DIY Bread Proofer[Short URL]

Bread requires a warm, moist place to rise. I find that out on the counter, it can be too cold in the winter. Putting it in a cold oven with a bowl of boiled water works, but requires a cold oven, so it tends to be impractical when I have many batches of bread to do. I've also used my bed with a hot water bottle, but this is less moist than I would like and doesn't work overnight. What I want is a place dedicated to rising bread: a proofer. The moment that cemented the need for a proofer was when I was raising calzone in the oven and two fell off the rack and splattered all over the inside of the oven.

Bakeries use specially designed ovens that can maintain a constant warm temperature and high humidity and then can switch to very high heat to bake. These ovens start at several thousand dollars, require 3-phase power, and can do at least 20 loves of bread simultaneously. My quest was to duplicate this functionality for considerably less (bread and dollars). The basic idea is this: create some kind of box where I can hold two or three trays of bread at high humidity and about 30°C. There must be a safety mechanism to kill the heat if the temperature rises above 40°C, since that risks killing the yeast. This is the result of that quest.

To regulate the internal temperature, I decided to create a controller using a PIC micro-controller, two temperature sensors, and a heater. Basically, the PIC tries to maintain the temperature of the water in the pan and the warm water makes the air in the chamber humid.

The schematic for the circuit and the code to run the controller follow.

Cabinet Plan Follow Link
Controller Program Follow Link
Controller Schematic Follow Link

For simplicity, I used the Small Device C Compiler to write in C instead of PIC assembly. It seg-faulted when I did division, but happily multiplies by fractional amounts.

It was important to me that this thing be robust and safe. To that end, the micro-controller can sense if the door is open and can shut off power to the heating element. A circuit breaker protects the whole mess and it is rated to trip at 10A since the band heater requires about 6A.


I got most of the electronic parts at A1 Electronic Parts. Not shown in the schematic are the following parts:

The clips allowed me to connect the circuit breaker and switch without soldering. I was not able to get the PIC at A1, so I ordered it from DigiKey. Combined, that's about 80$ in parts. I also looked at The Source (formerly Radio Shack) and found a magnetic proximity sensor which works as a door open sensor for an over-priced 10$.

As for the cabinet, I got some stainless steel from Metal Supermarkets for 150$, which hurt in the wallet. From Real Canadian Wholesale I was able to get a small deep steam table pan to boil the water after being fit with a heater. The heater I found by the road side in an electric water boiler left out after garbage day. It is a mica band heater, so I could simply wrap it around the exterior of the pan and tighten it down. The remainder of the cabinet is from Home Depot and Rona and basically consists of wood strapping, butt hinges, a handle, weather stripping, Styrofoam, and some caulk. I had some one professionally make the outer cladding and it saved me all kinds of headache and it cost about 150$. It's shiny.

Finally, I got some half-sized sheet pans from JFS Restaurant. Each was about 10$. HomeSense provided a SilPat, which is designed to fit a half-sized sheet pan and not your random cookie sheet. I also got some linen from Fabricland to make couches for rising.

[Picture of a Half-sized Sheet Pan with SilPat]

Constructing the Electronics

I basically soldered everything to the board and it worked...after much pain, suffering, revision, and fried components.

The AC power worries me slightly. Although I am not scared of AC power, he combination of AC power, my wiring, a large metal box and a pan of water seems like a dangerous one. The power switch is double throw so that when the power is off, there is no power near the heater. There is also a circuit breaker to prevent total destruction. The AC ground is connected to the cabinet and not to any DC components.

[View of Controller Circuit]

I was originally going to try to use E&CE Department equipment. After about 10minutes of dealing with the lab staff, I gave up and bought a programmer over the Internet.

Programing the PIC was infuriating. Every I/O pin servers triple duty on the PIC and assigning pin usage to not create conflicts and strife was virtually impossible. The PIC data-sheet needs more footnotes that read Remember, this pin is not available if FOO is set. See page 123., because otherwise, it is very, very confusing. I definitely had trouble with power. Driving two LEDs at once using the PIC is just not possible from a current perspective. Hooray, active low. I also regret not building my circuit to do in-circuit programming.

Initially, I was going to forgo a display, but debugging was difficult and I was never sure what temperature the micro-controller calculated. So, I built a small daughter board with a 7447 BCD-to-7-segment converter with an LED numeric display. Coupled with the fault and heating lights, I had a nice little display to tell me the state of the world.

Constructing the Box

The sheet pans are of standard dimensions. This is wonderful. Pans for the home baker come in a wide array of sizes and shapes, but professional pans are all with in a few fractions of an inch tolerance. The pans should not exceed 13" by 18". The width of the cabinet is 13.5" to accommodate the pans with enough gap to make them slide easily. The depth is 20" to accommodate the pan and enough space to allow moist air to circulate around the pans. A loaf pan is about 3" tall. Normally, baker's racks have the pans spaced every 2.25" or 5". To accommodate a loaf pan, I went with 3", bearing in mind that I can remove trays to create more space. Normally, if I'm really packing things in, I'm doing pretzels or calzone which are less than 3" tall. I added 1" at the bottom to make sure that there was a gap between the bottom most pan and the hot water. The wooden box needs to be taller since the heating pan hangs out of the bottom.

To actually build it, I used the Engineering Student Machine Shop. While working, many people would look at me with this expression as if to ask are you so weak you can't work that piece of metal? and I would simply say It's stainless steel. and then they would say Oooooooh. The blind cut at the bottom I did using a milling machine. I could drill the runners and the small end panels on the drill press, but the large sheet had to be done by hand. Bending a box is not possible on their equipment, so I bent all but one of the corners and the shop machinist helped me set up a jig using angle iron to pound the last corner into shape.

Next was to construct the wooden frame that the metal box was to live in. It's basically a skeleton made of strapping with bolts going through the steel to hold the runners and the box itself in place. Panels of Styrofoam fill the spaces for insulation.

[Photo of Cabinet Interior]
[Photo of Cabinet Exterior]

One temperature sensor is attached to the underside of the hot water pan and the other is in a copper probe that can be inserted into bread or tucked under a loaf.

Sticking the sensor down was a challenge. I did not want to glue them since glue does not stick to stainless. I asked a welder at the UW shop and he was able to help me solder nuts to the stainless. I then made small steel plates to hold the sensors firmly against the sides of the box. I did try epoxy, but it would snap off much too easily.

[Photo of Temperature Sensor]

The heater I salvaged is a band heater and requires only 6A which is convenient since the relay has a maximum current rating of 10A. I used a solid-state relay, an opto 22, because electromechanical relays were drawing too much current and causing the micro-controller to reset. The second advantage to the solid-state relay is that I can pulse it frequently without fatiguing it or hearing that annoying clicking.


Since there is no real user control, the only thing to do is switch it on. The green light indicates that the oven is heating and the red light indicates one of the door is open or the temperatures are above limits. It also displays the two temperatures, alternating every few seconds.


There are one thing I might add: different settings for temperature and humidity controlled by some kind of button. Something simple that cycles through a bunch of presets. Since I've had much variability using other rising methods, I'm not even sure what kinds of different programmes I would want. I am out of I/O pins, which makes this impractical.

Doing It Again

If I were to do it again, I would keep the stainless steel parts roughly the same. I would add a lip with nuts so that I could create an steel case that could screw onto the lining and insert Styrofoam into the gap, eliminating all the wood.

On the electronics side, I would definitely have added an in-circuit programmer to remove the need to constantly rip the PIC out of its socket. The digital output uses a lot of I/O pins. I originally had a design which made use of a shift register, but I had a lot of problems with it. I would have tried something else along those lines. Perhaps do away with the LEDs and have a single LCD display.


This would have all ended in tears if it were not for the guidance of Joe Megyes and my father. The machinists at the Engineering Shop have been extremely helpful in providing tools, know-how, and advice. I would also like to thank Zdenko at Smaak Enterprises for making the outer case.

Sun, 6 Dec 2009 09:25:41 -0500 View History