Primary flow > Secondary flow

Lost_aussie

Hi all,

I have a in-floor radiant hydronic heating system with a combination of in-slab (level 1), subfloor with emitter (level 2, under new flooring) and staple up (level 3, under existing flooring). I have zone valves and a single Taco VR1816 ECM circulator set on constant high. The house is split into 3 zones (L1 has 2 Airbnb rentals that have independent zones, L2 & 3 are a single zone).

L1 and L2 are 1/2" but the L2 zone consists of 3/8 tubing and I struggle to get enough flow on a few of the loops. I made a bit of a mistake when piping and the longest zone ended up at 245'. With the circulator set on constant high I have 4 loops which free-flow and still struggle to get enough flow, requiring me to throttle back all the others and run higher temps than optimal to get enough heat. I also made the mistake of running my boiler to manifold piping in 3/4" so I get additional pressure drop on that due to a penny pinching design decision.

My primary system (Noritz CB199DV, soon to be replaced with a NRCB199DV) runs higher primary flow than the secondary, all the time, based on how I have flows throttled.

I'm wondering whether I could change from the current primary secondary system (with closely spaced tees) to a variable primary only system, putting both pumps in series and giving me extra pumping juice to push higher flow? I could initially do this by putting a full port ball valve in the closely spaced tee and trying it, but are there any red flags I should be aware of?

Alternately, can anyone recommend a secondary ECM pump that would deliver better performance (higher head) for my situation?

For reference, pump curves are shown below.

VR1816


Boiler circulator



The Taco also has other modes which can be selected. Proportional mode on high could help minimize extreme flows when only 1 zone is calling and they are in series.

hot_rod

It could be that the heatload is higher then what the tube can deliver. If so you would not be able to pump your way out of that.
While not ideal, a 250’ 3/8 loop is workable. It was a common Heatway design as their rolls were 500’ 😁

Do you have transfer plates, or bare tube stapled to the floor?

Was there ever a heatload calculation performed?

Jamie Hall

Putting pumps in series is always an interesting problem. Their flows will match -- they have to -- and so the combined head developed will not necessarily be the sum of their individual performances.

A much better approach is to keep the primary/secondary, and select pumps for each which will deliver the flow you need at the head loss in the loops. You can construct a similar head vs. flow curve for the loops -- head loss vs. flow -- and then simply see where that curve will intersect the various pump curves. Pick the pump which gives you what you need.

hot_rod

the pump head is related to the gpm you are trying to push. The gpm is related to how much heat you are trying to move through the tube.

Typically .35 gpm for 3/8 tube which would be a bit over 5' in a 250' loop.

Trying to push .50 gpm takes the heads to 10'

Increasing temperature is another way to increase output. But not knowing what the load is, it is a guessing game.

Dave Carpentier

How about split off the 3rd level onto it's own closely-spaced-tees and circ pump ? You can have more than one secondary loop off of a primary.
If you do a heat loss calc on the 3rd level, and with the known emitter properties, you might be able to specify a single speed (cheaper) circ pump for that dedicated loop.

Lost_aussie

hot_rod said:

It could be that the heatload is higher then what the tube can deliver. If so you would not be able to pump your way out of that.
While not ideal, a 250’ 3/8 loop is workable. It was a common Heatway design as their rolls were 500’ 😁

Do you have transfer plates, or bare tube stapled to the floor?

Was there ever a heatload calculation performed?

Thanks for the insights. There was a heat load calculation performed and below is the design for L2


L3 is staple up in omega style aluminum heat transfer plates.
L2 is installed snug in channels between the 2nd layer of subfloor (3/4" plywood) on top of 0.005" foil, thermally coupled to the foil with a bead of silicone. It seems to work pretty effectively.

The main manifold (serving L2 & L3) is currently running a ~14F deltaT (100F supply, 86F return at this moment, based on outdoor reset temp) on the main floor as compared to 10F in the design, which makes me assume insufficient flow. The flow indicator for the longest run is only showing 0.2GPM as compared to ~0.5GPM design. This is with all zones currently calling for heat (so high flows), it gets a little more flow when there are less zones running.



Also, for information sharing sake here's a pic of my pump and closely spaced tees. Note that I did shift the pump after this photo was taken as I realized I had violated the minimum straight pipe length prior to the pump. The boiler is just to the right of this photo - I can grab some more later today.

Lost_aussie

Jamie Hall said:

Putting pumps in series is always an interesting problem. Their flows will match -- they have to -- and so the combined head developed will not necessarily be the sum of their individual performances.

A much better approach is to keep the primary/secondary, and select pumps for each which will deliver the flow you need at the head loss in the loops. You can construct a similar head vs. flow curve for the loops -- head loss vs. flow -- and then simply see where that curve will intersect the various pump curves. Pick the pump which gives you what you need.

Thanks. I guess I should sketch out what the combined head of the pumps would look like - will do that later this afternoon and post for comment.

Lost_aussie

hot_rod said:

the pump head is related to the gpm you are trying to push. The gpm is related to how much heat you are trying to move through the tube.

Typically .35 gpm for 3/8 tube which would be a bit over 5' in a 250' loop.

Trying to push .50 gpm takes the heads to 10'

Increasing temperature is another way to increase output. But not knowing what the load is, it is a guessing game.

My flowsetter is showing 0.2GPM flow on the ~250' loop of 3/8" tubing.



Based on your screenshots I'd need to add ~5' of additional head at the manifold to double the flows on this loop. My current pump delivers ~17.5' of head, so there's clearly a lot of pressure loss over the 3/4" supply system. That has me thinking I should add comfortably more than 5' pumping capacity to get flows on these weak loops up to where I want them (holding all other flows constant).

Everything is now behind sheetrock, so there's no way to solve this with upsizing my pipe unfortunately. Lots of first timer lessons learned on this project! As I see it, my options are:
1. bigger pump
2. change to variable primary flows (pumps in series), or
3. accept a bigger deltaT and bump up my HW setpoints (and throttle back flows on the other loops even further)

Jamie Hall

Water is terminally lazy. It will always take the path of least resistance. In your situation at least for the moment I'd consider throttling the flow to the loops with higher flow rates and see if that diverts more water to the loops with the lower flow rates. In an ideal world (yeah, right) the flow rate would be more or less related to the length of the loop -- with the longer loops having greater flow rates, which is exactly what water does NOT want to do.

hot_rod

Did the design include a pump spec?
But you are probably correct in the 3/4 pex to the 9 port manifold is part of your constipation. The head calc is for the entire circuit as you know, not just the manifold. PPI has a nice pressure drop calculator foe use at their site, but that ship has already sailed it sounds like.

Ideally with every zone/ loop running you want to get to design flow on every loop.

You could keep upping the pump head, but be careful at some point the actuators may start leaking past? I'd guess 25- 30 psi shutoff delta P on those helmet heads?

A large enough delta P circ may be in your future, I think you are getting into Magna territory

Lost_aussie

hot_rod said:

Did the design include a pump spec?

The design had a single 1" from the pump to the manifolds and spec'ed the VR1816. I intended to run a dedicated 3/4" to each manifold instead as getting 1" PEX through the joists/walls was a **** (I'm still kicking myself for not putting this in or under the slab). Had some material wastage and didn't have material on site to do the independent runs, was in a major time crunch, and cheaped out by running them both off a common 3/4". Lots of lessons learned from a first time DIY'er, and I would not make the same mistake again.

hot_rod said:

But you are probably correct in the 3/4 pex to the 9 port manifold is part of your constipation. .

At design flows, all zones calling for heat and with all installed fittings, I'm calculating 15' head just in the 3/4" line between the pump and the main floor manifold, not including any restriction from the air separator. It seems obvious that this is the cause of my restriction (constipation as you say). With just the L2L3 zone calling, the calc is showing 9'. That correlates to what I see in the real world - right now I have just the L2L3 zone calling for heat and I'm seeing 0.3 GPM on the longest 3/8" zone as compared to 0.2GPM earlier on the same zone when all 3 zones were calling.

hot_rod said:

You could keep upping the pump head, but be careful at some point the actuators may start leaking past? I'd guess 25- 30 psi shutoff delta P on those helmet heads?

Thanks for the heads up. I've no desire to get up to that kind of pressure in PSI. Feet of head, yes.

hot_rod said:

A large enough delta P circ may be in your future, I think you are getting into Magna territory

I'll do a search on that. Are pump flanges typically compatible between brands?


For **** and giggles, here is a summation of the 2 pump curves, with the ECM on various settings. As I understand, this would represent a variable primary system design (ball valve in closely spaced tees closed).



I'm eying the grey curve (VR1816 on proportional mode high setting + boiler pump in series). That would give me runout at 35' (about double what I am getting now) and wouldn't drop below the existing secondary pump curve until 4GPM (which I'm not achieving with current operation, even with all zones open). My hope would be to see appreciably more flow under single zone operation and smaller flow gains under multi-zone. Thoughts?

hot_rod

I have never tried two circs in series running a delta p mode. I think that is what you are proposing.
What is the total gpm load of the system. From the design. Is it gpm at 9’ head?

So everything is 3/4 boiler loop, primary loop, and 2 or 3 manifolds on that 3/4 loop that disappears into the wall?

I think you want that boiler kept as p/s, and try to find enough circ capacity for the actual distribution side load, gpm and head. Does an Alpha 2 the 26-99 version get you closer. Sounds like you are looking for a very high head, low gpm circ?

Lost_aussie

hot_rod said:

I have never tried two circs in series running a delta p mode. I think that is what you are proposing.

More asking the question than proposing. If no-one's done it, it's probably a bad idea.

hot_rod said:

What is the total gpm load of the system. From the design. Is it gpm at 9’ head?

Per the design with all zones calling for heat it was 6.98GPM. But it's run satisfactorily for 2 years at the current heavily restricted flows, so I think the home's envelope is performing a lot better than was allowed for in the design.

I'm aware that getting to 7GPM is out of the question, I'd just like a bit more to get more even heat distribution, hopefully some better response times and less boiler cycling when only some zones are calling.

I currently run a wide deadband on the boiler temperatures that allows it to run up to ~15F above setpoint. That + the thermal mass in the system keeps minimum run times at approx or close to 30mins for either of the 2 larger zones. The small zone cycles if it's calling by itself, but I've tried to minimize that with schedules.

hot_rod said:

So everything is 3/4 boiler loop, primary loop, and 2 or 3 manifolds on that 3/4 loop that disappears into the wall?

Correct. The 3/4 that disappears into the wall serves the 2 manifolds. Any other comments on the layout? Is my closely spaced T setup ok?

hot_rod said:

I think you want that boiler kept as p/s, and try to find enough circ capacity for the actual distribution side load, gpm and head. Does an Alpha 2 the 26-99 version get you closer. Sounds like you are looking for a very high head, low gpm circ?

The Alpha 2 looks like it would be a great fit, thanks for sharing. I think it would need to be on CP1 or CP2 mode to deliver enough head, although PP3 could be worth a try.

hot_rod

You might try removing the blue actuator from the loop with only .2 gpm. There have been reports of those actuators that have adapter rings not stroking fully With
the actuator removed the valve is off. If you still have the white cap screw it on. Or use a screwdriver handle and push the pin all the way down, see what flow you get.