These are made from various rare earths and historically were single layer, where the dielectric is between two parallel plates, as the iconic capacitor model represents. Discs with radial leads are typical, and the value accuracy and stability depend on the specific materials used. The best ceramics for audio are C0G or NP0 rated, and are good up to values of 100pF.
Distortion is otherwise quite high with higher values and adds "grit" to the sound. In a guitar amp, this may be welcome, but in hifi it adds distortion that smears the sound.
MLCC Caps
Multi-layer ceramic capacitors are used in switching power supplies and have some quite terrible characteristics. Their capacitance varies with the voltage across the cap. The capacitance varies with frequency of the current through the cap.Their voltage rating is typically unusable, with half the rating being a common safe maximum. Even though they are designed to nominally handle high switching currents, this rating is often overstated. Unsuitable for audio.
Mica Caps
Mica is a natural material that can be shaved extremely thin. It has high voltage withstand and can be cut precisely for reasonable capacitance accuracy. Distortion is much lower than with ceramic and higher values can be used, although anything above a few nF may be expensive and impractical.
Plastic Caps
There are a few common plastics used as dielectrics in capacitors.
Polyester has a few trade names, such as Mylar, and is the most common plastic cap type. It has an unfortunate attribute of adding odd-order distortion to series audio signals. The higher the signal voltage is across the cap, the higher the THD. Compared to ceramic, a polyester cap will sound cleaner in a guitar amp, but it is best to use other polys for critical applications. placing caps in series can reduce the individual capacitor distortion but the composite effect may be not enough to justify the use of multiple capacitors.
Polyethylene is about equal to polyester.
Polycarbonate was favored for a while, but has its own distortion mechanism. Lower THD than polyester. Not as low as others.
Polypropylene is distortion-free! The best poly cap and its reach and breadth of available sizes, voltages and values expands every day. Although safety caps of the X and Y types use PP for its self-healing characteristic, the caps use impure PP and are unsuitable for audio.
Polystyrene is also distortion free but only available in very low values and at voltages <100V or so.
Paper Caps
paper dielectric is ancient obsolete technology that is entirely unsuitable for audio use. Yes, vintage equipment contains these caps, which decades ago were already dried out and behaving more like inductors than capacitors. Replace them with any type of poly cap and the amp will be loud and clear again.
Oil-filled Caps
There are usually meant for motor-start applications and are typically in a metal can with faston-style connections. Not suitable for audio.
Electrolytic Caps
Electrolytic capacitors have a dielectric that is formed using voltage and they require regular application of voltage to remain healthy. If you install a high-voltage cap in a low-voltage circuit, the cap will eventually reform to the lower voltage and not be able to meet its rating unless you slowly reform it by carefully applying higher voltages to it. In general, electrolytic caps have just about the worst combination of DF, DA and ESR possible, but this depends very much on manufacturing quality.
The rated voltage of an electrolytic capacitor is also its WORKING voltage, which means that you can use the cap at the rated voltage without concern.
Life expectancy is listed in 1,000s of hours at the rated maximum temperature and voltage. Realistic life doubles with each 10-degree drop in temperature, so that a 3,000-hr cap rated for 85C should be within spec for 64,000-hrs at 25C. If this were a 105C rated cap, its life would stretch out four times longer. For high-fidelity audio, capacitor life is considered over at 14 years, or when DF begins to go crazy (scientific, ya?)
High-values of capacitance are achievable and electrolytics are the mainstay in power supplies of all (conventional) types. Because of their poor audio qualities and wide value tolerance, electrolytic caps should never be used to define frequency roll-offs, such as in the feedback loop of a solid-state power amp.
Polymer Caps
This is a newer dielectric that is a sort of "enhanced electrolytic". Current surges are handled better and reverse voltages withstood better. They work well in stacked power supply situations, for example, full-wave voltage multipliers.
Tantalum Caps
tantalum capacitors have good value stability and can be made to tighter tolerances than electrolytics.king them an upgrade in old-style security systems that relied on fixed-frequency oscillators. On every other count, they are unsuitable for use in audio circuits.
You cannot build a linear circuit without resistors, and there are some less known parameters that need to be considered when selecting these essential devices. For our purposes building audio circuits, we tend to use axial-lead carbon-film, metal-film, metal-oxide, wire-wound, types in different circuit locations or to achieve different sonic results.
Resistors in small-signal areas will be carbon-film or metal-film and may be 1/8W up to 600mW. Power resistors from 1W to 3W tend to be metal-oxide, with 5W+ being wire-wound. London Power does not use carbon resistors.
Resistance & Tolerance
The resistance value is the most important parameter and you will always see this on schematics. Resistance is in 'ohms', denoted by the Greek letter Omega. There are historic ranges of values that seem odd today, with newer ranges popularised as manufacturing methods improved allowing greater accuracy in achieving the expected value.
The value charts tell you how many values there are per decade. The decade is denoted as 'E' and the number of values plainly as a number. For example, E12 has 12 values per decade - not very useful. E24 has 24 values per decade allowing the huge gaps of E12 to be filled in a bit better. E96 offers the best overlap and complete coverage of common and less-common values. Within each chart, the basic values are given and these can be multiplied by any exponent (hence E) of 10 over a range that covers fractions of an ohm to millions of ohms.
With each of these value ranges, the resistance tolerances improves, becomes tighter and is represented by an ever-smaller percentage.This percentage represents how far off the value can be. For example, a 1k resistor with a 10% tolerance can be as high as 1,100-ohms or as low as 900-ohms. If its tolerance is 1%, then its value can be between 990 and 1,010 ohms.
E12 is 20%
E24 is 10%
E48 is 5%
E96 is 1%
Material
Resistors can be made from various materials that give them unique characteristics.
Cracked-carbon resistors have the worst characteristics, including: poor tolerance, high noise, high distortion, high voltage sensitivity causing distortion with varying voltage.
Carbon composition has finer particles and all of the problems of cracked carbon to a lesser extent.
Carbon-film takes the carbon issues down another level.
Metal-film has low-noise, very good temperature stability and low-distortion for varying voltage.
Metal-foil is slightly better than metal-film all around but with higher price and greater bulk.
Metal-oxide has higher power capability than metal film but is slightly noisier.
Thin-film and thick-film resistors have characteristics between carbon-film and metal-film.
Wire-wound uses a metallic element usually covered in ceramic.
Power Rating
Self-explanatory, the power rating tells how much heat the resistor body can dissipate in free air. Like most devices, most of the heat is dissipated by the component leads. Smaller body sizes will have a higher surface temperature for a given power than will a larger body size.
Some ranges of resistors with power rating of 1W+ may have shock thermal ratings allowing their continuous power rating to be exceeded briefly.
Voltage Rating
Yes, resistors have a voltage rating. Resistors have two leads and are constructed either axially or radially. The space between the leads has a voltage limit based on the exterior coating of the resistor and the shape of the body. Ridges on the body can increase the effective distance (creepage distance) between the leads, providing better voltage withstand.
Package
Through-hole resistors have leads arranged either axially or radially from the resistor body. The leads provide support for the device and the through-hole mounting on a PCB provides large solder connections. power devices should be elevated from the board by 6mm.
Through-hole resistors are the historic format used in hand-wired assemblies, on circuit cards, on turret boards, and on most printed circuit boards even today.
There is a type of power resistor that has its own heat sink with solder-lug leads and attachment holes. The smallest type is rated for 20W provided it is bolted to a proper sized heat sink.
Surface-mounted devices allow smaller overall circuit assemblies and are very difficult to handle manually, and thus are the mainstay of highly-automated assembly, particularly for computers, monitors, cell phones, portable devices, and generally every consumer electronic item considered to be disposable. It is an irony of our times that such sophistication is so devalued.
Come in where it's warm!
A warm welcome to tube amp modding fans and those interested in hi-fi audio! Readers of Kevin O'Connor'sThe Ultimate Tone (TUT) book series form a part of our population. Kevin O'Connor is the creator of the popular Power Scaling methodology for amplifiers.
Please remember these three principles: respect, sharing, community.
![[-]](https://theultimatetone.com/images/collapse.png "[-]")
