I have a tankless water heater that is just a little bit to small for our new house - the temperature rise isn't quite enough for the winter months. Our basement is big and relatively warm so I struck on the idea of using a prewarmer, either in the form of a radiator (think automotive radiator, say off a F150 size truck) plumbed in before the water heater or a 100 feet of pex coiled in a 55 gallon drum filled with water.
How should I proceed?
The typical way to pre-warm water is to install a standard tank water heater before the tankless, but leave it turned off.
Using an automotive radiator is a terrible idea. You don't want to connect something not designed for potable water to your plumbing. That's just asking for trouble.
The pex idea is interesting. The only reason that I can think that this wouldn't work is that 100 feet of pex introduces quite a pressure loss. Put a meter on the end to make sure you have enough pressure going in to your water heater. If you have the stuff to do that, you can try it out. But I wouldn't buy 100 feet of pex just to do it.
Using a tank water heater has a couple of advantages over the pex-in-a-bucket-of-water idea:
You could add a smaller tank-less heater in front.
Some of them are even rated to be plugged into an outlet so you won't have to do additional electrical work.
In Europe I've seen it done with a bigger unit to supply the whole house and a smaller one at the shower for use during the winter.
In either case you're going to get a lot further with this solution as far as performance and you still get the energy benefits of a tank-less system.
Do NOT use an automotive radiator - it's not meant for nor suitable for potable water use.
If you have a "relatively warm basement" you can either use a plain, uninsulated pressure tank (a "tempering" tank in this application - cold in the bottom, warmed out the top) or run a long run of large-ish diameter PEX (to minimize pressure drop) around the basement ceiling to collect heat. I see little advantage and several possible downsides (the primary one being potential rust-out or "plastic fatigue" depending on drum material) to "a coil of pipe in a 55 gallon drum of water"
There are a few problems with your plan.
First off. If you're going to make a heat exchanger, you should use copper instead of PEX. Copper has a thermal conductivity of 401 watts/meter kelvin (W/mK), while PEX is closer to 0.51.
The next problem, is the volume of water you're preheating. 1" PEX has an inside diameter (I.D.) of just 0.865". Which means 100' of PEX, can only hold 3.0527478 gallons of water.
Volume = pi * r^2 * L
V = pi * 0.4325"^2 * 1200 in.
V = pi * 0.18705625 sq. in. * 1200 in.
V = pi * 224.4675 cu. in.
V = 705.18544896966691375461340363629 cu. in.
One cubic inch can hold 0.004329 gallons of water.
X = 705.18544896966691375461340363629 cu. in. * 0.004329
X = 3.0527478085896880696437214243415 gallons
In the US, shower heads are limited to a maximum of 2.5 gallon per minute flow rate. Which means if you're taking a shower, the preheated water will be used up in just over a minute. Even if you were to use 100 ' of 1" Type M copper tube, you'd only have 4.5411268397 gallons.
V = pi * 0.5275 in.^2 * 1200 in.
V = pi * 0.27825625 sq. in. * 1200 in.
V = pi * 333.9075 cu. in.
V = 1049.0013489785338857719650955024 cu. in.
X = 1049.0013489785338857719650955024 cu. in. * 0.004329
X = 4.5411268397280731915068368984299 gallons
That would almost give you a two minute shower. Though even with the higher thermal conductivity of copper, I doubt the incoming water would be able to pick up enough heat to make a difference.
This is why using a 40 or 50 gallon holding tank is preferable. At 2.5 gallons per minute, you could take a 16 minute shower with just the water in the 40 gallon tank.
Modern hot water tanks are designed to retain heat, and are well insulated to do so. If you're using an off the shelf hot water tank as your holding tank, this insulation is going to work against you. You'll want to remove as much of the insulation as you can.
To prevent contamination into the water supply, you're going to want to install a check valve on the supply line feeding the holding tank. Because heating water increases its volume, you're going to have to install an expansion tank between the holding tank and the heater. Finally. Standing water is a breeding ground for bacteria, so you'll want to make sure you heat the water to at least 140°F (60°C).
Did you insulate your pipes beyond the water heater so that they are not losing the heat rise per foot? The hotter they stay before appliance the less demand for rise on the heater. Make sure you tape all the seams.
Also do a tune on your heater -- often with both on boil and forced style heaters the efficiency and heat rise are below spec. I've seen claimed 95& with X to Y rise that were much lower out of the factory due to fuel quality, atmosphere/elevation, etc. Tuning them up you can get to about 98% and sneak out more heat rise.
A tank or radiator is not a good idea. Tank insulates itself preventing circ or convection and a radiator is delicate and will get clogged quickly. Pex is not a good idea either (since its not a good heat conductor). Instead get baseboard heating fins by the foot. You don't need the covers, just the fins. They are normally meant for copper pipe but you can slide whatever through them. You could cover your whole cold supply line going to the heater with the fins.
Another thought, if its not high efficiency vent, you can coil copper tube around the exhaust in order to leech some heat from the flue. Most on-demand heaters dont have the old (metal/single wall) exhaust style, but its worth a look. There is probably some massive warranty void while doing this though.
I guess another thought would be to leverage the bottom of things like the tub (or even the output hot line itself), assuming your supply is close enough (or touching), it could pre-heat the supply. Seems weird but in concept it could work -- preheating itself by heat that is normally leeched into the air.
Whatever sort of water heater you have, make sure to maintain its efficiency: If you have hard water, install a scale inhibition system (small water softener that resembles a filter), adding new pellets every 6 months and flush the system with vinegar or similar acid once a year. If it has a tank, replace the sacrificial anode every four years or so to prevent leaks.
Tim I've been in the industry for 35 years and to answer your question directly. Undersized tankless heaters are very common. Even professionals don't know your personal habits and needs. To increase output you should definitely put a tempering tank before your tankless. If passive heat isn't sufficient then a small 120v electric water heater on a timer is a fantastic solution. I recommend setting the 120v at 94deg F. Just below the ideal breeding temp for legionella.
You can box in the electric water heater, and run dryer exhaust in there and then out the box. to the outside.
otherwise, if the dryer vent is not powered, a 12 volt CPU fan to circulate basement heat down there.
open up the water heater to remove insulation.. because you WANT ZERO insulation. You want heat from the room to radiate into the tank.
you will have 50+ gallons of water warmed up to 70+ degrees overnight... and only have to heat from there. Your tankless will have no problem.
If you dont have a tank, you could just have incoming cold water pipe coiled through a tub of water/ This serves two purposes.. warms up incoming water... acts as emergency water supply.
A tankless water heater should always be combined with a counterflow heat exchanger, especially all local tankless heaters. Can save 20 to 70% of energy, in theory next to 100%. Many instructions for DIY are available, e.g.
Easy to build, not expensive. But some local codes require to have at least 2 separating layers between fresh and drain water, which is not the case with those tube-in- tube solutions.
Another tube-in- tube solution would be to run a (coiled) copper tube inside an enlarged long siphon.
The tankless heater must be specified for higher input temperatures (> 25 degree Celsius), which is not always the case if the power electronic's cooling (triacs, IGBTs etc.) relies on cold input water.
The shower head resp. water nozzle should have reduced throughput. Closing 30 to 70% of the small holes with silicon (from inside) helps.
The comfort is normally nearly unchanged - and it saves water as well.
But the resulting higher dynamic pressure must not be raised over the maximal specified output pressure of the heater.
