|
|
|
PS3 Fan Fix V4.2
|
|
|
mzhussein
Newbie
|
14. January 2011 @ 13:19 |
Link to this message
|
Originally posted by KillerBug:
When you said you had "forgotten the second cap" did you mean C2 from his diagram? If so, it should not be needed...it was just to stabilize the unreliable "power on" signal we were using back then. Even the ne555 should be stable with a 5v vreg.
I'll have to try this out over the weekend...funny, once I moved past it, I never thought of applying the improvements from my pot-based design to his thermistor design. I might even give it a go on my slim...at 3.3V of course.
BTW...you seem to have access to a few more PS3s than I do, so perhaps you could answer me a question...
Are all the fan connectors on the units that you have worked with the same? I mean, are the plug/sockets identical, and the color scheme/pinout the same? The reason I ask is because I found a source for the connector for the fan in my old 40GB model, and I am going to check if it is the same thing as in the slim...if I can confirm that the 60GB and the 40GB have the same connector, and, if the slim has the same connector, then it is relatively safe to assume that all PS3s that are not within the 1-year warranty use the same plug, and thus I can make a plug-n-play design.
i forgot the cap labled C3, it WONT work without it. if you remove C2, it'll just mean you wont be feeding clean 5v into the NE555 (all tested on a scope)...yeh, ive got access to some 40's and 60's, they are all the same. i havnt had a look at an 80 yet. but when i do i will. i cant see a reason why sony would change the connector? its just a standard 3-pin
In your opinion where is the best place to put the thermistor...bear in mind it is sensitive, so dropping it on the RSX would just cause the fan to whack up, because thats always going to be "warm". i was thinking of using thermal paste and sticking it to the heat sink somewhere? i cant imagine the ps3 will ever need more than 40% duty provided there is good heat transfer. and for the most part 20% will be sufficient, so im looking for somewhere in the ps3 that is at about 30-40deg under normal load..
|
|
Advertisement
|
|
|
|
AfterDawn Addict
1 product review
|
14. January 2011 @ 22:51 |
Link to this message
|
I have already mounted thermistors inside of my system as part of my (now aborted) Auto-JB project. I used Arctic Adhesive to attach one to the center of the base of each heatsink (on the fin side of course). Yes, it will be a lot warmer than most places, but that can be adjusted for by changing resistor values...and it should give a nice reliable sense of how hot each chip is getting. Plus, being mounted on the other side of the heatsink base means that the temp should be somewhat stable (assuming that chip heat output stays the same).
Why would sony change the connector? Just because...sony does a lot of things "just because".
Oh, as to C2...that is the reason I always use TLC555s, even on 5V setups...it is designed to run on dirty power, low voltage, high voltage, even changing voltage in some cases. The price difference between the NE555 and the TLC555 is the cost of the cap, so you get all those other features for free plus you save board space.
|
|
mzhussein
Newbie
|
15. January 2011 @ 09:15 |
Link to this message
|
Originally posted by KillerBug:
I have already mounted thermistors inside of my system as part of my (now aborted) Auto-JB project. I used Arctic Adhesive to attach one to the center of the base of each heatsink (on the fin side of course). Yes, it will be a lot warmer than most places, but that can be adjusted for by changing resistor values...and it should give a nice reliable sense of how hot each chip is getting. Plus, being mounted on the other side of the heatsink base means that the temp should be somewhat stable (assuming that chip heat output stays the same).
Why would sony change the connector? Just because...sony does a lot of things "just because".
Oh, as to C2...that is the reason I always use TLC555s, even on 5V setups...it is designed to run on dirty power, low voltage, high voltage, even changing voltage in some cases. The price difference between the NE555 and the TLC555 is the cost of the cap, so you get all those other features for free plus you save board space.
ahh i see, im not clued up at all on 55 chips, infact this is my first endevious using them....
if i wanted to alter GoToChkns design to make the duty cycle lower at room temp, what compenent would i need to adjust? mainly because i want to "fine tune" it to get it as quiet as possible. atm it never speeds up, which is telling me its never getting hot, so is running too high. id like to slow it down a bit. also i dont think ill ever need the higher end of the duty cycle range....ive glues my thermistor inbetween the heat pipes on the heatsink, seems to be a "cool" place rather than a hot place.
This message has been edited since posting. Last time this message was edited on 15. January 2011 @ 09:16
|
AfterDawn Addict
1 product review
|
15. January 2011 @ 09:53 |
Link to this message
|
I'm not much of an electrical engineer, but I think that a change to R3 should change the range as well...but I'm not really sure on that; as you see, I don't use pin 5 at all in my design...and this might be the first time I have ever seen someone use it with a resistor instead of a capacitor. He is almost using the chip in reverse...I used the output pin as the output, and he is using the discharge pin. I connected the discharge pin to variable resistance...if I didn't know that it works, I would think it doesn't.
Here is the manual for the TLC555...Pages 9 & 10 might appeal to you...page 9 has a poorly designed PWM circuit that is easily adaptable to almost any task (as long as you don't need square waves). Page 10 has the instructions for adapting it to any task.
http://focus.ti.com/lit/ds/symlink/tlc555.pdf
[edit]
just looked at his design again...I think I know what he is doing...maybe.
Try putting a resistor in line with the thermistor. Something in the 30k-50k region should do the trick.
This message has been edited since posting. Last time this message was edited on 15. January 2011 @ 09:58
|
|
mzhussein
Newbie
|
15. January 2011 @ 10:52 |
Link to this message
|
Originally posted by KillerBug:
I'm not much of an electrical engineer, but I think that a change to R3 should change the range as well...but I'm not really sure on that; as you see, I don't use pin 5 at all in my design...and this might be the first time I have ever seen someone use it with a resistor instead of a capacitor. He is almost using the chip in reverse...I used the output pin as the output, and he is using the discharge pin. I connected the discharge pin to variable resistance...if I didn't know that it works, I would think it doesn't.
Here is the manual for the TLC555...Pages 9 & 10 might appeal to you...page 9 has a poorly designed PWM circuit that is easily adaptable to almost any task (as long as you don't need square waves). Page 10 has the instructions for adapting it to any task.
http://focus.ti.com/lit/ds/symlink/tlc555.pdf
[edit]
just looked at his design again...I think I know what he is doing...maybe.
Try putting a resistor in line with the thermistor. Something in the 30k-50k region should do the trick.
thanks, ill give it a try when i can do some more proto-typing at work.
I should be getting a look at an 80gb this week, and ill post up.
ive got some pics that ill upload aswell at some point.
thanks for the replies
|
AfterDawn Addict
1 product review
|
17. January 2011 @ 23:24 |
Link to this message
|
I have now confirmed that my 40GB Fat PS3 and my 250GB Slim PS3 both have the same fan connector, and the same pin-out. The only difference is that the fat uses 5V pulses of PWM, while the slim uses 3.3V pulses.
Also, I found that it is easy to get at the fan connector on the slim...two screws remove the power supply, and then it is out in the open and easily accessible.
I will have a couple new PCB designs within a few days to allow for the creation of plug-n-play units.
[edit]
Just submitted the design for a plug-n-play unit to BatchPCB; it is going through the bot now. It is the fully-internal design that is just set and forgotten; I'll design the version with a remote POT soon, and maybe even something using a thermistor (although the thermal epoxy is so much more difficult than soldering that plug-n-play would make little difference)
[edit2]
Here is a new design...this one is plug-n-play (or it is once it is soldered together anyway). This design is all internal...set the trimmer and forget about it. Full parts list and instructions are on the batchPCB site. I can make these up, but I need a lot of lead time, as I don't have any of the PCBs on order yet...and they tend to take 3-5 weeks to get once ordered. Similar design, but with a remote POT coming soon.
http://www.batchpcb.com/index.php/Products/51161
This message has been edited since posting. Last time this message was edited on 19. January 2011 @ 00:43
|
|
mzhussein
Newbie
|
20. January 2011 @ 05:21 |
Link to this message
|
|
ok, done more testing, seems that putting anything in series with the thermistor just alters the range response, ie put a big resistor in series and you get very little chnage with heat....starting duty cycle is unaffected...
BUT if you change R3 on GoToChkns design you can alter the starting duty cycle, after some playing the correct value for R3 is NOT 4.7K but ~6.7K THIS IS ONLY FOR A 40GB FAT MODEL, this gives about 35% duty cycle, which for some reason on the 40gb fan is quite quiet, ranging to about 60% duty cycle, which should keep anything cool, im gonna do some extended gaming tonight to test the usefulness of these circuits... i havnt tested it with any other fans, am going to crack open a 60 now and do some playing and come up with some values that work....
for the 60GB model the R3 value should be about 2.7k, but is ok upto 5k....for that reason, the ones i make for the 60gb version will have a tiny 5k pot in place of R3, so you can adjust a bit....
when my 80 arrives ill let you know how that differs...
This message has been edited since posting. Last time this message was edited on 20. January 2011 @ 08:13
|
AfterDawn Addict
1 product review
|
28. January 2011 @ 13:24 |
Link to this message
|
|
|
Reptile_16
Newbie
|
29. January 2011 @ 12:59 |
Link to this message
|
|
Killerbug, is there still an option to buy one of these pre-assembled? (I'm terrible with a soldering iron)
Cheers,
Reptile_16
|
|
Oddious
Newbie
|
1. February 2011 @ 20:40 |
Link to this message
|
|
Killerbug, is there still an option to buy one of these pre-assembled?
would love to be able to get one for my 60g Phat.
send PM and let me know the details. thanks
This message has been edited since posting. Last time this message was edited on 2. February 2011 @ 21:02
|
AfterDawn Addict
1 product review
|
2. February 2011 @ 02:52 |
Link to this message
|
I am still waiting for those PCBs...with luck, they will be here by the weekend and I will have 16 more pre-made controllers available for people (the big shipment of PCBs is still at least 3 weeks away).
At the moment, I don't have any preorders scheduled because of the delays that batchPCB keeps giving me...but I will start a reserved list right now:
Reserved orders from the next batch of 16:
Reptile_16
Oddious
The general rule of thumb is that shipping inside the USA is $5.20 for priority mail, and shipping abroad by air mail is usually about the same cost, just much slower (however, some countries cost a lot more). The controllers are $20 each for either slim or fat design.
|
AfterDawn Addict
1 product review
|
7. February 2011 @ 09:35 |
Link to this message
|
|
Update -
I got 28 boards, but is looks like most of them are bad. I will be finding a new supplier...two months to get a batch of defective boards is simply unacceptable. I made up 6 controllers using a few of the boards that looked perfect, but only two of these worked. Reptile_16 & Oddious get first dibs on these. I can't make any quality promises. They worked through testing, but I really don't trust these boards and I will be tossing the rest of them out I think. If you would rather wait about a month for me to find a new supplier, submit the design, and wait for shipping (or if you just want to take the design to a local nerd), I would understand.
Very annoying...I might have to jack the price up to $25 to cover the losses on this order
|
|
floppie90
Newbie
|
8. February 2011 @ 09:02 |
Link to this message
|
Hey guys i am following this thread for awhile and tested the 555 circuit on an breadboard with wires hanging out of my ps3 , it works great, but i just love testing and had another idea.
i use the ps3's pwm output as input on my circuit and have c3,r1 and P1 to control how much those pulses are extended, this way if the ps3 thinks it's too hot it will adjust the pulsewidth and my circuit will just follow.
with P1 short, the output pulse is same as input, and with P1 at max, output voltage was measured at 2.2V (ps3 just turned on)
all components are small, including the electric capacitors.
tested it on my 60GB phat
REG = TS2940CZ 3.3V (regulator)
C1,C2 = 10microF/16V electric capacitor
C3 = 470pF
R1 = 4K7
P1 = 47K
Diode = 1N4148
IC = cmos 4093
i was allready told i was overdoing it, so don't try this at home ;)
|
AfterDawn Addict
1 product review
|
8. February 2011 @ 11:56 |
Link to this message
|
|
It is an interesting circuit; if it works with other systems than the 60GB, it might make a good alternative to the thermistor-based fan controller designs. It looks like it should work with other systems, although it may need some slight tweaks to component values...it seems like the circuit you describe would be better for a slim; since the slims run on 3.3V fan signal pulses, while the fats all use 5V pulses (at maximum duty cycle of 66%, this reads as 3.3V on a multimeter). Then again, the fans in every fat I have seen are made by delta, and they may have used some of the circuitry from their retail 4-wire fans that support pulses anywhere from 3.3V to 12V. Of course, I have no way to confirm this without buying one of every fan model and then risking burning them out with testing...the datasheet was not published.
Personally, I am sticking with the manual control; the YLOD is a result of changing temperatures, and it seems that the whole idea of PWM control is to have changing temperatures. I prefer to just keep the system as close to room temperature as possible without the fan being so loud that it is annoying. Granted, your design should reduce the severity of temperature fluctuations, but they will still be there...and I am taking no chances. In fact, one of the new designs I am going to order as soon as I find a trustworthy PCB maker is going to have an internal trim pot...just set the fan speed as high as you can tolerate and leave it there. Of course the other design still has a POT in case you want to be able to turn down the fan while watching netflix or whatever. BTW...both designs will be plug-n-play...well, open the case, plug, close the case, and then play...but still better than soldering, and 100% reversible. :)
|
|
floppie90
Newbie
|
8. February 2011 @ 13:27 |
Link to this message
|
i looked at the leaked schematics from the ps3 and IC4002 that outputs the pwm pulses is running on +3.3V_EVER , don't see +5V MAIN on the powerport.
that's why i used 3.3V
and like i said, don't try my idea, it's over the top
|
AfterDawn Addict
1 product review
|
8. February 2011 @ 21:46 |
Link to this message
|
LoL...that is why I want to try it if I ever get some free time.
I didn't look at the schematics; I just used a 'scope to read the full speed using the "fan test" that the 40GB models have...multimeter read 3.3V, 'scope read 5V peaks, and a bit of math came up with a 66% duty cycle. I then confirmed that the fans could be swapped between fat consoles, so I guess I just assumed that all fats have the same 5V pulses...it might just be that the fan does not care about the voltage of the pulses.
|
|
floppie90
Newbie
|
9. February 2011 @ 14:47 |
Link to this message
|
i don't even have an scope ,and with an binary counter/divider ic i tried to calculate the pwm frequency from the ps3 and your circuit , i thought the ps3 must be around 24khz, and your circuit around 1.5khz.
not that this info matters anyway :)
|
|
bitbang
Newbie
|
12. February 2011 @ 10:40 |
Link to this message
|
|
KillerBug, would you mind posting the schematic for the all internal version? Or both.
This message has been edited since posting. Last time this message was edited on 12. February 2011 @ 10:41
|
|
floppie90
Newbie
|
12. February 2011 @ 10:47 |
Link to this message
|
if you mean his tlc555 version it looks like this
R1=27K
R2=80K
P1=100K
C1=100nF
IC=TLC555
|
AfterDawn Addict
1 product review
|
12. February 2011 @ 14:20 |
Link to this message
|
Thanks Floppie90...the diagram is perfect...but I think you missed the red text at the start of the first post...
R1 and R2 are both 10K.
The 7805 is a L4931CZ33-AP for slim units or a L4931CZ50-AP for fat units. It only needs to be a 7805 if you are using a low resistance pot (and the only reason to do this is if you really like the appearance of some pot you happen to have or something).
100nF = 0.1uF; not really a correction, just a tip for those having trouble finding 100nF caps.
For the completely internal design, the 100K single-turn mountable POT is replaced with a 100K 12-Turn trimmer POT. R1 and R2 are both removed; The R1 side of P1 connects directly to Pin 8 of the TLC555, and the R2 side of P1 connects to pin 2 of the TLC555.
If floppie90 would like to draw that, I would appreciate it (as you can see, I am no artist).
[btw]
Anyone looking to use a pot other than 100K should adjust resistors accordingly. They should be 1/10 the resistance of the POT. I tried this with a 250K POT, and it worked buggy, so I would avoid anything over 100K. Also, to minimize the parasitic drain on the fan power line, I recommend keeping the resistance of the POT over 10K. 100K is best as far as I can tell; all the units I make are 100K.
[also]
I am going to order a batch of 100 PCBs from Pad2Pad as soon as my check clears (probably Tuesday). I had to completely re-do the design to fit their silly rules, but hopefully it will be worth it if they:
A> Take less than 2 months to ship
B> Don't screw them up
I still have an order of 50 of the Plug-n-Play design with BatchPCB...god only knows when I will get them or it they will be done correctly.
This message has been edited since posting. Last time this message was edited on 12. February 2011 @ 14:35
|
|
floppie90
Newbie
|
13. February 2011 @ 07:20 |
Link to this message
|
i build my own diagram and used an experimental pcb , it's not adviced to use by others because it's easier to make mistakes, even i made some ...lol
i even soldered an wrong resistor in, found out when i tested the ps3 , had to take it apart again.
anyway here is your internal design , and i'm not an artist either ,still using an older photoimpact program to draw , not that smart but it works :)
about the fan wires , i was not sure if the 12v wire was brown, it could have been near purple.
REG = L4931CZ33-AP (slim) or L4931CZ50-AP (phat) (TO-220 package)
IC = TLC555
P1 = 100K 12 turn trimmer
Diode = 1N4148, but any diode will work i guess
C1 = 100nf
|
AfterDawn Addict
1 product review
|
13. February 2011 @ 22:55 |
Link to this message
|
|
Looks good; I can draw up a circuit board, but I can't seem to make something that is easy to understand like that. Note that the L4931CZ33-AP and L4931CZ50-AP are both TO-92 parts...the TO-220 uses too much space.
My first several of these were on experimental PCBs. I was actually able to make it smaller with an experimental PCB than with a printed PCB...but it took forever to make each one, and it is easy to make mistakes when you have so many bits and pieces overlapping. Plus, everything was soldered on; making external POT mounting a bit of a problem.
|
|
bitbang
Newbie
|
14. February 2011 @ 00:12 |
Link to this message
|
|
Thanks guys.
|
|
Oddious
Newbie
|
14. February 2011 @ 14:42 |
Link to this message
|
|
Great Looking work guys.
Going to wait on the controller Killer. Hate to hear you had trouble after waiting for them and paying for them. not to mention all the work you put into them.
I will follow the tread and keep an eye on the progress.
|
|
Advertisement
|
|
|
|
floppie90
Newbie
|
14. February 2011 @ 16:56 |
Link to this message
|
killerbug in my opinion there's enough space where the fan connector goes , so there's no problem using the TO-220 package , a picture from my build below , maybe it's the first philips ic that's inside an sony product? :)
|
|
