Simple 230V AC to DC converter

[tooki]

you are very likely to burn the houses of your victims down. Please stop.
Alternatively, I think we should start gathering some information about this guy and report to authorities to prevent loss of life.

i saw this alot people typing. is there any reason to be excessively hypothetical or paranoid? we test every circuit in our home first. so at worst its his house that will burnt down. he may making a post out of ignorance.. so you are here to teach..

You will get no sympathy from the vast majority of people here when it comes to compromising safety, and that’s exactly as it should be.

It’s not paranoia or hypothetical: the dangers and challenges of high voltage DC are well known. OP clearly does not appreciate this (since he was advised of them months ago and just ignored it), and is on a path to build a death trap. It killing him would be bad enough, but you don’t know who else lives in his house, apartment building, or neighborhood — depending on the construction, a fire in one home could easily engulf any of those.

[NiHaoMike]

Bridge rectifier between the thermostat and heater and a large inductor in series with the heater.

[Zero999]

I remember seeing many years ago an extremely simple circuit that was providing very smooth DC output and at the time I thought it was very clever solution.

You could be referring to LM317/7812 circuit variants:

Fro 230VDC at 5A? Are you joking or just didn't read the original post properly?

Hi,
I need to build a 230V AC (mains) to 230V DC converter. It needs to be simple and reliable so the obvious initial thought was just a bridge rectifier with massive capacitors to smooth it out but the output would not be flat DC, it would obviously have smoothed out peaks every cycle.

I remember seeing many years ago an extremely simple circuit that was providing very smooth DC output and at the time I thought it was very clever solution. But cannot find it anymore. It was not just a rectifier and massive caps. It had a few extra components but less than a handful. I googled it for days but cannot find that circuit anymore. Any suggestions?

It is to be used only in the UK and in controlled environment by an experienced engineer So the only safety required is just a fuse for over-current protection.

The specs are as follows:
- 230V AC to DC
- 5A
- when the load is on the voltage should be as steady as possible. This is the most important spec.
- the load is only resistive (heating elements)
- the heating element itself has a thermostat integrated in it to turn it on and off. It switches on/off every 20-30 seconds
- voltage dips transients when turning on and off are acceptable (perhaps an inrush current limiter on the output might help improve those too?)

Does anybody have any clever suggestion/idea?

Thank you as always

Why aside, to answer the question, I wouldn't dream of designing such a thing.. I would buy it ready made. Here's an example. The nominal output voltage is 250V, but it can be trimmed by 5% down to just below 240V.
https://uk.rs-online.com/web/p/switching-power-supplies/1860192
https://uk.farnell.com/xp-power/hdl3000ps250/power-supply-ac-dc-250v-12a/dp/3525443

[dobsonr741]

In summary, the OP is trying to eliminate AC power due to perceived human health impacts of electromagnetic radiation.  The load is an infrared lamp for a health sauna. we

OP should consider building a Time Machine instead the complicated rectifiers, go back to 1886, convince Nicola Tesla to give up the idea of AC current, and let Edison standardize on DC power distribution.

Here, on the forum then we can start to have a more fruitful convo on Time Machine building, using various AliExpress bought parts. I already have a few ideas.

[/parody]

[rew]

The load is an infrared lamp for a health sauna.

In that case, just rectify and big caps. The voltage will be about 325V. Heating elements can usually handle that. If you can't afford to try plan B AFTER plan A has failed, just using them on 325V is plan A.
Plan B is to put 1.5 in series, or maybe 2, but then you might have to increase the number of lamps used because they won't reach rated power.

Edit: To REALLY approach the "requirements" I'd use a transistor in "linear regulator mode". Simply make the base voltage with a capacitor that should stabilize to "just below the lowest voltage on the caps in the last cycle". This can be done with a diode. Also you don't want it to take a few hundred cycles for the voltage on the cap to reach that 325-10V. Then the power transistor will be blown. So add a zener to make it at least 10V below the actual peak voltage. Could be a handfull of components. The current in/out of the "voltage setting" capacitor should be small, so the transistor will probably be a darlington(configuration).

[Zero999]

The load is an infrared lamp for a health sauna.

In that case, just rectify and big caps. The voltage will be about 325V. Heating elements can usually handle that. If you can't afford to try plan B AFTER plan A has failed, just using them on 325V is plan A.
Plan B is to put 1.5 in series, or maybe 2, but then you might have to increase the number of lamps used because they won't reach rated power.

Remember, as the voltage is increased linearly, the power dissipation in a resistor goes up exponentially, doubling by every factor of √2, so 325V would result in double the power dissipation in a resistor, compared to 230V, which will quite likely blow it up.

Edit: To REALLY approach the "requirements" I'd use a transistor in "linear regulator mode". Simply make the base voltage with a capacitor that should stabilize to "just below the lowest voltage on the caps in the last cycle". This can be done with a diode. Also you don't want it to take a few hundred cycles for the voltage on the cap to reach that 325-10V. Then the power transistor will be blown. So add a zener to make it at least 10V below the actual peak voltage. Could be a handfull of components.

That's still a massive amount of power dissipation in the transistor. Forget using a BJT because  it'll exceed the safe operating area. MOSFETs are more reasonable, but check the data sheet.

Perhaps a large inductor in series with the bridge rectifier + smoothing capacitor, could be used to reduce the output voltage to 230V. It will also have the positive effect of reducing harmonics and improving the power factor.

[tooki]

Edit: To REALLY approach the "requirements" I'd use a transistor in "linear regulator mode". Simply make the base voltage with a capacitor that should stabilize to "just below the lowest voltage on the caps in the last cycle". This can be done with a diode. Also you don't want it to take a few hundred cycles for the voltage on the cap to reach that 325-10V. Then the power transistor will be blown. So add a zener to make it at least 10V below the actual peak voltage. Could be a handfull of components.

That's still a massive amount of power dissipation in the transistor. Forget using a BJT because  it'll exceed the safe operating area. MOSFETs are more reasonable, but check the data sheet.

A MOSFET for linear operation?!? Isn't that one of the areas where BJTs are generally superior to MOSFETs (since modern MOSFETs are almost universally optimized for switching, not linear operation)?

Perhaps a large inductor in series with the bridge rectifier + smoothing capacitor, could be used to reduce the output voltage to 230V. It will also have the positive effect of reducing harmonics and improving the power factor.

The power factor... of a resistive heater? Isn't that 1.0 by definition?

How would a series inductor reduce the output voltage?

Does rectifying AC change the power factor the load places on the grid?

[Faringdon]

Does it have to be exact 230VDC.?
If so thats ok anyway..

As someone said, a Boost PFC can get you to mains peak or above.
Or...just a simple boost converter after the mains bridge can do it for you.

Or if you want 230VDC ....just put a buckboost or sepic after the mains bridge....or even a flyback

Or if you  want some voltage which is just above mains peak...you can use one of those boost pfcs which do that "a follower boost".

But if you give me firmer specs, i  can give you the solution you seek.

You have said 230VDC at 5A...but first i want to check is there any leeway in that.?..because 230vdc is below mains peak and you there4 need some extra stuff to handle that.

So if you need 230vdc and 5a, and as steady as possib;e, then whatabout a hysteretic mode flyback....that easy as pie......simple.....i can doo that for you no probz.
Hysteretic mode means it will be 230v +/-3v say, but will always be within that.
Or i can do nice flyback with fast feedback loop....i have tricks for that....and we can do it like thta...but the noload 2full load (and vic ver) transient dip coudl be a bit of a bu**er....unless we use lots of cout.

But give a shout as this is simple

[Andy Chee]

A MOSFET for linear operation?!? Isn't that one of the areas where BJTs are generally superior to MOSFETs (since modern MOSFETs are almost universally optimized for switching, not linear operation)?

There are definitely MOSFETs out there designed for linear operation

https://www.littelfuse.com/products/power-semiconductors/discrete-mosfets/n-channel-linear.aspx

(as a sidenote, take a look at another section of their website and application notes about another less commonly used MOSFET, the depletion mode MOSFET).

[tooki]

A MOSFET for linear operation?!? Isn't that one of the areas where BJTs are generally superior to MOSFETs (since modern MOSFETs are almost universally optimized for switching, not linear operation)?

There are definitely MOSFETs out there designed for linear operation

https://www.littelfuse.com/products/power-semiconductors/discrete-mosfets/n-channel-linear.aspx

(as a sidenote, take a look at another section of their website and application notes about another less commonly used MOSFET, the depletion mode MOSFET).

That's why I said "almost universally", not "universally". The percentage of modern MOSFETs that are designed for linear use is vanishingly small, but existent.

[Faringdon]

Also....dont forget the inimitable  and simple "swamped boost converter".
That could maybe do a job for you.

So for mains you get 339vpk.......so set your vout to 339vdc.......and shove a hysteretic mode boost converter  on it...it will spark up and shovel current into the output whenever it goes below 339vdc.
...and its just a cheezy simple boost converter. As cheezy as Rapunzel!

So, a good "el cheapo" way to get solid vout in face of severe load transients, is to do a hysteretic converter....but a hysteretic converter will mean you need to put up with a vout of say 230v+/-3v...........so...is that ok for you?...it will never go outide that....(unles of course a direct sht cct on output).
And doth you require isolation?.....or "you dont but it would be handy"...or "you dont, but if its cheap youll take it"

I guess you cant tell us the application as its a secret market area?

[Andy Chee]

I remember seeing many years ago an extremely simple circuit that was providing very smooth DC output and at the time I thought it was very clever solution.

You could be referring to LM317/7812 circuit variants:

Fro 230VDC at 5A? Are you joking or just didn't read the original post properly?

In the kooky context of the OP's ridiculous project request, the OP could've seen anything those many years ago, including a voltage boosted linear regulator.

[Zero999]

Edit: To REALLY approach the "requirements" I'd use a transistor in "linear regulator mode". Simply make the base voltage with a capacitor that should stabilize to "just below the lowest voltage on the caps in the last cycle". This can be done with a diode. Also you don't want it to take a few hundred cycles for the voltage on the cap to reach that 325-10V. Then the power transistor will be blown. So add a zener to make it at least 10V below the actual peak voltage. Could be a handfull of components.

That's still a massive amount of power dissipation in the transistor. Forget using a BJT because  it'll exceed the safe operating area. MOSFETs are more reasonable, but check the data sheet.

A MOSFET for linear operation?!? Isn't that one of the areas where BJTs are generally superior to MOSFETs (since modern MOSFETs are almost universally optimized for switching, not linear operation)?

It is true that modern MOSFETs are designed for switching, rather than linear operation, but at high voltages, MOSFETs are much better in the linear region than BJTs, which suffer from secondary breakdown. This is why all high voltage linear regulator ICs use a MOSFET for the pass device, not a BJT.
https://www.ti.com/lit/ds/symlink/tl783.pdf
https://www.onsemi.com/pdf/datasheet/ncp781-d.pdf

Perhaps a large inductor in series with the bridge rectifier + smoothing capacitor, could be used to reduce the output voltage to 230V. It will also have the positive effect of reducing harmonics and improving the power factor.

The power factor... of a resistive heater? Isn't that 1.0 by definition?

How would a series inductor reduce the output voltage?

Does rectifying AC change the power factor the load places on the grid?

Adding a smoothing capacitor to the output of a bridge rectifier does reduce the power factor, because it only draws current when the voltage across the capacitor is lower than the mains voltage. If a large filter capacitor is used, the current is drawn in huge spikes, which leads to greater I²R losses in the wiring, than the equivalent purely resistive load. An inductor would limit the current spikes, thus improving the power factor.

[Andy Chee]

I guess you cant tell us the application as its a secret market area?

Read page 1 of this thread.  This is the application:

https://www.eevblog.com/forum/projects/ac-infrared-sauna-heaters-driven-with-dc/

https://en.wikipedia.org/wiki/Electromagnetic_hypersensitivity

[JohanH]

Hi,
I need to build

No, you don't. If you have to ask such a question, you are not qualified to even try to do something like that. Stay on 5V DC and be safe.

[rew]

Re: power in the resistor (Does this forum link the post I hit "reply" on? Edit reacting to: https://www.eevblog.com/forum/projects/simple-230v-ac-to-dc-converter/msg5276722/#msg5276722 ... and found the quote button. Sorry.

Yeah. I did the math. max 10V * 5A = 50W. In practice about half that. Calculate the cooling solution for 50W and you'll be OK.

There is a trade-off between power in the transistor and the price of the capacitors. Capacitor formula is I = C dV/dt, so C = I dt/dV . So C = 5A * 10ms / 10 = 5mF. A 4700 uF at 400V or so is not going to be cheap.

You can go cheaper, e.g. half that, but then you'll have to deal with at least double the power in the transistor. This can be optimized if you actually decide to build it.