Printable View
Hello
this will be my first attempt to use injectors in a phase change system, today only pics of the main parts i will use (they are very high q- if someone wants i can send him 2048x1536 res pics, 1.3mb each)
Ok, first: the vital parts: evaporator base and injector
How will the injector be place on the base, it will be lifted a few mm so the freon flowing out of it spreads on a wider area
The injectors connection is a banjo joint, ill simply drill a hole through the screws top and braze a very thick cap tube to it
btw. its a startup injector from a ford escort rs turbo
And my favorite pic (the best pic a took in the history of my digi cam, finnaly learned how to use it)
Tomorrow ill post pics of the rest if the evap and probably tomorrow, the day after ill test it with my lil danfoss sc10b.. shuld own my cap tube configuration ;]
Oh and one more thing ^^
Me and the guys from Poland work on projects with injectors quite hard, were more advanced on it than with cap tubs etc. and probably soon there will be a small flood of refrigeration systems from poland using injectors ^^
Some pics of my friends work, Radek_Marko:
http://www.antyksklep.com.pl/FC/wtrysk/IMG_5664.jpg
http://www.antyksklep.com.pl/FC/wtrysk/IMG_5665.jpg
The software were writing to controll injectors:
http://www.antyksklep.com.pl/FC/wtrysk/program.jpg
This a alpha 0.001 software, the more advanced will read temps from multiple sensors and adjust to the conditions based on scripts we will write, were planning with Radek to use cfg files and a "console" (quake 3 style :P) to load certain cfg filed (like: thermal paste burn in, low temp reading, 100W load, 150W load etc)
Very cool berkut (no pun intended)!
I like the idea if the practicality works out...you'll definitely be able to utilize the interface you guys are working up. Best of luck.
Looking forward to hearing more.
..... yes this injector system is interesting.
I am really following this with utmost interest. I also need another way of injecting refrigerant than a captube (and I need throttling over a short distance for a reason I won't explain yet), but I am investigating the road of converting a TEV into a heat-motor operated EEV and such. Also, the Alco EX2 pulse-width valve with orifice X might be useful, though it's recommended cycle time (6 seconds) is too long for our small evaporators.
Would a system like this (if it were capable of detecting cpu load with the software) be able to be adapted to allow a system to be charged for optimum temps @ load conditions?
what I mean is, what it allow you to charge the system for full load conditions and then reduce the flow when under little / no load so you didnt slug the compressor.
Im not quite sure how it would behave tho. Would it behave like an overcharged system (when under little to no load) or, would it behave like a system with a too-restrictive captube. or both?
Obviously either way idle temps are going to suffer somewhat, since there will be too much refrigerant in the system for the load, but would that adversely affect the system in a negative way? (cause increase in highside pressure?)
Oh, just to add, I actually work for a ford dealership and I just checked the price of these injectors new.
Not good news. We're talking just under £130 + vat.
Something I have been wondering, some of you may or may not know, but nitrous systems use a similar design. They have a solenoid valve which allows the NOS to run through to a small injector. 1 benifit of this system I can see is that they are significantly cheaper than these injectors.
If I remember correctly, they are about £40 - £50 for the solenoid and the injector itself can be anything you want. As simple as using a small carburettor jet screwed into the end of the line from the solenoid.
Im not sure however, if these solenoids can respond quick enough for them to be a viable option as opposed to an injector, but its worth a thought.
What you basically want to control is the superheat. Thus, you measure the temperature of the refrigerant directly after the injector (or evaporator pressure), and the temperature of the evap outlet. You want this a few degrees higher since that indicates just no liquid returning from the evaporator.Quote:
Originally posted by Devilman
Would a system like this (if it were capable of detecting cpu load with the software) be able to be adapted to allow a system to be charged for optimum temps @ load conditions?
If you do this correctly, the system will always adjust itself perfectly for the condition.
TEV's work on the same principle, but TEV's are in fact a proportional controller. An electronic expansion device with a suitable PID controller can deliver faster reactions to load changes and better superheat control, leading to lower evaporator temperatures in our case. Even more important, they can work independant of the refrigerant.
Hmm, perhaps I interpreted berkut's idea incorrectly, but I had asumed from the picture that the injector itself would be used directly in the evap, rather than having a small piece of pipe from it, to the evap.
If that is the case, it would be impossible to messure the temp (or pressure) of the refrigerant after the injector, since the injector is mounted directly in the evap and spraying the refrigerant into it.
Also on another note, I would question the ability of these injectors to handle the pressure involved. Bare in mind that these injectors were designed for fuel systems, which at most, usually only have about 3bar pressure. It could be quite possible that once the system is charged and under pressure it will blow the "guts" out of the injector, straight into the evap :eek:
No, you did not misinterpret his idea. I have no idea on what principle berkut bases his control system. What I mentioned is how a simple EEV controller could (and does) work.Quote:
Originally posted by Devilman
Hmm, perhaps I interpreted berkut's idea incorrectly, but I had asumed from the picture that the injector itself would be used directly in the evap, rather than having a small piece of pipe from it, to the evap.
I am not too worried about the injector blowing, but more about the ability of the coil to lift the plunger (correct english?) with much pressure behind it. Also lubrication of the injector, coil heat dumped into the evap and resistance to extreme cold worries me. These are the reasons why I am awaiting Berkuts results first before jumping on the injector bandwagon.Quote:
Also on another note, I would question the ability of these injectors to handle the pressure involved.
That is indeed a very valid concern. these type of injectors were never designed to operate at the pressures we would be putting them under. they may not open correctly, they may not open at all. It may blow the "plunger" of the injector straight out of the injector body.Quote:
Originally posted by DaBit
I am not too worried about the injector blowing, but more about the ability of the coil to lift the plunger (correct english?) with much pressure behind it. Also lubrication of the injector, coil heat dumped into the evap and resistance to extreme cold worries me. These are the reasons why I am awaiting Berkuts results first before jumping on the injector bandwagon.
No, at most it just won't open. Those injectors are built for a harsh environment, so they won't blow due to extra pressure.
The opening problem can be solved by applying extra current, btw. But then you are pushing the limits of the coil.
Personally I would prefer a piezoelectric injector instead of a magnetic one. These are much quicker and more accurate (4 pre-injections and 1 main-injection per cycle on some modern engines. At 6000rpm this means 50 * 5 = 250 injections/second.). Also, they are smaller, lighter and capable of handling extreme pressures.
But they are new, expensive and not yet available at the scrap yard.
Im hoping it will work, if im not mistaken its rated for 6.5 bar, but will stand 10 easylly (the safety margin), i think they work @ 5V so if they dont want to open i can always overvolt them to 12V (remember its cooled down to ~ -40C beacause it has direct contact with the evap, even without that it shouldnt burn) + its pulse operated...
Right now i feel preety damn misereble and i have a 2maths lesson in ~20min, after that ill start making the rest of the evap if i feel strong enough..
You are ill pretty often, aren't you? Too bad :(
Anyway, I expect the best performance with peak-and-hold type injectors. Literature (Bosch) I just read talks about leakage above 65 psi fuel pressure. That's bad.
This page sums some injectors and their factory flow rate. To get flow rate and required duty cycle at our conditions, keep in mind that flow increases with the square root of the pressure drop over the orifice. Thus, a 4 times increase in pressure doubles the flow rate. With a factory designed pressure drop of 3 bar, our refrigerant flow at 10 bar pressure drop will thus be approximately sqrt(10/3)=1.82 times the given flow rate in the table. Assuming that we need 4kg/hr of R404a (will be close, but I didn't calculate it) at load, we pass about 3240cc of refrigerant per hour through the injector, which is 54cc/minute. A fairly standard 240cc/minute injector will thus operate at approximately 22% duty cycle. This can be done. At idle our load drops, and the duty cycle drops below the 10%-15% range, where it gets hard to control.
I wonder if you are better off with an (eventually modified with stronger spring and overvolted coil) solenoid valve made for refrigeration and a piece of captube. That would solve the temperature rating problem, the pressure rating problem, and the additional heat load on the evaporator problem. This way you can also keep duty cycle into the 80% range for full load.
this could be very powerful if you could write software that changes the throttle of the injector according to cpu load. it would be like a txv but more practical to our use. i cant wait to see some results.
kinda off topic, but how did you make that copper base?
i think he used a dremel
OK, one of our systems was testes for the first time with a injector...
Short summary: injectors tend to "freeze up" when they spray freon so we had to give them a shock @ 22V. They also dont like to close, thats all i know right now from my friend Radek, he'll perform additional tests today if he didnt kill his compressor.
So u guys, pray for his compressor to live :(
Compressors don't die so easily. But I would have added an accumulator anyway while testing, and so should you.
As soon as I finished my apartment I will try to repeat this, but then with an Alco valve. Let's see how that works.
Ok, got some stainless steel banjo connectors and copper orings today:
This is a preety bad pic but you can see the custom made connector (its made of epoxy glue and cyanoakrylate glue)
A shot of the banjo joint ready to assemble ;]
And the last, unfourenetly crap pic showing all the gunk, crap, oil and everything you can find in gasoline was in the injector (its like tar and i could wash it out only by flushing the injector with acetone... on the begening before i discovered whats inside it didnt want to even flow but after i blowed through it with air (before cleaning) suddenly a drop so very thick oil / tar dipped out on me- it was cleaning time :D
Oh and here's a desert for u guys, a snapshot of my friends Radek system running :
http://www.antyksklep.com.pl/FC/Test5/IMG_5975.JPG
wow, crazy!
Yes .. crazy, but it's realQuote:
Originally posted by La1kr0diZ
wow, crazy!
I can't speak English very well, so berkut has posted photos of my system ...
I guess - me and berkut are the first in using car injectors in refrigeration systems
Very wild Berkut. Is that a run-of-the-mill fuel injector you're using? Also, how are you going to construct the rest of the evaporator around your base?
:toast:Quote:
Originally posted by radek_marko
Yes .. crazy, but it's real
I can't speak English very well, so berkut has posted photos of my system ...
I guess - me and berkut are the first in using car injectors in refrigeration systems
Its pretty much run of the mill tachyon. Its an Escort RS Turbo cold start injector.
Unfortunatly (as I posted previously) these retail new @ around £100, so the chances of using them as a legitimate item may not be ideal for a lot of people, unless you know of a scrapyard with lots of old RS Turbos in it :)
about the nitrous oxide solenoid deal, i dont think they would work, because they aren't designed to wok over extended periods of time. In both my N.O.S. brand Cheater kit, and my Nitroux Express Stage 6 kit, the manual says do not enguage the solenoids for longer than 17 seconds or they will burn up. So if its firing 500ms then off 500ms, yeah, u get where i'm going.
Plus nitrous soleniods work by having a plunger lift up that lets the nitrous oxide through the opening, and the opening is about 1/3 inch in diameter, which would defintly not work with freon. right?
What are the advantages of using the injectors vs. a standard capillary tube setup?
Tachyon:
Simply that you would be able to perfectly meter refrigerant flow to load conditions to get optimum temps all the time. (Thats the theory anyway)
How do you guys wire up the injector/how are you currently metering the refrigerant flow through the injector?
Injection is nice, but I'd rather be blown :)
whats that stuff growing off the copper tubbing? looks like mold;) jk nice work man ne eta on the main project now?Quote:
Originally posted by berkut
Oh and here's a desert for u guys, a snapshot of my friends Radek system running :
http://www.antyksklep.com.pl/FC/Test5/IMG_5975.JPG
can you explain how the injector is interfaced with the controller and how you rigged it up? btw why is the liquid line frosted?
how can you control throttle of the injector when your computer is not off? if it is connected to your main computer there would be no way to control it on startup becasue the computer wouldnt be able to function, do you have some sort of circuit made? im sure a 555 timer woud be good for somthing like this
variable throttle interests me greatly, please share what youve done in detail if you could.
ok, answers:
I got mine for free ;] And they cost about 10$ here on junk yardsQuote:
these retail new @ around £100
Heavy subcooling ;] The system wasnt correctly calibrated and it flooded the suction line and compressor a bit but fourtenetly Radek didnt kill his compressorQuote:
btw why is the liquid line frosted?
Ill answer that for Radek:Quote:
can you explain how the injector is interfaced with the controller and how you rigged it up?
The interface is the LPT port, Radek also build a controller wich is made of few tranzistors/mosfets, diodes and resistors + few caps, its very easy to make but i dont have a plan of it @ the moment and i dont know if Radek agrees to post it right now.
Radek also wrote a program to controll the injector and he runs it off a laptop, ill use that way also.
We use laptops to controll everythingQuote:
how can you control throttle of the injector when your computer is not off
To be honest we would like to work on it a lil so you guys will have a ready "manual/ walkthrough" how to make these systems.Quote:
variable throttle interests me greatly, please share what youve done in detail if you could.
Today ill be mounting my system (R22/R22) cascade and ill test how it works... it has a very high throughput so i thing it will work preety nice even with 1.5- 2 bar of pressure (aldough its designed for 6.5 bar nominal)
And about pressures and the injectors not opening up: the injector Radek uses is a Bosh designed for 3 bar but it works preety nice with a 11 bar pressure
Ok, Radek was first to use a injector, im second ;]
The system works INCREDIBLY good, exept a leaking oring its simply hard to belive that w/o any vacuum before running it worked...
The injector works @ 5V w/o any problems (Radek had to overvolt his Bosh from 12V to 22V) and closes with out any problems even with a -65C SST (i think i can assume that the needle and injectors head was @ -60C ?), when it closes its completely sealed, no leaks with a 4 bar head pressure (later i turned off the I stage to see how it works with a higher head pressure- also perfectly)... My system was filled with R22 on the I and II stage, kindoff dramatic but worked ;] Tomorrow ill mount the rest of the system...
Nice work berkut! Looks nicer then Radek`s setup!
Waiting for the temps now! :)
Thats a test setup, the evap is a hollow type... My "main" evap is being constructed and tomorro ill try to find a 2nd injector like that one... ~monday ill buy some cap tube, untill friday i should have my flex hose cutted and R23 @ home ;]
I dont see any sensors mounted any where
How are you (or will be) controling Superheat?
currently its all about making it work, when everything works - ill be happy and start making everything else, now ill be using a pulse generator / laptop to generate impulses...
If the injector seals the 'bridge' (for lack of a better word) between the high and low side, how quickly does the high side pressure rocket? Fairly quickly I should imagine? You wont want to keep the injector closed for too long right?
]JR[
nope, it acctually lowers a little as theres no load on the HX and when the temp lowers the SCT is also lower = lower pressures
but unfourtenetly there are problems with the freon flowing back from the cap tube i use to supply the injector so if you arent pulsing it a little all the time it takes freon about 1-2s to flow to the injector, but i think that a thinner, ~1.5mm id cap tube should help as there would be less capacity and thus quicker response
theres still some work left even with such simple things :\
nice work.
How to yo plan to mount it?
It looks very long and stiff.
it sounds like you will have to use that circuit (555 timer it sounds like) so that your computer is independant from cooling. no sense in having an extra computer just to control the throttle on pulldown. what is possible is using the circtuit on pulldown and somehow overriding the circuit once the computer is on and control everything from software on the local computer. i know the 555 timer puts out variable pulse depending on a resistance, such as an adjustable pot, this would be a nice thing to have to vary the throttle depending on load and desired temperatures. keep it up, i might go this route for my next setup.
I wont controll by pulsing it... It will be more like this:Quote:
it sounds like you will have to use that circuit (555 timer it sounds like) so that your computer is independant from cooling. no sense in having an extra computer just to control the throttle on pulldown.
Pull down:
Fully opened untill it reaches a certain temp and "freezes" up the whole system
Operation:
When the evaps are fully flodded i only open it once a few moments for a longer time to fill the evap so its nicelly filled up- no pulsing
Shutdown:
Injectors opened so the pressures can gently equalize in the system, filling the relief tanks... After a while in this state they are turned off and the soleneoid opens to fully equalize pressures and keep ot that way
Ok, and the pic's of the new injector:
Its a bad pic but its visible that on the new injector (on the left) the nozzle is slightly smaller (~0.6 vs 0.5mm or 0.7 vs 0.6-0.5mm) and that injector will be used on the VGA card
And here's a family picture:
Oh and for those who are searching for these things:
This is a KE-JetTronic (by Bosch) cold startup injector, it costs 10$ on a scrapyard in Poland (buy the banjo joint connection type)
And heres a short description on how to clean it:
1. If it opens:
Blow air through it (when connected to electricity- 5V DC should be enough), a small drop od "nagar" (dunno the english word- its like very thick oil mixed with gasoline and carbon and comes from gasoline) should flow out (if it doesent - your'e lucky).
Next, take something you can use to fill the injector with Acetone / Chloroform (you must do this out side so you wont hurt yourself), fill the injector with it, shake a little. Turn it on so the gunk can flow out. Now, fill it up when its on- it should flow freely though it. Finnaly, take a rubber hose, fill a part of it with Acetone (connect it to the injectors inlet before) and with the injector turned on blow acetone though it. Finnaly, fill the rubber hose completelly and with everything positioned so it can flow out through the injector- cycle it so the parts inside move and it cleans itself mechanicly) - you can repeat the last part a few times.
After that treatment it should work flawlessly.
2. If it doesent open:
Heat it up with a hair dryer to ~60C, after this it should open as the gunk inside isnt as thick as it was. Now, with the injector turned on, fill it with acetone and leave it for a few seconds, remember that it evaporates beacause of heat so you'll probably have to filler up at leas once. Now, after this it should work preety nicelly so you'll be able to use the method in point no.1
are u powering those with a pc psu or something else
yes, currently i connect them to my main psu (antec tp 480W) but ill use a dedicated psu
At the risk of sounding like a newb, what is the purpose of using an injector?
very nice berkut. when i first started my system my father asked it it would be possible. i told him i wasnt sure, and if it was the programming would suck to do. but i see you seem to be having good luck with it so i might give it a try. have you wrote a program to control flow or are you just varying the pluse yourself? be nice it you could program a program to adjust to perfect temps under varied loads.
The purpose of using an injector is being able to keep the CPU temperature completely steady (which arguably improves stability, reduces risk of cracking, and DEFINATELY prolongs life).Quote:
Originally posted by Chris_F
At the risk of sounding like a newb, what is the purpose of using an injector?
OK, theres been nothing from you Berkut or Rad in here for awhile,,,how about an update on your project? Hows it going, how'd it turn out??? R you still working on it? Any progress?
The project is halted for a while but you'll see some info when i finish the II stage of my cascade... Its a matter of 1/2 days
Cool, thanks Berkut. Looks like I butted in just in time to see the results! LOL....hope its going good. Cant wait to see! :)
http://paranoya.org/~berkut/Kaskada/Parownik/21.jpg
http://paranoya.org/~berkut/Kaskada/Parownik/8.jpg
Tomorrow's the big day..... FINALLY! On R290 dough ...
Good Luck!!! Cant wait to hear the results!!! :)