Trying to do the math on an affordable HPA system

Larry3215

Well-Known Member
The nozzles have the correct part number stamped on them 301AFD0.7-80.

The way I determined the opening pressure was to slowly turn up the pressure on the pressure reducer. They start slow dripping at around 30-35 PSI, and I start to get decent looking spray at around 50 PSI. There is no flow to get reduced initially, so the pressure should be equal everywhere in the system when the drips start. With just a single nozzle, I wouldnt expect huge pressure drops even at full flow.

The gauge is T'ed in right after the reducer with maybe 3 ft of 1/4" tube going to the solenoid, then maybe 2" between the solenoid and the nozzles.

Im wondering if air pockets in the system or some other slop - tubing pushing in/out of fittings, or expanding/contracting - could cause this? I didnt have all the lock rings installed, so Im going to repeat the test with them in place. I will also move the gauge to the solenoid as see what that says.

Speaking of solenoids, it also occurred to me to wonder if slow opening/closing could be part of the issue? They dont seem slow, but I would never be able to detect a few miliseconds of delay.

hmmmmm maybe I can do that after all. I will look at the video in the editor software. I can play it back one frame at s time with time stamps.....
 

Larry3215

Well-Known Member
Looking at the video one frame at a time, it looks like the time from the first appearance of any water until the spray looks like its fully 'on' is roughly .1 seconds on average. The shortest time was .07 seconds and the longest was .17, but most were right at .1.

The time for the water to go from full on to off was longer. The shortest time was .07, and the longest was .2, but two others were .17.

Those times are rough. Its hard to judge exactly when the spray is full on or just starting to diminish in the small preview window. Im going to see if I can rig up an LED to turn on/off with the solenoid. That will give me much better times I think.
 

Larry3215

Well-Known Member
Here is a second test I did with an LED panel I took out of an old underwater light project.


Looking at this one frame by frame, the mist becomes visible in the same frame that the light turns on up to at most .03 seconds later. Each frame takes up .03 seconds. That means the nozzles starts emitting spray some where around .03 seconds, or less, after the solenoid turns on.

It takes a lot longer to turn off, but it doesnt seem excessive to me. The average is .17 seconds with a single outlier at .2 seconds - exactly 1 frame extra. So it takes almost three times as long to turn off as to turn on. Total lag time is around .2 seconds +- .03 seconds.

I have no real clue, but that dosnt seem all that bad to me.

The key question is how much water am I wasting during that .2 seconds? I wont know that until the timer gets here and I can run some output tests - shooting into a plastic bag.

The new accumulator will be here tomorrow, and the timer the next day, so I can run the rest of the tests with the entire system up and in place.

Well, that assumes my luck changes and the parts arrive in good shape! :)

One other thing that I am very relieved to notice is that there are fewer drips now than when I did the first test. The only reason I can think of for that, is maybe trapped air - which might add to the lag - is getting purged? No matter why, Im glad to see that. There are still drips, but not nearly as many.
 

Atomizer

Well-Known Member
You`ll find a freshly charged accumulator needs a little time to settle down, a sample taken directly after charging may appear milky compared to a sample taken slightly later which is crystal clear.
 

Larry3215

Well-Known Member
I dont have a tank hooked up yet, but thanks for the heads up. I was planning to run at least a few gallons through the system before ever putting any plants in it.

Any comments or thoughts on the other stuff I posted above about lag times, or why my nozzles seem so different from yours?

Also, do you know if the ADV pressure is adjustable? I can see the ball part of the valve in the stem. I was wondering if the seat portion can be pressed in further to increase the spring pressure on the ball? I only have a single spare nozzle, so Im reluctant to experiment without some idea if its possible or not.
 
Last edited:

Larry3215

Well-Known Member
I just realized I forgot to add in the time it takes for the spray to reach full strength at the start, so the total opening lag is around .1 seconds on average, and the closing lag is an additional .2 seconds on average = .3 seconds total lag.

So, out of a one second pulse, that means somewhere around 30% of the mist will be out of spec, plus the drips that get wasted. I wish I knew if that was good, bad or in the ball park :)
 

Larry3215

Well-Known Member
Im bored waiting for parts, so I decided to do a 'dry run' on the flow test. There is no tank, and no timer, so this is very rough. I mainly wanted to see how well it works.

So I started by weighing a couple of 1 gallon ziplock bags. They weighed 13.3 and 13.2 gms each. My scale can only do .1 gm.

First, I ran the nozzles for 15 seconds with no breaks. Total weight of bag + water was 30.8 gms - 13.3 = 17.5 gms of water over 15 seconds for 1.16 gms/second. That works out to roughly 1.11 gallons per hour. Thats a little higher than the spec sheet says it should be at 80 PSI - maybe 10% or so, but in the ball park.

Second run was for 60 seconds with me watching the seconds on my phone to start/stop the solenoid. I did my best at timing 1 second ON and 1 second off for the full 60 seconds, so a total of 30 seconds of ON time. This time the water in the bag weighed 35.7 gms. That works out to 1.19 gms/second. Thats a lot closer than I was expecting it to be to the first 15 second continuous run. That run works out to 1.13 gallons per hour which is not nearly as bad as I was afraid of as far as extra water from drops.

Of course, my rough timing by hand might be off a good bit, so I will have to test again after I get things put together properly, but this is encouraging.
 

Attachments

Larry3215

Well-Known Member
By the way, what the heck happened to the weight of water???

Last time I checked - dont ask how long ago that was - 1 cc of water weighed exactly 1 gram! It was an agreed upon international standard - or so I thought.

I just looked it up, and 1 cc of water now weighs .9971 grams??????

Did an asteroid knock a chunk off the earth when I wasnt looking so it has less mass and less gravity? Or did the moon speed up the rotation period thus reducing the weight? Or has global warming caused all water to expand from the extra heat, so it has less density?

After I calm down I will Google it to find out who fucked this up, but in the mean time - WTF??? ;)
 

Larry3215

Well-Known Member
Ok, I figured it out. Some idiot(s) did fuck it up!

For some silly reason, some group of idiots decided to build a "Standard kilogram" and store it in France somewhere.

The only problem is, they fucked it up, and got the size wrong! So, their stupid 1kg "standard" weight is off! They apparently cant measure length x width x height any better than weights, and got it wrong. So now every drop of water on the planet weighs less than it should.

sheesh! :)
 

Atomizer

Well-Known Member
Flowrate testing is best done using plain water, not nutes cos they weigh more for obvious reasons ;) If the timer has a led indicator you can use that in the vid editting software to figure the timer accuracy. If the vid editting software has an visual audio stream you can use the spikes created by the solenoid on/off to figure the open duration. The typical solenoid response time is 20-50 milliseconds.
 

dstroy

Well-Known Member
By the way, what the heck happened to the weight of water???

Last time I checked - dont ask how long ago that was - 1 cc of water weighed exactly 1 gram! It was an agreed upon international standard - or so I thought.

I just looked it up, and 1 cc of water now weighs .9971 grams??????

Did an asteroid knock a chunk off the earth when I wasnt looking so it has less mass and less gravity? Or did the moon speed up the rotation period thus reducing the weight? Or has global warming caused all water to expand from the extra heat, so it has less density?

After I calm down I will Google it to find out who fucked this up, but in the mean time - WTF??? ;)
1cc=1ml=1gram

This is for pure water. Impurities have mass.

As far as I know, there has only been one slight correction to the standard and that was a long time ago.

If your results are different it is most likely due to experimental and measurement error.

Any definition that we assign to a quantity of something is arbitrary. Mass is an intrinsic quality, weight is not.
 

Larry3215

Well-Known Member
1cc=1ml=1gram

This is for pure water. Impurities have mass.

As far as I know, there has only been one slight correction to the standard and that was a long time ago.

If your results are different it is most likely due to experimental and measurement error.

Any definition that we assign to a quantity of something is arbitrary. Mass is an intrinsic quality, weight is not.
Thats exactly what I thought, but if you Google "How much does 1 cc of water weigh", you get this answer:

Capture.JPG

If you Google for why it changed, it comes up with the "Standard kilogram" which is being kept in France someplace.

I didnt bother to read all the details, but no - 1 cc of water no longer weighs 1 gram.
 

Larry3215

Well-Known Member
Flowrate testing is best done using plain water, not nutes cos they weigh more for obvious reasons ;) If the timer has a led indicator you can use that in the vid editting software to figure the timer accuracy. If the vid editting software has an visual audio stream you can use the spikes created by the solenoid on/off to figure the open duration. The typical solenoid response time is 20-50 milliseconds.
Good tips, and yeah, this is all with pure tap water for now.

The software Im using (microsoft Movie Maker) only has a resolution of 30 miliseconds, but I can get in the ball park with that.
 

Larry3215

Well-Known Member
Yeah, I finally realized that 30 miliseconds = 30 frames/sec. I know my phone will do 60 f/s if I can find it in the settings....
 

Larry3215

Well-Known Member
The new timers arrived today and they seem to work well. Once you get past the crappy instructions, and actually see how to do the programming, its pretty straight forward.

They have some features that I like too - like a "STOP" button, and an external signal input to trigger the timer or the switch if you need that. I may end up using the trigger to switch between day and night timing modes, but that is still up in the air.

In addition to being solid state instead of using mechanical relays, the other big difference is that these timers feed input power directly through the timer. In other words, the relay portion isnt an isolated switch. It is integral to the timer circuit. You feed 12 volts into the timer, and that same 12 volts is switched at the output terminals. It simplifies the wiring somewhat, but you need to keep that in mind if anyone is considering using these in another project. There are times when the power you want to switch needs to be isolated from the supply power, and these wont work in that situation.

After re-doing the flow tests with the timer, I am no longer worried (very much) about the drip issue.

I re-ran the flow tests with longer times. That should reduce any measurement errors to some degree.

I did one run with continuous flow for 30 seconds. That resulted in a flow rate of 1.103 grams/second.

Then I did a 2 minute run with 1 second ON and 1 second off, for a total of 60 seconds of ON time over 60 cycles. That resulted in a flow rate of 1.196 grams/second.

So, the drips are costing me roughly 6% extra, wasted flow that wont ever get to the roots. I can live with that easily I think.
 

Larry3215

Well-Known Member
Back to the mist quality. Im worried that Im not getting properly sized droplets. I re-did the mist test at 80 PSI. The timer was set to 2 seconds ON with a 1 second pause. I also moved the camera closer to get a better view.

As near as I can tell, Im only getting about 1.25 seconds of "hang time" after the nozzle shuts off.


I sure hope you guys can comment on the quality, and hang time of the mist Im getting. I am afraid these Hypro's are not up to the standard of the ones Atomizer had.
 

Atomizer

Well-Known Member
The testing gives you a real flowrate of 4-4.3LPH so i would use a 4.15LPH nozzle flowrate for throughput and tank run time.calculations. 1.25 seconds hang isnt bad for hpa. If you make the room pitch black and shine a bright light (10-20w led floodlight) into the chamber it will reveal the droplets you cant normally see ;)
 

Larry3215

Well-Known Member
Whew. At this point, I will take a 'not bad' with a huge sigh of relief! I will re-do the mist test with the room dark, and using my LED panel. Its the closest thing I have to what you suggest.

If Im understanding this correctly, that 4.3L/hr would be reduced by the relative on/off cycle times? If I am running say 1 second on, and 60 seconds off, thats an actual flow rate of 4.3/60 = .071 liters/hr?

The new tank claims to be 4.5 gallons, but they dont list an 'acceptance' amount. Until I do some testing to verify, I will assume that its going to be in the same ball park as the earlier calcs we did. So assuming aprox 3.8 liters total usable volume, thats 3.8/.071 = 53 hours between pump cycles for a single nozzle running 1 second ON and 60 seconds OFF, or 26.5 hrs for two nozzles, etc.

That 1 sec/60sec ratio also gives me a daily through put of 1.7 liters per nozzle.

So, I think the next thing to determine is how many nozzles it will take to get good coverage. Then play with ON times until Im filling the chamber. Then re-visit those calcs. Its obvious that one nozzle isnt going to work at all no matter how long I let it run. I have a feeling its going to take at least three, or maybe four, but I am hoping 2 might work.

Even if I end up with 4 nozzles, thats a weekly through put of around 48 liters or 12.6 gallons. I can live with that. Its about the same as I was using before with weekly rez changes.

Of course, those numbers may change a lot depending on how the tests go and what my actual cycle times end up being.

Thanks again!
 
Top