It's a 5 to 10 gal. RIMS using a rectangular picnic cooler for the tun, a slotted manifold, sight gauge and a unique controller and return manifold.
A Drawing and Photo:

Drawing of the RIMS - 16k

Photo of ugly Rims - 43k

The Tun & Fittings
I used a ~50 qt. Rubbermaid cooler which has worked well-  very good insulation and, after 20+ mashes, there's no warping.  To make it into a tun, after the drain valve was removed, a 1/2 copper coupling fit nicely (with a bit of keg lube) through the rubber grommet that remained in the hole in the cooler.  On the inboard side, a ~1" length of 5/8" ID vinyl tubing fits over the coupling and is secured to it with a hose clamp.  The 1/2" copper tube from the manifold slides thru the vinyl tube and into the coupling.  Well, before fully inserting it, note on the drawing that there's a hole in the tube from the manifold.  It's for releasing trapped air from the manifold after the foundation water is added.  After air is bled-off (shaking the rig helps), the tube from the manifold is fully inserted into the exit fitting.   .On the outboard end there's another ~1" length of  5/8" ID vinyl tubing- much like the inboard end except another hose clamp is installed further outboard.  This acts as coupling for the modified tee fitting at that end.   Both of the above lengths of vinyl tubing are install tight against either side of the grommet- this prevents the coupling from moving when the manifold is remove or installed.

Photo - Outboard end of Tun Plumbing
The inboard end is shown in the photo below of the interior of the tun.

The Pump
Photo - Pump Installation
The pump is a March MDX-3.  It a plastic impeller housing and 1/2" OD barbed suction and discharge fittings.  I initially ran it at full speed and used the ball valve downstream of it to set the recirculation flow.  For the last year, I've used a ceiling fan speed controller in the 120VAC power supply to the pump to vary its speed and hence the flow.  Not all fan controllers work with March pumps- I tried a couple before I found the Lutron FS-5F which works.  This is better than twiddling the ball valve since the controller allows for finer adjustment of the flow.  I changed from a discharge throttling valve to a motor speed controller since it'd been posted in the HB Digest that pumps can damage enzymes.   I've not noticed any changes (e.g. faster conversion or changed f.g.) since the change, so the controller hasn't been that much of an improvement.

A drain was added to the bottom of the  impeller housing to allow for complete drainage of the housing.  This is needed since the rig is located in an unheated garage.  It's a 6-32 thumbscrew with a O-ring that screws into a drilled and tapped hole in the impeller housing

Short pieces  of 1/2" ID vinyl tubing with hose clamps marry the suction and discharge fittings to the adjacent 1/2" copper plumbing inside and outside the tun.  They are secured with hose clamps.   Fitting the 1/2" ID vinyl tubing over the 5/8" ends of the copper tubing is easy if the vinyl and copper tubing are first heated in hot water.  The vinyl tubing functions as unions (permits easy assembly and disassembly) and also provides a bit of flexibility so that thermal expansion of the copper plumbing or bumping the plumbing doesn't over stress the pump's plastic suction and discharge fittings.   (Thanks to Rick Calley for the tip!).

All rigid tubing/piping and most of the fittings exposed to wort are copper.  The fittings used to mount the thermistors and thermometer and the 1/2" valves on either side of the pump are brass.  A past Home Brew Digest thread indicated that many brass valves have lead in the alloy to enhance the machineability of the brass. At the temperatures and acidity during mashing, it's surmised that some of this lead leeches into the wort.  Brewer/metallurist John Palmer posted that a soak in a peroxide solution will remove the surface lead.  Along the same line, only no-lead solder was used.

The Sight Gauge
I use a sight gauge to assess how much pressure is available where it attaches to the plumbing or, looked at another way, how much suction there is across the grain bed and manifold.  It helps avoid a stuck mash and pump cavitation.  To me, it's invaluable!   With no flow, the level of liquid in the gauge is the same as the level of wort in the tun (duh!). When the pumping, the liquid level will drop by an amount equal to the friction loss of the wort traveling through the grain bed, the manifold and plumbing to the point where the sight gauge attaches.   If there is alot of friction loss, the liquid in the tube will be sucked from the gauge and into the pump followed shortly thereafter by the pump attempting to suck in air-  not a good thing! Always start the pump at a low or no flow or and work up from there. I try to maintain a pump speed which gives a level in the gauge about 2-3 inches or so below the bottom of the tun but it goes lower with a sticky mash.

The Exit Manifold

The manifold at the bottom of the tun is a very important part of the RIMS since low recirc. flow is the bane of RIMSs.  Although the grain bed is more important, alot of pressure loss occurs at the bed/manifold interface.

Photo - Bottom of manifold

Photo - Manifold Installed
I much prefer a manifold over false bottom.  Based on playing around with various falsebottosm and manifold designs, they seem to allow for much higher recirc. flow.  The current manifold is made of 1/2" copper tubing.  It has LOTS of hack sawn slots!  To easy the tedium of sawing them and to make them uniform, a jig was used.  The Manifold page explains it.

The inboard tees of the manifold were hacked to serve as crosses.  Here's a photo
Tee hack into a cross
They were made by drilling a 1/2" hole thru the back side of the tee via the bull outlet then filing and grinding the hole and the pipe-end stop on the bull end of the tee out to 5/8" so the traverse tubing could pass through it.   The tee outlets along the run were left long since I'd planned on not soldering those joints to allow for disassembly for cleaning.  Assembled, it had to be handled too carefully and the mash had to be stirred carefully also so I abandoned that approach and soldered them.

In the photo above, there's a piece of vinyl tubing with a hose clamp installed where the pickup tube from the manifold exits the tun.  The thing serves several functions: to allow for removal of the manifold, to prevent the pickup tube from dislodging the copper coupling it fits into, to seal the joint air-tight  and to provide a air vent.  The air vent is important since the pickup tube forms an air trap when the tun is filled with foundation water and the resulting air "bubble" will greatly impede flow to the pump.  After foundation water is placed in the tun, the pickup tube is not fully inserted in the coupling thereby exposing a 1/8" or so hole in the  pickup tube which provides the air vent.  Once the air has been dislodged (shaking the tun helps), the pickup tube is fully seated into the coupling and the vinyl tubing then covers the air vent hole and forms an air-tight seal.  An air-tight seal is also needed to allow for complete drainage of the tun.

The Return Manifold
Return Manifold- business endReturn Manifold- top end
The return manifold returns wort to the top of the grain bed in a gentle manner. To accommodate differing grain bed heights, the level of the manifold is adjustable via the screw thing shown in the second photo above.  The business end of the manifold has a cap at each end and a tee in the center that are secured via hose clamps and 6-8 slots sawn in the fittings- tightening the hose clamps compresses the fitting (due to the slots) and hence it grabs the tubing.  This allows for easy assembly/diassembly for construction and cleaning although I've seldom had to clean the manifold since not much makes it thru the small slots manifold at the bottom of the tun and what does readily clears the 7/64" holes in the return manifold.

The Heater
The heater chamber is made along typical RIMS line using 1.5" diameter copper pipe and fittings except for the fitting the heating element screw into- it's the end female end of a 1" bronze hose fitting.  It has straight threads which makes threading in the heating element easy.  Details are are shown in the photo below and the drawing above.
Photo - Top of Heater
The heating element is 240 VAC, 4500 W low watt density hot water heater element.  Whatever you use has gotta be of the low watt density type to avoid scorching of her wort.  A gasket that comes with the heater fits between the element and the end of the chamber- it's this that makes the joint leak tight and not the threads.  The photo above shows the PVC elbow over the heater terminals which insulates it.

The Temperature Probes
They are made from thermistors stuffed into and sealed in 1/4" copper tubes.  One is shown installed in the drawing above.  Details on the probes and how to make them are on the Thermistors for Brewery Temperature Measurement page.
The thermistor as well as the thermometer fittings are modified 1/4" compression couplings.  Basically, drill out the fitting to 1/4", turn down one end until it's straight/smooth and solder it into a hole in the plumbing.   Nylon bushings are used in the fittings.  The electronic thermometer has a 1/8" stem, so a bushing of 1/8" silicon tubing is placed over the stem before inserting in the compression fitting.

The Controller
Block Diagram of Controller

This is the part of RIMS design I enjoyed the most! Major features of the controller are a 2 line x 16 character LCD and 4 keys for an operator interface, solid state relays made from discrete devices, a real time clock, thermistors as temperature sensors, a piezoelectric sounder for alarms and additional feedback to the operator and finally an optional serial link to a PC which functions as a data logger. FWIW, others have built the controller with good results.

The controller is my own design. A schematic appears below. It's brain is a $49 Basic Stamp II made by the fantastic folks at Parallax Inc. This gizmo is  computer on a 24 pin IC.  It's connected it to a PC via a serial port and programed and debugged in Basic. Once programmed, it can be disconnected from the PC. It has 16 i/o lines can be configured via the programming to do all sorts of things like serial i/o for talking to other another Stamp or PC, reading resistances, counting pulses or measuring their width, generating DTMF for dialing phones, controlling X-10 wireless 120VAC control modules, pulse width modulation, etc. It's an amazing device!!!
Controller Schematic

More Controller Info

How the Controller Works
The controller controls both the RIMS and the hot liquor tank temperatures. It monitors (via thermistors) the temperature of the recirculating wort 1) where it exits the RIMS tun and 2) on the discharge side of the heater. It then controls the heater based on these two temperatures. The heater is turned on ONLY when the temperature at 1) is less than the set point AND the temperature at 2) is less than the set point +2 degF (this dT can be changed via changing the programming). The high limit cutout is needed (at least IMHO) so that the recirc isn't overheated if the flow rate is low. Additionally, since both the RIMS and the hot water heater tank are plugged into the same 120 VAC circuit, the controller programming ensures both are not on at the same time and trip the circuit breaker. Preference is given to the RIMS- when it's heater is on the one in the hot water tanks is turned off. Since I usually heat the water in the hot water tank before starting a mash and the hot water tank is well insulated, this is not a problem. Even on occasions when I've forgotten to preheat the water, tap water was brought up to sparge temp. during the mash.

The display shows the temperature at both RIMS thermistors, the RIMS set point, the elapsed time and the on/off status of both of the heaters. The temperature of the water in the hot water tank and it's set point can be displayed via pressing one of the keys. The set points for both the RIMS and hot water tank can be changed at any time via the key pad. When the set point is changed, the controller asks if the elapsed time should be reset (that's what the Dallas DS1302 real time clock chip is for). Until I incorporated a timer, I often forgot to keep track of rest times. This function is also available separately via the key pad is is handy for timing sparges as such. The piezo element doesn't do much at this point- it just emits soft beeps indicating when the heaters are on- different beeps for the two heaters. Since the STAMP is programmable, it's possible to program the entire mash schedule and let the controller take it from there. As the textbooks say, this is left as an exercise for the student :-).

The controller also includes an optional serial transmit only link to a PC with the PC acting solely as a data logger. I've written a simple little program for the PC that captures the data stream and writes it to a file for later analysis (holler if you want a copy). Some of the resulting time/temp graphs appear below. The data was imported into 123, distilled to 1 sample/sec. from the 2-3 samples/second in the raw data and graphed. Both mashes were done with the heater programmed to turn off with a wort temp. > 2 degF above the set point and a flow of about 0.5 GPM. You'll note that the temperature at the heater discharge continues to rise for another 2-3 seconds after the heater power is killed before peaking at 5 degF or so above the set point then pretty dropping rapidly. These peaks are much higher with lower wort flows- another good reason for making the best false bottom you can. OTOH, one could use a proportional or PID control algorithm rather than the simple "bang-bang" one I used...

For those of you who want a simpler electronic thermometer or controller, Ken Schwartz has written a very nice page here.

Controller Programming (Stamp2 BASIC source code)

Here's the main piece of the Stamp programming.

This is the other, supporting piece of the programming which loads memory with lookup values the Stamp uses to calculate temperature from the thermistor readings.
UPDATE  ______  UPDATE  ______ UPDATE
The controller info above is a bit dated and needs revision, so, until then, a brief description of some of the modifications:
Changed the controller schematic and programming so that I/O pin 15 is free (it had been used for sending mash data to a connected PC via a 2 wire serial connection). The new serial link is 2 way and uses the programming port (physical pins 1-4 on the Stamp). I've dispensed with the Maxim RS232 driver with great results- communications are excellent at 1200 baud over 100' of cheap flat type telephone cable.
I now use one of Scott Edwards serial LCD "BackPacks" instead of the parallel interface on detailed on this page.   This frees up several I/O pins (which I've not used yet...) and, more importantly, it greatly simplifies programming and resulted in smaller a program, so I...
Added a function to the controller program which allows the dT to be changed "dynamically" via the keys on the controller.  (The dT is the difference between the set point and the temp. at the thermistor located downstream of the heater- if the programmed dT is exceeded, the controller cuts the heater off).   Kinda useless for me tho' since the old 2 degF dT gives good results with no worries of scorching or denaturing the enzymes.
Added center off, DPDT switches to provide Hand-Off-Auto control for the heating elements.
Added a controller function which allows the RIMS heater to be turned on via some keystrokes thereby passing the automatic control.  The program makes the beeper emit a terrible noise so I won't forget!  

Commissioning and Using a RIMS
Here's info on commissioning a RIMS

Lessons Learned

False Bottoms   A phil's phase bottom (which phloats!) did not work in the first RIMS I built- it caused way too much flow restriction. Dion (web site) uses one tho'.   I think manifolds work better.  I ran an experiment with a previous incarnation of my system to determine where most of the friction loss on the suction side of the pump takes place. I had a sight gauge attached to the piping between the tun and pump and to the tun just above a false bottom and ran a mash noting the difference in levels (i.e. pressure) between that gauge and the one in the pump suction piping. Most of the loss was in the grain bed- 8" there and another 2" in the false bottom and piping (all 1/2" copper) to the point where the lower sight gauge attaches. Almost all of the later loss is in the false bottom/grain bed interface since the flow was only about ~1/2 GPM. A similar mash with a manifold had only a 1/2" pressure drop in the manifold and piping and a higher flow rate (about 3/4 GPM).

Stoppers   I used rubber stoppers for mounting the thermistors in the plumbing on an old RIMS. Do not try to fill the system without the one in the suction piping inserted or operate the pump without the other one inserted (DUH!). Yeah, I've done each of these bone headed things!

Heaters  Do NOT operate the heaters without fluid (duh!). If you use alot of sparge water, it's quite easy to do with the heater in the HLT. Consider putting in a level switch affair which'll cutout the heater at a low water level. Do NOT operate the RIMS heater without flow through the heater chamber for it'll rather promptly scorch and perhaps even boil the wort in the heater chamber. If you are the type that's sometimes forgetful, consider electronically interlocking the heater with the pump.

Cleaning   Run at least a batch of hot water through the system as a flush immediately after mashing. I usually flush twice, once with tap water then recirc with a gallon or so of 168 degF water for ~5 minutes. Drain well while the system is still hot and leave all the valves open to allow a bit of air to circulate an dry the innards. When system cools, close the valves and otherwise seal up the system. Before using, I recirc. a gallon or so of warm tap water for ~5 minutes and sometimes bring it up to 168 degF or so.  This has served well - I've not had any deposits on the heater element other than a very thin mineral looking layer which gives the element surface a whitish appearance- much like water has dried on it (which it has).

Future Improvements

I'd like ot someday add some hard and software to automatically control the level of water in the tun during sparging and to automatically control the RIMS pump speed based on the level of wort in the sight gauge.  A flow meter would be nice too.

Use an old PC for the "brains". The system will still a the Stamp, but only as a interface to the RIMS.  I like this 'cause I could keep my brewing notes/log on the PC rather than on scraps of paper.   At least that's the theory!

There's an old saw about standing on the shoulders of others who've gone before you... here's some of the very kind fellow homebrewers who've shouldered the previous work in RIMS design. I'm deeply indebted to them! THANKS guys!

3/03- First posted

Comments, Questions, etc...
If you've questions or suggestions I'd really like to hear them! Please email me

Legal Mumbo-Jumbo:
Use any and all of the above for your own use for FREE! Use the stuff and make money with it and I want some of the $$$! Use any and all of the above at your own risk. It works for me but may not for you, i.e. YMMV.