Hello everyone.

I was trying in the past few days to find a way to build a cheap temperature logger that I can use to document my roasting experiments. My goal was to build a cheap and a scalable (with regard to the number of channels) logger that can be very easily interfaced to a PC.

To optimize the cost, the PC can be used to implement most of the functionality of the logger, hence using software on the PC to substitute hardware (i.e. electronics) in the device. The simplest such arrangement is interfacing a "temperature sensing device" to the PC directly, and letting the PC control the device and poll it at equal intervals to read the temperature.

Furthermore, having more functionality implemented in software on the PC means that this part of the functionality is upgradable and can be implemented in neat ways. For example, the PC can not only be used to record the temperature profiles and save them into files on the disk but also to display the temperature, as well as other metrics (such as the first derivative) in real-time, which is a cool feature.

The great news here is that the hardware part of such a device already exists. It was designed and implemented by "Tienmu Taipei" from Taiwan. Here is the page for his "USB Thermocouple" project:

http://www.chinwah-engineering.com/USBThermocoupleProject.html

The hardware in this design is quite minimal. It is composed of two electronic chips, one for sensing the temperature using a K-type thermocouple wire and the other for interfacing the temperature sensing chip to the PC through a USB bus. The interfacing chip needs minimal support, for example, no crystal and no termination resisters for the USB bus wires are needed, instead, they are integrated INSIDE the chip, which makes the device even simpler, using less electronic components.

Seriously, the only components needed beyond the two chips are the connectors (thermocouple and USB female connectors) and some cheap capacitors. The interfacing chip costs 4.37 dollars (Newark.com) whereas the temperature sensing chip costs 13.86 dollars (Newark.com), totaling around 18 dollars which is okay (not including the price of the connectors, the PCB board which is to be used to assemble the circuit, the thermocouple wire, and the USB cable).

My research lead me to a similar design, but using a more-expensive chip for interfacing the temperature sensor to the PC. This design needs lesser components, is better, and cheaper.

The only thing we need is a modified software designed specifically for roasting coffee, as opposed to a simple temperature gauge.

Furthermore, this design can be easily modified to include many channels, with little effort.

Is this something you guys already know about or is it a cool idea that I have just come up with? (to use such a cheap design for coffee roasting)

see THIS thread for hardware and software solutions..

Also THIS one for some interesting software.

Wow, that is a looooooong thread, with an impressive build!

I did not get the very details, but I would like to write something commenting on the differences between what I had in mind and the "TC4 Shield" project referred to by Randy G (click here for a brief and an easy-to-read account of the invention).

To the untechnical person, the "TC4 Shield" is a daughter board that can be plugged to the famous development board named "the Arduino". It extends the capabilities of the Arduino board, making it possible to attach 4 K-type thermocouple wires to the board and use it to log 4 temperature channels simultaneously. In the latest version of this "daughter board" (i.e. version 3) it is even capable of controlling the heating element of an electric roaster, hence functioning as a PID controller, with direct control of the rate-of-rise (in degrees per second) of the bean temperature. As you can see, this is MUCH more than what seems to be a 4-channel temperature logger. It is a powerful, open-source product with endless possibilities.

The whole unit (composed of the Arduino board and the TC4-Shield daughter board) can be used in two ways. Either as a peripheral connected to a personal computer, with the ability to log and display different metrics in real-time on the PC, or as a battery-powered unit for logging the temperatures and storing them in an on-board memory (whose contents can be viewed later from a PC, off-line). Again, this is a very versatile unit whose functionality is self-contained and can be extended. The extendability comes from the fact that the product is built around an extendable device, which is the Arduino board.

However, it is not very cheap and not very easy to assemble and bring to function, due to the fact that the Arduino board is needed (around 30 dollars) and that the daughter board needs to be assembled by the final user (i.e. components soldiered into place on the bare PCB, which can be purchased from the project launchers, for 6 dollars only). Furthermore, the fact that the components needed in the daughter board are surface-mount (with legs lying on the surface of the PCB board rather than passing through it via holes) makes assembling the daughter board harder for the unexperienced person, because it requires precision in the soldiering and a heat gun (as compared to a simple soldiering iron in the case of through-hole components).

To bring this unit to function, the final user needs to be able to do the following tasks successfully: (1) assemble the Arduino board in case it was bought unassembled, (2) program the Arduino board, (3) assemble the TC4-Shield daughter board (which means buying the bare PCB and the list of electronic components needed -totaling 12 different components), and finally (4) plug the two boards together to have a working product.

The software in the case of this unit is split into two parts: (a) the software that is to be run on the device itself (this program is transferred to the device in the second step above "program the Arduino board"), and (b) the software that is to be run on the PC. The first of these two parts is called the firmware.

Following the project's very spirit, which in my opinion is to produce a very "modular" design, some different software options (firmware + PC software) are available which differ in their capabilities and functions (e.g. the type of PID control algorithm), which is a fascinating concept (the concept of extending modularity to software).

Now I wanna talk about what I had in mind and how it is different from the TC4-Shield project.

What I had in mind is an extremely-cheap, extremely-simple device that is PC-based (i.e. Arduino-free). It is a dumb device, i.e., it does not possess the ability to log the temperature on its own, and hence must be connected to a PC while being used (which plays the role of the actual logger). It needs no programming, with a much easier assembly than the TC4-Shield daughter board, and with most of the functionality implemented in software on the PC. Another way to view this approach to temperature logging is "connecting a dumb temperature sensor to a PC".

Assuming that the software used to control this "dumb device" could run on a handheld device (e.g. devices running Windows Mobile), the device can possess some mobility, hence simulating a standalone unit but while keeping the spirit of very cheap design based on shifting the "logic" from the electronics to the PC, as much as possible.

Hence I can summarize the differences in the following points:

The TC4-Shield Project (an existing project):

- Philosophy:
A very modular design with many possible upgrades, tweaks, and software options.
- Pros:
4 temperature-logging channels, standalone functionality with a battery, can function as a PID controller with direct rate-of-rise control, extendable hardware functionality, many software options (including both firmware and PC software).
- Cons:
Not that cheap, composed of many parts, not very easy to assemble, hardware needs programming prior to use.

A cheap PC-based logger (an idea for a project):

- Philosophy:
A very cheap design with minimal hardware and hopefully the ability to incorporate more than one channel.
- Pros:
Very cheap, very simple, very easy to assemble, no need for programming the hardware prior to use, functionality mostly in software, software can be available for handheld devices.
- Cons:
Cannot be used as a standalone unit in a literal sense, non-extendable hardware functionality.

I am wondering if it's a good idea, I mean, if some people would prefer such a simple design to the Arduino-based TC4-Shield unit.

Hi
Interesting idea.

I was wondering about the cost of the PC version, so I looked at the sparkfun page, and the first three items on his list add up to $30, for one TC channel. I am also not sure if you can connect two 6675's to one FT232R, to get a 2nd channel.

It would obviously be cheaper is you bought the parts from digikey (I did not cost this), and built it yourself, but then you need a board for it, and the FT232R does not look that easy to solder.

I think the TC4 ends up costing around $30 in parts (don't hold me to it), plus another $30 for the arduino. And you do need to solder some SM parts, and put the thing together in some kind of box.

Full disclosure. I wrote the PID for the Arduino/TC4, so I am bias towards that project.

Randy

Does anyone know of a source for a pre-assembled TC4 shield? There should be enough of a market for someone familiar with surface-mount soldering to build and sell a few for a small profit.

I know I'd buy one right away...

I used to have the time and loved building my own pcb-based hardware, but these days a few $$$ are easier to come by than time not to mention that my eyesight is getting marginal for some of those tasks

robca wrote:Does anyone know of a source for a pre-assembled TC4 shield? There should be enough of a market for someone familiar with surface-mount soldering to build and sell a few for a small profit.

The volume is miniscule, so small is the operative word

Kidding aside, TC4 has always been a not-for-profit project. When we accidentally made a few dollars on the first batch of boards, we shipped it off to CoffeeKids.

If you want a board with just the surface mount components pre-soldered, send me a PM. I still have a few bare boards left and can pop one in my toaster-oven-reflower for you. I'd only charge you for the materials.

Jim

A new shipment of bare TC4 boards has just arrived. Information on the boards, including a link to order a bare PCB, can be found on the project's googlecode site.

A limited number of boards is also available with the surface mount devices pre-soldered (you would still have to solder the through-hole items yourself).

More information on the TC4 project is on this very long Homeroasters.org thread.

Jim

Bill has just added a circuit for the variac volts readings to my soon to use TC4 16x2 display. Time, RoR, V, BT, ET
The Smithsonian has been asking me for the first RoR board JimG made.

farmroast wrote:The Smithsonian has been asking me for the first RoR board JimG made.

Don't accept their first offer. You should hold out for the big bucks

Jim

JimG wrote:Don't accept their first offer. You should hold out for the big bucks

Jim

They first offered one of the space shuttles for it but I said the mileage was tooo high on those darn things.