Point to Point Versus PCB Amps - What's the difference?

Ever hear this one?

"Get a point-to-point wired amp - those PCB amps are no good. They just don't have the tone they did in the old days." 

The deeper you dig into this issue, the more repair techs you'll find that say that PCB amps are no good. They'll give explanations involving things like tiny traces run too close together, and the capacitance leaking signal from place to place and then make the leap that this somehow spoils the tone of the amp. Is this bit of folk wisdom right? 

Yes and no. The short version of the answer is this:
There is no technical reason that PCB amps *can't* be every bit as good as point-to-point wired amps - or stripboard wired amps, which is what the grizzled techs usually refer to. However, there are some serious reliability (not tone!) shortcomings in most of the PCB based amps that have been put on the market; this is the origin of this folk myth. 

At the bottom of both sides of this question is self-interest. On the side of the amp makers, most of the time the PCB conversion is done as part of a cost saving measure; the same motive to change to PCB's also drives changing to cheaper practices and parts all around. This most often gives worse tone. They're doing less than the best they could, for money.

On the amp tech side,  PCB amps that have been done poorly have frequent breakdowns. They are not as easy to repair as hand-soldered terminal wiring. They require some delicacy in fixing, and they can be a true pain to get the boards in and out. So to the techs, they break more often, and are fiddlier to repair. The techs usually can't charge more for PCB amps, so - no surprise - they don't like them much.

So - PCB based amps, from what we've seen so far, have poorer reliability and techs don't like to repair them; but that is because they were poorly designed from a reliability standpoint in the first place. The tone is not necessarily poorer. Here are some comparisons

PROS and CONS

Most PCB amps that have been produced *have* been poorer than tube amps, for reasons having nothing to do with the PCB's. Like plastic bobbins in transformers, this does nothing to change the tone. The *other* poor practices that go with a cost-cutting attitude that were introduced at the same time may, but PCB's are unfairly indicted.

If you're interested, here's some more information about the differences in point to point, stripboard, and PCB construction practices that DO make a difference.

Point to Point (PTP) - What is it and what's good about it?

This is one version of real "point to point" construction.

The view is from underneath the chassis surface, looking at the bottom of a 9 pin miniature tube socket and two terminal strips that the circuit is built on. Good point to point layout makes the actual component bodies bridge most of the distance from the socket lugs to any other connections. Really, really good PTP construction will have the components arranged in almost a star pattern, all of them leading radially away from the socket. 

What's important there is that (a) all the component leads are pretty close to being as short as they can be (b) the component to component wiring lengths are as short as possible (c) neither the components nor the wires criss-cross one another (d) the lengths of paralleled wires are about as small as they can be.

The reason that all these are important is that there is an unavoidable bit of capacitance between any two electrical conductors that are at different voltages. You can think of this as a little "ghost" capacitor between every combination of two points on the schematic - ugly! These ghost capacitors are often referred to as parasitic capacitors.

This capacitance is proportional to the facing areas of the two conductors, and inversely proportional to the distance between them. For maximum coupling between two conductors. they need to be as close together as possible, and have the maximum area exposed to one another. 

For two wires, the closest they can get is with their insulation touching. For the maximum area, they'd have long parallel runs; for minimum area, they'd cross at right angles. In true PTP, the components and wires are as far from one another as it's practical to get, and where crossovers are needed, they can be at almost right angles. Also, the wiring between amplifier sections can be minimized by laying out the tube sockets in a manner that the signal flows directly from one tube circuit to the next with no long runs of parallel wires. Also, the terminals themselves space the wires a goodly distance away from the presumably grounded metallic chassis.

The worst case for cross coupling in wiring of any kind is where the wires run parallel and with their insulation touching. I spend part of a Saturday morning digging through my collection of electronics and physics texts to come up with some rules of thumb for how big the capacitive coupling is. Of course, this varies all over the map depending on the layout and wire sizes. However, for two 22 gauge hookup wires with kind of nominal insulation thicknesses, the capacitance between the two wires is about 0.381pF per inch. Ten inches of parallel run, the capacitance is 3.81pF. With the wires crossing, there's only about 0.025" of wire "parallel", so the capacitance is on the order of 0.008pF.

Oddly enough, the capacitance of two PCB traces under similar conditions is actually less. For two 0.025" wide traces spaced a conservative 0.075" apart, the capacitance is 0.317pF per inch. Finer traces even closer together do have a higher capacitance per unit length, but much less than you'd expect - it goes up much less than linearly with smaller, finer traces. We have the curious result that wires and PCB traces have very similar parasitic capacitances, and PCB traces have a slight edge over wires. That kind of means that it's how you use them, not which you use, huh?

Note that a poor PTP layout can be a nightmare of criss-crossing parts and wires that can loose a herd of electronic demons. What makes a PTP wired circuit good is (a) careful planning and (b) very skilled labor doing the wiring neatly and carefully. For reason (b), true PTP is very expensive, and has been even as far back as the late 50's.

True PTP construction is good for frequencies up to UHF (several hundred MHz!!) if properly done. This is the reason all that old tube radio equipment was laid out that way. Good PTP layout has wires that almost never come close to one another, and when they do, the cross instead of running parallel. That's good, but it's a long way from audio to UHF, so maybe there's something else that can fill the bill. 

Stripboard Construction

The kind of construction in amps that is most often referred to as "point to point" is not really point to point at all. Most amps use some kind of insulating board with either eyelets (notably Fender) or turrets set into the board, and the components strung in neat little rows on the board. 

All the connections to tube sockets are made with wires from the stripboard to the tube sockets and controls. Note that unlike true PTP, the wiring lengths from the components to the tube sockets are no longer as short as they were. Also, the wires are often stranded wire, and free to move around, where the component leads in PTP were solid, and pretty much held in place by being short and stiff. To get easier wiring, we have sacrificed the very short interconnections that are a characteristic of good PTP wiring.

The freer, stranded wires can now move around relative to one another. This means that they have variable capacitances among them! The coupling of signal from place to place by accident (if it was on purpose, we'd use a wire or capacitor) is now a variable that depends on how you push the wires around after they're installed. The placement of wires in this kind of construction is referred to as lead dress. Not too surprisingly, poor lead dress can cause the wrong signal to get picked up and amplified. This can cause hum if the wires get too close or too parallel to those green/black filament heater wires, or oscillation if an output signal wire gets too near a sensitive input.

Good stripboard construction can approach the performance of true PTP. However, stripboard construction abandons the electronic elegance of true PTP for easier, faster manufacturing by less-skilled (and cheaper!) workers. With stripboard, assuming a minimum of good workmanship on putting the parts on the stripboard, the differences in quality will depend primarily on the wires that lead away from the stripboard to the outboard controls and other components.

This is an illustration of a good style of stripboard layout. The input and power transformers are at opposite ends of the chassis. The tube sockets, stripboard, and controls are laid out so that the wires from the stripboard to the sockets and controls are as short and direct as possible - no long runs of parallel wires.

Doing this is not all that easy. It takes a good deal of foresight and skill at the processes of wiring and building things in this style to get the wires as short and direct as possible. 

There is also a built-in service problem with this style. Notice that the stripboard is "tied down" by the wires coming off both sides of the stripboard to tube sockets on one side and controls on the other. To lift the stripboard itself requires unsoldering all the wires from one or the other side. Fortunately, most of the parts are easy to get to and replace from the top side if the board has NO components or wires running under the board. Sadly, this is not always true. Poor layout is not limited to PCB's.

Stripboard has a dark side, too. Consider the following picture.

In an attempt to be neat and tidy, the wires have been gathered into neat, tidy bundles where the wires from some mechanical area are gathered into a paralleled bunch, often tied into the parallel bundles by lacing ties (beeswax impregnated cotton is the old standard!) or nylon cable ties. The advantage here is that you can drive nails into a dummy amp baseboard at the corners of the wiring, to pre-cut and lace in place all the wires before ever putting them in an amp. What's wrong with this picture?

That ugly word - parallel. This is how you maximize coupling, remember. Putting the wrong components side by side can cause performance problems in high impedance circuits. Putting the wrong wire in the wrong bundle can make this construction oscillate uncontrollably, or can introduce treble loss. Conversely, not every signal wire coupling makes a difference to every other wire, so there's usually some bundling arrangement that can make this arrangement work OK. Sometimes a signal wire must be shielded to run in a bundle without unwanted coupling.