• You are viewing the forum as a Guest, please login (you can use your Facebook, Twitter, Google or Microsoft account to login) or register using this link: Log in or Sign Up

DIY Inline Heater - Trapped Air

somail

New Member
Joined
13 Oct 2020
Messages
18
Location
USA
Hi,

I have been building my own filter system that has separate enclosed sumps which then feed into an inline heater and then the pump. I am beginning to test the system and I am having a problem in which the inline heater is not fully filling with water. The inline heater design is based on the typical designs I have seen online and no one mentions an issue with trapped air. The waterflow is top to bottom, which from what I can tell is standard with this type of design. Since the tube is not fully filling with water a lot of air is getting pushed into my pump causing it to cavitate.

Can some please look at my design (pictures attached) and let me know if you see an error in my build or if there are any hints/tricks to get the tube to fully fill. I am at a loss on how to accomplish this.

Thanks,

Somail
1.jpg
2.jpg
 
Looks like a really nice setup, very tidy!

The space for the water to flow seems quite restricted, which I guess is by design, to ensure faster heating? Where is the air becoming trapped, at the top of the heater tube, or is the flow into the tube so fast that it cavitates on entry?

Can you prefill the heater with water before assembly or turning on the pump,. This may help to ensure the cavity is full of water. Or run the pump very slowly when first turned on, to get the heater cavity to fill gently. This may also help you observe where the problem may be arriving i.e as the top of the cavity floods, air gets trapped.

Might also be worth trying to play with flows, the heater just might be causing too much restriction, starving the system.
 
Last edited:
Assuming you mean the tube that houses the heater. Looking at the pictures I would suggest the issue is caused by the verticle section of pipe above the top elbow that is trapping the air. In plumbing terms it would be called a dead-leg and would always collect any air bubbles. You would need some form of bleed valve to allow the trapped air to be removed. You may find the water flow could be able to pull the air through if left running long enough, similar to a canister filter expelling air.
 
I am beginning to test the system and I am having a problem in which the inline heater is not fully filling with water.

Since the tube is not fully filling with water a lot of air is getting pushed into my pump causing it to cavitate.
Assuming your pump is "pulling" water through the filter, sounds like your pump is too powerful for your system and/or your filtration is causing too much flow reduction. The pull of your pump and associated "reduced flow" through the filter media is causing reduced water pressure and air to come out of solution and collect in the heater. This is quite often seen in the marine keeping world with their complex filtration systems, air collecting due to running pumps too fast.

As a test try running the system with no filter media in the 4 containers, so as to not impede flow, and I bet you will find no air collects in the heater.

If this is so, solutions are:
- Reduce the pump flow, smaller pump or variable speed pump. Marine boys use variable speed pumps to prevent this type of situation.
- Reduce the amount of filter media impeding the water flow.
- Increase the size of filter media containers so as to impede flow less.
- Place the pump at the inlet so water is being pushed through the filters & heater. More frequent cleaning of pump will be required as pumping dirty water, so quick release connections are essential.
- Place the pump after the first filtration container so only pulling water through one filter medium.
- Some combination of all of the above.

Hope this helps.:thumbup:
 
Can you see the trapped air within the heater? Or are you inferring this from the fact that the pump burps every now and then as a load of air hits the the impellor?

This is a variation on Ian Ms post, but another possibility is that the heater seal is acting like a venturi under negative pressure. I am assuming that flow through this is high, and that the heater (and filters) are acting like a restriction to that flow.

As water flows through a restriction there will be an increase in velocity, which creates negative pressure proportional to the velocity.

The seal will have a barrier pressure within which it can resist leaks. If the negative pressure created by being on the suction side of the pump (with flow restrictions upstream) exceeds the barrier pressure exerted by the seal it will pull air across the seal, like a venturi.

Is there enough space to syringe a small bit of water around the sealing face where seal meets the heater? You only need enough to see a little rim around the top of the seal, probably only 1-2ml. Switch the pump on, and see if the little rim of water you've put around the seal goes down. If it does, the seal is sucking in air.
 
Thank you everyone for the replies, I will do my best to answer all of your questions:

- The pump is pulling the water at the end of the system and is an EHEIM 1260, which I oversized based on what I think I need as I was trying to be conscience on flow through the various sumps. It's hard for me to tell what my actual flow rate is right now due to the amount of air going through the pump, but I estimate probably around 400GH/1,500LH. +/- 10%

- I can definitely see the trapped air in the heater tube. The heater tube is filling up to about half way up the main clear section with the top half looking like a fast waterfall which then end up pushing air into the heater outflow.

- The system itself is using 3/4 inch everywhere (pipes, tubing, sumps, tank inlet/outlet) with the exception of the Heater tube. The Heater tube is 1.25 inches and the heater itself is a bit under 3/4 inch. So capacity of the heater pipe is 1/2 inch, although I am not sure if that is really constraining the flow and based on what I see. Given the cavitation in the pump, I do not think I am hitting 1/2 inch worth of flow.

- I can prefill the heater tube and I did plan for this as well. See the various white valves in the pumping. These were meant as ways to fill/prime/drain various parts of the plumbing. The issue I think I am having is if any air is not completely removed from the system it is going to collect in the heater tube and I think that is what I am experiencing. My fear is that it is going to be an endless air trap (a dead-leg per papa_c). Even if I put a bleed valve on the top I am worried the issue I have will be chronic and could impact the effectiveness of the heater. However, maybe there is an infection point where flow will increase due to less air flowing to the pump, which would pull excess air out as papa_c said.

- I have tried playing with the flows (slowing down to a trickle, etc.) and it doesn't seem to change the issue. The water just wants to go out of the bottom outlet rather than build up. I feel like this is a design flaw more than anything else.

Future Testing:

- I am using a uniseal at the top of the tube (where the heater connects) and will check for air being sucked in. This is an interesting theory as a lot of air is clearly leaving the the pump, yet the amount of air in the Heater tube is not decreasing. The air must be coming from somewhere. Some air (very small micro bubbles) is coming into my intake, but no where near the amount that is existing.

- I will operate the system with empty sumps to see if that increases flow and solves the issue.

Follow up Questions for Everyone:

- Does my existing design make sense or is there an inherent flaw? Maybe a bottom to top flow for the heater section? I really do not know if top to bottom or bottom to top matters. The design seems popular so maybe I am just

- Could this whole issue be caused by improper priming of the system. I am filling the water via the white valves in the picture, which eventually fills the in/out flow tubes to the tank as well. Should I just dry out the system and prime by sucking the outflow until water flows from the inlet all of the way to the outlet? Similar to how Eheim says to prime their classic canister filters. However, I think this would still cause the air gap.

Thanks for the help. You have given me a few things to think about and test.
 
The other thing you could try, that I have seen before is a venturi air removal pipe. Connect a length of 4mm tubing as in my picture below. It relies on the "venturi" effect at the bottom connection to pull water/air through the tubing, pulling any collected air from the top of the heater. Ensure pipe is removable for cleaning.

1602664640629.png
 
Nice build, but I would redesign it.. :) Simplify it and place the heater at the pumps outlet side... So it is fed from the bottom to the top, then there will be no air to trap.

Then you also get rid of all those unnecessary 90° knees that only restrict flow. Also rather use a 45° T with a 45° knee as heater outlet. And do not extend the T, but place the uni-seal and heater directly in the T. And make the gap above the T outlet small as possible.

Naamloos.jpg


Then if you are also using an external thermometer to control the temp it's advisable to put an inline thermistor in it as well. See red dot, where it should be placed.

In case you temp control uses a NTC 10K look at this one. It's dril a hole tap M8 thread and screw it in sealed with some Teflon tape. In case the pump shuts down and the thermistor is in the tank, the heater keeps heating. The thermistor should be close to the heater inline - cold side.
Amazon product ASIN B07N69NP4L
 
Thanks again everyone. Based on the advice, I am going to try to redesign the things a bit. I am worried about vertical space, but I do like the idea of the heater on top of the pump (or maybe I place the pump sideways so the output is horizontal and inlet is vertical). My biggest concern is that I may always have a vertical drop in the water's path as the sumps always need to have adequate clearance off of the ground so I can unscrew them and I can't go much farther up with the plumbing. Meaning this problem could still manifest itself in other forms.

Here is the probe I am using: LINK The website says it uses an NTC sensor, but I am not sure if that means it is compatible with an NTC 10K probe. Any advice?

Ian_m - I might try your idea first if I can not figure out a possible rework of the layout. I think your idea can easily be accomplished with two threaded RO push connects (like: THIS) and tubing. I already have these parts.

The concept of the venturi effect is still new to me as this is the first time I have come across the name. But if I understand correctly, by attaching that bypass tube ("venturi air removal pipe") it will automatically bleed (suck) the air out of the top via the force of the water at the bottom. So essentially, I would have an air trap that is continually drained. Is that correct? That just seems too good to be true.

An an aside, my fourth sump will act as a C02 reactor and I have worried that even if I fix my current air problem some C02 could get trapped in the tube if not fully dissolved. So the venturi route could fix this concern if I understand the concept correctly.
 
The website says it uses an NTC sensor, but I am not sure if that means it is compatible with an NTC 10K probe. Any advice?

So for the specification they give, the only one is better to ask them? :) Then to assume it a common 10KΩ sensor.
I also doubt that they can not say, but in case they won't and or can't provide an inline thermistor.

For example, the thermistor may be 100Ω, 1KΩ, 10KΩ, 100KΩ, etc. NTC stands for Negative Temperature Coefficient Resistance. It can very easily be checked with a multimeter set to measure KΩ, at relative room temperature the measured value should be close to 10K when the temperature rises the value should be lower. Thus if you measure 9.2 KΩ with a multimeter at warm room-temp then it's a 10K thermistor. If the value is x 10 higher or less than it's a ntc sensor with a different Ω value.
 
The concept of the venturi effect is still new to me as this is the first time I have come across the name. But if I understand correctly, by attaching that bypass tube ("venturi air removal pipe") it will automatically bleed (suck) the air out of the top via the force of the water at the bottom. So essentially, I would have an air trap that is continually drained. Is that correct? That just seems too good to be true.
Yes the water flowing past will draw water/air along the 4mm loop tube. A bit of experimentation may be required.
 
I might have missed something obvious and I'm not familiar with the seal you've used for the heater but is there a reason you can't mount the heater horizontally and have the T connectors in opposite directions? That way the water would force the air through when you prime the system and any residual air trapped around fitting connections etc would be pushed out when you turn the pump on.
 
I might have missed something obvious and I'm not familiar with the seal you've used for the heater but is there a reason you can't mount the heater horizontally and have the T connectors in opposite directions? That way the water would force the air through when you prime the system and any residual air trapped around fitting connections etc would be pushed out when you turn the pump on.


Yes, that is one of the options I was thinking about. The T Connectors on my current tube are sealed with cement and so I would need to remake the heater tube.
 
Nice build, but I would redesign it.. :) Simplify it and place the heater at the pumps outlet side... So it is fed from the bottom to the top, then there will be no air to trap.

I think zozo's suggestion will be the most fool proof. We don't know if this air is coming in via the uniseal or coming out of solution, but both are driven by negative pressure/suction.

Pumping it through the heater and filters will be positive pressure and avoid the issue. Downside would be back pressure on the pump and possibly a small bit of air trapped in the dead leg section at the top of the heater chamber.

The venturi bleed looks interesting, not something I have any experience of.

Did you try putting a few drops of water onto the uniseal?

Some bed time reading on the Bernoulli principle, which underlies the venturi effect. Guaranteed to put you to sleep:
https://en.m.wikipedia.org/wiki/Bernoulli's_principle
 
Last edited:
Hi Everyone,

Thanks for the continued help. I did a lot of testing the last couple of days and discovered a lot and at the same time identified what I think is causing the issue. The problem is that I do not know what the cause is. :banghead:

A few discoveries from testing:

1. The uniseal for the heater was not 100% effective. I was using a 3/4 inch uniseal, which I thought was the recommended size for a 300W ehiem heater. It turns out 1/2 inch is much more snug. I replaced the seal with a size smaller and so far everything appears good. Water spayed/dropped on the seal area appears to nothing (which is good).

2. Sump #4 (the purigen reactor) was causing some flow issues. Purigen is small enough that when it bunched up it was acting as a major flow restrictor. I put some angled holes in the tube of the reactor to cycle the purigen so it doesn't bunch up.

After these fixes there is no air in the system and all is well. EXCEPT...
 
Sorry, I posted before I was done and the time limit expired to edit the message. Please read this one:


Hi Everyone,

Thanks for the continued help. I did a lot of testing the last couple of days and fixed a lot. At the same time identified what I think is causing the issue. The problem is that I do not know what the cause is. :banghead:

A few discoveries/fixes from prolonged testing:

1. The uniseal for the heater was not 100% effective. I was using a 3/4 inch uniseal, which I thought was the recommended size for a 300W ehiem heater. It turns out 1/2 inch is much more snug. I replaced the seal with a size smaller and so far everything appears good. Water sprayed/dropped on the seal area appears to do nothing (which is good).

2. Sump #4 (the purigen reactor) was causing some flow issues. Purigen is small enough that when it bunches up it can restrict flow (at least in my system). I put some angled holes in the tube of the reactor to better push the purigen around so it doesn't bunch up.

After these fixes there is no air in the system and all is well. EXCEPT... I was doing all my testing using an submersed pump to make priming of the system easier.

However, once I turned on my main pump (which is not submerged) a lot of my issues returned. Mainly, the pump immediately begins to cavitate (even though it is fully primed) and air appears out of nowhere throughout the system. Attached is a picture of the system I was using to test.

Basically all I did was continue the system into a filled bucket and used the exact same pump (submerged) in the bucket.

So the test setup was: Bucket ->Submerged pump (ON) -> inflow of system (top of picture) -> inline pump (turned OFF) -> outflow of system (bottom of picture) -> Bucket. It runs perfect.

And the problem setup is: Bucket -> Submerged pump (OFF) ->inflow of system (top of picture) -> inline pump (turned ON) -> outflow of system -> Bucket. I am now a magician and creating air from nothing.

The only difference is which pump is turned on. The path of water is exactly the same. The inline pump is fully primed as all I am doing is unplugging the submerged pump (once all air is bled) and plugging in the inline pump.

Can anyone offer advice on what is causing this?

3.jpg
 
From what I understand the difference is the pump sits before the filters, or after. So again it's a difference between negative pressure/suction (pump after) and positive pressure (pump before).

The resistance within the system will be roughly the same, and its the resistance to flow that's making the pump cavitate.

Have you tried removing all the media to determine whether it's media resistance or the system design itself? If the pump still cavitates with no media in it you'll have to have a rethink. The media is also clean now, as it clogs flow will further reduce.

It might help to understand where these filters fit within the overall system. Is this being fed via gravity? Or is it coming pumped from a sump/somewhere else? Is the pump returning to the tank and part of your main flow in the tank, or returning to a sump? What are the pipe diameters?

I haven't run a sump since keeping marines in the late 90s, but there are people on here with more recent experience. But to my eyes the inflow pipe top right doesn't look much bigger than the return. It may be that if its gravity fed, the flow cannot keep up with the flow rate of the pump.

On your test you may have gotten the inflow -> pump -> filters orientation to work without cavitating as the inflow is likely shorter (less resistance than long tubes) and lacks the resistance from the filters, so flow will be less impeded on suction side, but will instead be impeded on the pumped side.
 
Hi Everyone,

Based on everyone's advice I went back to the drawing board to redesign the system. After about a week of testing I concluded that the pumps (Eheim 1262 & Danner Supreme Aqua-Mag 12) were probably too much pump for the project. I also stripped down part of the system to decrease the pipework to the pump so the pump had less negative pressure to deal with. In the end I decided to go with a budget DC pump (Jebao DCP 500), which appears to have a similar flowchart to the Eheim and Danner I was using. Additionally, it is adjustable so it will allow me to fine tune the pump to the flowrate I can provide from the tank. This pump combined with the reworked system layout appears to be working really well. I attached a picture of the new setup if anyone is curious.

I did my best to incorporate everyone's feedback and so overall I made the following changes:
- Reduced travel time to the pump by moving it directly under the initial mechanical filtration and before the heater and all other filters
- Removed two elbows using the new design by using some tubing instead of PVC
- Changed to an adjustable pump
- Moved the temp probe to before the heater
- Removed about half a meter of tubing to the tank

As of right now I am running the pump at 60% power and this is generating 1,200 L/H of flow which is about a 6x turnover rate and was my original target. The new layout primes considerably easier and the pump is near silent. The best news is that the inline heater is working well, which I guess brings the story to a conclusion since that was the original topic of the thread to begin with. The very top part of the heater (above the elbow) does have a small air pocket, but I expect this behavior based on the change in pipe size of the heater tube.

The only thing I haven't fully figured out is the Purigen reactor (center of pic). It is working well, but the purigen is so small I have a hard time figuring out how to contain it. Right now I am using very fine filter foam with filter floss stuck in between. This appears to hold the purigen in place but isn't my favorite solution. I might just break down and buy a few purigen bags, but of course that won't have the cool circulation look. :bored::bored::bored: Maybe that is a thread for another day.

Thanks again for all of the help.
 

Attachments

  • 4.jpg
    4.jpg
    1 MB · Views: 298
That looks great! any chance you can give some more information on how you built it, where you got the components from etc?
 
Sure, see below for the main item list.

Canisters: 4 x https://www.freshwatersystems.com/p...water-filter-housing-black-clear-w-pr-3-4-fpt
Pump: 1 x Amazon product ASIN B01MTVIZSFHeater: 1 x Eheim 300 Heater Amazon product ASIN B003I5UC0WHeater Controller: Amazon product ASIN B07QWTJNX2Lilly Pipes: JARDLI Glass Lily Pipe Inflow and skimmer inflow (3/4") Glass. I got both off of Amazon, but I am having difficulty finding the link
Heater Tube: https://www.homedepot.com/p/Formufi...Sch-40-PVC-Pipe-in-Clear-P114FGP-UV/206028405
Valve Unions: https://www.pvcfittingsonline.com/fpv-3-4-pvc-true-union-ball-valve-socket-threaded-ends.html
Normal Unions: https://www.pvcfittingsonline.com/9897-007-3-4-schedule-80-cpvc-union.html
Uniseal (for heater): https://www.aussieglobe.com/12-Uniseal-U050.html
Sealant: Lots of Teflon tape (usually 10 - 12 times around)

Everything else (mainly PVC pipes & connections) came from a local hardware store or https://www.pvcfittingsonline.com/. If black/grey colors were not available then I just taped off the connection and spray painted it black.

The heater was the only part that actually needed to be built (compared to just screwing together). It is basically a 1/2 unseal, the clear PVC (linked above), two 1.25" elbows, one 1.25" cap and two 1.25" -> 3/4" adapters that are glued to the unions. There are a lot of youtube videos with similar builds. The only difference in mine is that I added the unions. One things to note is that some videos mention using a 3/4" uniseal for the Eheim 300 Watt. 3/4" did not provide a tight enough seal for me and I did have leakage. I ended up going 1/2" inch. You need to soap up the heater's glass tube a lot to get it to fit, but the seal is much better against the 1.25" tube it is sitting in. Note that I also added a 3/4 cap at the very bottom of the tube that sits inside the 1.25" cap. If you look at the picture you see it at the very bottom of the heater tube. This is the 3/4 cap. The Eheim heater tube fits perfectly in the 3/4" cap and acts as a holder for the glass tube so the flow of the water doesn't knock it around.

Not pictured is the C02. The C02 connects to the white RO Water fitting sticking out of the 4th stage filter. This was probably the trickiest part of the build. I needed to "Tap" 1/4" threads in to the top of the canister. This required me to buy something like this: https://www.homedepot.com/p/Drill-America-1-4-in-18-Carbon-Steel-NPT-Pipe-Tap-DWTPT1-4/304632091. I tapped the threads into the pressure release hole that was already in the canister top (you just need to drill out the hole a bit bigger) then use the tap. Once tapped I just screwed in a RO water quick disconnect fitting (1/4 male NPT and 1/4 quick disconnect). The 1/4 fittings connect to the C02 line, which is very close in size. If this part is scary you could just hook up a inline atomizer into the return line or just use an intake diffuser. I can tell you that the current setup appears to be diffusing the C02 really well and not a single air bubble makes it to the tank.

One thing to be aware is that the canisters I bought had a pressure release valve built in (just a spring loaded button). The problem I was running into is that when running a negative pressure system (pump sucking water) these valves were letting air in. Therefore, you should either purchase ones that do not have a valve, remove them and silicone seal the holes (which is what I did), or have your complete system have positive pressure (pump goes before all canisters).

Hope that helps. It probably doesn't paint a full picture but if you are serious about replicating this the above should send you down the right path. I can answer more if needed (just can't type up a how to book right now).
 
Back
Top