C1 (film)

C2, C8 (ceramic or film)

These are bypass capacitors for the op-amps. While they are not necessary with all op-amps, they can't hurt, they're cheap, and they're easy to add. The standard type to use here is an 0.1 µF ceramic; the exact ceramic dielectric type doesn't matter here. If you're using an extremely fast op-amp (say, over 100 MHz) you might try a lower value instead, as smaller caps are more effective at high frequencies.

If you want to be an audio snob, use a film capacitor here instead. Panasonic offers an 0.047 µF PPS film cap in the proper size, for instance.

Optional? Yes, do not jumper.

Largest Part Size: 1206 package

C3 (C0G/NP0 ceramic)

C4 (electrolytic)

This is the main power reservoir capacitor, also called the "rail capacitor" because it spans the power rails.

Use a 100 µF to 1000 µF capacitor with a voltage rating higher than that of your power supply. For example, use a 25V capacitor if your power supply is 24V.

I recommend the Panasonic FM, Panasonic FC, and Nichicon PW lines. The Panasonic FC and the Nichicon PW are identical. The Panasonic FMs are a little nicer than the FCs, but there are fewer values and fewer case size choices in that line. If your chosen distributor doesn't carry one of these lines, try to find a cap line that features long life and low ESR.

If you want to choose your own power capacitors, there are two main rules to keep in mind:

1. Bigger is better, within a particular line of capacitors.

2. It's usually better to use a lower-capacitance part from a better line of capacitors than a higher-capacitance part from a poorer-performing line of caps.

Sometimes you must compromise on quality (rule 2) to get a sufficient amount of capacitance (rule 1). For a PINT, I recommend that you use at least a 220 µF cap, and I'd much rather see something more like 470 µF. If you're looking at caps over 1000 µF you're probably compromising too much on quality; try looking for a line of capacitors that will let you trade some of that excess capacitance for higher quality. If you're already looking at the best capacitor line available to you, you may simply choose not to buy that 2200 µF capacitor, but instead get the 1000 µF one and save some money. I doubt you can hear the difference.

Now to more specific advice.

First, decide on the capacitor's dimensions. For the cap to set down on the board properly, the diameter needs to be 12.5mm or 10mm. The height will be limited by the amount of space inside your case. You don't have to use the tallest cap that will fit, you just need to keep this in mind as a limit.

C5 (ceramic or film)

C6 (tantalum)

C7 (tantalum or ceramic)

These caps improve the stability of the virtual ground voltage divider. The default value is 1 µF, for which you'll probably need to use tantalum.

I haven't tested it, but I suspect you couldn't tell the difference with values down to as low as 0.1 µF. If so, you could just order extras of the same caps you use for bypassing the op-amps. There's no question that it will work, but using too small a value may turn the voltage divider into a crosstalk path by passing too much rail-modulated ripple into the virtual ground circuit. That's why I don't call these caps optional, even though the amp will work without them.

Optional? No.

Largest Part Size: EIA 3528-21 package

R1

This resistor sets the current used by the NiMH battery charging circuit.

The formula for computing the proper resistor value is R=1.25÷I, where I is the desired charging current. I recommend that you use the standard trickle charge rate: 0.1 times your battery's mAh rating. (The notation for that is "0.1C".) Since the default resistor value gives a charging current of about 17 mA, it's suitable for NiMHs of 170 mAh or higher.

Charging current is a trade-off between charge time and battery cycle life. You can choose to charge faster than 0.1C, but your battery will become useless after fewer charges than if you didn't abuse it so. If you charge much greater than 0.1C, you must use some sort of intelligent charger, which the PINT does not offer. Contrariwise, you can charge slower than 0.1C to get more cycle life from your battery, but I wouldn't go lower than 0.05C. Below that, the battery's self-discharge rate becomes too significant: the battery won't charge as fast as it "should", and it may actually fail to charge.

Optional? Yes. Leave it out.

Largest Part Size: 1206 package

R2

This is the input impedance resistor. It must be at least 10× the value of your pot for the amp to function correctly. Making it higher than necessary won't improve performance much, and can raise the amp's noise floor too much.

Optional? No.

Largest Part Size: 1206 package

R3, R4

These are the feedback resistors. They set the amplifier's gain.

The simplest path is to just use the values given on the schematic. This gives a gain of about 6, which should be enough for most any headphone and source combination. You might want to lower the gain, though, in order to get more useful travel from the volume control knob, and to lower the amp's noise floor. You must understand the issues discussed at the top of this page before you do that, however. The PINT is not like most of the other amps on this site, where you can change the gain by changing only one of the gain resistors.

Optional? No.

Largest Part Size: 1206 package

R+, R-

These resistors divide the power supply voltage in half, setting the virtual ground point.

Smaller values will lower the amp's noise floor. However, because this divider is directly across the power rails, lowering these values wastes more current. If you want to use a different value from the default, you will have to trade off between noise and power supply drain.

Another concern is due to the current flowing in or out of U3A's +IN input. This current goes across one of the two divider resistors, pushing the divider's midpoint away from the centered ideal. The largest bias current you'll see from any of the recommended op-amps is 7 µA. (That's with the LM6172.) Following the standard rule of thumb for this sort of thing, we must ensure that the current through the divider is at least 10× the bias current. The lower the supply voltage, the lower the current through the divider, so let's set an arbitrary limit of 5 V. (That's near the clipping point for all of the recommended chips with typical headphone loads.) At that voltage, we can push over 70 µA through the divider if the resistors are 33 KΩ. (The formula is R=V÷140 µA, where R is the resistor value, V is the supply voltage, and 140 µA is twice the minumum divider current, since we're dealing with the midpiont of two equal resistors in series.) The default value of 47 KΩ assumes that you are either not using the "most biased" of the recommended op-amps, or that if you are, you are not so exceptionally unlucky that it has the worst-possible input bias current.

Optional? No.

Largest Part Size: 1206 package

L1

This position is for adding a part to increase the output impedance, which can prevent oscillation with some op-amps. You can put a resistor, a chip ferrite, or a chip inductor here.

If you need a part here to prevent oscillation and don't have access to an oscilloscope, the safest path is to use a resistor in this position. A value as low as 10 Ω will quell most oscillation problems, but you may need to go up to as high as 100 Ω in some situations. The lower the value, the better.

If you do have access to a scope, you can get better performance from the amplifier by using a ferrite here instead. The reason is, ferrites have low impedance at low frequencies, and rise in impedance as frequency rises. This is exactly the behavior we want, because we don't want any extra impedance at audio frequencies: we only want it at high frequencies, where oscillation can occur. Just as with the resistor, an impedance of between 10 Ω to 100 Ω at the oscillation freqeuency will work best. But again, I must stress that if you don't have an oscilloscope so you can determine the oscillation frequency, you don't have enough information to properly pick the right ferrite. Even then, it can take some trial and error to find the right part.

You can use a chip inductor instead of a ferrite, but they have higher DC resistances. I only mention this because it may be that you cannot find a suitable ferrite, but you can find an inductor. This is why L1 allows for 1210 size parts instead of 1206 like most of the other passives: 1210 seems to be a more popular size for chip inductors.

RLED

This is the LED current limiting resistor. Use Ohm's law to figure the appropriate resistor value given the LED's voltage drop (Vf), the power supply voltage (Vs) and the desired current (I). For example, let's say the power supply voltage is 18V, and we want 2 mA through our LED, which has a 1.8V forward voltage drop:

 	R  = (Vs - Vf) / I
	R  = (18 - 1.8) / 0.002
	R  = 16.2 / 0.002
	R ~= 8.2 KΩ

Most LEDs require 1mA to get minimum acceptable brightness. More current gets you more brightness, but of course uses more power, which mainly matters with battery power supplies. You don't want the LED to be too bright, because that's annoying.

Typical values are 1 KΩ to 10 KΩ. I personally use 2.2 KΩ and 4.7 KΩ most often.

Optional? Yes. Jumper only if your LED has an integral resistor. Most don't.

Largest Part Size: 1206 package

Semiconductors

D1, D2

These diodes perform several interlocking functions.

In the simplest case, where you have just a single power supply, you can use either D1 or D2 to provide reverse power supply voltage protection. (Use D1 if your power supply's positive leg is connected to the W+ pad, and D2 if it's connected to the B+ pad.) If you connect a circuit like this up to a power supply and the voltages are backwards, you'll blow up the op-amps almost instantly. That's a small price to pay for 0.7 V or less of voltage loss.

U1, U3

These two chips are dual-channel SO-8 op-amps. One chip handles the left and right audio channels, and the other handles the input and output halves of the virtual ground circuit. Some manufacturers make chips in packages that aren't exactly SO-8, but are close enough to work on the PINT, such as New Japan Radio's DMP-8 package.

The op-amps have the single biggest effect on sound and power draw of any component, so it behooves you to pick these parts carefully.

The right and left channel op-amps do the actual amplification in the PINT circuit. This should be the highest quality chip you can find, while not going over your desired current budget. Unfortunately, it's generally true that the higher the op-amp's quiescent current draw, the better it will sound. The rule isn't infallible, but it's a good indicator.

There's a good argument that you should use the same chip for handling ground as handles the output channels: the ground channel effectively sits across the headphone from the left and right channels, so using a matched circuit can result in more symmetric handling of the music, resulting in less distortion.

In practice, however, there are a couple of good reasons to choose a different op-amp for each position. First, mixing op-amp types can give sonically interesting results. Second, mixing op-amps can allow you to use a chip in the left and right channels that you would not want to use in the ground channel for some reason, or vice versa. One such situation is that the chip you're looking at for the left and right channel draws too much current to use two of them. As long as you use a ground channel chip capable of sinking enough current, it's okay to use a different chip type here. Another situation is when you have a chip with input bias currents so high that they cause too much DC offset in the left and right channels. The ground channel is not nearly as susceptible to problems from input bias currents, so such a chip could work just fine there.

Most any op-amp can be made to work in the PINT board, but some are more suitable than others. Any FET-input op-amp should work on the board without any trouble. If you use the resistor values on the schematic, any bipolar input op-amp with a maximum input bias current in the low single digit microamp range or lower should work fine. If it can have more than a few microamps of bias current, you may have to choose different part values to get DC offset down to acceptably low values.

The canonical part for this board is the AD8397, which is rebadged as the AD45048. It has high output current, sounds great, runs to low supply voltages, and will work from relatively high supply voltages. The downside of the 8397 is that it drinks current like an Arizonan drinks water after a bowl of hot chili. It's not uncommon for a PINT with a pair of 8397s on it to draw over 50 mA! The longest lasting "9V" style NiMH batteries you can buy today will only give you about 5 hours of run time with this configuration.

If you need a longer run time than that, I recommend first replacing just the ground channel op-amp. (U3) Try something like the LM6172, or if the 12 V power supply limit isn't a problem, the LMH6643. (The 6172 will also work in U1, but the 6643 will not, due to its high input bias currents. High bias currents don't matter as much in the ground channel.) Not only will the amp run longer from your battery with one of these chips in the ground channel, it's more likely to be stable, as these chips are happier than the 8397 when running in unity gain.

Other suitable chips are recommended in the parts list.

U2

Choosing a Power Supply

Choosing an Enclosure

The standard PINT enclosure is a regular mint tin, such as an Altoids or Penguin Mints tin. The board is designed to be just large enough that you can just barely swing the board into position with the pot attached. (That's why it doesn't actually go competely from front to back in the tin.)

I expect that another popular enclosure will be the Hammond 1455C801. The PINT board is just barely narrow enough to fit top-to-bottom in the case and still leave enough room for a single 9V battery, a DC power jack and the audio jacks. I haven't attempted this yet, though.

Because the PINT is so small, there are myriad other enclosures you can use, from standard electronics project boxes and enclosures, to found objects more creative than a mint tin.

If you want to use a metal enclosure, read the next few sections, as this affects which panel components will work.

Choosing the Audio Jacks

If you're using a metal case, at least one of the audio jacks needs to be isolated from the case. Isolation, in this instance, means that none of the audio signals are connected to any metal parts of the jack that touch the mounting surface. Usually, this means the jack body and mounting threads are plastic, but not always. The most common way for a jack to be non-isolated is to have a metal shell that's connected to the ground signal.

It's okay if either input or output ground connects to the enclosure, but if both do, the ground channel will be shorted out. At best, this makes the ground channel useless, and at worst it can cause the amp to become unstable. If you're going to connect one of the grounds to the case, it should be the input ground.

The easiest type of jack to deal with is fully isolated, which some of those mentioned in the parts list are. Again, this usually means that the jack is mostly plastic. This can indicate a lower quality jack, so there's some incentive to put up with the problems you get by choosing a different type.

Some jacks have plastic mounting threads but have a metal tab connected to ground that touches the inside mounting surface. These can be made isolated simply by adding a plastic washer around the jack's snout before mounting it.

If the jack's threads are metal, you can still isolate the jack from the case by using shoulder washers, but finding or making them in the proper size isn't easy.

If your amp will use a DC power jack, read on. The isolation issue gets more complicated in that instance.

Choosing a Power Input Jack

If you are using a metal case and you will be adding a DC power jack to your amp, you have a similar isolation issue as brought up for the audio jacks. A metal bodied DC jack usually has no isolation between the barrel of the plug and the jack body. Therefore, using such a jack with a tip-positive power supply would tie the case to V- when the supply is plugged in. That's hardly the worst thing in the world from an EMI shielding perspective, but if both of the audio jacks aren't isolated, this will tie V- to virtual ground, which will collapse the virtual ground.

You have three choices: 1) use an isolated power jack; 2) isolate your audio jacks; or 3) don't use a metal case.

In my recent builds in mint tins, I've been using a non-isolated DC power jack and isolating the audio jacks. The reason for this is complicated. First, the only isolated DC jack I'm aware of is too big to fit on the side of a mint tin without hackery. Second, while I could use a more compact metal bodied jack and isolate it, I don't have the proper shoulder washers for this. Third, I'm not dogmatic about tying V- to the case instead of input ground; some people insist that this is evil.

Speaking of fitting jacks on the side of mint tins: only the one marked "compact" in the parts list is small enough for this. The others could be made to fit, either by grinding down the mounting flanges so the lid can close, or by cutting a notch in the mint tin lid's edge.