Convert ATX PS to Benchtop PS project problem

@jeffberk

Maybe the old car radio needs more current than the 12v from the ATX power supply can provide.  How much current does the ATX power supply indicate it can supply on the 12v line?  Should be a label somewhere on the ATX supply that might indicate that current.  How old of a car radio?  Do you have something else that needs a 12v source but might not be needing as much current?

There are actually two radios that are selected by a DPDT switch. The dash radio is a 1974 Hitachi KM-1520ZC that draws 200 mA. 

https://www.manualslib.com/manual/1027630/Hitachi-Km-1520zc.html?page=3#manual

There is also a Sony CDX-F5500 I'm installing in the glovebox that is being powered to retain its memory all the time but both radios cannot operate at the same time because of the DPDT switch. For some reason, I'm unable to locate the power draw for the Sony.

https://www.sony.com/electronics/support/res/manuals/3261/32619321M.pdf

The power supply is also powering two relays that switch the speakers between the two radios.

The ATX label says 14 Amps at 12 volts.

Jeff

@jeffberk Change the relays to MOSFETS, that will save some power, then use your VOM or a clamp on meter to determine the current of the second radio.

I've never looked into using MOSFETS nor do I understand how they work. I'll need to read up about them. 

I ended up using the car battery as a power supply because I needed to get this car back on the road. There's no fun in owning a 260z and just letting it sit in the garage.

I still need to get this power supply working so I might try various loads (12v light bulbs?) to see what the tipping point is. Would running the power through resistor be an option or will it just act as a short?

Hi @jeffberk,

   Sorry, but diagnosing problems 'by mail/forum' is far from easy.   Certainly, putting a 'known' load on the power supply is a good start.

------------------

I assume you are 'switching the power supply on' ..

Sorry, I am not being facetious .. PC power supplies start in a low current state  which ensures there is a low current 5V supply available .. the 'front panel' PC on push button two pins on the ATX connector to 'activate' the power supply from the 'standby' to 'full power' mode .. typically this PC switch connects the ATX pin with green wire to the ATX pin with black wire, but not all ATX power supplies are the same, especially those fitted by some well-known PC manufacturers, who commission their own variants.

Although, the low power state is essentially a 5V system, I don't know whether 12V would also be present at a very low current in the 'off'  state.

--------------------

It is possible your PC power supply is faulty, they seem to have quite a high fail rate, and sometimes fail in weird "half working" ways. I have a feeling, this might apply in your case, but lets try to be optimistic!

Unfortunately, fixing them is not generally an easy task, and if you are not used to handling mains power voltages, they can be lethal, sometimes even after they have been unplugged, so please take care if you open the box.

-----------

To do a 'useful' test you need to draw an appreciable current ... drawing a few milliAmps from a 14A supply will prove nothing.

Say you aim for 2A to start with ... then Ohm's law   R = V / I , becomes R = 12 (V) / 2 (A)  = 6 Ohms

But remember the wattage ... W = V * I = 12 * 2 = 24 W ... not a job for the usual 1/4W or 1/2W resistors!

Of course, you can get high powered resistors,

Something like (picture is only an illustration ... not a recommendation of a specific product):

But, for a 'rough, one-off' test, perhaps it would be easier to find a car 12V filament bulb ('traditional' type ... not LEDs, etc.).

In the UK, 21W bulbs (for cars built in the 'pre-LED' era) are commonly used for brakes and direction indicators. I presume that, or something close, applies internationally. You probably have one already, but if not a local car spares shop should be able to help.

And 21W is conveniently just under 2A at 12V.

The resistance of a filament bulb varies hugely with temperature .. a cold bulb will momentarily draw a higher current. I seem to recall up to 4 times, but even 2A * 4 = 8A, so your 14A power supply should not be upset by it.

Then, assuming the lamp glows brightly, check the voltage.

Unfortunately, if the lamp does not glow brightly, etc., perhaps your power supply has expired.

Good luck and best wishes, Dave

@jeffberk If you have a car with a 12V battery, why do you need an ATX PS. That would mean converting the battery 12VDC to 110VAC via an inverter, then the ATX plugged into the inverter converts 110VAC back to 12VDC. That is a lot of conversions, all with losses. It's a car radio being put into a  car, isn't it? Just connect it.

@davee @jeffberk It sounds like he is in 'standby' mode. Bill covered it very well. Do you know if he has a 'Standby' light when he plugs into mains? Then when he toggles the 'Power On' switch, does he have a 'Power' light? If he does, what does the spec sheet say for the max amperage and the 12V amperage? Connect the radio to a 12V battery thru the VOM in current mode to determine if he is in fact, overloading the 12V supply. I recall either the 5V or 12V was low amperage. My ATX PS is rated for 432W/36A of 12VDC. That should be similar to yours and is enough to power any car radio. Could you remove the two relays from the circuit as a first test? I won't be surprised if that is the cause of your overload. Either redesign that as one latching relay so there is no wasted power or use a MOSFET. That is a solid-state switch with virtually no holding current. The MOSFET needs specs of a 5V gate and maybe 20A current, although going higher causes no harm.

Here is link to Bill's MOSFET article LINK

Hi Ron @zander & @jeffberk,

   I agree there is a significant chance the power supply is in the 'standby' condition ... hence my section that started "I assume you are 'switching the power supply on' .."

I think there is also a significant chance that the PSU is faulty ... most of them are built to a price, and contain a lot of parts that are highly stressed and frequently fail, plus the many electrolytic capacitors which 'dry out' and age.

--------------

Unfortunately, all of the PSUs I have seen do not include any LEDs, they are part of PC case switches, etc., and/or on the motherboard, so a more direct measurement is required.

No 'sensible' and functional car radio would require enough current to overload an ATX power supply. ( i.e. I am excluding any 'over the top', hundreds of Watts devices that might be available for custom car mods.) I am also assuming the radio is known to be fully functional.

So my advice is to  start by checking out the power supply with the simplest load possible ... no switches,relays, Mosfets or whatever should be present for the test, as they add extra unknowns... as I described, just a resistor or a lightbulb acting as a resistor .. to make sure it is being switched to 'full power' (from standby) and can deliver at least the modest current of say around 2 Amps.

If it passes this test, then optionally increase the loading to say 10 Amps, but at lowish volume levels, I would expect a car radio to use less than 2 Amps ... so testing for about 2 Amps should at least be enough to establish that the radio should show reasonable signs of life.

When the power supply is shown to be fully functional, then the other advice, including using MOSFETs to replace relays, could indeed be useful,for the whole project, but start with ensuring the basic parts are functional before making the more elaborate plans.

Best wishes, Dave

It seems the OP has abandoned this thread but I want to make a stab in the dark based on my work with hundreds of PC power supplies in the last 40 years, work that predates the ATX standard for PSUs.

Some of them, I would estimate a small minority, require a minimum load on the 5V rail to keep running.  They will startup and produce voltage on all the rails for a short while and then shut down exactly as the OP describes.

@cecil Hard to say as the OP never answered the questions put to him. Since he was using Bills ATX video as his guide I assume he had an ATX supply and if he did either he did not connect the standby  correctly or indeed he has a bad PSU. The attached pics show where Bill mentioned as you say the need for a +5V small load (8-20 ohms 10W power resistor) plus the appropriate part of the wiring diagram. I will be surprised if we ever hear from him again.

@jeffberk Per Bill's video/Blog, here is how you provide the load, see post https://forum.dronebotworkshop.com/tools-equipment/convert-atx-ps-to-benchtop-ps-project-problem/#post-41667