How to limit HV capacitor charging current?

I have a voltage doubler, which is connected to mains (without galvanic isolation):

It enticed me with its simplicity, configurability and option between charging speed or desulfatability. The problem is that in real practice lamp eats 20-50% of wattage, what i want to avoid.

How could i modify scheme to get rid of lamp and raise its efficiency to 85-95%?

PS. It needs to be done by limiting current but not dissipating extra power in air as with lamp.

UPD. Nominals on upper scheme are wrong. here is scheme for one of simplier devices i have by hand, with exact soviet partnumbers:

UPD2. Before posting how inefficient and rude it is, and to understand what circuit do, and how low frequency high current SCRs works, you may use following working simulation, by tony EE and me:

• This looks like a really bad idea. Where did you get it? – Transistor Sep 22 '18 at 15:59
• @Transistor It is not idea. I had used this for years. This is very good and working solution. – user197437 Sep 22 '18 at 16:00
• @xakepp35 It looks very nasty. This is not a topology I'd ever use. – Harry Svensson Sep 22 '18 at 18:29
• I'm confused, why would you want to double the 220V to charge/desulfate a 12V battery? – BeB00 Sep 22 '18 at 18:36
• It was well before year 2000, I got a job where a great part of machines, electric end electronic equipment and other tools had Soviet origin. They seemed laughable. Coarse looks, all sucked 200% more gasoline and electricity than western counterparts. All, except the most basic tools, needed talented+well trained pros to keep them running - just, like your circuit needs a pro electrician to use it without causing funerals. Bit by bit the laugh changed to respect. Winter freezed nothing, everything worked in snow, sand and mud. Your solution to reduce losses = Inductive current limiting . – user287001 Sep 22 '18 at 23:23

Conclusion: Ultimately, this is NOT the best way to make a pulse charger. (Design too simple, compromises too complex. ) 1st make specs for input and output pulse current, rise time, duration. The best simple way uses the battery itself to make negative spikes then a simple battery charger to make up the 5W to 500W you choose for desulphation rate.

Anecdotal

You have something that works with a similar efficiency of any unregulated bridge diode-RC charger with a lot of low rep rate (100Hz) line pulse noise too. It ought to pulse at 20k to 50kHz rate with microsecond pulses and current depends on design but with 100W rapid and 5W maintain desulphation without acid electrolyte evaporation at > 14.2V. Basically, as we heard from an investor who wanted to buy my client's patent for \$1m, they had a blow torch but once trigger auto accelerometer sensor airbags nearby, we had a Bic Lighter that worked using only 5W to trigger sulphate crystal breakdown from harmonics.

=============================

Old info

I see better that you know what you are doing. Newbies, do not try this yourself without details on each component. Risks of doing it wrong are severe.

The bulb characteristics are 1:10 R ratio e.g. 10 to 100 Ohms (cold to hot) at rated power but heat rise time and cool delay time affects the pulse characteristics of this bulb pulse SCR charger. In some ways, the bulb heat modulation with the diodes mimics a "PTC semiconductor", current limiter, but not efficiently.

Conclusion> The only way to improve efficiency is to reduce ESR of expensive caps and insert bulky ferrite core very low DCR coil (LC filter) to reduce current and store energy. That is how Buck regulators store energy.

Reducing bulb resistance is the only way to reduce losses but that affects pulse width and current so they run cool.

This is my sim. You must change the values to your model.

My switch allows you to charge an 11V ideal voltage source with battery capacitance with a 1.3V charger using a modelled convenient rapid (fake) charger)

I support the theory and practice of pulse de-sulphation for. My former employer used to manufacture them and I verified the design and I know someone with 20kHZ 300A ??ns pulse battery rejuvenator that works extremely well.

This work ONLY for proven experienced technical people that understand battery and Cap ESR and that a short-circuited battery cell fuming hydrogen can be dangerous.

I cannot optimize your design, off the cuff, without these characteristics. But with this tool and your actual model, you might.

You can even change the battery 10mOhm ESR with an ideal pot of 100mOhm value and adjust it.

I've done a similar thing in my past life ;) in Winterpeg, Manitsoldhere, Canada with cold dead batteries and not having a charger at the time in the late 70's and even later.

To reduce 120V/60Hz to12V means I was 10% efficient so I used a 1500W toaster in series with "HOT" = Line1 120V. It got me to work on time in frigid weather. After a toaster, I used a single 25A diode in a glass of snow after the ~100V voltage drop with a careful connection of my neutral (near 0V earth) to chassis ground.

At least I could heat a bagel while waiting 10 minutes to get enough charge to start the engine.

• Please recheck schematics. Its manually controlled "smart" charger (as opposed to "classic" transformer 14.8V boilers). It outputs sawtooth current, with exponential raise and falloff, and aimed at dissolving sulfates. Rate is tuned by capacitance and bulb resistance. A light bulb limits capacitor charging current, giving typical spike rates of 4-20hz. And SCR unleashes current from capacitor almost immediately. Circuit is tweakable, it has 4 tunable parameters. Its MEASURED efficiency typically lies between 50 and 80%, typically 72% – user197437 Sep 22 '18 at 17:01
• OK. What is the bulb V,W rating? Is 10 Ohms cold? or warm? – Sunnyskyguy EE75 Sep 22 '18 at 17:10
• Its efficiency comes from PWM nature. When measured practically, it average voltage drop on the lamp is 50V when voltage drop on the diods is 180V, when 230V in outlet. And its known that lamp and device in series shares same current, 180/230 = 78% efficiency and rest is wasted on flickering lamp emission. Measured power consumption was around 50W, so its 40W charger. Its like boost converter, but generator, transformer and nmos replaced with a lamp, capacitor and tyristor :-) – user197437 Sep 22 '18 at 17:16
• Circuit inefficiency comes when we start to charge the capacitor and bulb glows bright, wasting watts. But capacitor charge law is exponential so the bulb dims fast and much of the time less energy is dissipated. I think same principe applies to conventional PWM converters: when device starts and mosfet opens - many energy flows through transformer and mosfet, heating them up, only to increase flux.. but because the law is exponential we get high efficiency. – user197437 Sep 22 '18 at 17:34
• OK. What is the bulb V,W rating? 300W 230V? If you supply, I will show a simulation – Sunnyskyguy EE75 Sep 22 '18 at 17:35

How could i modify scheme to get rid of lamp and raise its efficiency to 85-95%?

You can't, this circuit needs the lamp to limit the current. Without the lamp too much current will flow and battery and/or diodes and/or SCR will blow up.

This circuit is very crude and also unsafe and inefficient like you mentioned. That's because it doesn't do anything to convert the high mains voltage efficiently to a lower voltage (while raising the current) to charge the battery efficiently.

To safely charge the battery mains isolation is required. To efficiently charge the battery a transformer is needed. Although this circuit is "very good and working solution" (your words, not mine) it is a nasty and cheap solution.

I have seen proper (but cheap) car battery chargers which aren't expensive but are safe and (somewhat) efficient. In any case they will be more efficient that this solution.

So in conclusion: no there's nothing that can be done to improve the efficiency of this circuit, it is a crappy circuit which cannot be improved.

OP complained that this doesn't answer his question. I think it does because when I wrote this answer the question was How could i modify scheme to get rid of lamp and raise its efficiency to 85-95%? I explained that it is not possible.

Also the question was changed, the posted circuit was incorrect, however the "new" circuit worked more ore less in the same way.

Also it was not made clear (in the first version of this question) that this circuit was for desulfation of the battery.

The fact that a circuit "works" doesn't make it a "good" design. I already mentioned the safety issue. Imagine using this circuit to revive a dead car battery. A car battery has a direct connection to the car's chassis so that chassis might become mains live. Also the discharge cycle is quite "violent" and will be hard on the capacitors limiting their lifetime.

• Downvoted for not providing solution or alternative (not an answer), and using personal negative oppinion in words like crude, nasty without any justification. Have you assembled it? Used it? Made measurements? At least.. simulated it? So.. what are you talking about? – user197437 Sep 22 '18 at 17:24
• @xakepp35: Upvoted for correctness. Nobody says it won't work. It won't work well because of the inefficiency you've noticed yourself. And, that thing is waiting for a chance to kill you or some unsuspecting person. – JRE Sep 22 '18 at 18:50
• @JRE Was the question about safety? Nope! it was about how to limit capacitor current more efficiently. Could we use DC-DC boost converter with transformer for isolation and microcontroller for PWM and charging control, instead of a lamp? – user197437 Sep 22 '18 at 19:07
• I understand the votes, but disagree , but this user obviously must have the necessary skills. The way to reduce losses is reduce bulb power and increase stored energy by series L to reduce dI/dt=V/L on caps but add a pulse switch to the load. The high burp charges are to desulphate battery without boiling electrolyte. – Sunnyskyguy EE75 Sep 22 '18 at 19:07
• The real problem with this design is the low voltage high current battery pulse is to some extent shared with the grid and complex EMI issues are more relevant than inefficiency Most battery smart chargers are expensive with high efficiency. THis one works but compromises several parameters. The risks of exposure is also a factor with 2 line EU power vs 1 line + Neutral North American power. – Sunnyskyguy EE75 Sep 22 '18 at 20:54