Saving $20,000 USD With A Single LED


[N8Mcnasty] is a HVAC tech who works on some big machines. One of his charges is a Carrier 19EX Chiller, rated at 1350 tons of cooling. 1 ton of cooling = 12,000 BTU. This particular chiller contained an odd LCD screen. It used a fiber optic bundle and a halogen light for backlight illumination. The system worked fine for over a decade. Now though, the halogen bulb has begun melting the glue on the fiber bundle, causing a dim display. The display in question shows some very important operating parameters, such as oil temperature, current draw, and process temperatures. Since they couldn’t easily see the display, the machine’s operators weren’t running the machine, placing stress on the other chillers in the building’s physical plant. [N8Mcnasty] tried repairing the bundle, however the glue kept melting.

A replacement display was no longer available, meaning that the entire chiller control system would have to be upgraded to a newer system. The new control system uses different sensors than the old one. This is where things start getting expensive. Replacing the sensors would also require draining the 15-20 gallons of oil, 4500 lbs of R134a refrigerant, and bringing the whole system down for almost two weeks, a $20,000 job. Rather than go this route, [N8Mcnasty] found an alternative. LED’s have come a long way since 1996, when the chiller was built. He simply replaced the halogen bulb with an LED and appropriate resistor. [N8Mcnasty] was even able to reuse the halogen bulb bracket. A bit of heat shrink tube later, and the fix looks like it was a factory option. He’s documented his fix here on reddit.

81 thoughts on “Saving $20,000 USD With A Single LED

      1. No they couldn’t get another bulb… The bulb was fine. It was doing what halogen bulbs do; heating the crap out of the surrounding area with a waste by-product of bit of light. The article explains that the lamp was melting the glue for the fiber optic bundle.

        This is what hacking is all about. You’re stuck between a rock and a $20,000 solution but you come up with a $5.99 solution. Awesome work!

        1. I assumed the bulb was “bad” and was overheating because the surface became contaminated. Anyways as the article says, it was specifically designed this way and worked fine for over a decade. Either the lamp was bad, or the circuit driving it was. I would assume a HVAC tech could check the voltage of the line powering the bulb to ensure it was in spec, and the amount of current the bulb was consuming.

          Halogen bulb drivers arent exactly difficult to source. There were plenty of more solutions to this problem than a $20k upgrade. As I said below, this would be similar to saying changing a $3 oil filter is a hack to prevent a $10k engine replacement.

          1. Nothing has to be bad except for the original engineering. Its quite common in large scale industrial equipment. A 10 year life span isn’t exactly amazing for a unit such as this. It’s likely that there was no issue other than a design flaw, and the fix for the design flaw is to replace the control system with a new design.

            As for the plenty of other solutions, you’ve clearly never dealt with vendors of such equipment. There were options which involved massive upgrade costs, everything else had to be made up on the spot. The root cause of the problem was heat from the halogen, the change fixed the problem. Nothing much more to it than that.

            Also a better analogy would be that the person changed a $3 oil filter with one that he made himself, put together from other bits which were not compatible with the design all because the original vendor used a non standard part and no longer offered replacements. Oh but the new $10k engine uses a standard filter that’ll be much better.

            Just because the fix is simple doesn’t mean it isn’t a clever workaround for a flaw that could have otherwise cost a small fortune.

          2. If it worked for 10 years, and then started melting the glue over and over again, even after being re-glued then it isnt a bad design, there is a fault in the system somewhere. Either in the bulb or the driver which is causing it to overheat.

            Regarding the other solutions, I was referring to replacing either the bulb or driver circuit. Yes I know you cant buy a offical Carrier halogen lamp driver, but there are plenty of available aftermarket ones at your local home improvement center.

            And no, your analogy isnt better because this guy didnt make his own LED, he bought COTS components.

            This isnt exactly clever, performing root cause analysis would be.

          3. I think a better analogy would be for someone to adapt a different filter to a car other than the stock one. (and I don’t mean after market identical, more like different form-factor)

          4. No, I think it’s a design flaw based on the technology of the day. If it were designed today there’s no doubt in my mind that they would use led lighting or another low heat/low wattage solution. It just makes sense.

    1. If the LED defers a $20K maintenance bill, I’d say it’s worth a headline. Though this hack is missing some over kill. How about a color-changing LED or a 10watt LED on a heatsink? Maybe an RGB LED and an Arduino to change the backlight color based on machine status?

        1. Spending ~$10 in parts that will work for pretty much forever versus spending who-knows-how-many man hours sourcing an obsolete part, not finding it, and throwing away a $20K machine is a hack in that it saves a ton of money with a simple refitting of a different part. It’s not like saying changing a $3 filter to save a $10K engine. It’s like changing out that rusty factory steel fuel line that would cost $100 from the dealer (preformed) with one made of stainless steel from the local AutoZone for about $10. A little elbow grease and you have a fuel line that will last years longer than the factory part.

      1. Overkill would be a massively over-powered LED module and optics such that the display info would be projected on the wall opposite the machine…8′ wide by 4′ high and clearly readable with the lights on. Of course, long before getting to that point, he’d probably be back to having the fiber optic adhesive melting, even from the waste heat of the LED…

  1. 4500 pounds of r134a

    just be glad it was not r22 either way for sure you have to recover that you could not get away with venting that much.

    maybe you could vent a few ounces from a consumer refrigerator but 4500 pounds would need to be recovered.

        1. I just checked the manual, It doesn’t have a receiver, And your not going to be able to valve it off fast enough for refrigerant not to back flow through the centrifugal compressor. recovery is really the only option. It would take 45 100 lb canisters to hold it all. That number seems a little high, According to the manual it’s maximum charge should be no more than 3539 lbs. Not playing the BS card, but i’m getting a BSer alert going off somewhere. But the biggest system I’ve worked on is a 15 ton chiller, this thing is just insane.

    1. Venting more than a minute amount (when I was an HVAC tech it was defined as the amount left over in the charging manifold hoses) will result in tens of thousands of dollars in fines. Recovering even e few pounds of refrigerant is a time consuming process that requires specialized equipment. I can’t even begin to imagine how difficult it would be to recover 4500 pounds of the stuff.

      Is this a hack? Absolutely. Is this worthy of an HAD article? Of course. Will the naysayers come out in droves in a ridiculous effort to prove that their opinions are correct and no one else’s are? As always, yes.

      1. I have worked for organizations that would of gladly paid the $20k to upgrade the equipment as the LED would of been an “unauthorized modification” and would of required 5 layers of approval.Then they would of laid-off 5 people to boost profit margins.

        Now I work for a place where you are expected to come up with hacks like this. And as far as I am concerned, this “hack” is that defines a hack. Simple, quick, and too the point! Anything more and its good, solid engineering. Not a bad thing, but misses the point of what a hack is all about.

    1. i came here to say the same thing. I’m an HVAC engineer and most people in my industry say BTU. the problem is that if the units are incorrect (and BTU is incorrect) they don’t work out in equations and you can’t derive equations.

          1. @RP No, in a 24 hour period. If you had a cold storage building that required X tons of ice delivered per day, you would use a X ton refrigeration unit.
            The specific heat of fusion of water is 143 BTU/lb.
            So: (143 BTU/lb * 2000 lb/Ton * 1 Ton) / (24 hr) = 11917 BTU/hr or about 12000 BTU/hr.

      1. You have no way of knowing that. The employer might raise wages, or it might put the savings into its stocks so it raises their value. Which can have a positive effect, as more stocks sold means more funds available for investment in the company. Might get a few more jobs out of it too. You have no way of knowing, but it seems you just want to blame others anyways. So yeah, fuck it…

        1. And unicorns may fart rainbows across the sky too.

          We’ve got no way of knowing so we’ll just go with the explanation that seems most likely, that any extra cash will go to the Chairman’s new Ferrari.

          1. Maybe instead of being bitter we should see what Nate has to say?

            > My employer pays very well. I average about 100k a year in wages but I drive an all expense paid company vehicle with mostly company tools inside. I don’t own a vehicle for myself just a family vehicle and I have no commute cost. I get 10 days paid vacation as well as my union benefits that include a family health insurance, 2 pensions and an annuity. The company also does a 401k savings plan with company match up to $20 a week. I am able to live comfortably and allows for my wife to stay home and raise our kids. I love working for Carrier.

            Looks like he’s doing just fine. But it’s so nice of you to be concerned!

      2. It may show up in his annual review…
        and even if it didn’t, a good supervisor will remember his initiative and willingness to think out of the box. Saving the company $20K, just might be the straw not placed on the camel’s back… therefore saving his job from layoff.

    1. actually his company “lost” 20k by doing this “hack”.
      He’s not working for the owner of the HVAC, but for the HVAC company, so if he did not do this ‘add a LED” trick they could charge the customer 20k for the “upgrade”…
      Quite surprised there are still company’s where management even allow you to do this, I know plenty where this would get you very pissed off boss because he “missed” 20k on his balance…

      1. Lots of companies are willing to do this because the people running it understand that keeping a customer that’s going to be paying for upgrades and service fees is worth passing up a quick buck. It’s nice when market forces actually work.

        1. Yep, I see it all the time. You eat some cost that you could legitimately charge for because it makes life easier for the client, or charging would seem petty. If management is smart (the ones you seem to know apparently are not), the contractor would make sure that the client knows the situation; Do this the regular way and it’d cost you 20k, but we have this clever fellow on staff who figured out another option that’s going to cost you a couple hundred bucks. The client will be thrilled and you’ve cemented your position. Certainty in contracts is worth more than an easy buck.

      2. Yea, but try explaining to the customer that because a light bulb isn’t working you have to pay $20k, I’d figure your ripping off my company and it’s by-by at new equipment time.

      3. That 20k savings probably got them a die hard customer for life, if they were even willing to pay that to begin with.
        I’m in HVAC controls and we do small things like this all the time, never for a 20k savings, but I’ve directed at least a dozen people to buy a $600 computer and a $300 expansion card to their old system instead of a $7k controller for a better end result.

    1. It most likely does. But given the option of spending $20k on a new controller designed for the unit, or spending $10k on a workaround for a part which is clearly showing signs of failure, which road would you go down?

        1. “Optional Ex$tra”. That’s how.

          The latest & greatest scam is that manufacturers sell the same common unit cloaked in “basic”, “silver”, “gold” and “platinum” – functionality and the functionality is enabled by flipping a bit in the software. AUDI cars does that.

          It is also normal to use a proprietary interface so the system manufacturer can gouge the customer for an “interface option” and a separate license for the interface management software.

    2. it does. the problem is that you have to program the front end to display all those control pointe which costs big bucks. typically, they monitor only a few points on the front end and then use the unit’s control panel for displaying more specific points when there is an anomaly noticed on the front end.

  2. I agree that its a hack but i fail to see how he got the stage of having to talk to the manufacture/ supplier about it, I’m not the most qualified Guy ( im going to have to read this article again just to figure out what the machine does) but while reading the article as soon as i saw the words “Halogen” and “melt” i thought to myself ” he should put a MR16 LED replacement lamp in there”. a cheep one is $2 NZD and a good one $40 NZD. how this took more then 10min to diagnose and repair (and that he had to discuss this with someone ) is beyond me.

    1. he’s an HVAC technician, not an Electronics technician. this is no different then a person coming up with a clever “hack” for a car repair that may have been obvious to a mechanic. If you were put in a position to have to modify a HVAC system’s controls to get it to work right, you’d probably call it a hack, but he’d probably think it was standard operating procedure.

  3. I was replacing a rocker switch at a fast food place every 6 month because of heat damage (it was the switch for the fry lights and it was about 6 inches from a hot location). Until I got tired of doing it. So I relocated a heavy duty light switch in a metal box 2 feet away and had to change it once in the past 10 years. saving my customer money like that ensures them as customers even though the replacing it every 6 months would be more profitable. keeping your customers happy in the long run is how you survive in this economy.

          1. Oh right.. I lost context and was referring to the modification in the article but yes yours is definitely a common sense solution.

            The ironic thing about your solution is that I’ll bet there are thousands of instances with the same problem and only a small percentage with be resolved by simply relocating the switch.

    1. Depends on what “tonnage” actually is. Cooling power? Lots of industrial cooling equipment is measured in cooling “tonnage”, how many tons of air it can cool.
      Not actually the “weight” of the R134a

  4. Having dealt with multiple vendors using odd “one off” designs that were clearly half baked, and having to hack fixes to those problems, I appreciate that you call this a hack. Because it is. If not, the people that manufacture the control system would have a retrofit kit of this type already available.

    HVAC and power distribution systems are notorious for proprietary designs. Forcing the customer to upgrade to a newer version or electronics, normally at a high cost. Using common sense to extend the life of a product is worth the effort, espc at a $19990 savings.

  5. I say great fix, The upgrade uses less energy, produces less waste heat.
    Sadly In my line of work I can not do suck changes to the equipment I service since most of it is on medical systems.
    And I would get into major trouble for changing the design and if that unit failed and caused a injury or worse yet a death I would be in some real shite!

  6. I think this is a great hack (though I agree that the root cause analysis was not complete or accurate – something else was going on to cause the melting).

    There is a “political” cost and risk to a project like this that I do not think has been mentioned.

    Whenver you mess with an expensive machine like this, and do unauthorized non-standard mods, it is a big personal risk. Because if there is a failure (even non-related), you may take the blame. And they will quickly forget how you were trying really hard to save a bunch of money, even if you were given permission..

    Repeat: you will potentially be brutally punished if something fails, but you will likely not be significantly rewarded if your efforts are successful.

    I have also learned that a scarce resource can create political power. When a/c or power are short supply, the person who provides those resources can enjoy an important status. When those resources are no longer a major concern, then those resources are taken for granted.

    Also, when you go above and beyond, you risk alienating your peers. Even if they are lazy, it does not matter. And the senior guys who are supposed to know it all? They may also have grudges against you.

    Those were tough lessons for me to learn. I wish I’d learned them sooner.

  7. To all those people out there: Yes it’s a hack. A hack is solving a problem in a non-conventional way. So the original bulb was heating the glue to the point that it needed to be repaired/replaced on a regular basis. This guy recognized the problem and came up with a solution that the original manufacturer did not think of. In fact it appears that the non-hack solution was to replace everything. It may not be as cool as building a 32 node cluster of Pi’s but it’s still a creative solution to a problem that most people wouldn’t have thought of. I could easily see a company replacing the whole thing because it would be SOP but this guy didn’t and saved the company $20k.

    Hanging a tennis ball on a string so you know how far to pull your car forward may not be complicated or beautiful but it solves a problem by using resources in a way they were not intended.

  8. I think the idea of replacing the system was a great one, yes he could get a replacement bulb or driver, but may run into the same problems, and both i could guarantee would be more costly than what he did do.

    It keeps it simple. always a good thing. off the shelf parts are always better to use since (specially this small) replacement of it or additions to other units is cheap and easy to keep up with. why over complicate it?

    Though i light the RGB led setup that would turn red when a fault comes into play, though that would require more cutting into the electronics if there isn’t an easy way to get there is a fault otherwise.

    1. I will have to disagree with you on one statement. Off-the-shelf parts are NOT always best. First off, they’re not always reliable. They may be cheap initially. But, if they constantly have to be replaced you’re wasting time and money. For instance, I work on heated adhesive systems. I have a customer who kept buying cartridge heaters from a standard supplier at around $10 each because they didn’t want to pay $30 each from us. They changed the heaters 5 times in the course of the year and finally came to blame the design of the system (simple thermostat control, not a lot to go wrong). We have our heaters made to more exact specifications. The customer finally bought them from us and they’ve been going perfectly for the last 10 years. So, sometimes cheaper is not always better.

      Sometimes, however, cheaper can be much better. Like our competitors who come out with a new, more complicated system every 5 years and force their customers to upgrade when they come out. Their new systems made with all sorts of proprietary parts have all sort of things that hardly anyone uses and cost almost $15k. The machines barely last more than their 1 year warranty, forcing the customer to buy another new product or pay exorbitant repair costs. I can sell them a new system that does the exact same thing using commercially available products that are slightly modified and last a lot longer than the 2 year warranty for around $4k.

      1. And sometimes off-the-shelf works better than OEM.
        My Datsun diesel (may it Rest-in-Pieces) is a case in point. The original glow-plugs burned out within 4 years. The J.C.Whitney (yes, I hear you laughing) replacements lasted the remaining 10 years of its life.

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.