Limited range voltage doubler circuit

[gunrunnerjohn]

I have a little module that has a switching power supply module that operates from AC or DC (model train track voltage) of around 9 volts up to 18 volts. I'd like to be able to power this module from track voltage as low as 5 volts. It draws a maximum of 100ma at low voltages and drops to less than 50ma at higher supply voltages. The switching module is rated at 28 volts maximum input, and I can attest that it'll croak if you go higher!

The logical plan seemed to be a voltage doubler to raise the low voltage from 4-5 volts AC to a usable level. That appears to work great at low voltages, I can operate down to around 4 volts AC input power. The problem comes when the input power goes to 18 volts, the voltage doubler hits 40 volts on the output and smokes the switching power supply!

What I need to do is have a voltage doubler that operates normally at low voltage, but limits the maximum voltage to around 18-20 volts regardless of the input voltage. A key requirement is low cost and small size as this is going inside the model locomotive.

I've considered a simple resistor and Zener to limit the voltage, but that's the brute force method, and generates quite a bit of heat at high track voltages. I'm looking for something a bit kinder and gentler.

The floor is open for any ideas.

 

[EinarA]

The best solution is to replace the existing regulator with one that has a wider input range. If that is not possible, a step up converter can made be made that boosts low voltages and passes higher ones straight through. I do not have a circuit handy tho.

 

[gunrunnerjohn]

I've looked for a suitable module, but I have not found anything that isn't prohibitive in cost. This is not an "unlimited budget" type project, and I'll be building a number of them. Right now my module costs a bit over $2 and is a 500ma switching module in a very small footprint. I have not found a better choice.

The step-up is one choice, but anything I've looked at runs up the cost more than I can justify.

 

[EinarA]

I will need to know much more about what you have now and what is being driven before I could create a schematic. I assumed this was a one-time hobby project.

 

[gunrunnerjohn]

Actually, it's a little project that I built for command control O-gauge locomotives. The board works great in the target environment, constant AC track voltage up to around 18 volts maximum. I'm trying to adapt it without making a new board to run in a conventional environment where the voltage varies from around 5 volts to 18 volts.

The existing board uses a RECOM R-78E5.0-0.5 DC-DC converter as it's power source. It's DC input comes from a diode and 200uf of filter capacitance. This works dandy in the proper environment. However, it needs about 9 volts AC on the input to properly produce a full 5V output. I need the board to operate down around 5V AC coming in. The AC current required by the load is around 100ma maximum at the lowest operating voltage of 9 VAC.

 

[EinarA]

I assumed you used a bridge rectifier. Half wave and 200 uF is too skimpy for 100mA. Build a bridge from schotky diodes and use 1000 uF and the RECOM unit may work down to 5V AC.

 

[gunrunnerjohn]

I couldn't use a bridge on the main module as a requirement was the common of the power supply input is also common for the DC on the board. It's also why I can't put a bridge in front of the module. Much of the logic on the board drives stuff that has a common power return.

 

[EinarA]

I thought that might be the situation. I have not worked with model trains in almost fifty years so I am rolling blind, I will have to see full schematics before I can give good advice. A regulator with an isolated output might be required.

 

[EinarA]

It is still the case that 220 uF is too small. It will allow up 7V of ripple and the reg will drop out. Try wiring in a larger filter cap, 1000 uF, on a unit and try it out and report back.

 

[gunrunnerjohn]

I was thinking some sort of pre-regulator, I just didn't want it to be more expensive than the original board! I think I'll have to do the voltage doubling to generate enough voltage, and then limit it. FWIW, the Zener diode does work, but I feel like I'm killing flies with a sledgehammer.

 

[gunrunnerjohn]

The regulator drops out at about 8.25 VAC with a 1000 uf cap across the 200uf capacitance, that still doesn't solve my problem.

 

[EinarA]

That you for the quick response. Not what I hoped; will need more info before I can offer any more advice.

 

[Misterbenn]

I'd use a boost regulator with a shut-down voltage at 10V. With a boost regulator shut-down the diode just lets the standard voltage flow through...

 

[EinarA]

Considering that I have a homemade 5V .75A supply that is powered by a 6.3V 1A transformer the performance you report is very poor. Your time will be better spent finding a regulator that does work at a lower voltage. It is likely to be the same size and not much more expensive.
Misterbenn: I suggested that in my first post but the OP thought it too expensive.

 

[gunrunnerjohn]

If you have any suggestions for something that size and roughly the same price, I'm all ears. I didn't find anything. It pretty much needs to be a switcher, as the 100ma from a linear supply takes too much heatsinking at higher input voltages.

 

[EinarA]

Looking through catalogs is too much like work so I won't volunteer to do that. I do have an idea for a voltage limited doubler though. It probably won't meet your price/size limits so don't hold your breath.

 

[gunrunnerjohn]

I'll try not to turn blue.

 

[EinarA]

I have a basic circuit that uses two transistors and four other parts, untested as yet. It would double below a selected voltage, say ten, and then add ten volts to everything above that. I am trying to make it decrease this added voltage as the output gets higher.

 

[EinarA]

Here is a possible circuit. I don't have a variac so cannot do a good test therefore it is untested. (Tested now; will work if you put diode across C1, same direction as D2.) With AC voltage below ~7V peak Q2 is off and Q1 is on and the voltage doubler of C1,D1,D2,C2 works normally. Q2 turns on when the input drops below -7V, turning off Q1 and stopping charging of C1. As V+ goes above 15V, D5 turns on, raising the voltage on Q2's emitter and making it turn on earlier and so reducing the voltage on C1. Around 22V Q1 should be always off. Perhaps you will be willing to try it out.

 

[EinarA]

Above has an error: there needs to be a diode across C1, same direction as D2. Will do some more test tomorrow. ( Tests with various input voltages indicate it does work as intended.)