What is a part called, that rectifies powerful PWM (10 A)?

You don't "rectify" PWM (the term you are looking for is to "smooth" it) - you just use it to control your heater.

That's the whole idea - the heater doesn't care - being fed with full power for half the time is exactly the same as being fed with half power (which would incidentally, correspond to 1/√2 times the voltage).

I'm not sure if you can buy a module like that. What you're looking for is a "PWM controlled switching voltage regulator".

As Paul says, you don't need a linear voltage for a heater. You don't need PWM either. You turn a heater on 'till it hits the target temperature, then you turn it off. Then on again when i falls below the target.

Why do you require a smooth DC current. If you're
just using it to heat a heat plate, a FET power switch
would be fine.
If it is a noise problem issue, you'd need a heavy inductor
in series. The size and current handling would depend on the
parameters you are using. It would average the PWM voltage
for you.
Dwight

If you need 10A at a variable voltage, your best hope is probably a digital potentiometer replacing the voltage setting resistors in a buck or boost converter. What voltage do you need? If you need more than 5v, your options for suitable digipots is limited.

I did something like this a while ago, using a digipot to control the output voltage of a buck/boost converter. I don't know if I have the notes anywhere. It wasn't terribly sophisticated.

I think a MOSFET is the correct device for this. A Peltier will work just fine with a PWM waveform. It won't work well with a simple on-off thermostat. That causes thermal stresses in the Peltier.

It's not a heater, yes, it's a Peltier-unit (thermoelectric cooler). I know that it can be controlled through just PWM. And i do have such an assembly now - a MOSFET controlled by PWM. But i've read that TEC-s don't like PWM very much - not only slow cycling, but also PWM. That they operate with better efficiency on direct current (and wear out slowlier).
I need to maximize the efficiency. The assembly i have approaches limits otherwise. Hm, it's strange that such a module isn't widely known.
I need either 10 A at a variable voltage, or 12 V at variable current.

Making a digitally controlled resistor - maybe i'll manage to do it. Will look in it. (So the reaction of regulators will be acceptable, they don't expect that resistance to be set only once, at least i know now.) It needs to be 0 to 12 V. But i don't need many steps. 6 or 8 steps would be enough.

Slow cycling (without PWM - turn on, turn off for seconds) - hm, if in some opinions it causes bad effects to the cooler, better not to do it. It seemed bad to me. Maybe because i will have big temperature on the expel side - once off, the temperature will crawl back very quickly, hardly this will be efficient.

PWM controlled switching voltage regulator
will search yet, quickly it now returns the same regulators with PWM inside.

And that motor driver - doesn't it produce smoothened current? (guess no because it doesn't have big coils, and they usually have)

Peltiers are more efficient with constant voltage/current drive than PWM: Peltier Element Efficiency

The device the OP was flailing for is a controllable DC-DC converter (which could be an LC filter
after a PWM power stage, like class-D amp)

You are of course correct that TECs do not like shock stress, so that "simmerstat" switching is indeed undesirable and PWM is much better. The frequency (several hundred Hertz) is such that thermal shock is not a concern.

So what is the problem with PWM?

Well, Peltier devices are essentially semiconductors like LEDs - they have a threshold voltage and are essentially current driven as they tend to maintain that threshold voltage. In effect, they behave as a voltage "sink" which performs their "work" in series with an internal resistance which is simply a power wastage and of course, interferes with their cooling effect (but of course, benefits somewhat, their heating effect). This means that feeding more voltage disproportionately increases this energy loss, so driving more current for less of the time, is indeed, less efficient.

What you are looking for is a constant-current driver. A simple inductor in series with your PWM driving FET, and a flyback diode across the combination, will suffice. It is not a module as such unless you obtain a complete constant-current driver at your required specification. Someone else may care to suggest appropriate parts for the inductor and diode. :grinning:

MarkT:
Peltiers are more efficient with constant voltage/current drive than PWM: Peltier Element Efficiency

The device the OP was flailing for is a controllable DC-DC converter (which could be an LC filter
after a PWM power stage, like class-D amp)

Oh, so I needn't have bothered writing all this! :astonished:

I'd thought he was using it to heat.
He is right that they work more efficiently cooling with a constant
current.
It is because If you turn the voltage off, The difference in temperature
makes the peltier a generator, and a leaky one at that, as some of the
current it generates flows through the device, making it warmer.
It is something like paralleling solar panels without blocking diodes
and having one panel in the dark. The one in the light will waste
power in the one in the dark.
He is, most likely, more worried about ultimate temperature delta
than power source efficiency.
You can add a large inductor in series with your PWM power source to
smooth it out. You might try some secondary winding on some scrap
HiFi equipment.
The primaries may have too much resistance and inductance but
at about 500Hz something like a 24V winding might do a good job
of averaging a PWM signal into a smooth DC. A small amount of ripple
shouldn't be too much of an issue. Anyway, give it a try and see
if it is close.
If your going to make hundreds of these, you can calculate the inductance
and a reasonable IR drop for optimal performance.
The other option is to go with a D/A and a linear power source ( possible
modified old power supply ).
Dwight

Do remember that a (power) transformer will not work as a DC choke!

Paul__B:
Do remember that a (power) transformer will not work as a DC choke!

WHy?
Dwight

I think because if you leave the other winding on the transformer it will act more like a transformer than an inductor.

Losses will be higher because the transformer core won't be efficient at the higher PWM frequency. It is also possible to reach saturation with a high DC current through the transformer. Then the core has no additional benefit.

Just saturation.

DC chokes require an air gap.

(No, the other winding is irrelevant. If both windings are similar, they can be put in series.)

so I needn't have bothered writing all this

No, it's good to read, because what i've now found from “controllable dc-dc converter”s, like IR3475M or Semtech SC190, is when the chip has some easier scheme than external resistors for voltage setting, they all still require a ton of components as drawn in typical applications. Coils, capacitances, resistances. It's weird why they don't make circuits with all introduced. I think i won't be able to gear such a chip up. Or like in the solution like from this article: arduino-digitally-controlled-step-down-buck-dc-converter on eevbldg it's the same too.
That way the digital resistor way also becomes too complicated.

And special Peltier-controllers (i think i also searched for them before, but this page is the most clear), these are too expensive. Like 235 E. So maybe i'll unevitably have to wire up the inductor and a diode. If it's just as much, maybe i won't mess it down. But i'll have to calculate them.

Also i can't fully get this idea: you both, dwightthinker and Paul_B, write that constant current is better. Is it just “than PWM”, or “among the constant current and constant voltage“ still constant current is better? I know that an inductor will do the constant current way, but if i ever can implement any of'em fully, just what to choose.

The other option is to go with a D/A and a linear power source ( possible modified old power supply )

Is it possible to use a DAC here? (i don't understand why linear source though) From the parameter searcher i got the feeling they are small power devices.

Sorry about misleading you with the AC coil.
I should have known better, having designed switchers in the past.
Still, most of these down switchers shown could be used for
what you have in mind without using a digital resistor.
You can take the PWM output and clean it up through a RC filter
to control the switching regulator control pin.
It can be treated just like a linear regulator.
If you don't mind a big heat sink, a linear regulation setup
would also work fine.
Dwight

Well, you mean the regulator itself will be dissipating much heat, and will need the big heatsink – not the RC-circuit? And the power dissipated will be on orders of the same 100 W? If that, it's not doing to suit. Such a size (10 cm3) for just regulator is too big. And of course the system mustn't lose all power that's not powering it.
And you see, applying the digital resistor is not the hardest – all the harnessing is the hardest. So linear entry to that point i guess isn't simplifying it.

I'm now wondering what must the inductor be, if it's just a flyback. The parameters must be like: diode: >= 10 amperes, and inductor >= 10 amperes, right? But how much henries? Perhaps what formula is used.
I get the notion, eeeeh, like it was about RC here at 25 minutes; the more L, the more time it'll be settling the curve (at any change of PWM), but the less L, the less the compensation will be. And i mustn't go over some critical scale. But around what. (I can't pretty much try variants, because big inductors are hard to find.)

I didn't say the linear would be practical for your application.
For the switcher the size of the inductor depends on the frequency
of the switcher and the ripple you are willing to tolerate on the output.
Even though, you don't want a full turn-on/turn-off, I would suspect
for your application a 2% or 3% ripple is well within the range you
can tolerate.
You don't want to run the switcher at the PWM frequency of the
adruino. That would require a really large inductor. You just want
to make a relatively smooth DC to control the switching regulators
input control. Let the switcher go to a higher frequency.
Dwight

I think i don't intend to control a switcher regulator – but a transistor, and for that need the inductor. If i need to raise the frequency, maybe i could use the driver of the Mosfet. But sure it isn't critical. I could even do without the smoothener, so i guess even 10% ripple will be not so bad, ?

The size of the inductor should be ultimately fittable in the 5'25 computer bay together with the Arduino and all other small parts with it. Also i read there can be several (smaller) coils combined.

So i found this formula – is it it? ind. for bucker
L = (V0 × dt) / (dI × F) = (V0 × 10) / (I0 × F) for 1/10 ripple dI/dt = I0/10, if i need that oscillation on the current to be 1 10th of the main amplitude. It means L ≈ 12 / 62500 (the max. freq. of Ard.) ≈ 0,0002 H? And for 2% it would be 0,001 henry. These coils don't seem very big, isn't it n't?

200uH is fairly big, 1mH is large. Remember you need a high efficiency ferrite-cored gapped
inductor for switching usage, and if you look for 1mH 10A rated inductors you'll find they are
not small or cheap. High current means thick wire means fewer turns means larger core, and
the energy-storage requirement indirectly means larger cores too (even though the energy is
mainly stored in the magnetic field outside the ferrite, its vital the ferrite doesn't saturate, and
ferrite doesn't need much magnetic field before it saturates.

The need for a gapped core is to store energy (a transformer doesn't store energy, so its
usually much smaller for a given inductance). Highly magnetic materials are less able to store
energy than air/vacuum (it turns out), hence the gap. The core serves to guide the magnetic
field so that it nearly all passes through all of the coil and delivers the field efficiently to the gap.
This allows much more compact (and thus low-resistance) winding.

Where do you get 62500. The PWM is only 490, 980 if using an Uno.
Dwight