Dangerous Prototypes

I really like this new breakout board.  Building another PSU power supply is on my short list of things to do and you just made it very easy.  However, one suggestion I'd like to make is that in the next revision please consider moving the PSU connector (and possibly all the other components to the back side of the board.  This would make it considerably easier to build an enclosure that houses the PSU & breakout and all the other wires that dangle from a PSU.  The way the connector is now this would be difficult and would have the unsightly connector on the front of the enclosure.  Thanks! Keep up the great work!

It occurred to me today that the banana plug connectors can probably be mounted to the other side of the PCB...  I can't find any pics of the back side though.  Can anyone confirm?

You can mount them on the backside. The terminals are screwed to the board.

However visual indicators and on-off switch will be on the 'backside' :-)

[quote author="ianlee74"]I really like this new breakout board.  Building another PSU power supply is on my short list of things to do and you just made it very easy.  However, one suggestion I'd like to make is that in the next revision please consider moving the PSU connector (and possibly all the other components to the back side of the board.  This would make it considerably easier to build an enclosure that houses the PSU & breakout and all the other wires that dangle from a PSU.  The way the connector is now this would be difficult and would have the unsightly connector on the front of the enclosure.  Thanks! Keep up the great work![/quote]

there is a simple hack for what you want to do...simply unsolder the atx connector and solder the atx wires directy to the beard from behind...this is what I did on my ATX BB,...I cut into the PSU case, placed the board, and solderd the wires....:) it looks like a normal PC PSU with the BB as a front panel...I also unsoldered many of the extra wires from the PSU PCB, to keep the cluter down

pic of what I meant, the board is a previous version :)

---

Your fuses should really be set at something more than 1.25A, WAY MORE than 1.25A.
I have a 500W ATX power supply. The specs on it are:
+3.3V at 22A
+5V at 16A
+12V (output 1) at 15A
+12V (output 2) at 16A
-12V at 0.8A
+5V VSB at 2.5A

So the result is the fuses for the voltages should be
+3.3V fused at 22A
+5V fused at 15A
+12V fused at 15A (for all outputs of this voltage, since my PS doesn't identify which color code is output one and which is output 2 for 12V)
-12V fused at 800mA
+5V VSB fused at 2.5A

Unfortunately your box does NOT have user-customizable parts. This means your low-amperage fuses are permenantly attached. a VERY LOUSY product. I have a Ham Radio unit that draws 10A at 12V, and I was going to use your device as a way to let my computer ATX powersupply power my Ham Radio.. For this too work, I need access to the full 15A of the 12V output of my power supply, to make sure I have enough current to run the Ham Radio. Your fixed but resetable fuses (I assume these are actually circuit breakers of some kind because they are resetable), that allow only 1.25A of current, render your device (which is otherwise a superb device) useless for the SPECIFIC APPLICATION I'll be using it for. Now if you can get some 15A breakers for this product, that would be great. But until you do get such high-amp breakers, consider yourselves to have lost a customer. I will check OTHER COMPANIES for similar products who DON'T use too conservatively rated fuses in their product. If I can't find any, I'll make my own ATX power supply breakout box, out of parts from ebay, radioshack, etc.

This is primarily intended for running small embedded projects we build. And in case of a short the 1.25 amps will still fry the board, but 30A would fry the board, the table it was on and probably cause injury to the user...

While the ATX BB has fuses on it, the rest of the PSU connectors are free for you to use in any way you like...just connect your HAM radio to the CPU or GPU 12v rails and use the ATX BB as a overly complex switch. If you really want to increase the juice on the board, you can remove the fuses and replace them with a wire... I would not recommended this how-ever, as you wont have any short circuit protection. I also don't think the traces have enough copper to handle 15A without heating up...

[quote author="BenHutchinson"]Unfortunately your box does NOT have user-customizable parts. This means your low-amperage fuses are permenantly attached. a VERY LOUSY product.[/quote]
Any examples with user customizable parts? Kits will be kits, most do not come with different parts to keep the costs down. And this does not make them VERY LOUSY.

[quote author="BenHutchinson"]I have a Ham Radio unit that draws 10A at 12V, and I was going to use your device as a way to let my computer ATX powersupply power my Ham Radio.. For this too work, I need access to the full 15A of the 12V output of my power supply, to make sure I have enough current to run the Ham Radio.[/quote]
Hey, nice to see another ham here! :) I wouldn't use a 15 A fuse if the device is rated for that. There is a fuse in the device already. You know, overkill.

[quote author="BenHutchinson"]Your fixed but resetable fuses (I assume these are actually circuit breakers of some kind because they are resetable), that allow only 1.25A of current, render your device (which is otherwise a superb device) useless for the SPECIFIC APPLICATION I'll be using it for.[/quote]
Yep, resettable fuses. Here (http://http://en.wikipedia.org/wiki/Resettable_fuse) is how they work. It was not designed to support your SPECIFIC APPLICATION. But what you can do is get some higher current ones and replace them. You are a ham, you should have some soldering experience.

[quote author="BenHutchinson"]Now if you can get some 15A breakers for this product, that would be great. But until you do get such high-amp breakers, consider yourselves to have lost a customer. I will check OTHER COMPANIES for similar products who DON'T use too conservatively rated fuses in their product. If I can't find any, I'll make my own ATX power supply breakout box, out of parts from ebay, radioshack, etc.[/quote]
15 A breakers will cost more money, both as a parts cost and as possible lawsuits as people fry themselves and get angry. 1.25 A is used to protect the user, you will see sparks, but you will not be surrounded with sparks. Also it is real hard to put traces for 15 A on a board. That's why they use multiple wires on the ATX connector. Multiple paths for the current, less resistance, less power loss.

If you require a SPECIFIC APPLICATION, I advise you to make a SPECIFIC BOARD for that application. These are hobbyist kits for GENERAL USERS. There are schematics online for these type of boards, or just use the schematic here (http://http://dangerousprototypes.com/docs/images/9/9a/ATX_BB_SCH.png). All you have to do is connect "power on" to ground, maybe add a high wattage resistor to +5 V rail for stabilization and that is it! Cut all wires for +12 V (output 1), group them together and you have 15 A, easy right. I designed your board in 1 minute or so. And without criticizing your project and saying it is LOUSY because it does not fit my needs.

PS: Welcome to the forum BTW.

[quote author="arakis"]pic of what I meant, the board is a previous version :)[/quote]
I have this version. With 10 A fuses! (Yikes!!!) I like this design more somehow, not production friendly with all those extra connectors but more eye pleasing maybe? :) Do I see wires on PTC pads? Also where is your switch?

yup you see wires :D I didn't have any resettable fuses around, and no local store stocked them...I grounded the green wire inside the case, and I use the 220V switch on the back of the PSU to turn it on/of.
I also like the older board, it had 5 Grounds :D that made me happy, as I always seem to be missing grounds when testing projects, and then I have to use alligator clips in the most impossible places...

Thanks for the suggestions, especially from Ben, negative comments are always welcome and valuable feedback too.

The original design made by Sjaak did have the underside connector. It was really a tough decision which side to use it on, and we compromised on the right angle connector.

Ben, hello from another fellow ham (KB0TFS) :)

I wouldn't recommend this board for your project, or high power projects in general, because there's not enough copper on the board to carry those currents. We specifically chose a low current limit because PSU can be so beefy and most projects don't need that element of risk, we don't in our workshop where it was born.

I admit the fuses are lower than my ideal. They are polyfuses and the highest current that was widely available for the 12volt rail was a 1.25A fuse at 18volts. I would have preferred 2.5amps to 3amps, but the parts weren't available.

For a ham project I'd recommend just shorting the PSU on pin and tapping the wires directly.

A power supply that CAN put out 15A should in fact be protected by fuses RATED AT 15A, not 1.25A. Seriously 15 A fuses shouldn't be that much more pricy than 1.25A fuses.

I can't IMAGINE that a device that would burn out at a given amperage (such as 15A) would have fuses any LESS than that. (or any more than that either).

You see, typically fuses are selected for EXACTLY the same amperage that is the "damage threshold" of current for the device, so that current up to that level can be used. That way the MAXIMUM POSSIBLE USES for a given device can be realized. It is poor engineering practice to fuse a device at significantly LOWER current levels than the "damage threshold" Fuses should be no less than 90% of the "damage threshold". So for a 15A power supply, the fuse should be NO LESS than a 13.5A fuse. And the closer to EXACTLY 15A, the better.

You wouldn't burn the PSU, but you can cause much damage to the circuitry you are working on an yourself. That's why 1.25A fuses.

Imagine, if you will, you are working on a uC project that draws a maximum of 450 mA, hence the ability of being USB powered. You decide that you don't want to use your computer as a test subject for the smoke test as you want to eliminate the possibility of damaging your USB ports. So you use this ATX board. Lets say that the power traces are good for up to 1A and there is a short on the board. With 1.25A fuses on the breakout board, the maximum current will be 1.25A through the traces, and possibly they'll not be damaged. But if there were 15A fuses, suddenly there is 15A flowing through the power traces, blowing the short, burning the traces. And possibly because of the close proximity of the traces (as the board was not designed with enough isolation as it was a low power one) damaging some other traces and blowing some other components on the way until the polyfuse trips and no current is flowing.

So the fuses are there to protect you and the circuit you are working on, not the circuit inside the PSU. It already has fuses in place.

<sarkasm> I must have a LOUSY DESIGNED house. ^_^  The mains wiring into my house can probably support 100+ amps but still the fuses in my power distribution central is only fused with wimpy 10 amp fuses on the circuits.  I must hurry down and replace them with more beefy fuses so I can power my linear accelerator ring the day I finish it.... </sarkasm>

Why not just replace the polyfuses with a piece of wire if you want to max the PSU out? The PSU got internal protections over- or under-current.  But is there really a need to use the ATX breakout at all in a fixed installation like this?  Just cut off the connector and solder on new wires either directly into the PSU PCB or onto the already existing wires.

A power supply that CAN put out 15A should in fact be protected by fuses RATED AT 15A, not 1.25A. Seriously 15 A fuses shouldn't be that much more pricy than 1.25A fuses.

I can't IMAGINE that a device that would burn out at a given amperage (such as 15A) would have fuses any LESS than that. (or any more than that either).

I have to disagree here. The fuse is a poly fuse and pops when the rated current is exceeded for a certain amount of time depending on temperature, current, and voltage. It is to protect the project from the full brunt of the PSU.

I couldn't find any bigger (let alone 15A and cheap) surface mount poly fuses on Mouser, neither could our sourcers in Shenzhen. We could explore through hole for a future version. If you a link to an alternate SMD part I'd be happy to check it out too.

The fuses are ment to protect the circuitry and _not_ the PSU.

When you are using fuses it is common to use selectivity (1:1 translated from the dutch selectiviteit) The idea is that the fuse the closest to the problem will pop and the remaining of the stuff keeps wokring. For example normal households over here are fused 25A, 50A or 63A (depends on the connection to the grid) where the mains-cable enters the house, each electrical circuit is fused 10A or 16A. Electrical equipment are fused even lower. Even when the electrical powerplant is capable of supplying several MA (mega amperes).

If you need more Amperes from the PSU, please design an alternative pcb as the tracks on this aren't desgined for lots of current (they are beefy but not calculated/rated for large currents). Also the wires (mine are less teh 1mm2) cant supply large currents. PLease be wise enough to use fuses in your design though.

The tracks on the board are about 5 mm wide and those short runs in open air would be able to handle 12-15 amps without going more than 30 degrees C above ambient.  A bit hot yes, but not too bad. 

According to Daves experiments on his blog this could be increased to considerably more (double?) by removing the soldermask over the tracks and put a good helping of solder on top of the tracks.

Still, most people only need an amp or two so the PCB fulfills that need.  For larger currents cheap banana plugs and the most likely thin wiring is not suitable anyways.  For really large currents a proper PSU should have sense-wires that will compensate for the voltage losses in the cables and contacts - something that the PC PSUs are sadly lacking.

Can someone offer a good inexpensive source for the red and black terminals? Digikey or Mouser? Everything seems pretty pricey.

My best bet has been ebay, aliexpress, etc

I'd like to see a 9v section of the board.
I'm not an electronics expert, but I can solder.
Is there a voltage regulator to provide 9v without much heat?

[quote author="Royalblu"]Can someone offer a good inexpensive source for the red and black terminals? Digikey or Mouser? Everything seems pretty pricey.[/quote]

i have ordered a few double (black + red currently on a homemade power supply) from newark/farnel they were <$3.00 ea pomona brand too.

A 9 volt linear voltage regulator in the 78xx series  would not generate too much heart if connected to the 12 volt rail on the PSU.  But a small heatsink out just a piece of metal might be required if you will pull more than half an amp from the 9 volt.

The standard output voltages of the 78xx is of course 5 volt using the 7805 and 12 volt with the 7812. But there are also 7806, 7809 and 7815 for 6, 9 and 15 volts.

An example of a suitable regulator is this: http://my.element14.com/stmicroelectron ... dp/1087089 (http://my.element14.com/stmicroelectronics/l7809cv/ic-v-reg-9-0v-7809-to-220-3/dp/1087089)

[quote author="ian"]I couldn't find any bigger (let alone 15A and cheap) surface mount poly fuses on Mouser, neither could our sourcers in Shenzhen. We could explore through hole for a future version. If you a link to an alternate SMD part I'd be happy to check it out too.[/quote]

Digikey has a 3A 24V PTC fuse in SMD in stock, but I can't link to it since I have a new account.

Digikey P/N: MF-LSMF300/24X-2CT-ND

The through-hole stuff has somewhat higher holding currents, and a fair number are available if you're willing to drop the voltage spec from 18V to 16V, up to about 15A. MF-R1100, 16R1100GF, AHRF1500, etc.

-R C

Thanks. I found this at mouser and will pick some up on my next order:
http://nl.mouser.com/ProductDetail/Bour ... 22Y%252bnU (http://nl.mouser.com/ProductDetail/Bourns/MF-LSMF300-24X-2/?qs=sGAEpiMZZMuh1LHWSSg6K7ppsV22Y%252bnU)

Hi Ian I couldn't find specs for the V2, but from what i see from the Kicad screenshot you are going to add VADJ, are you going to use lm317/lm337?, there will be current measurement points or a micro to do that?

If I may, i would suggest to replace the lm317 with a LT3083 or LT1083, and instead of using a potentiometer(assuming that is what your going to use) use a digital potentiometer. Why all this?, the LTx083 will let you shutdown the Vadj output just shorting to ground the ADJ pin. So if we also are able to measure the current(i'm assuming that a current sense resistor will be there) we can drive the LDO and have CC and CV using an external micro to drive all.
Otherwise you could add and leave unpopulated a low pass filter +opamp to ADJ pin so if someone want can drive the LDO using an external micro.

seems like it exists some 3A fuses in 1812 packaging also:
http://www.ebay.com/itm/100-PCS-1812-3A ... 0940014947 (http://www.ebay.com/itm/100-PCS-1812-3A-PolySwitch-SMT-SMD-Resettable-Fuse-/110940014947)
3121 seem to be very big?

I am fairly committed to the 1.25 amp limit as a feature - about the same as your project gets from a TO220 7805. However, clearly the community wants more :) I guess I will pop a through-hole fuse footprint on the next version to make it easier to pop on a big cheap polyfuse.

Polyfuses that are used for this are 15volts 1.25 amp, I havent understood yet how these works, if I buy 30volts 1.25 amp polyfuses instead, will that be bad?
i would guess the fuse reacts to the wattage, 15x1.25, so if I buy 30volt it would be 30x1.25, which is considerably higher.
Will that be seriously bad if i do that? (I found cheap ones on ebay... :) )

It is the maximum breakvoltage (when the fuse is tripped)

It shouldn't matter. These fuse are basicly ptc (positive temperature coefficient) resistors. When lots of current flows through them, thry get hotter, temperature gets up inside. When the temp is too high (because of the current) the resistance will be infinite which is an open connection.

Since the ATX PSU connector and the board have holes to bolt the connector down, would you care to suggest appropriately-sized hardware for this?  Or zip ties, perhaps?  Or do you recommend against this?

Should the load resistor be flat against the board?  (The power supply I took out of an old Gateway Pentium 2 machine needs the load resistor to start up.)  I am planning on putting it in the sick-of-beige case for it.

I'm trying to resist suggesting a small breakout board for the PSU's additional connectors that plug into the disk drives. :-)

Anyway, thank you for this design, it looks nice and will be very useful.

... and I just noticed, as someone else has pointed out, that the standoffs for the case don't seem to be taller than the resistor. :-(

Since the ATX PSU connector and the board have holes to bolt the connector down, would you care to suggest appropriately-sized hardware for this? Or zip ties, perhaps? Or do you recommend against this?

I would suggest not to use zipties but nuts and bolds instead. lots of stress in put on the connectors when connecting and disconnecting the cable. M3 or m2.5 would work I guess. Although zipties are better then nothing :)

Should the load resistor be flat against the board? (The power supply I took out of an old Gateway Pentium 2 machine needs the load resistor to start up.) I am planning on putting it in the sick-of-beige case for it.

for the best heat dissipation you should mount it a cm away from the pcb. The resistor get quiet hot. Dunno if the sick of beige case can withstand the heat of it and if there is room for mounting it this way.

I'm trying to resist suggesting a small breakout board for the PSU's additional connectors that plug into the disk drives. :-)

I have bought a couple of bananna plugs and soldered different plugs to it (barreljack, .1"female header, probes) Depending on the DUT i use the cable needed. I have plans to make one with the drive connector too :)