This story starts with a trip to my corner hardware store to replace a toilet that was not repairable. They had just two models for sale, both made by American Standard. The first was a basic model called the Cadet Pro and the second was a somewhat more expensive model called the Champion that was billed as having a more powerful flush due to its 4 inch flush vale. (The Cadet like most modern toilets has a 3 inch flush valve). My usual instinct would have led me to the basic model which would have been the wiser choice in this case. However the Champion was on sale and because of that it was only $15 more expensive than the Cadet. A bigger flush valve is got to be a good idea, right? So I brought home the American Standard Champion toilet.

With all toilets I have used, pushing the lever lifts the flapper valve which sends the water in the tank to the bowl. When you release the lever, the flapper takes a few seconds to descend giving time for most of the water in the tank to flush the bowl. My first flush after the installation was with the lid off and I could see something different right away. As soon as I released the lever, the flapper immediately fell onto the flush valve seat shutting off the water entering the bowl. To get a complete flush I would have to hold down the lever until all the water in the tank was used. With other toilets the lever would have to be held down like that only when something in the toilet was broken or misadjusted, or perhaps all the time with some very old toilets. So I went online to see how to fix the problem. That's when I saw the reviews. Some customers didn't realize you had to hold down the lever and just complained that the toilet had a very weak flush. Most customers figured out you need to hold the handle down for a couple of seconds but didn't like that at all and were frustrated that there was no way to fix it. A contractor had written that once he advised a client about the need to hold the lever down during the flush the clients always chose some other model. A plumber had written that the only way he had found to fix the problem was to replace the whole flush valve with a standard three inch valve. (He didn't say how he accomplished that but I think it would require replacing the whole tank.) I thought replacing the flush valve was beyond my expertise so I just used the toilet as is. I did get used to it but after a few weeks I realized it annoyed me a little more every day. I really didn't want to think about how long I was holding down the lever ... too long and wasting water and time, or too short and not getting a complete flush. A solution occurred to me that perhaps the plumber who wrote in didn't think of. I tied a float onto the chain that connects to the flapper thinking the float would slow the flapper descent giving time for a complete flush. I found that the flapper closes with a fairly high force requiring a surprisingly large float to make only a marginal improvement. I thought I could continue to go to larger and larger floats, but no. At a certain point the float was so strong that when the tank was full it would pull up on the flapper strongly enough that it wouldn't create a full seal with the value seat and the tank would start leaking. Once I saw that I realized a float would never work as a solution to this problem. Perhaps that plumber tried and failed with this approach as well. So I went back to the hardware store to see if I could return the toilet, but I was one day past their 30 day return policy.

It was then my electrical engineering brain kicked in and it occurred to me that a solenoid could do the flush for me and it could be timed to hold down the lever for the optimal time. It didn't take me long to come up with this circuit. There is no electrical outlet near my toilet hence the battery powered design. I estimated that it would take a fairly high powered solenoid to pull the lever so I chose a 44 watt solenoid from the DigiKey listings. I decided to use super-capacitors to store the energy used to actuate the solenoid. This allows me control the duration of the flush by changing the capacitance and also reduces the amount of current that needs to be sourced by the voltage regulator. (The solenoid I chose draws about 8.9A when driven at 5 volts).

To see how the circuit works, first imagine the simpler circuit without the relay. The power bank charges the capacitors thru the resister. After about 20 seconds there would be enough energy stored in the capacitors to engage the solenoid when the pushbutton was pushed. The solenoid would disengage as soon as you let go of the pushbutton, and so the circuit would suffer the same problem of the original lever. This is were the relay comes in. When the pushbutton closes, at the same time current begins to flow thru the solenoid, current also flows thru the diode to power the relay. Once the relay closes, the solenoid then gets its current from the capacitors thru the relay and stays engaged until most of the energy in the capacitors is depleted.

Parts

SOLENOID: DigiKey #1144-DSOL-1351-05E-ND ($23.49) 44W pull type solenoid with 1 inch stroke: spec sheet

RELAY: DigiKey #J107F1AS125VDC.36 ($1.09) General Purpose SPST 12A 5V relay: spec sheet
                Alternative relay from Amazon:     spec sheet

CAPACITOR:   DigiKey #`-CHP5R5L405R-TWX-ND ($3.91) 4F 5.5V super capacitor, ESR 0.135Ω @DC: spec sheet
This series/parallel configuration of four 4F capacitors has a capacitance of 4F so you might be wondering why I didn't use just a single 4F capacitor. The main reason is that the supply voltage will need to be close to the 5.5V rating of the capacitor and supercapacitors tend not to last very long if held continuously near their max voltage rating. The capacitor's series resistance is .135 ohms which will result in about a 1.2V drop when the solenoid engages (8.9A). For this reason the power supply will have to be somewhat above 5 volts to provide sufficient power to the 5V solenoid. (Experimentally I determined that about 5.25 volts was sufficient although this can be adjusted upward to get a longer flush.) Another reason for the series/parallel combination is that the 8.9 amps is close to the current limit for this capacitor. (In this configuration the max current thru each capacitor is only 4.5A). It's conceivable one could find a single capacitor to replace this set of four, but it is likely to be expensive and difficult to find.

PUSH BUTTON: Any momentary pushbutton will work and I found one in one of my parts boxes that was just the size I was looking for. But if I needed to buy one I would probably choose one such as this.  

DIODE:   1N5340
The shutoff voltage for the 5V relay is fairly low so this diode (as well as the resister) is needed to make sure the relay would disengage. This is a Zener diode which I chose because it's forward voltage drop is slightly higher than it is for a rectifier. It doesn't have to be a 1N5340 however. You can just use any Zener diode you happen to have on hand.. If for some reason the relay still doesn't disengage, two diodes could be used in series which would surely be enough of a voltage drop to ensure the relay disengages.

RESISTOR:   2.7Ω 1W
This resister must have a high enough value to give a voltage drop (in conjunction with the diode) to allow the relay to disengage after the energy in the capacitors has dissipated, but low enough that the capacitors charge quickly enough for another flush after the tank has refilled. The other purpose of this resister is to serve as a fuse. If you accidently leaned a mop or something against the pushbutton holding it permanently closed, without the resister the voltage regulator could supply enough current to burn out the solenoid which is a lot more expensive to replace compared to the resistor. This 1 watt resistor would probably burn out in about 20 seconds. A 5 watt resistor would last a lot longer but that would be counter productive because you want to make sure that the resistor fails before the solenoid.

BATTERY:   LiFeP04 Battery.
Pretty much any battery with a voltage between 7V and 45V will work here. I chose this 12.8V LiFePO4 battery for it's high capacity and low cost. This one is 12AH which will last a very long time before it needs to be recharged, but you could easily use a smaller battery to lower the cost.

VOLTAGE REGULATOR:   Pololu D30V33MAL.
Since I didn't know the exact voltage needed to charge the capacitors with enough energy to engage the solenoid, this regulator was chosen for it's adjustability (1.4V to 7V) and also for it's very low quiescent current which allows the battery to last as long as possible before requiring a recharge. You can order this part here

My first idea for powering this circuit (instead of the battery & voltage regulator shown above) was to use the 5V delivered by a USB power bank. The nominal 5V USB spec would not be a high enough voltage but most power banks are set to around 5.2V (near the high end of the 4.75V - 5.25V range allowed by the USB spec). 5.2V is just high enough to allow the circuit to work. The other problem though is that most power banks shut themselves off when it doesn't detect a device plugged in that needs to be charged. The leakage current of the super capacitors is in the micro-amp range which is not high enough to be recognized by the power bank so soon after the capacitors are charged the power bank shuts itself off causing the capacitors to slowly discharge due to the leakage current. However a few power banks (such as the Voltaic Systems V50 that I purchased) have an "always on" mode that provides the 5V source even with no device connected to it. This seemed to work for a while but occasionally it would randomly switch from the "always on" mode to the normal mode and therefore it would not keep the capacitors fully charged. The voltage regulator solution shown above doesn't suffer from that problem since it never turns off and has the added benefit of an adjustable output voltage.

SWITCH:   SPDT Switch
When this switch is in it's down position, you can monitor the battery voltage on the digital volt meter. This is useful to give you an idea of when the battery should be recharged. When the switch is in it's up position, you can monitor the regulator output voltage. This is useful primarily when you are turning the voltage regulator potentiometer to set the output to the desired voltage. To turn the digital volt meter off, move this switch to it's center position. Leaving the volt meter off will allow the battery to last much longer between recharges.

VOLTMETER:   I purchased this set of 8 digital voltmeters. Even though I only needed one meter for this project, buying eight doesn't cost much more than buying one and I've found quite a few situations where a meter such as this comes in handy, so I think I will eventually use all of them.

ENCLOSURE:   Plastic project box (Amazon).
To test out my circuit I didn't use the enclosure. Instead I just secured it to the side of the toilet using packing tape. Surprisingly, the circuit worked the first time and I only needed to retape it once to adjust the timing by adding one more capacitor. There are hundreds of boxes I could have chosen on Amazon.

CONNECTORS:   Barrel Connectors>
You will notice from the pictures below that I connected the voltage regulator (mounted to the battery) to the main circuit board in the plastic project box using a USB cable. That made sense for the first iteration which was powered by a USB power bank. If I were to build it again I would connect a cable to the circuit board containing the voltage regulator with a male barrel connector (shown in this picture) on the other end. This would plug into the female panel mounted barrel connector screwed into the plastic project box. One could use this same male/female barrel connector pair to connect to the cable coming from the pushbutton. The disadvantage of that is you could accidently connect the battery and pushbutton cables the wrong way around. At least that would do no damage to the circuit and it would start working properly once you swapped the cables back to their proper positions. The the pushbutton connector, I actually used a 3.5mm audio jack/plug combination since I happened to have one laying around and I didn't have any barrel connectors on hand. A 2.5mm audio jack would also be a good choice.

Construction

I took the plunger out of the solenoid to attach this braided wire to one end. (There is nothing holding the plunger into the solenoid.) I also added a thin layer of felt to the pointed end. Without the felt, the solenoid makes a sharp clacking sound as it engages. You may find the sound satisfying, although with the felt you will barely hear the solenoid engage. As you will see in the next picture, I attached a screw to the other end of the braided wire. Then on the other end of the screw I attached a solid copper wire which connects to the toilet flush level. I used the screw for this purpose because it allowed me to easily adjust the tension on the wire.

I constructed the circuit on a small piece of vector board. This picture shows the solenoid in its fully extended position (i.e. when there is no current in the solenoid coil). Then tension in the wire connected to the flush lever keeps the plunger in this position (touching the plate) until the solenoid is engaged. The black cable going to the right edge of the picture goes to a connector leading to the pushbutton and the fat silver cable (which has a male USB connector on the end) connects to the power source. Actually this photo doesn't show the current design (with the four 4F capacitors in series/parallel) but rather my initial design which used four 1F capacitors in series. So the total capacitance is the same and the circuit worked as well as my current design. The reason I abandoned that design was because I was driving the circuit with about 5.2V but the capacitors were only rated for 5V maximum. The lifespan of a supercapacitor when it is usually left charged to it's maximum voltage rating is already somewhat marginal and any higher than that will shorten its lifespan even more. (I wanted a circuit that would work for several decades.) I would have replaced the 5V capacitors with ones rated at 7 or 8 volts or more, but I didn't find any that were readily available. Actually this photo doesn't show the current design (with the four 4F capacitors in series/parallel) but rather my initial design which used four 1F capacitors in series. So the total capacitance is the same and the circuit worked as well as my current design. The reason I abandoned that design was because I was driving the circuit with about 5.2V but the capacitors were only rated for 5V maximum. The lifespan of a supercapacitor when it is usually left charged to it's maximum voltage rating is already somewhat marginal and any higher than that will shorten its lifespan even more. (I wanted a circuit that would work for several decades.) I would have replaced the 5V capacitors with ones rated at 7 or 8 volts or more, but I didn't find any that were readily available.

The previous picture (above) shows the position of the solenoid when it is not energized. The weight of the water holds the flapper valve down which holds the solenoid in this position thru the tension in the wire connecting the screw to the flapper valve. This picture however shows the position of the solenoid when it is fully energized. The screw is now about 1 inch closer to the circuit board which is enough to fully depress the flush lever. The metal plate is needed to prevent the wire tension from pulling the plunger completely out of the solenoid. Conceivably you could use the wall of the enclosure for this function but then you would need to be precise about where the circuit board was mounted.

Click on any picture on this page to see a larger image.
(Then hit the back button to return to this page.)


This shows the circuit board mounted inside the enclosure. (Actually I just epoxied it in.) The connector on the side of the enclosure is for a cable going to the pushbutton. I originally put that connector on the smaller side of the enclosure (near the right edge of this picture) but that was when my plan was to glue the enclosure to the side of the toilet. But soon I realized that was difficult and I decided to mount it on the floor next to the toilet. This meant I had to relocate the connector. I plugged that hole and covered it with epoxy. I used a file to carve a small notch in the enclosure so the silver power cable would have room to sneak thru.


After screwing the cover on, the enclosure is ready to be mounted to the floor.

(This picture shows the enclosure with the solenoid in it's fully engaged position.)

Here are two views of the battery with the voltage regulator and digital volt meter mounted on top. When the switch is pointing away from the USB connector (as in the picture on the left) the meter reads the voltage of the regulator output. When the switch is pointing toward the USB connector (as in the picture on right), the volt meter displays the battery voltage. With the switch in the middle position, the volt meter is turned off. As I mentioned before, if I were to build this again I wouldn't have a USB connector here for the power source. I would simply connect one end of a 2-conductor cable to the circuit board and put a male barrel connector on the other end of the cable. This would then plug into a female panel mount barrel connector on the side of the plastic project enclosure.


I attached the enclosure to the floor directly onto the floor so that the solenoid was aligned under the tip of the flush lever using exterior quality Velcro rated at 15 pounds. It's hard to see from this picture but the wire from the flush level is attached to the screw using a crimp terminal. The tension on the wire can be adjusted by screwing the nuts towards or away from the screw head. If you have too much tension, the flapper valve won't go down far enough to seal the flush valve and the water will continue to run without filling the tank. So back off the tension to allow the flush valve to close and then back off a little more allowing a small amount of wiggle on the flush lever. The two conductor white cable is plugged into the connector on the side of the enclosure and runs behind the toilet to the right side of the toilet where the pushbutton is glued to the side. (See next picture). The battery with the attached voltage regulator is not secured to the floor or the enclosure. It just sits there on the floor making it easy to remove when it needs to be recharged. The battery I have is plenty heavy enough that it isn't going to be wandering around, so it stays put wherever I place it. While you are recharging the battery, you can still use the toilet by pushing on the flush lever the old fashion way.

The button can be attached to the toilet in the most convenient spot for your preference. I liked this spot on the right side just below the tank lid. I attached the button to the toilet with a small dab of epoxy.

I drilled a small hole on the underside of the flush lever and threaded the bare copper wire thru that hole and twisted the wire to securely attach it to the lever.



Success:
Click on the picture to see the video showing
what happens when the pushbutton is pressed.

Now that I have completed this design, I do like this toilet. It is satisfying that a feather light touch on the pushbutton completes the perfect flush every time. And unlike most toilets, I never have to worry about how much toilet paper I'm putting into the toilet. I haven't seen a clog yet.

That said, unless you like to tinker (like me), you should probably choose some other model the next time you are in the market for a new toilet.

After I finished testing this design I thought of this alternative configuration for the relay, this time using both the normally open and normally closed terminals. The advantage of this configuration is that the power from the voltage regulator is not connected to the solenoid while it is activated. This means you could lower the value of the resistor to allow faster charging of the capacitors without worrying about whether the capacitors would discharge enough for the relay to disengage. You wouldn't want to eliminate the resistor altogether since it is still useful as a fuse to guard against the mop accidentally leaning against the pushbutton scenario. Because the voltage regulator is no longer contributing to the energy used to keep the solenoid engaged, the capacitor would probably have to be increased from the 4F up to perhaps around 5F. I didn't bother testing out this circuit since my other design was working so well.

I've been using the free version of the Claude app lately, so I thought it would be interesting to ask Claude to design this circuit for me. I didn't give it any hints other than I estimated that it would take a solenoid with a one inch travel and a power of 40-44W to be strong enough to pull the lever. It came up with this design which actually is pretty similar to what one of my friends suggested when I showed him my design.



U1 LM555 timer Runs fine directly off 12V
R1 2.2MΩ + 200kΩ series trim pot trim pot lets you dial in exactly 2.7s
C1 1μF film capacitor Timing capacitor (low-leakage/tantalum improves accuracy but isn't essential)
R2 10kΩ Pull-up on trigger pin
R3 220Ω Gate resistor (limits inrush into MOSFET gate)
R4 10kΩ Gate pull-down (keeps MOSFET off at power-up before 555 settles)
Q1 IRLZ44N or IRF540N (TO-220) Logic-level N-channel MOSFET, handles the solenoid current easily
D1 1N5404 or similar (3A+) Flyback diode — critical, across the solenoid, cathode to +12V
C2 10nF Control voltage bypass (pin 5), standard practice
SW1 Momentary N.O. pushbutton Your trigger button


I think Claude did a respectable design and the idea is similar to what one of my friends suggested when I showed him my design. Claude's design is easier to adjust the length of the flush and would be cheaper to build. The MOSFET is inexpensive and doesn't require the voltage regulator in my design which is the most expensive component. Perhaps my design has some advantage over this one, but if so, I can't think of what it would be :) I was tempted to clean up Caude's schematic, but I decided to leave it as it was presented to me so you get a feel for what Claude is capable of doing. I also made no edits to the logic flow diagram on the left. Claude generated the parts list (including the comments) pretty much as you see it here, except that in this case I did edit it somewhat to make it into a legible table.