Homemade air filters - Good indoor air quality, cheap

[Poindexter]

This is pre-pub. I have a lot of documentation to do after I collect a bale of data. I wasn't expecting to post here for another 6-8 weeks, but check this out.



This is in my living room, about 10 feet from my wood stove that has been running 24/7 for days. I also have a geriatric long haired cat in the house, and some 12 year old carpet that was on the block to be replaced "next month" in December 2019.

This is the second of my two particle counters that came in today. My first particle counter, a Dylos 1100PRO, came in Friday and the numbers I was getting kinda made me wonder if the Dylos didn't get a bump on the noggin knocking it out of calibration. I just couldn't quite believe the filter system I built is working this good, but it seems to be working that good.

So how did I do it? Meet the Poindexter box:



Straight up WYSIWYG. 20 inch box fan on low, wafting cool air from the bedroom floors down the hall towards the stove room. Top and bottom panels are cardboard from the box the fan came in. The other three sides are 20x20x1 inch MERV 13 furnace filters. MERV is an ASHRAE (hvac industry) standard. If you are shopping at the team orange homestore look for filters at level 10. If you are shopping 3M filters you want either 1900 or 2200. Assembly, all of it, about half a roll of duct tape. Tools required, something sharp. I used a drywall knife and a pair of scissors.

Assembly instructions: Gather materials. Cut the two largest possible panels from the box the fan came in. Assemble with duct tape, trim cardboard to fit as you go.

I did use 3M 2200 for mine. As a clinican at the bedside for a quarter century now I think MERV 13 is a very reasonable household filtration level for particles. If MERV 13 isn't enough, your kid is still sick or etcetera, then look at HEPA and money. Baseline healthy people in a baseline healthy house, my opinion, do not require household wide HEPA level filtration.

I have been collaborating, loosely, with a couple guys on a different website. Alan and David. One of them is a cardiothoracic anesthesiologist with a professional interest in whole life air quality. The other is an MSEE who got interested in airflow dynamics as a hobby. No idea if either will register here. Also I must give credit to Comparetto of "Comparetto cube" fame, and Corsi and Rosenthal of "Corsi-Rosenthal box" fame. The Poindexter permutation or Poindexter box is clearly an evolution. I don't give a hoot if the Poindexter Box is patentable or something that could be monetized. I want you and your kids breathing the cleanest air you can get so there will be somebody healthy to take care of me when I am old and feeble. Older and Feebler.

I gave @begreen a heads up a bit ago. He said this research is going to end up in the green room. I asked him to leave it in this section for at least a few days so you guys could see it first. @BKVP , this should be the first you have heard of this unless begreen spilled. My research is not supported by any commercial entity and I will publish my findings no matter what they are.

For now I am not prepared to exhaustively defend this thing I made, but now I have something to do this winter I can win at doing.

I do recognize MERV 13 filters don't do anything about VOCs. I have a plan for that, hopefuly something to say by Christmas.

 

[ABMax24]

I like the idea, pretty practical and affordable for almost everyone.

Step one for me is air quality monitoring, where did you get that meter from? I'd like to get one for our house.

Edit: Found it on Amazon, will have it Friday.

 

[Poindexter]

I like the idea, pretty practical and affordable for almost everyone.

Step one for me is air quality monitoring, where did you get that meter from? I'd like to get one for our house.

Edit: Found it on Amazon, will have it Friday.

Yup. My goal wasn't to see how good I can get my air quality. My goal starting out was how good of air quality can anybody have.

I did put the cart before the horse. I do not know "how bad" my house air quality was when I started running the filter. I was running a Corsi-Rosenthal filter in my wood shop in the garage. One night watching TV upstairs with the wife I went down to the shop to get a beer out of the downstairs fridge, went back upstairs, put my beer down, went back downstairs, brought the C-R upstairs and started running it.

The Corsi-Rosenthal and Comparetto models (configurations?) both break the convective loop I used to have with the box fan on the floor, therefore horizontal output, et viola.

There is consistent data out there that the $60 particle counter as pictured above runs pretty well neck and neck with the more expensive Dylos. I will be building my own data set to compare the two with somewhat more robust data sets than I am accustomed to seeing online. The pictured AQ meter is available on Amazon, and also from a company named "banggood" that will send you many many emails I am told. I ordered mine from Amazon because I get enough email already. Also be aware the more expensive Dylos counts particles per cubic foot, there is some assumptions and some math to get to micrograms per cubic meter from the raw Dylos data.

A brief word about VOCs. There are four methods to measure VOC in an air volume. Gold standard is to get an air flask in the mail, follow the directions, mail the specimen back to the lab and they hook your sample up to a gas chromatograph machine. About $200 for a total VOC count, about $400 if you want the lab to print out every little thing to four decimal places. Among reputable field devices the Flame Ionization Detector (FID) and Photo Ionization Detector (PID) sensor types retail in the $3-10k range but can sometimes be rented, with a hefty deposit.

Finally, the Metal Oxide Semiconductor (MOS). You may have heard of MOS-FETs, a metal oxide semiconductor field effect transistor. A MOSFET is a thing that acts like a transistor but uses the electrical fields generated by the metal oxides instead of physical PNP or NPN junctions in the dope inside the device. You can get a MOS sensor to hook up to an Arduino or Pi for about $20,and they appear in full featured "air quality monitors" up to +/- retail $500. I am not a fan of this sensor for VOC detection for a number of reasons beyond the intended scope of this thread.

I look forward to finding out where @ABMax24 particulate air quality reads before he builds a filter.

 

[kennyp2339]

Your side projects always amaze me, solar kilns, anti-smoke re-loads on the stove, to clean in-door air, way cool.

 

[EbS-P]

I’m going to borrow my dads air quality monitor that does pm 2.5 and voc and formaldehyde ect. and get a weeks worth of data before I start burning. I think that’s an important number. I usually run two air purifiers with a poodle, 4 kids and 100% hard wood floors the dirt is always stirred up and in the summer I need to clean my ac filter very frequently like every two weeks. All the articles about wood stoves and indoor air quality I would like a personal data point.

 

[Poindexter]

I’m going to borrow my dads air quality monitor that does pm 2.5 and voc and formaldehyde ect. and get a weeks worth of data before I start burning. I think that’s an important number. I usually run two air purifiers with a poodle, 4 kids and 100% hard wood floors the dirt is always stirred up and in the summer I need to clean my ac filter very frequently like every two weeks. All the articles about wood stoves and indoor air quality I would like a personal data point.

My wife and I haven't had to dust again since this thing has been running. I am sure we will have to dust again someday, but I can't tell you what the new interval is yet.

 

[peakbagger]

So next thing to do is incorporate a UVC lamp or two with appropriate shading and interlock to deal with airborne pathogens. I did a lot of work over recent years in various hospitals HVAC systems and I am always running into UVC banks in the supply air ducts. I think they were put in years ago to deal with TB.

 

[stoveliker]

about UVC, just be careful that you measure the ozone concentration. It produces ozone, and too much is not good for your lung (lining) either. Never let that UV light on when air is not flowing (local concentration goes up too much).

 

[Poindexter]

So next thing to do is incorporate a UVC lamp or two with appropriate shading and interlock to deal with airborne pathogens. I did a lot of work over recent years in various hospitals HVAC systems and I am always running into UVC banks in the supply air ducts. I think they were put in years ago to deal with TB.

I am not encouraging healthy adults in healthy homes to take this step at this time.

1. MERV 13 filtration will catch 75% of virus sized particles and 90+% of bacteria and mold sized particles on the first pass.

2. Hospitals are full of sick people, none of the beds are for healthy people to take a nap on their way to the golf course.

3. As @stoveliker already mentioned, ionizing radiation like UV or an "ozone generator" is actually a pollutant on its own.

For the operating rooms it makes sense to kill every possible living organism in the air ducts to reduce the number of living organisms that get through the skin at the surgical incision. It also makes sense to not be pumping Legionella bacteria into a room with a sick sleeping elder in it. But the air coming out of the ducts now has 'some' nonviable organic debris in it, as ionized particles. So no infectious materials, but ozone as an introduced pollutant and ionized particles as an introduced pollutant. At the hospital level adding on another filter to deal with the ionized particles and ozone is probably regulated at the federal level.

At the household level the "cure" may be worse than the disease. The most common exception would be a home with so much mold in the ductwork (I don't have forced air at all) that a UV light or ozone generator brings the mold counts down far enough to be tolerable, but now the homeowner should be doing something about the ozone and the ionized particles. - and looking at this as a temporary AQ fix while the mold in the duct work is remediated.

I should reiterate that I do NOT advocate sticking a MERV 13 filter in an air handler that is not rated MERV 13. It will throw your air handler out of balance, just like a super dirty hasn't been changed in 3 years MERV 11 filter might. I actually joined an hvac forum for this project because I was tired of proving I wasn't a robot everytime I wanted to use the search button. With an overrated or very dirty filter the most common service call is the condensor coils ice up because of diminished air flow and then the compressor blows up. Hundreds of dollars to repair. Or an $80 standalone filter cleaning the air you are actually breathing inside the house might work.

 

[ABMax24]

Yup. My goal wasn't to see how good I can get my air quality. My goal starting out was how good of air quality can anybody have.

I did put the cart before the horse. I do not know "how bad" my house air quality was when I started running the filter. I was running a Corsi-Rosenthal filter in my wood shop in the garage. One night watching TV upstairs with the wife I went down to the shop to get a beer out of the downstairs fridge, went back upstairs, put my beer down, went back downstairs, brought the C-R upstairs and started running it.

The Corsi-Rosenthal and Comparetto models (configurations?) both break the convective loop I used to have with the box fan on the floor, therefore horizontal output, et viola.

There is consistent data out there that the $60 particle counter as pictured above runs pretty well neck and neck with the more expensive Dylos. I will be building my own data set to compare the two with somewhat more robust data sets than I am accustomed to seeing online. The pictured AQ meter is available on Amazon, and also from a company named "banggood" that will send you many many emails I am told. I ordered mine from Amazon because I get enough email already. Also be aware the more expensive Dylos counts particles per cubic foot, there is some assumptions and some math to get to micrograms per cubic meter from the raw Dylos data.

A brief word about VOCs. There are four methods to measure VOC in an air volume. Gold standard is to get an air flask in the mail, follow the directions, mail the specimen back to the lab and they hook your sample up to a gas chromatograph machine. About $200 for a total VOC count, about $400 if you want the lab to print out every little thing to four decimal places. Among reputable field devices the Flame Ionization Detector (FID) and Photo Ionization Detector (PID) sensor types retail in the $3-10k range but can sometimes be rented, with a hefty deposit.

Finally, the Metal Oxide Semiconductor (MOS). You may have heard of MOS-FETs, a metal oxide semiconductor field effect transistor. A MOSFET is a thing that acts like a transistor but uses the electrical fields generated by the metal oxides instead of physical PNP or NPN junctions in the dope inside the device. You can get a MOS sensor to hook up to an Arduino or Pi for about $20,and they appear in full featured "air quality monitors" up to +/- retail $500. I am not a fan of this sensor for VOC detection for a number of reasons beyond the intended scope of this thread.

I look forward to finding out where @ABMax24 particulate air quality reads before he builds a filter.

I'm definitely curious to see what I end up with for air quality. I already run MERV 13 filters in the furnace and do notice an increase in buildup on the filter when the stove is operating. It would have been nice to see a baseline reading before starting the stove, but that will have to wait until spring. I'm also curious to see what the indoor air quality is like when wildfire smoke blows into the area.

 

[Poindexter]

I am posting my first data set now that I have two particle counters telling me the same thing. I will come up with a more organized way to present data in the future. All particle counts are expressed in micrograms per cubic meter.

Briefly I did a cold start on my stove last night at 2200 with no visible smoke in the room and trivial readings for air quality in continuous mode. Once I had the loading door shut I set the Dylos to sample one minute of every hour and store the data. The $60 thingy from Amazon doesn't record data, just real time display. So cold start last night at 2200, then hot reload around 0830 this morning.

Sensor was on my chairside table 11 to 12 feet away from the closest vertical corner of the stove on the hearth. Today I only was able to harvest outdoor PM2.5 data for 0800, 0900, 10, 11 and 1200.

Worst measured air quality (particles only) was at 0900 with indoor AQ 2.56 mcg/m3 of PM2.5 with indoor PM10 at 0.6 mcg/m3, simultaneous outdoor PM2.5 at the NCore sensor seven miles from my house was 13 mcg/m3.

There are many many variables to chase down. I know perfectly well on hot reloads I am gettng both VOCs and particles coming out the open loading door. I will eventually set up a tripod near the loading door and hang sensors on it, but my hypothesis is the 'stuff' coming out of the stove on hot loads is a chaotic rather than smooth flow. Sort of like pouring cream into a cup of coffee, I expect to get perhaps a real high reading here, and then 5 cm away a very low reading.

I am not interested in peak values for particles and VOCs coming out on hot reloads, I want to know total particles and total VOCs. I have so far three ways to trial for hot reloads, two of which should produce lowest possible totals, the third should produce max.

 

[woodgeek]

You could get a cheap pressure gauge off eBay and see what kind of pressure the fan is developing.

Usually, fans don't move a lot of air when there is a back-pressure, and a squirrel-cage blower is more effective. This translates to higher cfm per watt.

 

[peakbagger]

30 years ago I owned a house in Wisconsin for a short time, the prior owner pulled a fast one and hid a hot air furnace issue and I ended up replacing it. I kept the belt driven squirrel cage blower and it came with me to NH. When I was building canoes and kayaks in my basement I generated lots of dust mostly cedar and it definitely floated around the house. I ended up building a wooden box out of scrap plywood with filters on either side and the squirrel cage blower inside. It exhausted upwards into the room. I got a case of filters for cheap at a close out sale of a Home Depot clone (anyone remember Home Quarters?). I had a regular prefilter on top of highest efficiency pleated filters they had (no doubt not up to current standards). It made a big difference when I was doing wood working. My guess is slowing down the blower and running it continuously would improve air quality if I had an issue. I run a wood boiler with a very aggressive draft so I expect I just do not have the air quality issues a stove operator would.

 

[Poindexter]

I just saw this, like ten minutes ago, on youtube.

To view this content we will need your consent to set third party cookies.
For more detailed information, see our cookies page.

Accept third party cookies

Now I got to register on youtube so I can message Troy Gardner and link him here and ask him if I should name this thing a Troy Gardner box.

Somebody figured out a couple years ago my Poindexter username was acutally a cartoon character. Of the little eggheads, Poindexter is the coolest name, but my favorite character was Egghead Junior:

To view this content we will need your consent to set third party cookies.
For more detailed information, see our cookies page.

Accept third party cookies

 

[Poindexter]

You could get a cheap pressure gauge off eBay and see what kind of pressure the fan is developing.

Usually, fans don't move a lot of air when there is a back-pressure, and a squirrel-cage blower is more effective. This translates to higher cfm per watt.

I could also put a water tube manometer on it to see when I am getting more pressure drop, I am not too worried about that. I am using an infrared pyrometer to check the motor temp, it has been stable between 69.5 and 70.5 degrees F so long I don't even check it any more.

One thing I will have to do is turn the fan off for a few hours while tracking temperature in the stove room and the back bedroom, to establish there really is a convective loop going. Every house is different of course, but I get the smallest deltaT within the conditioned space with the fan on low. My idea is to piggyback the filters onto the fan that is already running.

 

[woodgeek]

I've played around with manometers and the filter on my whole house return.... I also found that you can buy (used) gauge manometers cheap on FleaBay if you shop a bit, and they are much more accurate/repeatable than water manometers at the required pressures. I found that a higher quality (and MERV) 3M filter had a lower pressure drop than a cheaper, lower MERV filter. This confirmed what 3M specs online said, and that my HVAC guy who said my blower would blow up if I put anything heavier than a floss filter in (that I can see through), well, he was full of chit.

I was suggesting that you would want to verify that you were getting the design pressure drop on the filters, to indicate that you were getting reasonable CFM, or reasonable CFM per watt.

I wasn't worried about the motor overheating, I am sure residential fans are thermally over-engineered for things like being laid flat on the floor, etc. And when a fan is moving less than design cfm, the mechanical load usually goes down (bc output mechanical power is cfm*DeltaP).

If I was going to tinker with your system, I would perhaps make a guard on the intake of your box fan, like a circular hole on the intake side which was a bit smaller than the diameter of the blades. If the current rig is open over the whole square of the fan (can't tell) you could be getting a lot of recirculation of the air from the blades back into the box due to the back pressure. Improved CFM could be verified by manometer.

 

[Poindexter]

@peakbagger , @woodgeek , I am glad to have the extra eyes. I do agree a baffle behind the fan is likely to improve filter effeciency. Any kind of manometer to estimate filter usage/ indiacte filter replacement would be inexpensive and possibly useful.

However, my goal is to put the best possible air quality into as many homes as possible, as cheaply as possible. I got to zero PM2.5 without a baffle on the fan. When the AQ readings on the $60 meter start climbing, it is time to replace the filters. When the filters get cruddy and the baby's nose is running again, time to change the filters.

I really don't want to make the filter array more complex or more expensive unless I just can't get to and maintain excellence.

But I do appreciate the pointers. You both clearly have a better understanding of this process than the average bear.

 

[Brian26]

I've played around with manometers and the filter on my whole house return.... I also found that you can buy (used) gauge manometers cheap on FleaBay if you shop a bit, and they are much more accurate/repeatable than water manometers at the required pressures. I found that a higher quality (and MERV) 3M filter had a lower pressure drop than a cheaper, lower MERV filter. This confirmed what 3M specs online said, and that my HVAC guy who said my blower would blow up if I put anything heavier than a floss filter in (that I can see through), well, he was full of chit.

I was suggesting that you would want to verify that you were getting the design pressure drop on the filters, to indicate that you were getting reasonable CFM, or reasonable CFM per watt.

I wasn't worried about the motor overheating, I am sure residential fans are thermally over-engineered for things like being laid flat on the floor, etc. And when a fan is moving less than design cfm, the mechanical load usually goes down (bc output mechanical power is cfm*DeltaP).

If I was going to tinker with your system, I would perhaps make a guard on the intake of your box fan, like a circular hole on the intake side which was a bit smaller than the diameter of the blades. If the current rig is open over the whole square of the fan (can't tell) you could be getting a lot of recirculation of the air from the blades back into the box due to the back pressure. Improved CFM could be verified by manometer.

I use the old school Dwyer Mark ii manometer to measure pressure drop on my hvac filter. Its only like $45 on Amazon and if its helped me save a ton of money on filter replacements. I also found the 3m 1900 Merv 13 to be an incredible filter. It has an extemely low pressure drop and can last for a really long time. I was also quite surprised how restrictive all the cheap low merv filters were as well.

 

[woodgeek]

@peakbagger , @woodgeek , I am glad to have the extra eyes. I do agree a baffle behind the fan is likely to improve filter effeciency. Any kind of manometer to estimate filter usage/ indiacte filter replacement would be inexpensive and possibly useful.

However, my goal is to put the best possible air quality into as many homes as possible, as cheaply as possible. I got to zero PM2.5 without a baffle on the fan. When the AQ readings on the $60 meter start climbing, it is time to replace the filters. When the filters get cruddy and the baby's nose is running again, time to change the filters.

I really don't want to make the filter array more complex or more expensive unless I just can't get to and maintain excellence.

But I do appreciate the pointers. You both clearly have a better understanding of this process than the average bear.

I love a good project, and I am glad that it works for you, and think that many folks without a forced air system could benefit from it. Waay better than those tiny $$$ 'Molekule' filters in the viral ads.

I just think most people with forced air can simply put a higher MERV (and rated low pressure drop) filter on their system and voila! Better IAQ. That was all I needed to do. Checked the pressure drop against the rated max spec in the air handler manual, and never looked back.

I think this is a case where the 'old guard' HVAC dudes saying 'Don't do it! you will destroy your system' are doing a massive disservice to the general population (lower IAQ and all it implies). Kinda like the same HVAC guys that talk people out of heat pumps, mini-splits and HPWHs bc 'they don't work' or 'are too expensive' doing a big disservice to the climate.

I have been told all of those things by most of the HVAC people I have let into my house. Including one old feller recently that I told I was (1) happy, (2) comfortable (3) saving money and that the systems (ASHP and HPWH) had been running happily for years with zero maintenance...and then he told me that I would never be able to sell my house, and that I should tear it all out and put an oil boiler and indirect tank BACK IN. And then told me that I didn't need to call him again for a '$50 checkup' bc my systems didn't need it!

 

[clancey]

Now that's confusing..lol poor guy--he is retiring I bet --good for YOU....your house price will even be higher have faith..lol clancey

 

[begreen]

Our heat pump air handler was setup with the factory 4" MERV13 filter and I have the 1" 3M MERV13 filter as a prefilter. This is as much for house air filtration for wildfire smoke as it is for winter use when the heat pump runs only occasionally

 

[ABMax24]

Got my monitor delivered today, so far I'm pretty impressed with it, was sitting at 1-5 ug.m^3 for PM 2.5 when I first turned it on for about an hour, then went outside and lit the pellet grill to cook supper, a little bit of smoke came in the door and right away it hit 76 ug/m^3, basically as quick as you can smell smoke the sensor is detecting it.

Will be interesting to see what it shows over the next few days, and even tonight when I light the stove again.

 

[Poindexter]

@woodgeek, I am not liking this new reply feature with the last software upgrade.

I stand by my previous advice that typical homeowners should not "upgrade" to a higher filtration filter without professional advice.

However, I do also agree a homeowner with a manometer and a written spec for pressure drop across the filter can run any MERV rating they like as long as they are within the specified pressure drop.

 

[Poindexter]

@ABMax24 , awesome.

Be advised, the little number bottom right corner on the main screen, the "AQI" number on these items uses the Chinese AQI calculation, which is different from the USA EPA spec (more conservative) and the World Health Organization (WHO) calculation, which is even more stringent.

What I am seeing so far at my house is PM2.5 is almost always the controlling pollutant for calculating AQI. I am seeing more PM10 with cold starts than I do with hot reloads, but day in and day out PM2.5 is the thing to keep the closest eye on, at least seems to be so far.

In the US, PM2.5 from 0-6 mcg/m3 corresponds to 'excellent' AQI, with AQI numbers 0-25. 6.1 to 12 mcg/m3 @PM2.5 corresponds to AQIs from 26-50, "good" in the USA.


There is a PM2.5 to USA AQI calculator here: https://www.airnow.gov/aqi/aqi-calculator-concentration/ that returns 14 mcg/m3 of PM2.5 as AQI/55 - unhealthy for sensitive groups, aka moderate.

I haven't found a WHO calculator yet, there is a decent wiki article here: https://en.wikipedia.org/wiki/Air_quality_guideline

with this quote

The guideline stipulates that PM2.5 not exceed 5 μg/m3 annual mean, or 15 μg/m3 24-hour mean;

 

[woodgeek]

True @Poindexter .... I don't advocate everyone just go out and get better filters and install them blind. I also agree that MERV 13 is plenty for otherwise healthy people with IAQ issues/concerns.

I AM saying that the problem you are aiming to solve (getting higher IAQ in people's homes) is important but was solved many years ago by 3M engineers, who made low pressure drop MERV 13 filters. And the rollout of that solution, both for existing homes and new construction, has been largely blocked by HVAC technicians, out of obsolete experience and a blind desire to avoid callbacks for any reason. Those same techs on the HVAC forum that told YOU not to install a MERV 13 on your system.

Sometimes professionals prevent the adoption of simple solutions.

How could this be fixed? Well, many of those HVAC pros have gauge manometers, and know the nominal cfm and rated max pressure drop for every air handler they install. If the cfm (per square foot of filter) is below a threshold, and the max pressure drop above a threshold, then they KNOW that a 3M MERV 13 filter is aok and they could just tell the homeowner that. The govt could help by certifying high filtration filters that have a pressure drop below some threshold as 'gold rated' or something. The HVAC guy says, you can put any filter in that had a 'gold rating'. DONE.

I am telling you that 100% of the air handlers installed in the last 10 years can handle these filters....they have LOWER drops than the low MERV filters from 20 years ago.

Have an ad campaign 'Have an IAQ problem, call your HVAC guy for advice'. Your HVAC guy get s a damned callback and then googles your air handler part number, and replies 'Yeah, you can put in a gold filter' perhaps without even setting foot in your house, if he has the install info.

Folks doing an energy retrofit (often free from their utility)... well IAQ is ON THEIR LIST of improvements. They could do the checking of air handler specs, pop in a gold filter, give the homeowner a years supply free, and tell them 'only put in filters with the gold label rating'. As it is, they are giving some homeowners freestanding HEPA filters with expensive filters, that only treat one room, and which cost way more than a years supply of filters.

I am sure we will get some version of the above someday....when HVAC pros start to see IAQ as a serious, rather than nuisance, issue.

In the meantime, your box is a happy solution, and waaay better than the tiny boutique and super spendy 'air purifiers' being sold for hundreds of dollars.