24vdc to 12vdc converter

[RobertaME]

So normally I can solve most electronics problems on my own, but this one has me puzzled and I created this account to see if anyone else has some thoughts.

I have a 24-26vdc power source that checks for short circuits on the load line before applying full power by using a 3vdc testing voltage. I needed to convert this to a regulated 12vdc, so I bought a uxcell stepdown converter off of Amazon. After wiring, I powered up the unit and read the diagnostic display on the power supply which showed the load as being shorted out, so it refused to power on fully.

Trying to determine the cause, (and see if I got a faulty converter) I accidentally had the positive lead from the converter come slightly out of the wire nut connecting it to the power supply, which created some arcing between the contacts. Apparently, the extra resistance was enough to let my power supply think it had a good load on the line, so it kicked the power on and the 12vdc side of the converter started kicking out a steady 12.1vdc.

So here's my thinking... the converter when it first powers on is nearly a dead short until it starts putting power on the load side of the converter. So in order to get the power supply to start putting out power I have to add resistance between the power supply and the converter that will drop out of the circuit once the converter is running. I know there has to be a way to do this, but I'm just not seeing it ATM.

Attached is a simple diagram of the circuit. Any help would be appreciated. (I feel like there's a simple solution to this that is just escaping me)

 

[Nigel Goodwin]

I suspect you're right, with the 3V supply the converter is probably drawing too much current.

I'm presuming you've got no ground/chassis problems? - the negative in and out are most likely the same connection.

 

[RobertaME]

I suspect you're right, with the 3V supply the converter is probably drawing too much current.

I'm presuming you've got no ground/chassis problems? - the negative in and out are most likely the same connection.

No issues with grounding. I show 0Ω between the controller's -24vdc input, the controller's casing frame, and the -12vdc output. (everything in the circuit is negative ground)

Any ideas for putting a resistive load on the line until it kicks up to 24vdc?

 

[ChrisP58]

I would try using a relay with a 24 volt coil.

The coil goes across the power supply. The relay contacts are in series with the converter.

The converter doesn't get connected until the power supply output is up.

What is the current between the power supply and the converter?

 

[Pommie]

Could it be a capacitor in the converter that looks like a short while it's charging?

Mike.

 

[Beau Schwabe]

Dont use a resistor.... instead use an inductor in series with the 24V supply and the step-down converter.... something like 100uH rated for 10Amps or so.

Something like this .... https://www.digikey.com/product-detail/en/abracon-llc/AIRD-03-101K/535-11331-ND/2660621

 

[RobertaME]

Excellent suggestions, but after talking to my father (who was an AF avionics specialist) he gave me an idea.

He suggested putting a resister in parallel with a SPST switch, both in series with the stepdown converter.

I took that one step further and put a 24vdc 4.8watt incandescent light bulb in parallel with a SPST switch, both in series with the converter.

Now when the power supply checks for a short with the 3v test charge it sees the bulb and turns on the 24vdc, causing the light to turn on. Because the light is in series with the converter, as the converter charges its capacitors and starts drawing power, the light dims to almost nothing. Then you throw the SPST into the ON position and it short-circuits the light bulb, giving an open 20A pathway to the converter.

It's not automated, but the power supply should be running nearly all the time, so startup sequencing won't be too big of a hassle. (you also don't want too big of a load on the 12vdc side while in series with the bulb or you could max out its current handling and blow the bulb... with a 4.8 watt bulb you shouldn't exceed 1.6 watts of load on the 12vdc side)

Sometimes you have to stop overthinking and see what works. (it also only cost me about $6 and a trip to the hardware store... work done in about 45 minutes to an hour, including mounting the housing for the switch and bulb... the brain sweat took me all day!)

 

[AnalogKid]

Looks like a solution to me. If you want to automate it, 2 resistor, a capacitor, and power MOSFET can act as a power resistor that starts out large and decreases to almost zero over just about any timeframe you want. You need 2 resistors because most MOSFETs are rated for 20 V max. from gate to source. What is the 12 V output full load current?

ak

 

[RobertaME]

The 12vdc max load is 40 amps. (20amps on the 24vdc line)

This is a high-amperage system for driving 12vdc motors, lights, and heating coils, so I needed a solution that would stand up to those levels of current.

I considered an induction coil solution as mentioned above, but the constant changes in high current on the 24vdc side as the converter kept pace with changing loads on the 12vdc side would have made that problematic at best. (potentially catastrophic at worst)

The next idea was to set up a relay using the 24vdc to close the switch away from an in-line high-wattage resister and straight through to the converter, but that would add an additional load on the 24vdc line just to keep it open and the added expense of a high-watt resister. (those puppies aren't cheep!)

Considering how infrequently I should need to power-cycle the power supply and re-initialize the converter, I think this solution works.

I also hope that this lets people know about the uxcell stepdown converters and how they might show as a dead short on initial startup. They're inexpensive, compact, and well-built... but they have their quirks.

YMMV. ;^)

 

[rjenkinsgb]

I'd use a simple "soft start" circuit with a limiting resistor, plus a relay that pulls in once the 12V output starts up and both bypasses the resistor (something like your lamp & switch) plus a second contact that switches the 12V supply on to the final load.

 

[RobertaME]

I'd use a simple "soft start" circuit with a limiting resistor, plus a relay that pulls in once the 12V output starts up and both bypasses the resistor (something like your lamp & switch) plus a second contact that switches the 12V supply on to the final load.

That's not a bad idea. The only downside is that while the 12vdc is engaged there is a parasitic draw from the relay to keep the circuit closed. (more on that below)

I wouldn't bother with the second contact switching in the 12vdc load. That would require the relay to have to handle twice the current than it needs to as the 24vdc side is only 20 amps max while the 12vdc side is 40 amps. A 12vdc 40a DPDT relay is about $40, hard to find, and has a coil draw of 167 to 275ma. (4-6 amp-hours per day) In contrast, a 12vdc 20a SPDT relay runs $4.50 at the local B&M ($3 online) and typically only draws 75ma while engaged. (1.8 amp-hours per day)

Since the 12vdc loads should all be turned off when starting up the DC system (except for the 0.5 watt background draw of the converter) there's no need to isolate the 12 load. If you need to have the 12vdc load run to always-on devices that draw inductive loads of 12+ amps at startup, then a relay won't cut it anyway... you need a good Contactor to prevent arcing and contact welding in your switch.

 

[rjenkinsgb]

If power loss is critical, using a smaller relay on the input only is the best option.

A big power FET would have lower control current, near zero - but even an eg. 80A one rated at 20 milliohms on resistance would be dissipating 8W continuously at 20A, far more than a relay coil.

 

[RobertaME]

If power loss is critical, using a smaller relay on the input only is the best option.

A big power FET would have lower control current, near zero - but even an eg. 80A one rated at 20 milliohms on resistance would be dissipating 8W continuously at 20A, far more than a relay coil.

Power loss isn't critical, but it adds up as parasitic power over time. (something I'm trying to minimize in this circuit, which is why I got the uxcell stepdown as it's 97.5% efficient below 20 amps)

If I wanted to automate the function I think this circuit would work:

$10 and a few days shipping for the parts and I could forget about the power-up sequence entirely.

Kudos to you for the idea! You win the internet for today. ;^)

 

[rjenkinsgb]

That's basically what I was thinking of, but no need for the changeover contact - just short the resistor with the relay NO, there is then no break or surge.

 

[RobertaME]

That's basically what I was thinking of, but no need for the changeover contact - just short the resistor with the relay NO, there is then no break or surge.

Right... just the way I have the switch short-circuiting the light in my current setup... like this:

Much cleaner and no switch-over time.