littlecraft

10W Bluetooth Guitar Amp - Part 2 - Better Sound

Mon, 19 Dec 2016 00:00:00 +0000

Last time, in Part 1, we retrofitted an old guitar amplifier cabinet with a 10W bluetooth-enabled mono amplifier, based on the LM1875. The results have been nice, but the audio quality is noticeably lacking. The sound has a dull/muffled quality, and there is hardly any punch or body to the bass.

Here in part 2, to help improve the sonic experience a bit, I evaluate different speaker drivers, build a simple passive crossover circuit, and improve the acoustic properties of the guitar amplifier enclosure.

Things to Fix

The first issue, the dull muffled sound quality, is due to the speaker driver: the original Peavey 4Ω woofer that came with the amplifier. Woofers typically have a range of around 60Hz up to as high as 5kHz. Now I doubt that the original Peavey driver had a range nearly that wide, but probably still in the neighborhood of 100Hz-3,000Hz. Commonly, the human ear can hear between 20Hz to 20,000Hz. So, that leaves out a sizable chunk of the audible spectrum! Most of the mid/vocal range, and all of high range are left entirely out of the party.

The second issue, is that there’s really not much body or punch to the bass. Even though the driver is a woofer, not a lot happens with the sound waves it does manage to generate. This has to do with the enclosure. Open back guitar amplifier cabinets are designed to accentuate the mid and high range, and to throw sound broadly around a room. It also can help musicians hear each other on stage, in the absence of dedicated monitor speakers. The downside is that the low end frequencies end up feeling rather lifeless.

4Ω Woofer, and Open back

Part 1 of building the amplifier was a blast, and I learned a ton. So, let’s iterate and give this amp some punch!

Speaker Driver

The first thing to do is to find a speaker driver that is capable of covering a wider range of the audible spectrum. Normal speaker systems account for this by having multiple speaker drivers. Sometimes having two, three, or even more individual speakers built into the cabinet (you often see this as “two-way”, “three-way”, etc). The speakers are wired in such a way so that each covers a particular band of the audible spectrum. Woofers can cover the low and part of the mid range, and tweeters can cover the rest. In my case, the guitar amplifier’s enclosure is pretty small, so to cram multiple speaker drivers into it would ultimately require sacrificing the diameter of each speaker driver. A speaker’s ability to produce sound is proportional to its ability to move a volume of air (“Volume displacement”, or Vd). Volume displacement requires a large cone, and maximal inward and outward displacement (called, “cone excursion”), which can be achieved in two ways. Either have a really large cone, with marginal cone excursion, or a smaller cone, but with greater cone excursion. Generally, it is more practical, and less expensive to construct drivers that have a large cone, which consequently require lesser cone excursion. This is why you typically see woofers with large diameters, rather than small, very powerful ones.

A good option for the guitar amp enclosure, which can fit an 8” speaker driver, might be what are called “Full Range” speakers. These typically have more than one cone built into a single chassis, where each cone is responsible for a particular frequency range. They tend to be more expensive than individual drivers, but can be integrated into more space constrained enclosures. You can find full range speakers in a wide variety of sizes, and prices (from $5 - $300, or more).

Here’s one that might work. It’s by Speco, and designed for in-ceiling home audio installations. It’s a little on the small side, but seems like it could do the trick.

Speco SPCDC6

Specifications

1/2” Tweeter
6.5” Woofer
8Ω
65W
48Hz - 21kHz
Built-in crossover circuit

I was able to find one of these at a local speaker repair shop, for $20. After removing the mesh and mounting bracket, installation was simple (on account of the built-in crossover circuit. More on that later). While it sounded much better than the old driver, and it is certainly good enough, curiosity got the best of me. I want to see what I can get out of this little amp, and giving it the right driver could make a world of difference. Moving on up the price range…

Now, this one seems like it’d have some teeth! It’s made by Celestion, and is marketed as a “Professional Full-Range Driver”. Whoa!

Celestion FTX0820

Celestion FTX0820 Frequency Response

Specifications

1.4” Tweeter
8” Woofer
8Ω
200W
70Hz - 20kHz
No built-in crossover circuit

The novel thing about the FTX0820 is the way they designed the motor driver. It’s common for full range speaker drivers to have two independent motors that drive the woofer and tweeter. The FTX0802, on the other hand uses one, very large magnet to drive them both. The result is that the tweeter sits inside of the woofer, and the sound waves that they emit have a center point that very near each other, apparently making a composed sound that is more coherent. I wonder if I can sense the difference… I’m betting on this one strictly because of its weight–almost 10 lbs! Punch!

The one disadvantage of this driver is that it does not have a built in crossover. What does that mean?

Speaker Crossover

A crossover circuit is responsible for splitting an audio signal into two or more frequency ranges. Since a tweeter and a woofer both like to operate at different frequency ranges, we need to be sure and feed the correct frequency band to them, to eliminate distortion. Speaker cabinets typically contain what’s called a passive crossover that is specifically tuned to the speaker drivers it contains. In a crossover circuit, you define a specific frequency, at which point, anything under that frequency gets routed (or crosses over) to one path, and anything above it gets routed to another. A passive crossover does not actively re-amplify the signal, resulting in some amount of signal loss. Active crossovers incorporate amplification, and even sophisticated DSP analysis to precisely tune the crossover frequency response curve. Active crossovers are typically only found in professional or high end audio installations, and occupy a separate rack mounted device. For passive crossovers anyway, like practically everything involved in DIY audio electrical engineering, there are many ways to design and build such a circuit. So many that I don’t that I can fully appreciate the breadth of public opinion about the merits of each one. So, let’s pick a simple design, and see what happens.

The FTX0820 has one woofer and one tweeter, so it will require a two-way crossover to split the audio signal into the appropriate low-range and high range frequencies for each. The FTX0820 came with a handy, though cryptic app note “[recommending a] crossover of 2,200 Hz/12 dB slope”. Uh?

The 2,200Hz refers to the crossover frequency. E.g. anything under 2,200Hz will be sent to the woofer, and anything above it will be sent to the tweeter. The 12db slope, though? So, a crossover circuit isn’t exactly like a two-way valve allowing frequency ranges to go in strictly one way or another. By design and necessity, built into a crossover circuit is a frequency v.s gain response curve, allowing the frequency ranges defined by the crossover to slightly overlap. The 12dB refers to the signal decay per octave above or below the crossover frequency. You typically find crossover circuits in 4 slope varieties. The terminology for them is thus: A 6dB/octave slope is known as a First Order crossover, a 12dB/octave slope is known as a Second Order crossover, and so on. To achieve a particular slope (or order), requires varying degress of design complexity. A Second Order (or 12dB/octave) crossover, thank the heavens, is pretty straight forward.

Jumping to the chase, here is the circuit diagram, and component values for a 2-way, 12dB/octave Butterworth crossover, with a crossover frequency of 2,200Hz, tuned to drive an 8Ω tweeter and 8Ω woofer.

Second Order, 2-way Crossover Circuit Diagrams

Values

C1, C2 - 6.3μF
L1, L2 - 0.8mH

The values chosen were derived using the Butterworth crossover formula. There are other formulas which you can use to determine the appropriate component values, and each have slightly different characteristics. See Audio Crossover ², for more details about the tradeoffs between them. The Butterworth crossover is nice because it produces decent results, while minimizing the cost of the inductors (which can get spendy and huge). It also has a friendly name.

The calculations were carried out by this handy-dandy calculator at esraudio.

Here is the frequency response curve of this particular crossover. Notice that when considering the combined amplitude of the low and high end curves, near the crossover frequency, there is actually increase in amplitude. That’s not necessarily a good thing, but it’s a byproduct of the Butterworth crossover design.

Second Order Butterworth response curve @ 2,200Hz, 8Ω

Parts

2x - 6.8μF 100V bi-polar capacitors Parts Express
2x - 0.8mH 20AWG audio inductors (Air core, or solid core) Parts Express
1x - Veriboard
3x - 2 pole 5mm pitch screw down terminal blocks
2-4x - 0.205” female crimping speaker connectors

2,200Hz, 2nd Order Butterworth Crossover Parts

Layout & Assembly

Since this is such a simple circuit, the layout should be straight forward. To help minimize stress on the solder joints, I lay the inductors on their sides, and affix them in place with hot glue.

Crossover Layout - Front

Crossover Layout - Back

And that’s really all there is to it. I made sure to mark the + and - signal paths. For best results, the hi-pass path should be connected inversely to the low-pass path. Due to the nature of the design, the hi and low signals arrive 180° out of phase, so hooking up the two speaker driver components inversely to one another aligns their phases. This isn’t strictly necessary, but is apparently nice to do, especially since the acoustic centers of the tweeter and woofer are aligned within the FTX0820. Again, I’m not sure if I would know the difference.

Cabinet

The second order of business is to see what can be done about the cabinet. It has an open back, and I suspect that this has much to do with the flimsy bass response. Resonance is often employed to enhance the bass frequencies produced by a speaker driver. It’s the reason you typically see sealed cabinets with port holes cut into them. That structure allows the bass frequencies to resonate within the cabinet, and then to exit in a controlled way, either towards or away from the listener. This design is called a “bass reflex”. The sealed cabinet allows certain frequencies to resonate, making them more pronounced and richer. At a certain frequency the enclosure no longer resonates, the port holes then (when done correctly), continue the resonant effect, thereby extending the overall frequency response of the cabinet.

The diameter and length of each port hole is important, and depends on a couple of factors. 1) The volume of the cabinet, and 2) to some degree, the diameter of the speaker driver. I’ve read that, for an 8” driver, a total port vent diameter of 4” is recommended. That certainly won’t fit in this enclosure, but we can try to get as close as we can. An alternative is to have two port holes. Our goal is to get close to 4”. The maximum hole diameter that the enclosure’s front section can support is around 2” on either side of the speaker driver, so if we place two 2” holes, that would give us $d_{t} = \sqrt{d_{1}^2 + d_{2}^2} = \sqrt{2^2 + 2^2} = 2.82$. Not exactly the 4” equivalent we’re looking for, but it’ll do. Now, the length of each port needs to be proportional to its diameter. The guide I found suggested that a reasonable length for a 2” port is around 2.3”. Great, so two 2” diameter ports, each 2.3” long. No problem.

Since we’ll be making a new front panel to hold the speaker driver and port holes, let’s make it look nice, by covering it with speaker grill fabric.

Materials

1/2” plywood (roughly 10” x 10”)
1/4” plywood (roughly 12 3/4” x 12 x 3/4”)
1/2” - 3/4” thick scrap (roughly 8” long)
2” PVC pipe (roughly 6” long)
2-part epoxy (JB-weld works great)
1 yard black speaker grill fabric Parts Express
Black satin spray paint
12x - Drywall screws
1/2” silicone or foam weather stripping with adhesive backing

The first step for this cabinet is to cut out a 10 3/4” x 8 3/4” piece of 1/2” plywood. Cut out a 7.36” center hole for the speaker using a jig saw, a router, or using a Dremel with a hole jig. Then, using a 2” hole saw, drill out two holes adjacent to either side of the center speaker cut out. Be sure to leave space between the center hole, and the port holes to allow for the outside diameter of the speaker driver, and the outside diameter of the PVC piping.

Next, cut two sections of PVC pipe to length–2.3” long. Then, using the 2-part epoxy, glue the PVC pipe sections to one side of the plywood, and let cure.

Front Panel Cutouts

To allow the front panel to be fastened to the sides of the cabinet, take some scrap 1/2-3/4” lumber and attach it to the sides of the front panel, and paint it black!

Front Panel with PVC

Now it’s time to do some upholstering. Cover the front of the panel with the grill cloth.

Front Panel with Grill Cloth

Stretch, and fasten the grill cloth using a staple gun. As you go, make sure to keep the pattern even and level along both axis of the panel. Not so easy…

Front Panel with Grill Cloth Attached

Now, just try it on for size. A perfect fit, without any room to spare.

Front Panel with Speaker Attached

Seal it up! For the back panel, I chose some 1/4” plywood, cut to 12 3/4” x 12 3/4”, cutting out a notch for the power cable.

Back Panel with Cutout

To seal everything up, use 1/2” silicone or foam weather stripping with an adhesive back. Applying this to the inside of the back panel makes for a great seal to keep those resonating sound waves nice and happy. Also, paint the interior black so that it will not be noticeable from the vent ports, as well as the back side to give a clean look.

Back Panel with Seal

All Together

Putting everything together is straight forward too. The addition of the grill fabric seals the gaps between the front panel and the sides of the cabinet, and everything else fits nice and snug inside. If you’re wondering where I stuck the crossover, I attached it to the aluminum chassis, next to the amplifier. You can see it poking out at the top right. Same as with the amplifier, the circuit board is fastened to the chassis with nylon screws and 1/4” spacers.

The new speaker driver is huge. Everything JUST barely fits!

Inside, really filled out

Button up the back, and we’re ready to dance!

The back, sealed. Not the prettiest, but nobody's going to see it!

I now have a lot of appreciation for speakers and those who design and build them. The minutia involved with designing good sound equipment is extraordinary. I cut some corners here and there, but in general, this design covered a ton of fascinating details. And the results?

Results

Yes! This sounds amazing! These two additions made a world of difference. The sound across the range is extremely vibrant. The low-end punch that I was hoping is spot on, and I am enjoying the clean logo-free look as well.

Final result

The one thing I can say, is that introducing a true full range speaker, I can hear just how crummy compressed audio really is. It’s easy to hear the artifacts and warble. I’ve also noticed that the FTX0820 has a particularly pronounced and powerful high frequency end. It’s almost to the point of being irritating and jarring. Since I did not include any kind of mid/treble tone control on the amplifier, I’ve found that using a software EQ to roll off some of the high end, to be really helpful. Had I included tone control, to account to the wide variety of source audio content that we have at our fingertips, software EQ would be unnecessary.

Anyway, another great build with lots of interesting facets, and with a really stellar result!

References

Why Nominal Diameter Tells You Nothing About Tone - http://barefacedbass.com/technical-information/speaker-size-frequency-response.htm
Audio Crossover - https://en.wikipedia.org/wiki/Audio_crossover
Bass Reflex Speaker Design - Easy Explanation - http://audiojudgement.com/bass-reflex-speaker-design/

10 Watt Bluetooth Guitar Amp - Part 1 - Retrofit

Fri, 25 Nov 2016 00:00:00 +0000

It’s been pretty quiet in the lab this summer, and playing around with an audio project has always seemed like a lot of fun, though, until now I’ve been a little too intimidated to try it out. My AM/FM radio gave up the ghost a while ago, so perfect, I thought. Now would be a great time to shake things up and modernize the lab with a custom bluetoothy speaker system on the cheap.

Some design goals:

A self-contained amplifier/speaker system, that’d be easy to hack.
Something stylish to snazz up the lab.
Use an off the shelf bluetooth adapter that’d be easy to integrate into enclosure and the amplifier, and one that’s easy to supply power to.
Something in the 10-20 watt range.

Trial and Error

The first thing that I thought about trying was to take an old boombox or AM/FM radio, and hacking in an off-the-shelf bluetooth audio adapter either into an available AUX input channel, or somehow hijacking another available audio channel (like the one used for the tape input). After a few tries scrounging around, I wasn’t able to find a boombox or radio that met all of my needs. Either they were too complicated internally (they didn’t have easily accessible audio channel lines for AUX, or tape), or they just didn’t have the right form factor. But eventually, after a bit of digging, what I did find, was a 10-15 watt Peavey Solo guitar & microphone amplifier. Alright!

Peavey Solo Amplifier in need of some love

It had a broken potentiometer for the aux volume control that was rattling around inside of the enclosure. But that’d be an easy fix with a through-hole 50K pot soldered back on to the main board, and off to the races we go!

Poor little guy

Unfortunately, after plugging it in, and listening to the audio quality, the mood got a little deflated. Although it was loud, with 15 watts, and a 4Ω woofer, the fidelity was pretty poor, and I think that it was on account of the amplifier’s board layout, cost-reducing efforts, and by driving the power amplifiers beyond their specifications (producing clipping). The amplifier also had a pretty solid 60hz buzz, which at the time, I didn’t really know how to solve. Turns out, appropriate grounding and employment of bypass capacitors (to reduce noise on the power rails) are incredibly important to producing a clean amplified audio signal.

I had a second problem too. I wanted to have the bluetooth adapter’s power supplied by the amplifier board. After hacking the two together to test things out, I could hear really awful RF interference emitting from the amplifier. The reason for this had to do with the grounding technique of the amplifiers board layout. It definitely wasn’t in a star pattern, and it looked like power and signal grounds were intermixed, which I guess can lead to problems. The amplifier chips on the board also lacked proper bypassing, presumably to reduce the COGS and manufacturing costs of the product.

One half of the stock amplifier circuit; Comprised of two TI RC4558P chips, presumably in a bridge configuration⁴

I played around with adding some bypassing to the amplifier chips, but ultimately, trying to fix the problems started to seem less and less feasible for me to do in a reasonable amount of time.

Hacked in some bypass caps on the amplifier chips

Well, that didn’t work. So, short story long, let’s roll our own!

Disclaimer: There are many good quality bare-bones amplifier boards that are available pre-assembled, for a very low price. Just checkout parts-express. Some even have integrated bluetooth modules. But if you want a learning experience, and a special feeling of satisfaction for a job well done, then why not try doing it yourself!

New goals

Going back to the drawing board a little bit, I picked these as my new goals:

Build a self-contained mono amplifier/speaker system that will fit into the old Peavey guitar amp enclosure.
A circuit that’s easy enough for me to understand.
An off the shelf bluetooth adapter that can be integrated into the circuit.
Provides 10-20 watts through a 4Ω load (so that I can re-use the speaker)
A simple, yet snazzy chassis and face plate that fits into the cavity left by the stock Peavey chassis, and one which provides heat sinking for the amplifier chip.
An integrated 120V AC power supply.

Bluetooth

This off-the-shelf bluetooth adapter is from XINGDONGCHI, and is great for this project for three reasons:

XINGDONGCHI USB bluetooth Aadapter

It’s cheap (< $9 shipped).
The presence of the USB port means that it’s powered by 5v.
The presence of a 3.5mm headphone jack means that the audio is easy to intercept.

This device isn’t actually a USB audio adapter. It provides audio through a 3.5mm headphone jack, and only uses the USB connector for 5v power. It’s pretty ideal for integrating into an amplifier project.

To prepare the adapter for integration, first remove the plastic enclosure by placing it in a vice, along its seam, and gently applying compression until the two halves pop apart. Bingo. Then, to provide a more permanent integration point, depopulate the USB male connector, and solder on a 4x1 male pin header. That will provide power and grounds to the two outer pins, leaving the two inner pins unused.

Bluetooth adapter after depopulating the USB connector, and adding a male pin header, and white wire to the indicator LED

This particular adapter has an indicator LED, which blinks at different patterns according to whether or not something is paired to it. I chose to depopulate the LED as well, and run its signal (which is connected directly to the SoC via a 1K resistor), out to one of the unused pins of the male pin header. My goal for that is to incorporate the blinking pattern into the front panel of the amplifier somehow.

Bluetooth adapter after depopulating the USB connector, and adding a male pin headers

Now, since this project is going to be a mono amplifier (there’s only one speaker), you might think that you could forgo the 3.5mm audio jack and somehow route a mono signal through the second unused pin of the male header. But that would mean that the signal and the power source for the board would directly share a common ground pin. That may work, but to do our due “grounding” diligence, I decided to keep the 3.5mm jack, so that we would have separate grounds for the signal, and for the power source of the adapter.

Get down to business

Okay, so building an amplifier… The great thing is that there is no shortage of in-depth, detailed information out there on how to build DIY audio equipment. Unsurprisingly, that also turns out to be a downside as well, as it can be a bit overwhelming to wade through it all. Thankfully, there’s is a great article in MAKE that details pretty much exactly what I want to do:

They call it the “Skeleton amplifier”, and it’s based on the LM1875 chip, which has gained popularity for its ease of use, its quality, and power. The project also provides a circuit diagram for a power supply, and parts list, configuring two LM1875s as an 11 watt per channel stereo amplifier. Best of all, I only have to build 1/2 of the amplifier circuit, since I only need mono! Yes!

Parts

Here’s the parts list for the mono amplifier. These are mostly taken from the Skeleton Amplifier guide, and I was able to source almost everything from Mouser (as opposed to radioshack) reducing the cost somewhat. I include a couple additional components for the purpose of integrating the bluetooth adapter. I was able to salvage capacitors and other various things from other derelict amplifiers that I happened to have in the junk pile. In particular, from a fancy, yet busted Kenwood amplifier, I pulled two enormous 7500μF 80V caps that I then used for the power supply, and a large aluminum heatsink which I used to dissipate the heat form the LM1875. Below, I list the part numbers, where available, but feel free to find any part that meets or exceeds the specs described below.

If you don’t have a ready supply of capacitors and resistors, getting a pre-assembled kit from ebay or amazon that contains a wide variety of electrolytic, ceramic, and film capacitors is a great way to go. Resistor kits are also available, and relatively inexpensive. If you do buy capacitors or resistors from mouser, and you can spare a few extra dollars, buy more than you need, in quantities of 10 to reduce the per-unit cost. Having them around is really handy.

Note: the parts list below is for a MONO amplifier

Capacitors

2x - 4700μF to 7500μF (35-80V) electrolytic capacitors (ECA-1EHG472)
2x - 100μF 25V electrolytic capacitors
1x - 22μF 25V electrolytic capacitor (ECE-A1EN220X)
2x - 1μF 25V film capacitor (ECQ-E2105KF)
3x - 0.1μF 25V film capacitors (ECQ-E4104KF)
1x - 0.22μF 25V film capacitor (ECQ-E2224KF)
1x - 0.33μF 25V film capacitor

Resistors

1x - 1Ω 1W metal film resistor (CPF11R0000FKEE6)
1x - 100KΩ linear taper, through-hole potentiometer, 16mm shaft, don’t skimp, and get a nice one (P160KNP-0QC15B100K)
2x - 10KΩ any type resistors
2x - 22KΩ any type resistors
1x - 220KΩ any type resistor
1x - 1KΩ any type resistor

Power

1x - 25.2V out (115V in) power transformer (Triad F-41X)
4x - 40-60V 5A Schottkey Diodes (SB560-E3/54, 1N5822)
1x - 120V rated SPDT panel mount toggle switch
1x - 500mA 250V time lag fuse cartridge (0218.500MXP)
1x - Inline fuse holder (go to Radioshack and see what they have that will fit a 5mmx20mm fuse)
1x - 2-prong power cable

IC

1x - LM1875T
1x - XINGDONGCHI USB Bluetooth audio adapter (or something that seems like it’d be easy to work with)
1x - 2N706 NPN Transistor
1x - 5V voltage regulator (L7805CV)

Other

2x - Perf boards
1x - 1x40 breakaway male pin header
1x - A stylish knob for the volume potentiometer
1x - LED (I chose a neat defuse white)
6x - 2 pole 5mm pitch screw down terminal blocks (optional, but nice. Note: only 4 ended up in the picture below)

Chassis and heat dissipation

1x - 4x2x0.125” x 10 7/8” (or whatever dimensions you need) architectural aluminum stock.
1x - Aluminum heatsink (optional, you can alternatively find a way to thermally bond the amplifier chip to an aluminum chassis)
1x (at least) - TO-220 thermal insulator, to electrically insulate the LM1875, but still allow it to thermally bond to the heat sink (43-77-20G)
1x - Silicone heat sink compound

Mounting hardware

1x - Nylon 1/4” P-Clip (For tension relief of the power cable)
5x - #6 nylon screws
5x - #6 nylon nuts
5x - #6 nylon washers
8x - #6 1/4” nylon spacers
4x - #6 1/2” long pan head metal screws
2x - Small metal screws (secure the power supply)
2x - Small metal washers (secure the power supply)
?x - Small screws to secure the heat sink

Amplifier parts (most of them, anyway)

Power supply parts (well, everything except the mounting hardware and terminal blocks)

Chassis and mounting hardware

Power Supply

Be very careful with main power!

Also be careful with these large capacitors. They maintain high voltage even after they are disconnected from power. Properly discharge capacitors before handling the power supply circuit again.

This part is actually really satisfying. We get to make a bi-polar 20V DC power supply that can be integrated into the guitar amp enclosure. It’s actually surprisingly simple, and interesting.

Power supply schematic

So, the basic idea is that the transformer steps down the AC voltage to around what we’re after: 25V, and the arrangement of rectifier diodes (in a bridge circuit configuration) serve to efficiently convert alternating current into a positive 25V and a negative 25V DC output. The center tap of the transformer serves as our ground reference.

To get the diodes and capacitors to fit into a through-hole perf board, you may need to drill out the holes with a 5/64” or 3/32” bit. Below is my attempt at a power supply board using a strip contact board, where there are conductive traces along the horizontal axis of the board. I actually don’t rely on these traces exclusively for the power supply, and instead add jumper wires to better handle the current and voltage requirements. Maybe overkill, but better safe than sorry. To secure the capacitors, deposit hot glue on the bottom of them prior to placing them on the board.

By the way, if you do use perf boards with strips of contacts, you can use an x-acto knife, a drill bit, or a special tool to disrupt the traces where necessary.

Power supply, partially assembled

Power supply layout

Once everything is laid out and soldered together, attach the transformer. The lead with the stripe is the center tap, and is attached to the ground plane of the board.

Power supply layout, hooked up

Next, splice an old 2-prong power cable and solder the fuse holder to the hot lead (the one with the white stripe). Then, solder the fuse holder to your power toggle switch. Since this is carrying main power, be sure to choose a toggle switch that is rated for 120V. Solder a lead from the opposite pole of the toggle switch to one of the black input leads of the transformer, and then finally solder the neutral lead of the power cable to the other black input lead of the transformer. Here, I use black heat shrink tube around all of the solder joints to insulate them. Now, just plug it in, and test the voltage. From the 25V AC transformer, you will likely see somewhere in the neighborhood of +/- 20V. The drop in voltage is due to the accumulated forward voltage drop of the diode rectifier bridge. Be very careful, even after the power supply has been disconnected from main power, the capacitors will hold a voltage for quite a long time. If there is no load connected to the power supply, carefully, without actually touching any conductive surface, take a 1k resistor and short the poles of both of the capacitors. I taped a resistor to a stick, and used that to discharge each capacitor. You can use a screw driver with an insulated handle to short the poles, but this will most certainly cause arcing.

Power supply, with fuse and switch

This is how it ended up looking once installed in the enclosure. I drilled holes through the PCB and used #6 metal screws, and #6 nylon washers and #6 1/4” nylon spacers to raise the PCB off of the bottom of the enclosure. The transformer is held down by small metal screws and washers. To relieve tension on the power cable, it is attached to the side of the enclosure by a black 1/4” P-Clip.

Power supply, installed

Amplifier

The amplifier circuit itself is essentially identical to the LM1875 amplifier circuit detailed in the “Skeleton” Amplifier project. In that project, you of course build a pair of amplifiers, for stereo. Here we only need one. There are only a few main differences between the Skeleton circuit, and the one below.

It takes stereo audio input, and sums the left and right signal into a single mono source.
There is no external audio input source. The audio signal from the Xingdongchi bluetooth module is the only source.
There is an additional +5V power regulator circuit that is used to supply power to the bluetooth module.
There is a transistor that is used to control an external indicator LED.

Amplifier circuit

I want a single LED power/connectivity indicator for the front panel, but the LED indicator of the bluetooth module doesn’t quite behave in a way that I need. For instance, it doesn’t illuminate until the SoC boots (which takes a second or two). Secondly, I want to keep the design open for adding an AUX channel port, and allow for the bluetooth module to be powered down while the AUX channel is being used. This means that I cannot rely solely on the output of the bluetooth module’s LED output. To solve this, I use an NPN transistor, and create a not gate, which essentially inverts the LED indicator signal that comes from the bluetooth module. The result is that the LED is on by default, and blinks inversely to the signal of the bluetooth module’s LED output.

Before trying to come up with a board layout, I tried the circuit out on a breadboard, to see how it worked. This doesn’t actually give you a sense of how the amplifier will sound when laid out on an actual board. The contacts between components are not terribly secure, there are a lot of long leads that can pick up interference, and the grounding typically won’t be very good, causing ground loops, which can again, introduce interference. In particular, I noticed that the bluetooth module introduced quite a bit of interference, even though the amplifier is pretty well protected by bypass capacitors. Also, it is not a good idea to run the amplifier for very long without some kind of heat sinking. It gets extremely hot, even after a few seconds of activity. Still, this felt like an important step.

Aaaaah

Layout, in Theory

Probably the trickiest part of this project is coming up with a sensible layout. There are two main things to focus on.

Bypassing

Bypass (or “decoupling”) capacitors^5,6 connect from the power rail to ground and serve to dampen noise in the form of voltage fluctuations on the power inputs to our ICs. It’s important to place these capacitors as physically close to their IC as you can. Traces inevitably have some amount of impedance, so the longer the trace, the more likely it is to pick up noise.

Grounding

A “star” grounding configuration^7,8 is pretty important for this amplifier. This basically means that all ground junctions should run directly to a single point, and then that point becomes the common ground to the chassis. This prevents interactions between circuit features that could cause noise and instability.

Layout, in Reality

You may not end up getting a perfect star ground configuration, but give it a shot. I wasn’t entirely successful in getting a perfect star ground configuration, but it’s not bad for a first attempt.

Fig 1. Amplifier layout, top (1)

I used screw down terminals to connect all of the external parts. In Fig 1, starting from the top left, the screw down terminal closest to the bluetooth module is for the indicator LED, that will be attached to the chassis. The one below it is for the power. I only had 2-pole terminals available, so I had to use two of them for the power supply attachment point. These are wired as VCC- on the top, two grounds in the middle, and VCC+ at the bottom. The second GND terminal block will be used to ground the chassis using a lead with a spade connector. For the audio input, there are a total of 4 poles. Each channel has an independent ground, of which, I only use one. The terminal to the right is the output to the speaker.

The volume potentiometer is positioned flush with the front of the board so that it can be attached to the chassis’ front panel, and the LM1875 is positioned flush with the back of the board so that it can be bonded to the heat sink. I did not bother screwing down or heat sinking the L7805CV voltage regulator, since it is not going to get terribly warm, nor do I intend to move the finished amplifier around very often.

Fig 2. Amplifier layout, top (2)

Here is the layout of the underside of the board. In order to reduce the overall inductive surface area of the circuit, I was diligent about disrupting the on board traces, so that they were as short as possible. Again, you can use an x-acto knife, a drill bit, or a special tool for this.

Fig 3. Amplifier layout, bottom

Chassis and Front Panel

A single piece of 4x2x0.125” x 10 7/8” angle architectural aluminum serves as both the chassis and the front panel of the amplifier. I was able to source this from a local metal supply shop that sold 4x2x0.125” angle architectural aluminum by the inch. “Architectural” in this context means that it has a crisp right angle on the interior, which is perfect for this application. The heat sink and circuit board are both mounted to the 4” wide section of the angle aluminum, and the 2” section serves as the front panel. Because there is an inside angle of 90°, the circuit board can fit up flush against the front panel.

Aluminum stock 4x2x0.125 x 10 7/8 inch, angle architectural

To help make clean holes, and to more easily lay things out, put down a layer of masking tape over the outside surface, and drill inward. The front panel will have a power switch, a volume knob, and an power/bluetooth connectivity indicator LED.

Front panel layout

If you don’t have a drill press, use a center punch, a nail, or screw to make a centered guide for the drill bit.

Center punch holes

Chassis layout, circuit board footprint

To get the spacing correct, you can always try playing around with the boards orientation, and layout of externally accessible components, so that you can make sure that the heat sink, the amplifier, the holes in the chassis, and the volume potentiometer all align.

Chassis layout test

The heat sink I use has gaps in three locations across its length, and are easily tapped with a small screw. Once everything is in place, mark the location of the holes for the heat sink with a sharp nail.

Once you’ve drilled the holes, knock down any burs with a metal file, and affix the heat sink. To attach it to the chassis, just screw it to the bottom.

Attaching heat sink to chassis

To give the aluminum a nice finish, use a high-grit sand paper (i.e. 600 grit) to remove any swirls, and to produce a slightly matted grain finish by sanding in a constant horizontal direction.

Assembly

All that’s left now, is to bolt it all together. To mount the amplifier circuit, use #6 nylon screws, nuts, and 1/4” spacers. To bond the LM1875 to the heat sink, use a TO-220 thermal insulator, along with some silicone thermal compound (on both sides of the thermal insulator), securing it with another #6 nylon screw, washer, and nut. It’s important for the LM1875 to be electrically insulated from the heat sink, but thermally bonded.

Assembly - 1

Ground the chassis using a crimped spade connector, and attach the other end of the lead to the star ground of the amplifier circuit. I used the spare GND terminal block for this.

Finally, secure the power switch, the external LED, and the volume knob. Not pictured here, but I encase the bluetooth module in heat-shrink tubing to better protect it from an accidental short.

Assembly - 2

Results

In the end, things turned out pretty well. The amplifier sounds great, and is definitely loud enough for my small space. I also like the clean look of the aluminum face plate, and defuse white LED. I am especially happy about the volume knob. I found it quite a while ago in a stall at an electronics MEGA store in Akihabara, Tokyo, and have been holding on to it ever since; sometimes even daydreaming about what it might one day be used for.

Results, back

Results, front

It isn’t perfect. There is a bit of audible interference from the bluetooth module, but it is only detectable at high volume, and without the audio source playing. I also can hear a bit of DC feedback when the LED blinks, which I think is due, again, to improper grounding. Even so, it still sounds great to me!

Next Steps

I suspect that the sound quality of this amplifier can be improved. The enclosure has an open back, and uses a 4Ω woofer driver, which doesn’t seems to have a very wide range. It has decent bass, but muffled mids and highs. The next step for this project will be to hack the enclosure, sealing the back and air gaps, and opening up a bass reflex port so that the bass can resonate more efficiently. A full range speaker driver would also be a nice addition. These drivers often have two integrated speaker cones (a tweeter and a woofer) and built-in cross-over circuitry. In the mean time the lab is filled with pleasant tunes once again, and I had a great learning experience.

References

Power Supply Schematic - /images/posts/amp/Power-schematic.png
Amplifier Schematic - /images/posts/amp/Amp-schematic.png
Bare-Bones ‘Skeleton’ Amplifier - http://makezine.com/projects/build-bare-bones-skeleton-amplifier/
Bridge Amplifier Circuit - https://en.wikipedia.org/wiki/Bridged_and_paralleled_amplifiers#Bridged_amplifier
Demonstrating the Effect of Decoupling Capacitors - https://www.youtube.com/watch?v=UW_XFGGTh0I
EEVblog #859 - Bypass Capacitor Tutorial - https://www.youtube.com/watch?v=BcJ6UdDx1vg
Star Grounding in Tube Amplifiers - http://www.geofex.com/article_folders/stargnd/stargnd.htm
Solving Ground Loop Problems - Star Grounding - http://www.lh-electric.net/tutorials/gnd_loop.html
Grounding and Shielding for your DIY Audio Projects - http://diyaudioprojects.com/Technical/Grounding-Shielding/