Help with PSU (Temp control fan, load bank, & PWM circuit)

[()blivion]

OH! I have to do stuff? I'm cleaning 320GB of malware out of a 5200rpm HDD, (BTW, that means PAINFULLY slow going). Been at it for ~8 hours, 48% done.

PM's sent...


Edit: Oh yeah.... I'm paranoid. So I used THIS to give you guys my email. Don't take offense, it's me... not you.

 

[jocanon]

Sure ()blivion, don't trust us...no just kidding, I understand...I actually set up a dummy Yahoo email account long ago that I give out to everyone. I have a Gmail account that I ONLY give to family and friends and to this date (about 5 years running) it is 100% spam free Good luck btw with that HDD.

 

[ronv]

Think it's done. Thought of a few things:

No on off switch. didn't know what you wanted to do there.
No wire. I use 30 awg (wire wrap wire) but it's kind of small if you don't like that kind of thing.
No big wire 8 or 10 for the 50 amps.
Probably several things I didn't think of.

I used to build those nasty little disk drives.
So you speak PC. No wonder you know sound cards and such.

 

[jocanon]

I have 10 AWG wire already. Speaking of 10awg wire, does mouser or someone else you know of sell that silicon based 10awg wire that is really flexible? I have been buying it from an RC hobby supplier but I have never really shopped around other places to compare prices much.

I could use some more smaller gauge wire though, I'm almost out.

An on/off switch would not be a bad idea, at least for the op amps so I can leave them plugged in when not in use, but I was hoping I could wire that into the on/off switch on the PC case I ordered. They usually do come with one, right?

 

[ronv]

Cool,
I have never bought wire from them. It just seems to expensive. I ususally just do Radio Shack.
What kind of RC stuff are you doing that uses 24 volts at 50 amps?
Sounds like we are getting there. Need to see what ()blivion's got.
Can you read the schematics okay?

 

[jocanon]

I fly an Extreme Flight Extra 300EXP. It is a 48" wingspan balsa plane. It is very aerobatic, fast, and a lot of fun. It uses 4 cell lipo batteries and we charge them in parallel so we can charge multiple at the same time and get back up in the air quickly. Some guys that fly bigger planes and helis use even larger lipo batteries.

I am a newbie at schematics, but I think I am getting it. Thanks asking. I probably will have specific questions for you when I go to assemble it.

 

[jocanon]

I have been getting wire from ep buddy online, it's $1.25 a foot for flexible silicone 10 AWG wire. For the internal wiring I do not think I need to use flexible wire, that's more for the connection to the DUT so it can move more easily. They have 10 AWG wire at home depot that is stiff but it's only about $0.29 a foot or so if I remember right. Could I use that to make the internal connections? I think it is household wiring that is used for lighting or something like that.

 

[ChrisP58]

Auto parts stores are a good source for wire. Available in a wide range of gauges and more flexible than THHN house wire.

 

[jocanon]

Thanks ChrisP58, I'll check that out.

 

[()blivion]

Need to see what ()blivion's got.

I can't add anything, your schematic looks amazing. Wonderful work. If I HAVE to complain, it will be minor and mostly irreverent. 100% me being OCD I assure you.

(1) Cleaning up the schematic for 100% readability would be nice. Such as move R6-R9 + U1 + V3 left of R1 sense.

(2) make a labeled wire off of the FET's that the third meter attaches too. For his voltage measurement point.

(3) The "x10" next to FET M2 is kinda vague. Maybe "x10 MOSFET's and sense resistors" could be used instead.

I was hoping I could wire that into the on/off switch on the PC case I ordered. They usually do come with one, right?

They are usually momentary on type switches. Not like a light switch. They are only "on" when they are being held down.

I am a newbie at schematics, but I think I am getting it.

We will walk you through the whole process. Most of what it is is transforming parts on the schematic to what you actually buy. Not very difficult... THIS is difficult.

Other thoughts
I'm thinking we need to choose a prototyping PCB for this project now. Something **broken link removed** or larger is going to be needed. We *COULD* make it smaller in the long direction by putting your MOSFET's on both sides of the pipe. But (A) that's A LOT harder to solder and wire. And (B) as ronv had said, longer pipe is going to be better for good thermal interfacing with the water. Room is cheep, so I Would just get the longer board. You can always cut off the excess anytime you want to. it will more than fit into a PC case. Area wise I could fit 8 of those boards in my case. By volume, including the components and the pipe, I could fir about 40. An Xbox 360 would have been plenty of room honestly.

As for some of ronv's points about wire sizes and current. With protoboard, you can run wire along the bottom of the board, and solder it to all the eyelets in long runs. And you can double or triple up on them for even more current. Your likely going to need at least three rows for these levels of current. I would use a nice even four rows myself. The pipe is your positive, and will be an excellent current carrier for you. So you only need one high current path ran on the solder side of your board. If you want every connector coming off your circuit board, you will need at least two.

Anyway, If that's what your going to use for the projects board, this is how you would do the "hard parts"...

(1) Tin (wet with solder) the backs of your MOSFET's carefully. (USE LOTS OF FLUX, CLEAN THEM WHEN DONE)

(2) Put all your MOSFET's in the board, down the center, lined up so the backs all face one way.

(3) On the solder side, bend the first pin forward, the third pin backwards, and leave the middle alone.

(4) A few holes down, bend the excess leads again to go BACK through the PCB to the component side.

(5) Bend the excess leads flat on top, making them sort of "clamp" onto the circuit board.

(6) Go back to the solder side, and solder down the leads to the board.

(7) On the component side, make all the MOSFET'S lined up very very strait and flat.

(8) Flatten your pipe, clean it *WELL*, flux it, heat it, and tin it. CLEAN IT AGAIN. (solder on any fittings as well)

(9) Reheat the entire pipe, drop on the protoboard and MOSFET's on the pipe. let it cool.

(10) The rest is standard circuit building procedure.

Notes:
Keep from bending the leads too many times. They will break from fatigue after about 5~15 bending cycles. you will need to get the backs of the FET's lined up quite well too, so this will be somewhat challenging.

If you want, you can clip the center pins off of the MOSFET's, or leave them on. They are the internally connected to the metal backs we are soldering the the pipe. You can also solder them to their own run on the PCB if you want. Make sure you provide enough room for the MASSIVE run we are going to do for the source pins, and make enough room for the gates.

You can run the MOSFET's all on one edge, leaving at least 4 rows of holes from the exact edge. This will give you more room for building your other circuitry. You will not need to do this at all though. You have enough room on that protoboard that you could cut it down the middle long wise and make two of these projects. It will be more long than it is wide, as you have seen with mine.

 

[jocanon]

ronv, one thing that has me confused are the ohm values of R4 & R3, if you add these together the value is 14.77 and then you have to also add the value of the 1K pot. Should these sum to a value that is less than 12 since the supply powering them is only 12V? I am probably wrong because I am just learning, but I remember when we were powering the op-amps off of 24V the restor was only 23ohms.

 

[()blivion]

I remember when we were powering the op-amps off of 24V the restor was only 23ohms.

Don't put to much stock into my numbers, I'm not a very exact person normally. I literally just picked numbers that worked. A perfectly zen and symmetrical setup that was "by the numbers" would have the sense resistors at 0.1 ohms. And as you said, R4 and R3 would add up to exactly 12K.

Do note also that 12K is different than just 12.... about a thousand times different.... actually.... EXACTLY one thousand times different. He He.

 

[jocanon]

Right . I need to proof read better when typing on my iPhone should have said a resistor that was 23K ohms.

 

[ronv]

I think ()blivion made it so it came out right in the middle of the pot. I tried to make it so it could not be adjusted to more than about 5.5 amps. Speaking of that I made an eupps as Rick Perry would say. The way it was the alert would be on all the time.
Here is the change. Now a 5K pot and 71.7k resistor. I changed the bill..
(), does this look ok to you?

 

[jocanon]

I think ()blivion made it so it came out right in the middle of the pot. I tried to make it so it could not be adjusted to more than about 5.5 amps. Speaking of that I made an eupps as Rick Perry would say. The way it was the alert would be on all the time.
Here is the change. Now a 5K pot and 71.7k resistor. I changed the bill..
(), does this look ok to you?

I see. So does 5.5 amps on the op-amp translate to 55 amps of current load on the PSU?

 

[ronv]

Yep, 5.5 amps per each FET or 55 amps total. I'll make a little write up on what ech section does.

 

[jocanon]

Thanks ronv. I wonder if I should make it so that it can go to slightly over 55 amps since that is the over current cut off limit of the PSU I am testing...would it be beneficial to be able to test that the over current cut off is functioning properly? Or would you recommend against that?

 

[jocanon]

()blivion, I ordered 10 of those PCBs you put the link up to because they are so cheap. I am just going to cut them to shape and use them for building the temperature controlled circuit for the fans in each of the PSUs, but since I don't really want to wait for shipping from China to start building the load bank, would this work just as well (the guy is local to me so I could pick it up tomorrow). **broken link removed**

 

[()blivion]

I made an eupps as Rick Perry would say.
[sinp]
(), does this look ok to you?

Lol, It looked fine before honestly. I understood the little box thing. You gotta make stuff up sometimes when spice doesn't have the model for the part you want. You can substitute a 1k2 resistor for the alarm in the schematic. That will make a good approx for **broken link removed**


Edit: OH OH OH, NOW I see the resistor change. Yeah, that's fine. I would have to simulate to know if it would buzz all the time or work 100%.

 

[()blivion]

()blivion, I ordered 10 of those PCBs you put the link up to because they are so cheap. I am just going to cut them to shape and use them for building the temperature controlled circuit for the fans in each of the PSUs, but since I don't really want to wait for shipping from China to start building the load bank, would this work just as well (the guy is local to me so I could pick it up tomorrow). **broken link removed**

Yeah, those look fine. You just need to have at least 10 of the white grid spaces one direction, and three the other direction. You will put one MOSFET in the center of each grid space, giving one hole till a line on each side. This will put two holes between each MOSFET. Plenty of room. You only have to center the MOSFET on the long axis, you can put them all the way to either side on the other.

Edit:
Over current shut off.
I went 18.3 Amps, just 0.300 milliamps more than the rated nameplate Amps on the 12v rail of a PSU the other day.... destroyed it. But I would call that defective merchandise.