Amp Camp Amp Redux

Background:
In 2012 I helped put on an event for newbies who were interested in electronics or audio build their first amplifier. We called it Amp Camp. My largest contribution to the camp was to design and machine 50 (!) chassis for our 25 attendees. We were lucky enough to work with a very well regarded designer, Nelson Pass, who supplied the circuit design and many of the parts for that camp. Despite selling commercial amplifiers, Pass has been active in the DIY community for many years and extremely generous in sharing his designs, experience, advice, and even his personal stockpile of parts. He wrote an article describing the design and his process and reasoning for the design decisions he made.

Amp Camp Amp #1

Most of the constraints he placed on the design have to do with accessibility for new builders. Among them:
Switching PSU to prevent exposure to mains voltage
Low parts count for simplicity
Low parts cost

A thread on diyAudio quickly sprang up and is still very active today.

In true diy fashion, there have been many variations and changes to the circuit over the years. Since the inaugural camp, diyAudio has offered a kit in their store, where it has seen high demand. The current version of the diyAudio kit comes in a beautiful aluminum case with options to turn one stereo unit into a mono unit by connecting the channels in bridge or parallel mode for more power. It is clearly a hit in the diy scene.


Objectives:
A return to form for the Amp Camp Amp. Above all, ease of assembly and low cost. Accessibility is the name of the game for this Amp Camp Amp variation. My intent is not to make the best performing ACA, but the best design for someone who is interested in our hobby with little to no build experience.


Results to date:
An even easier to build ACA (1.6) for about $80 per channel in low quantities, if using non-premium parts. The cost comes down when building more as a batch. I built 2 channels and had an unexperienced friend build 2 channels for a first prototype run about 2 years ago and have been using them in my desk stereo system since. I have sat on this design for a while now and figured it was time to finally write it up for sharing. The amp you see below is a single channel.




The enclosure presents the biggest opportunity to cut cost in this amplifier (and most other DIY amps). There a few critical functions that an enclosure performs.
1. Heatsink
2. Finger-proofing
3. Aesthetic

I selected a heatsink meant for LED lights. Typical "grow light" heatsinks are now produced in large quantities and are one of the more economical large passive heatsinks available. The one I've selected is a circular forged pin heatsink 120mm diameter and 70mm tall (4.7" × 2.75"). I estimate its thermal resistance at around 0.9°C/Watt. It cost $19 with free shipping.

Another advantage of the LED heatsink is that the base plate is meant to be horizontal with the pins sticking up (and LEDs pointing down). This is what makes the "caseless" design a bit more practical. By placing the PCB below the heatsink it provides a bit of protection for the components underneath and IMHO, some visual interest. Standoffs hold the whole thing up off the table and no other enclosure parts are needed since the amp operates at a touch safe voltage.


Board layout and parts selection:
In the interest of beginner friendliness, I've elected to print component values on the board and line up all the small resistors according to their numbered designation. The large capacitors lay down flat to minimize the overall height (currently 96mm / 3.8" tall). The heatsink is "grounded" through brass standoffs to the PCB mounting holes. Nylon standoffs screw into the brass standoffs so they may act as somewhat less scratchy feet for the amp to sit on. I added a bit of camp themed artwork to the back side of the PCB along with the simplified circuit from Pass' original article. I chose white soldermask because I think it makes the art look good, but it's quite difficult to see the traces ¯\_(ツ)_/¯








In an attempt to make assembly as easy as possible, all the connectors and the power switch are mounted to the PCB. Look ma, no wires! PCB mount power jacks and and RCA inputs are commonly available. Binding posts are less common. This design uses a right angle screw terminal with a 5 way binding post screwed into it. The threads are secured with a bit of solder. Because the binding posts and screw terminals have a high thermal mass, this is the most difficult part to solder. But even a beginner with patience and a higher powered soldering iron should be able to make it work.




These LED heatsinks often come with pre-tapped holes with spacing for the most common LED modules and lenses. I aligned the MOSFETs with the furthest apart M3 holes and used the M4 holes near the edges for the standoffs.








Future improvements and changes:

• Add more info to silkscreen, e.g. bias voltage. Possibly sacrifice rear side art for complete build instructions.
• Add more output capacitance (this version has 2000uF).
• Replace LSK170 label with something a bit more generic.
• Consider repositioning the LED to shine up through the heatsink pins.
• Consider squaring up the back edge of the board so power and RCA connectors aren't angled.
• Move to a larger 140 x 70mm heatsink for more cooling capability and board area. This design doesn't need it, but I want to continue exploring this physical design while standardizing the heatsink size.

I'd love to incorporate feedback and improvements from the hive mind on diyAudio and gauge interest in boards / kits, no promises though!


Further developments:
A slow moving, inaudible fan above the heatsink will more than double the cooling performance, opening the door for higher power designs in the same form factor. I've started tinkering with some old F5 boards I had laying around and I've started thinking about a 2 channel, single board ACA for bridge / parallel operation. These future projects are the primary reason for moving to a larger PCB / heatsink.

Nintendo Switch Dock Modding - Integrated GameCube Controller Ports

The Nintendo Switch Dock should have shipped with GameCube controller ports. 

This post details the solution I created to integrate a third party controller adapter into the dock of my Nintendo Switch. I haven't seen any other projects that are this integrated but I would like to acknowledge Rated-e Mods who has made a few videos exploring ways to integrate an adapter with the Switch dock. Those videos got me thinking about how to take things to the next step.

Motor Controller Packaging

Quick note: This post is a work in progress. Some sections are incomplete or missing.

1. Objectives

The motor controller used in the MY2015 will be based off of the Unitek Bamocar D3 modules used in the 2014 car. The D3 controller meets the electrical requirements for our tractive system (voltage, power, control scheme) but there is an opportunity to make significant improvements for use in our application by reducing size, weight, and installation complexity. To make the most of a redesigned housing, both control modules and will be integrated into a single package.

The 6.115 Saga

I started work on this project almost a year ago. A majority of what follows happened over summer 2013, as of last week I finished the last task needed to get it to production.

This requires a bit of explanation.

There is a class taught by Prof. Leeb called Microcomputer Project Laboratory, or 6.115 in MIT speak. It runs each spring and students are taught how to work with microcontrollers and then are required to complete a sufficiently complex electronics project of their own design. To aid in this process, each student is given a briefcase size kit which holds protoboards, a microcontroller interface, power supplies, handheld tools, a bag of parts and whatever else they choose to use. It is highly encouraged that the kits be brought home so work can be done outside of class.

The 6.115 kit as used today and 10 years previous.

Constructing battery packs from prismatic cells

When the FSAE team decided to build an electric race car a few years ago, it seemed pretty likely that we could source our batteries from A123 Systems, an alumni company. They make very high quality lithium iron phosphate cells and battery packs. The packs we had hoped to get look like this:

It's a familiar story around here, but A123 went bankrupt and changed hands and we were never able to get any packs. However, we were able to get enough raw AMP20 cells for the pack we needed. A team member spent their thesis designing our custom pack based loosely on the A123 pack. I have somehow ended up on the battery sub-team and helped assemble one of our new modules. Bring on the photos.

Motorsports CAN bus

This year I've decided to rejoin the MIT motorsports team. The team designs, builds, and races a car for one of the various FSAE competitions held each year. I was on the team for a little over a semester my freshman year working on a temperature sensor network; it went okay. Now I'm back and the team has switched to the electric competition. My main task is to compact the motor controllers such that they actually fit inside the car. Because that's moving a bit slowly, and in the interest of rekindling the blogging habit, I have a smaller project to present first.

Modern cars have electronics and sensors stuck inside just about every functional part. These bits and bobs need to talk to the main computer and eachother to be useful. Just about every car made after 1996 (and quite a few before then) use the CAN bus protocol for communication. Our car is being made after 1996 so it's going to use CAN bus. To reduce the number of wires running all over the car there will be a breakout CAN node in the front of the car. Last Saturday I mounted and wired the connectors in a laser cut panel I made the week before. It needed to be a custom job because there were all sorts of problems from the board through holes not being big enough to a lack of keyways in the panel.

TL;DR: here are some pictures of a thing I did on Saturday. Just to be clear, I didn't make the circuit or boards for this.

"Branded" boards.

MIT kids for sure

There aren't many pictures of the electrical system. It was built in a blitz of excitement and determination to finish by CPW. In the long run it doesn't seem like it has paid off. The on again, off again nature of Zoran is like something from a teenage romance novel. Something a bit more... mature, would be preferable.

Before A123 declared bankruptcy you could buy their batteries from buyA123batteries.com. Yup, sounds about right. Recently they put up a notice regarding the modules we are using:

This is kinda what we thought the problem was. Our original wiring had three batteries in series. Now we've upped it to four batteries, but have them in a series parallel arrangement for 24 volts and 80 amps. This gives faster acceleration, but lower top speed.

The life and death of Zoran

We got tired of not having a name for the kart so we came up with Zoran. In Norse mythology he's the god of speed and the one who built Thor's Chariot. Charles (dear leader) thought we named our car Zorak, not quite, no. (Just to clarify, Zorak is the mantis in that clip and it's from a show called Space Ghost Coast to Coast).

We put together a logo type thing and engraved it on one of the leg plates. Here's a video showing the process, I suggest you click the little gear to speed up the playback. It prints rasters just like an inkjet, but at least thirty times cooler because lasers.