Interpreting radiator & near-radiator pipe temperature recordings

ariccio

I've been on a long and slow quest to convince the NYC multifamily highrise that I live in to adopt proper preventive steam trap maintenance practices. A few on this forum have probably followed this quixotic mission of mine.

I'm trying to be as thorough and specific as possible when I make the case to folks who categorically disbelieve the concept of failed-open traps. In addition to the stovetop steam trap tester that I'm about to assemble, I've been recording the temperatures at several points on a radiator.

I found some very convenient hose clamp thermocouples on mcmaster-carr, attached one just after the supply inlet, one at the outlet of the radiator, and one immediately downstream of the trap:



Yes, that is fiberglass insulation. I want the radiator on for these measurements, but at several points in march the heat in the building was cranked up so high that the apartment reached into the 80s with the windows open... so I had to do *something* to keep the place habitable. I would be a bit concerned about this throwing off the measurements, but you'll see in a second that that part of the radiator is a bit perplexingly cool. I'll move it to another radiator (without insulation) tonight anyways.

Since I couldn't see anybody on the forum has ever posted measurements like this, I figured it's an extra benefit to share them.

Here's the kind of data I'm collecting:


For reference, here's the weather around that specific cycle:


This is a vacuum steam system, the radiator is on the 9th floor of a 13 floor building.

Yes, I've triple checked that the "condensate return" channel is the one that's on the return, downstream of the trap. I've also checked and confirmed that the thermocouples are making good contact.

What can I learn from this recording?

I'm very slightly surprised to see the full five minutes between initial temperature rise and the first signs of warmth at the end of the radiator. Is that purely a function of the thermal mass of the radiator, or something else? In the past, I've noticed how long it takes (up to ten minutes) for a subjective feeling of warmth to come from any radiator in the apartment after the boiler kicks on.

How should I interpret the condensate return heating up slightly before the radiator outlet, staying hotter, and then that little bump of heating once the radiator starts to cool? My gut says "steam in the returns reaching this radiator before the steam in the mains". But I also know that I know very little, and should not be assuming.

Of course, this kind of reading also makes the "10-15 degree drop across the trap" test pretty useless. But I don't know if that may mean that this trap is leaking steam.

Finally, how does that trailing cooldown curve look? Too fast? Too slow? I'll say that for sure it does seem to contribute to overheating on a 60 degree night.

Big Ed_4

Are you a tech or a tenant ?

ariccio

A tenant.

ariccio

For reference: I live in a building which is doing the classic "coop board spends a million bucks replacing the boiler but not the traps" situation.

Complicated by the ever-so-common "building staff is well meaning but also categorically denies the very idea of a steam trap that can fail in a way that lets steam through and thus the radiator stays hot" along with "the building is often so overheated that apartments will be nearly 80°F on a cold winter day with the heat off", and "in the past year the radiators sound like they bought a drum set".

Jamie Hall

The time delay between steam reaching the inlet and the radiator heating up is quite normal. Lot of iron there...

I have never bought the 10 to 15 difference argument. The condensate when it enters the trap will be very close to steam temperature, and won't cool much going through the trap. If that trap were really failed open, you'd see the outlet from the trap much closer to steam temperature -- at least that would be my opinion.

Sylvain

Did you make some kind of calibration. E.g. do the sensors give the same reading when placed together on a single spot?
Probably yes but it is always better to verify.
Otherwise one doesn't know if the numbers are significant.

[edited]

Do all the working traps open and close simultaneously? I guess no.
How rapidly does a trap closes?

Jamie Hall

You mean working traps on different radiators? They will not all operate at once. Nor will they operate at the same interval. Too many variables!

Most trap designs will open and close relatively slowly -- they don't snap open or snap closed. A countable interval.

And on finding bad traps. I have always relied more than anything on just plain ordinary feel. I expect risers (or supply takeoffs) to be steam hot. YEOUCH! I expect returns to be warm if the heat has been on for any length of time, sometimes hot enough so you can't really quite grasp them for any length of time, but not steam hot. You can usually tell if there's a problem, without getting fancy.

ariccio

Jamie Hall said:

I have never bought the 10 to 15 difference argument. The condensate when it enters the trap will be very close to steam temperature, and won't cool much going through the trap. If that trap were really failed open, you'd see the outlet from the trap much closer to steam temperature -- at least that would be my opinion.

That has kinda been my feeling so far, and part of why I have gone so far down this rabbit hole. Whenever I see rules of thumb in an engineering context where many people give many different versions of said rule, I put up a little mental red flag. Some say 5-10. Some say 10-15. Some say 20. Obviously the conditions of pressure on both sides of the trap are in constant flux, so you couldn't even calculate with precision the temperatures you'd expect without measuring those pressures... And that means you can't have much of a good rule of thumb either.

My little mechanics stethoscope is not bad for $10, but I wasn't able to hear a clear whoosh-quiet-whoosh cycle. All I can hear is a quiet hissing/flowing sound (like steam rushing) and a distant but clear banging sound. The banging sound is slightly louder further past the trap. It does indeed sound like steam collapsing somewhere far away.

Jamie Hall said:

If that trap were really failed open, you'd see the outlet from the trap much closer to steam temperature -- at least that would be my opinion.

That is what I'm suspecting at this point.

This trap is relatively new (one or two years old) on that basis alone I'd expect it to be fine. I interpret the condensate return temperature at around 190°F as meaning that there's either little steam leaking through, or there's plenty enough vacuum to prevent it from flashing at a higher temperature? I'm not sure on that last part exactly how the system dynamics would work.

What has me looking into this more and asking these questions is that I recall Dan Holohan saying many many many times that failed traps in part of the system can destroy the remaining functional traps. For many many many reasons I'm confident there are a substantial number of failed traps in the system. Given the maximum severity of banging we had at this radiator in January (it sounded like a small bomb once or twice), I kinda expect that anything delicate would be broken by now. Are these traps special enough to survive what others can't?

If nothing else, I may not change any minds in my building, but I will learn a lot of fascinating things about steam heat.

Sylvain said:

Did you make some kind of calibration. E.g. do the sensors give the same reading when placed together on a single spot?
Probably yes but it is always better to verify.
Otherwise one doesn't know if the numbers are significant.

[edited]

Do all the working traps open and close simultaneously? I guess no.
How rapidly does a trap closes?

I know my FLIR doesn't quite read the pipe temperatures accurately. I was strongly suspicious (because metal pipes), but now I can measure with the thermocouple reading 213°F at the supply valve, and different parts of the pipe surfaces around it ranging widely from 206°F to 233°F:




I lack a more precise instrument to calibrate against it. I mostly trust the thermocouples themselves, as most things I buy from McMaster-carr are exactly what they say they are. Data logging meters for thermocouples were WAY overpriced on McMaster-carr, for which I certainly don't have the cash right now, so I bought some no name unit from Amazon. I'm betting on the precision/quality of the thermocouple itself being more important than the logger.

On a different radiator where I *could* hear the trap cycling with the mechanics stethoscope, I could hear it purge every fiveish seconds or so.

ariccio

Jamie Hall said:

I expect returns to be warm if the heat has been on for any length of time, sometimes hot enough so you can't really quite grasp them for any length of time, but not steam hot. You can usually tell if there's a problem, without getting fancy.

One of the first clues for me that something was amiss is that many apartments in this building are plenty warm in the depths of winter even with the radiators shut off. High 70s or low 90s.

Once I could afford a nice little FLIR, I started poking around and noticed the exterior walls that contain the pipes can get really warm, warm enough to be accidental radiators. And it's usually not just the supply, it's also the other side where the return is, and they're usually about the same temperature!

ARobertson13

ariccio,

I noticed that the steam is on for about 20 minutes and the radiator is located on the 9th floor. Assuming you are using older cast iron radiators and the apartment heat load is not great, I am not surprised that you would be experiencing over heating. Also, I always remove the rust from the pipe and radiators surfaces and then attach the heat sensor by strapping it down with plastic tie wraps. You might consider using a analog multichannel data logger like Onset because they allow you to use both ambient and contact sensors.

Sylvain

ARobertson13 said:

You might consider using a analog multichannel data logger like Onset because they allow you to use both ambient and contact sensors.

What do you mean by ambient sensor?
something like this: https://en.wikipedia.org/wiki/Wet-bulb_globe_temperature ?

Otherwise an additional thermocouple in free air in the room will measure the ambiant temperature the same way as an ordinary thermometer.

ARobertson13

Sylvain,

Sorry if you were confused,

There are different types of sensors: temperature, pressure, humidity ect. Different types of temperature sensors: Thermocouple, RTD, PTC and NTC. There are also different subgroups within each type Thermocouple: J, K, T ect. The manufacturer of data loggers will have an assortment of sensors for specific usage. Some data loggers cane only use a specific subgroup. Others can be configured to use different subgroups and even different types. The datalogger in my Coop measures Ambient temperature outside, Boiler amperage, Mains temperature.

ariccio

Ok, now this is interesting. Here's what it looks like over the course of a few days:



Zoomed in to the late night of the 15th, early morning of the 16th:




Now, it was a fairly mild night. It only bottomed out around 55 degrees freedom:



Not even lower than the target temperature for many evaporator coils!

If I were to assume zero energy loss between the boiler and the radiator supply valve, I think this would mean the boiler was running at about 4psi.

I find several things curious about this. The fact that the supply was bouncing between 216 and 223 for ten hours sounds to me like the controller was calling for heat from the boiler the entire time. This is not the norm for this system. Even since the new boiler was installed, it's cycled between "fully cool" and steam most of the time. I suspect this means the boiler was simply commanded to run the entire time, and was bouncing between two settings on the pressure control.

The other interesting thing is watching that condensate return temperature slowly the high of 209 degrees. On one hand, that seems nuts for a system that's supposed to be under vacuum on the return side. On the other hand, I can imagine that maybe after ten hours of constant heat, the surface temperature of these pipes might approach a high equilibrium anyways? I dunno. The early temperature of around 190 seems a lot more reasonable.

Jamie Hall

A few coments. First, this is residential or heat only larger building? In either case, you running pressures are too high. It should run just fine bouncing between 1 and 2 psig.

Second, yes, it appears that the boiler is being controlled by something -- and it's likely to be the pressure. This is also normal... sort of... but I'm going to guess that the boiler is somewhat oversize for the connected radiation. Is the system zoned in any way? That can cause an interesting assortment of unexpected problems...

Third, you say the system is "under vacuum on the return side". How is that vacuum created or maintained?

ariccio

Yup, residential building. 1960 NYC apartment building that's now a co-op. About 140 units and about 400 connected radiators. 13 stories tall. Vacuum steam. 175HP boiler.

I know the pressures are too high - I've been trying to convince the relevant decisionmakers for a while now! As I'm sure you know, the default instinct for many is to turn the pressure up or lengthen the cycle when someone complains they're too cold, never to chase down failed open steam traps. This is complicated when people refuse to believe that steam traps fail open in the first place. There is a pressuretrol, at least there was one last time I saw the controls.

I do strongly suspect it's oversized - what tipped you off?

What's infuriating is that this is a brand new boiler. The board decided we needed a new boiler, and they replaced the 60 year old 175hp boiler with a brand new 175hp boiler. I'm not sure how they could have done the sizing math correctly, since nobody went to every single radiator to calculate the EDR. As I'm sure you know one way or another, people skipping relevant calculations and measurements is a widespread problem in the HVAC industry. It does seriously irritate me that licensed professionals involved in this job seemed to have skipped this. Of course, if we just spent >$500k on a brand new boiler that's oversized, we've hamstrung our maximum efficiency for at least a generation.

Now, we do also have accidental zoning. What do I mean by that? The system is wildly unbalanced, so much so that for years many apartments have been warm enough on 30 degree days even though their heat is off. When on, those apartments will reach into the 80s.

We have a two pipe vacuum steam system, but management has in the past given one pipe advice to never turn the radiators partway. For example, an email from almost ten years ago:

> Please be advised that if you feel the radiators in  your apartment are supplying heat when you don't want it, you can turn the radiators off by removing the face plate to the radiator and turn the knob to the heating element all the way off.  Please note that the knob must be turned all the way off or all the way on.  Turning in only partially will not regulate the temperature like a thermostat and can cause your radiator to spring a leak.  Aside from turning your radiator off, please consider opening a window instead of running your air conditioner. 

As a result, something like half or a third of the radiators in the building are shut off year round! Nobody else has found it odd enough that they'll be hot in the middle of a cold winter with their heat off to spend a moment thinking about it. I strongly suspect some of the radiators that are shut off have failed-open traps, and this has let us get by for longer.

ariccio

Interesting differences when I pulled the latest data. I know they've been working on installing in-unit sensors in some apartments, and I suspect they've turned them on. In addition to my apartment being almost reasonably cooler from 6pm-5am, it's warm now from 6am through noon. Very loud water hammer now happens at 6am instead of 6pm. Charts look different too:

Focusing on some parts show interesting patterns. One of the more perplexing ones is this:

The fifth cycle appears to start with a truncated firing of some sort, where steam wasn't even present for long enough to heat the supply pipe surface to 212… which should be a warning sign for someone in the building how bad the imbalance is…

But I don't understand that little hump in the condensate return temperatures toward the end of each cycle. What is that reflecting? A short fire that lasts long enough to push some steam into the return, but not reach the supply?

Edit: the images are being scaled very strangely in the posted version but not the version in the editor. Click through to view them alone and reasonably formatted.

Erin Holohan Haskell

@ariccio we're investigating the image issue.

Thanks to all for your patience as we sort this out!

CLamb

@ariccio "But I don't understand that little hump in the condensate return temperatures toward the end of each cycle. What is that reflecting?"

I'm thinking it has something to do with the trap opening. Perhaps the trap opening wider.

ariccio

https://forum.heatinghelp.com/discussion/comment/1800085#Comment_1800085

No worries! Probably a CSS thing. Seems sorted now.

As a programmer myself, I happen to believe CSS is no better than assembly language, but more cursed.

ariccio

The second firing here looks a lot more like I'd expect a stable and healthy system to look, am I in the ballpark

?