Power Line Conditioner

[jacquesl]

Hi,

Does anyone know how to build a Power Line Conditioner?

Its used to compensate peaks and valleys in the power supply and reduce peaks in the power flow to what is needed by the machine. For instance, on a 120 volt service, if the voltage exceeds 130 volts, the output is switched to a transformer that reduces the voltage by about 8% to keep it closer to 120 volts. The same is true if the voltage drops below 110 volts, the output switches to a different transformer that increases voltage by about 9%. However, if the voltage drops below 90 volts, or exceeds 140 volts, the power is shut off and then after this the Inverter will probably kicks in, running from 12V DC source

But I’m in South Africa so I’ll make it 250V max //Standard

 

[Pano]

In North America Furman sell a range of power conditioners that do exactly this. I've used them plenty in pro audio and video.

They use a special multitap autotransfomer (toroidal). Triacs are used to switch to the winding needed to keep the voltage steady - up or down. You can see this happen in real time on an LED bar graph on the front panel. Each LED is driven from the same circuit that drives the triacs.

Beyond that, I can't tell you what they use as a voltage sensor, but there are a number of IC on the board. There must be a passage by zereo detector, too - so that the switching takes place at zero volts.

This kind of device is going to be hard to DIY, as it takes a special transformer.

 

[Eva]

A variac controlled by a servo may be another alternative. Sometimes there are variacs for sale in ebay for very cheap.

 

[wine&dine]

A variac controlled by a servo may be another alternative. Sometimes there are variacs for sale in ebay for very cheap.

I had suggested that one before in another thread - the servo mechanism woudn't be too hard for DIY.

Other "classical" methods are magnetic constanters using a transductor (aka mag-amp), or a motor/generator combination with a flywheel for energy storage - these were used in the old days of IT to filter computer center power (and before that for iron works ...)

 

[AndrewT]

tapped autoformer > constant Vac

Hi,
would a string of comparators each set up as window comparators measuring DC rectified from the variable AC give an acceptable output controller?

Each comparator output would pull in a different tapping on the primary side.

I see primary side tappings as the way to maintain ampere turns and thus VA at near constant value.

A little bit of hysteresis would prevent chattering of the relays.

The transformer could be an autoformer and the output treated as if direct on line from the safety perspective.

 

[martin clark]

That's pretty much exactly how the 'small' one I have at home works - 1.5KVA transformer with multiple taps on the primary, made by these people.

 

[AndrewT]

Hi,
nice to get the occasional confirmation that I'n not "off the planet".

 

[TechGuy]

An alternative method would be to rectify the power and generate a sine output from the rectified current. A PFC circuit could be used to regulate the DC voltage. A Sine wave output can be generated using a PWM controlled voltage regulator which would create half a sine wave followed by a full output bridge to flip the half wave output into a full sine wave. The response time would be a lot better than a varic controlled solution since a electronic controlled system can response in milliseconds or even microseconds.

To further reduce noise and distortion, a synthesized sine wave can be generated to reduce or eliminate harmonics. Google for "Magic sine waves" for more info.

Hope this was useful.

 

[jackinnj]

To further reduce noise and distortion, a synthesized sine wave can be generated to reduce or eliminate harmonics. Google for "Magic sine waves" for more info.

Hope this was useful.

Ya know, I talked to Don Lancaster about that...

Regeneration is at most 70% efficient, the rest goes up in heat.

 

[IVX]

jackinnj,
i've feeling, 70% efficiency it's rather shameful at moment, 2007 isn't it?
TechGuy,
A) IMO, flipping isn't good solution in the precision terms, i wouldn't be able to take the feedback after all. In the USA 120VAC, so just a 200V mosfets required, why the flipping is needed?
B) Pre-emphasis will work adequately only at the single type of loads, for what was adjusted one. BTW, such pre-emphasis nice way to improve precision of the cheap reference sine DAC, i had try it for ATmega8 11Bit PWM.

 

[AKN]

Hi,

70% efficiency seems low.
An online UPS (constant regeneration) can have obove 85% efficiency at distorsion less than 3%.
http://www.powerware.com/UPS/9120_Specs.asp

 

[martin clark]

3% THD is hardly worth the inefficiency though - I see about 4-5% out of the wall, see http://www.acoustica.org.uk/other/mains_noise.html

I suspect the only real difference will be the noise spectrum, and rather higher output impedance from the UPS - which virtually guarantees the UPS/regenerator has vastly increased %THD under load. Also, FWIW, some years ago I built a small 240v regenerator to have a quick look at the problems the approach brings. I can confirm power efficiency was dismal, though not the point of the exercise A short write-up is here:
http://www.acoustica.org.uk/other/mains_regen.html

More on DIY regeneraton in a previous discussion in this forum:
http://www.diyaudio.com/forums/showthread.php?s=&threadid=18066

 

[IVX]

3% THD is hardly worth the inefficiency though - I see about 4-5% out of the wall, see http://www.acoustica.org.uk/other/mains_noise.html

In the night i see 3% sometimes. So, most products on the market seems just useless, whatsoever efficiency, due to too high output impedance. Even stepup trafo free design, proud if had 500mOhm (e.g. Exactpower)! Simple simulation will show 1% at 500-700W (power factor ~.7, 120VAC) at 500mOhm in serial to IDEAL sine source.

 

[ak_47_boy]

Autotransformer?

 

[jackinnj]

Autotransformer?

A ferro-resonant transformer provides a modicum of regulation, but the output spectrum distortion is awful.

 

[TechGuy]

Regeneration is at most 70% efficient, the rest goes up in heat.

Where did you hear that? Its possible to get inverters to operate at 98% efficiency with low THD. Quality off-grid inverters have 98% efficiency over a signicant power load range.

>Ya know, I talked to Don Lancaster about that...
[/B]

And? It seemed like intriguing concept, although I don't know how well it really works.

 

[TechGuy]

[A) IMO, flipping isn't good solution in the precision terms, i wouldn't be able to take the feedback after all. In the USA 120VAC, so just a 200V mosfets required, why the flipping is needed?
[/B]

You would be rectifying AC-mains into 170VDC using a PFC. to get back to Full 120VAC, its 340V peak to peak. The feedback input for the controller is taken before the flipping bridge.

You could opt for a PFC to output at 340VDC instead but:
1. You lose some efficiency in the Buck-Boost circuit. Or you would need to use a step up transformer (more $$$)
2. You will need to use higher voltage Mosfets which have higher RDS ( lower switching efficiency and higher heat dissapation)
3. You will need to use higher voltage Diodes which usually have a higher built-in voltage and probably slower recovery which results in efficiency losses and higher heat dissapation.
4. You need higher voltage caps (more $$$) that usually have higher ESR.
5. Your feedback control will probably be more complicated since the voltage must swing across 340V Peak-to-Peak.

 

[Eva]

Ya know, I talked to Don Lancaster about that...

Regeneration is at most 70% efficient, the rest goes up in heat.

Linear sine synthesis tops at a theoretical 70% maximum efficiency. Switched-mode sine synthesis tops somewhere between 90% and 99% depending on design and component choice.

 

[jackinnj]

Yes, I should have side regeneration via sinewave is at most 70.7% efficient. The sinewave will have a minimum of spurious radiation.

If you look at the quiet switchers from Linear Tech they have a much higher conversion ratio but the waveform is more of a smoothly clipped sine wave --

 

[IVX]

You would be rectifying AC-mains into 170VDC using a PFC. to get back to Full 120VAC, its 340V peak to peak. The feedback input for the controller is taken before the flipping bridge.

You could opt for a PFC to output at 340VDC instead but:
1. You lose some efficiency in the Buck-Boost circuit. Or you would need to use a step up transformer (more $$$)
2. You will need to use higher voltage Mosfets which have higher RDS ( lower switching efficiency and higher heat dissapation)
3. You will need to use higher voltage Diodes which usually have a higher built-in voltage and probably slower recovery which results in efficiency losses and higher heat dissapation.
4. You need higher voltage caps (more $$$) that usually have higher ESR.
5. Your feedback control will probably be more complicated since the voltage must swing across 340V Peak-to-Peak.

I'm sorry, but I still don't see any reason to use flipping for halfbridge instead the full bridge 200VDC invertor, because mosfets, diodes and caps will use same type and quantity, so, flipping will have outstandingly poor precision and nothing more. I'm sure, that taking feedback from the outlet directly, it is most important feature for AC regenerator, without it any "AC regenerator for audio" just an expensive toy, they offer you 1-2% instead 3-5% from the wall. I can show here some pics, if you want, ~400W@230VAC real amp loaded regenerator with 4mOhm@50Hz vs 4+100mOhm@50Hz it gave .008% vs .02% THD, for 110-120VAC all things will be worse, obviously.