Thought I'd put a schematic out showing the results of the various tangent threads I've posted as I've worked on this project.

Background: My ATX unit is physically too small to permit the traditional way of modifying with components mounted in or on the unit itself. The obvious solution that occurred to me was a sub-panel that I could connect to the ATX cable. I simply got an ATX extension & removed the male end, leaving me all the wires needed to connect to the sub-panel components.

Two items in the schematic that haven't been put together: the snubber on the Panel Switch (S1) & the Variable Out (still wiring it together).

NOTE: Snubber circuit removed. See further along in thread.

Would appreciate any review/comments.

As always, thanks to all. Special thanks to SgtWookie, Audioguru, Beenthere & KMoffet who have contributed immeasurably.

 

What is the goal, besides the obvious. What voltages are you trying to bring out, and at what amperage?

I'm a fan of using them myself. You can't get a cheaper module, with built in protection for overvoltage to boot. I see a lot of people talk about spending big bucks for one, locally I can buy a 350W unit new for $16.

What is the .sch format? My computer wants to load something called Schedule.

 

Bill_Marsden wrote:

What is the goal, besides the obvious.

Non-obvious goal: Foremost, skill development. I got to learn a lot in this project & it's helping me for the other things I'm working on or planning.

What voltages are you trying to bring out, and at what amperage?

DC positive are 12, 5, 3.3 @ 6A, 18A, & 10A respectively. There's also 0.7A on the +5VSB.

Both the -12V & -5 are .3A.

This should give me plenty for the proto small circuits I'm interested in. I'm not cure what the amps will be on the 0-22V variable: the math got beyond me when we switched the LM3914N & LED bargraph to run on negatives - but Sgt Wookie understood it. I had originally set the bargraph to run in BAR mode, but dropped to DOT mode @ the Sarge's & Audio's recommendation to save current on those negative rails. That's why there's no fuse indicators there either (besides - got the bargraph, right? )

I've attached a document of the faceplate I laid out for the final put-together (everything's a bit scattered on the bench @ the moment, ).

What is the .sch format?

That's Eagle for schematic. Eagle's the free download I got connected to throught this site. I have both Vers. 4.16r2 & 5.3.0. SgtWookie can edit the ones in Ver 4x. What you're getting is Microsoft Schedule when you try to open. You have to go into your control panel to change the default program for the .sch file extension (the Microsoft site has a step-by-step).

I intend to upgrade my ATX also for more "oomph." Keeping my eye open at my dealer's (lol, sounds like I'm becoming an electronic junkie) for the right buy. What I like about the sub-panel is it's plug & play.

 

For most projects you have a pretty good omph. You might think about putting some outputs that are fused just in case. Protoboards will melt, you know. Even the most experienced among us do make mistakes (no one I know of course... ).

 

Hiya Doug,
I'm not quite sure what you're doing around S1, but I'm quite doubtful that it's going to work the way you intend.

D1, R4 and C1 just don't make much sense to me. I can't see how they'll help anything.

As far as LED1 and R1 being in the ground path for the PWR (green) line, have you tried it? Does it work? I just measured between my PWR and GND lines with the supply off, and I get 2.8v. It'll turn on with a 470 Ohm resistor between PWR and GND, but not when you add in a red LED.

All of your pads have the critical information in the wrong field. You added the critical info in the VALUE field instead of the NAME field. None of that will show if you create a PCB from the schematic, only the PADnn info.

R2 and R3 have their reference designators too close to the VALUE field, which makes it difficult to determine what's what; ie: R21 ohm 10W.

I don't think you'll get much out of the POST line.

In your panel art, you've mis-labeled the -12V line as GND. This will lead to unhappy smoke-filled adventures if you should ever forget what it actually is, and try to jumper your "GNDs" together.

 

Okay, here goes (lol, )...

Bill_Marsden wrote:

You might think about putting some outputs that are fused just in case.

I thought I had. That's why I fused all the voltage rails.

SgtWookie wrote:

I'm not quite sure what you're doing around S1,...D1, R4 and C1 just don't make much sense to me.

Well, I wanted to try to make a snubber circuit. I went to this site: http://focus.ti.com/lit/an/slup100/slup100.pdf & gave it my best shot. I gather I have returned to "you're gonna blow yourself up" .

As far as LED1 and R1 being in the ground path for the PWR (green) line,

I have no excuse - I didn't measure it, but guessed it was 12V. It's obviously 5V. Darn. Guess I ought to flip the correct sized resistor to the anode too. Maybe the same for LEDs #2 & #3?

All of your pads have the critical information in the wrong field. You added the critical info in the VALUE field instead of the NAME field. None of that will show if you create a PCB from the schematic, only the PADnn info.

That was deliberate because of length of field. As long as I have the pad number, I can refer to the schematic (or write myself an assembly manual, lol). Alternatively, I could swap the two fields & shorten the names. Problem is to save space I think I'd end up with acronyms that were meaningless to all but me.

...reference designators too close to the VALUE field,...

Good point. Fixed. Will edit my first post with corrected version after the other bugs are ironed out.

I don't think you'll get much out of the POST line.

All I want is a lit LED indicating the POST was good. (I like lights. )

In your panel art, you've mis-labeled the -12V line as GND.

Again, could catch. Lots of copying & pasting happening there & that one got by me. Fixed to "0V".

BTW, I relabled the +12V_(yellow) on PAD18 to +12V_(separate line) because I intend to use one of the extra's coming out of the ATX rather than splice off the one in the main plug.

Also, every ground is a separate line (again, coming out of the ATX molex). I know that may seem silly, but in the ATX, they & the main coming in are also grounded to the case. Just thought it's better than one line with multiple runs.

Thanks for the review guys. When you have the time, please let me know how to fix that snubber. I really think one should be there, & based on what I read in that link, the DRC version seemed "the cleanest."

I know, I know..."you're gonna blow yourself up!" Sheesh.

 

SgtWookie wrote:
(snip)
Well, I wanted to try to make a snubber circuit. I went to this site: http://focus.ti.com/lit/an/slup100/slup100.pdf & gave it my best shot. I gather I have returned to "you're gonna blow yourself up" .

Nah, you're not. However, there really is no need for a snubber circuit there. Switch on, run for awhile, switch off. Were you trying to drive the gate of a power MOSFET at 10kHz, then yes, you'd need a snubber. Not here though - it's just making things unnecessarily complicated.

As far as LED1 and R1 being in the ground path for the PWR (green) line,

I have no excuse - I didn't measure it, but guessed it was 12V. It's obviously 5V. Darn. Guess I ought to flip the correct sized resistor to the anode too. Maybe the same for LEDs #2 & #3?

My point was that I'm not sure your supply will even turn on if you have a resistor & LED in your switched path to ground for the green wire. It doesn't take a lot of current to turn it on, but I'm not certain what the threshold voltage level is. Before you go to the trouble of making a PCB, you should see if a red LED with a 330 Ohm resistor in series will pass enough current to not only power the LED, but power up the supply as well.

All of your pads have the critical information in the wrong field. You added the critical info in the VALUE field instead of the NAME field. None of that will show if you create a PCB from the schematic, only the PADnn info.

That was deliberate because of length of field. As long as I have the pad number, I can refer to the schematic (or write myself an assembly manual, lol). Alternatively, I could swap the two fields & shorten the names. Problem is to save space I think I'd end up with acronyms that were meaningless to all but me.

You're the only one who really needs to know what they're for
However, on many schematics you'll see wire colors abbreviated to three characters; violet=VIO, green=GRN, etc. As long as you're consistent, there should not be any confusion.

BTW, I relabeled the +12V_(yellow) on PAD18 to +12V_(separate line) because I intend to use one of the extra's coming out of the ATX rather than splice off the one in the main plug.

Sure, why not? I left all of the MOLEX connectors & harnesses on mine so that I could use it to power computer peripherals, should I want to test them or something.

Also, every ground is a separate line (again, coming out of the ATX molex). I know that may seem silly, but in the ATX, they & the main coming in are also grounded to the case. Just thought it's better than one line with multiple runs.

I have no complaints about that. Each of my ground binding posts (7 of 'em) have a separate wire to them also.

Thanks for the review guys. When you have the time, please let me know how to fix that snubber. I really think one should be there, & based on what I read in that link, the DRC version seemed "the cleanest."

Well, I really don't think you need a snubber there.
What or where is the "DRC version"?

 

Hiya Sarge -

Well, I really don't think you need a snubber there. What or where is the "DRC version"?

It's gone. As for "DRC," my apologies, I was hurrying. LOL, it's an RCD (resistor, capacitor, diode) clamping style I was trying to pull off. It's on pages 6, 7, 8 of the PDF for the link I posted. Don't know what a DRC is - new rap group?

I also made a few other oversight changes, i.e., corrected LED colors, added some notes, fixed the Pad vs. Value fields (just in case I do try to make boards ).

Which reminds me - when an item uses a heat sink, how does that fit into the schematic? I can't make the one for the 317 fit.

I re-calculated R1 for 5V which brought it down to 220Ω. I'll measure it to be sure, but I work the next few nights. If the circuit worked on 470Ω sans LED, the new numbers should work. We'll see.

Anyway, thanks again.

OT: Oh, they were E & G models (Wild Weasels baby, oh yeah!). You darn avionics types always got to work up top while us toads were spearing ourselves on pitot tubes under the belly. Yeah, I hate ya. LOL .

I've reattached the corrected version of both the png & sch by edit in the first post - save some bytes.

 

It's gone. As for "DRC," my apologies, I was hurrying. LOL, it's an RCD (resistor, capacitor, diode) clamping style I was trying to pull off. It's on pages 6, 7, 8 of the PDF for the link I posted. Don't know what a DRC is - new rap group?

LOL! OK, that makes more sense now.

I also made a few other oversight changes, i.e., corrected LED colors, added some notes, fixed the Pad vs. Value fields (just in case I do try to make boards ).

Good. Haven't reviewed the updated schematic yet.

Which reminds me - when an item uses a heat sink, how does that fit into the schematic? I can't make the one for the 317 fit.

Well, they're primarily for when you are laying out the board. You could just stack a bunch of heatsinks in a corner if you wanted, or put the heatsink symbol near the LM317. If you're going to use the Autoplace_v3.ulp program, put the heatsink near the LM317. Autoplace_v3.ulp is available for download from the Cadsoft website; you run it from the schematic program to automatically put components in the board layout in similar positions as they are on the schematic. Makes for a good starting place.

I re-calculated R1 for 5V which brought it down to 220Ω. I'll measure it to be sure, but I work the next few nights. If the circuit worked on 470Ω sans LED, the new numbers should work. We'll see.

Well, consider running your LEDs at well under their rated current; they'll last much longer that way, and won't take as much power away from your supply.

OT: Oh, they were E & G models (Wild Weasels baby, oh yeah!). You darn avionics types always got to work up top while us toads were spearing ourselves on pitot tubes under the belly. Yeah, I hate ya. LOL .

LOL!
The E's had something the B's, J's and S's sorely needed: guns. The MK IV gun pod was available, but was pretty much a dumb idea, as the first thing a pilot does in a dogfight is jettison external stores.

I've bonked my grape against at least my share of pitot tubes. along with taken numerous "baths" in Coolanol-25 and hydraulic fluid.