Using hot water tap to manage steam boiler water concentration

Water analysis, treatment, and mineral recipes for optimum taste and equipment health.
User avatar
Team HB

#1: Post by homeburrero »

Most of us with HX or dual boiler machines are aware that using the steam wand removes water vapor but not minerals, and as the boiler is refilled we end up with water in the steam boiler at higher and higher mineral concentration. Depending on what minerals are in the water it may become more scale prone, and if the chloride or sulfate is significant may even become more corrosive.

Vendors typically recommend periodic draining or flushing of the steam boiler to address this. Doing a little back-of-napkin calculating about that, consider a home machine that makes two large lattes per day, pulling maybe 100 ml per day of pure water from the steam wand. In a steam boiler with 1.5 liters of water, at the end of the week you will have built up a mineral content equivalent to 2.2 liter of feed water inside the boiler, a concentration factor of about 1.5. So if your feed water had a hardness of 50 mg/l, the hardness in the boiler should have risen to 75 mg/L by the time you drain/flush it. So I think for most of us a weekly drain/flush is probably reasonable.

For machines that have a conventional hot water tap there is another way. It's easy to do and also fairly easy to calculate and adjust for your particular situation. The key is to use the hot water tap to manage your 'steam ratio'. The formula for steam ratio and boiler water concentration was provided years ago in Jim Schulman's Insanely Long Water FAQ:
Actual mineral concentrations in the boiler, if saturation isn't reached, depend on the ratio of steam to water extraction and is as follows:
Boiler Concentration = Feedwater Concentration/(1 - Steam Ratio)
Steam Ratio here is the amount of water that is removed from the boiler as pure steam divided by the total amount removed from the boiler. In the case of an HX or DB machine it would be approximately the amount that is removed via the steam wand divided by the total amount removed by the wand plus the hot water tap. (Assuming a conventional tap that simply pulls water from the steam boiler.)

Knowing that, you can use a routine where you pull a pre-calculated amount of water out of the water tap before each steaming operation. If you're like me you'll want to always draw at least enough to trigger an autofill which may sometimes be a little more than the calculated amount. The calculated amount is based on how much concentrating effect you can live with, but in practice about 2 - 3 ml per each ml of water out the steam wand is good. You get diminishing returns going higher than that (see table below).

You will need to roughly estimate what your steamed water use is for your usual milk drinks. It will vary depending on the starting and ending milk temperatures, and the steam boiler temperature so I think it's worth doing a little measurement on your own machine. First get an idea how many seconds of steaming is required for your drink. Then get your pitcher nearly full of ice water, and tare that on a scale. Purge your steam wand into the pitcher, and then with the wand sunk deep in the water steam the ice water for the same number of seconds as for your drink. Follow that with a blast into the pitcher that simulates your wand cleaning routine. Then measure the water gain, which is your estimate of the steamed water use for that drink. Multiply that by 2 or 3 to get the amount of water you should draw from the tap before making this milk drink in order to keep your concentration factor down in the 1.3 - 1.5 range.

Sometimes an example is clearest:

I typically steam 120 ml of milk for a double small cappuccino, sometimes 240 ml for two split single caps. So I estimated my steam use for that 120 ml:
  1. I timed how long I typically needed to steam the 120 ml, which was 30 - 40 seconds. (I use a slow steam tip.)
  2. I filled my tallest pitcher within 1" of top with ice and water and tared that on my scale.
  3. I purged my wand into the pitcher, then with wand deep gave it full steaming for 35 seconds. Then gave it another few seconds of steam to simulate cleaning the wand.
  4. I weighed the pitcher and noted that it gained 20 g.
Then looking at the chart, I see that hot water tap use of 2x - 3x will give me a 1.3 - 1.5 concentration factor, which I'm fine with. So I make a habit of pulling 40 - 60 ml from the tap before steaming my usual 120 ml of milk, and twice that when steaming 240 ml. I always trigger an autofill when I do this, which helps makes sure I don't have an autofill kick in while steaming, which would tend to kill my steam pressure.
nínádiishʼnahgo gohwééh náshdlį́į́h
★ Helpful

User avatar
BaristaBoy E61

#2: Post by BaristaBoy E61 »

This is very intriguing and useful! I was just earlier this week measuring the weight of the steam I was adding to our pitcher but this is far more precise and insightful.

Thanks Pat for this post and for the countless others that you have posted that have helped immensely on this site over the years!
"You didn't buy an Espresso Machine - You bought a Chemistry Set!"

User avatar
Team HB

#3: Post by RapidCoffee »

Not to muddy the waters, but there's a potential complicating factor. If scale has formed in the boiler, you are replacing calcium-depleted water with fresh water containing a higher concentration of calcium. Couldn't this lead to increased scale formation? Ditto for chloride, sulfate, and other corrosive substances.
BaristaBoy E61 wrote:Thanks Pat for this post and for the countless others that you have posted that have helped immensely on this site over the years!
+1 :D

User avatar
Team HB

#4: Post by homeburrero »

RapidCoffee wrote:If scale has formed in the boiler, you are replacing calcium-depleted water with fresh water containing a higher concentration of calcium. Couldn't this lead to increased scale formation?
Good point -- this advice is best for people who use non-scaling water. If you are using scale-prone water then you expect that anything you do that increases the thruput would tend to increase the scale deposition. When doing scaling rate estimates, as per the Insane FAQ guidelines you would need to add the flushed water numbers to your thruput. This would be true even if you were to empty and refill the boiler every week or so. If you're using scale-prone water then you might postpone your need to descale by doing periodic steam boiler flushing with distilled or RO.
RapidCoffee wrote:Ditto for chloride, sulfate, and other corrosive substances.
For a situation with calcium sulfate deposition, maybe so. But for dissolved sulfate and chloride I think pulling water from the tap will always tend to reduce the concentration of those corrosives to be nearer the tap water concentration. The alkalinity will go up and down in the same proportion as chloride and sulfate*, so corrosivity may not be much of an issue with steam use.

* (edit addition) Worth mentioning that if you are at the CaCO3 saturation point, the alkalinity, along with the calcium, would hit a ceiling as the calcium and carbonate drop out into the limescale deposits.
nínádiishʼnahgo gohwééh náshdlį́į́h


#5: Post by DaveB »

Very interesting post! I never used my water spigot, even before I rerouted the line for a much more useful task. :D
Lately I've actually been increasing the mineral buildup in my steam boiler; about a month ago I started using the steam wand to preheat the cup, using 3 or 4 oz of water and bringing it close to boil (the idea is to get a hotter drink while steaming milk to an average of 137°F). So that's 2-3 lattes a day plus the preheating of cups. I use RPavlis water exclusively, and yesterday I drained my 1L boiler after 2 months and it was 470ppm. FWIW, I checked my fresh prepared RPavlis water and it's 55ppm.

I think I'll drain and check it again after a month to see what it amounts to. My question is, what's the worst that can happen with highly concentrated non scaling, non corrosive water in the boiler?
Von meinem iPhone gesendet

User avatar
Team HB

#6: Post by homeburrero »

DaveB wrote:. My question is, what's the worst that can happen with highly concentrated non scaling, non corrosive water in the boiler?
If you used the tap to make an Americano it probably would would make your drink taste flat and chalky, maybe briny. But even at your 7x concentration no minerals would precipitate and the pH would not get corrosively high.

Thanks for sharing those numbers. They tend to make the case for Dr. Pavlis' water with zero hardness, zero chlorides, sulfates, silicates, etc. He advocated it for his vintage levers (with single boilers) but when considering the steam boiler concentration issue with DB and HX machines it seems like an especially good water choice for those machines if you worry about failing to keep the steam boiler refreshed.
nínádiishʼnahgo gohwééh náshdlį́į́h

User avatar

#7: Post by hankbates »

Best thread I have seen on this surprisingly complex and very important subject.
I never considered concentration when I was using an LP lever machine, but I am always thinking of it now that I am using a HX machine.
I used rpavlis' water recommendations then, but his advice is so much more pertinent now.
Nice going, Pat! :)