12VAC to 12VDC 5A on 250ft 12AWG

[riwalker]

Please help.
Months of experiments came to naught.
we need 5A from 12VAC on 12AWG.
we cannot change 12VAC, its used in the industry exclusively.
Switching to 15VAC does get us to 4A max on 12AWG@250ft and maybe an option.
We could force customers to use heavier cable, but retrofit was a goal.

I am trying to pull 5A from 12VAC on a 250ft line, and max I can do is 2.5A with a BOOST circuit
The issue is the drop in volts after the bridge rectifier, even if I use ‘ideal’ bridge, or even 'bridge doubler', the volts drop dramatically such that not enough (>2V) to energise a buck-boost controller.
best from 12VAC is 2.5A@250ft, 1.5A@500ft

have heard that 'bridgeless' / 'totem-pole' / 'active full bridge' PFC may help, also BCM (Boundary Conduction Mode) would help, but all circuits out there are designed for 120-240V with ASIC minimums typically 22V.

also read the thread that bridgeless OFC is an EMC nightmare ? hence dotn want to embark on something that will never work

I sincerely appreciate any insight, as I am truly stuck.

 

[barry]

I don't quite understand your problem. It sounds like you are stumped by physics. 1000 feet of 12AWG is 1.6 ohms. So 250 feet is 0.4 ohms. 5 amps will drop 2 volts.

So, what's your problem?

 

[24vtingle]

If you cant do that volt drop then.....You will need to rectifiy it to DC, then use an inverter or similar to get it up to high enough AC voltage that your current goes down enough
to give less volt drop along the cable. But safety may become an issue if the customer may touch it and you make the ac a dangerously high voltage.
It sounds like 24VAC would be ok....why not do a "clipped sine" inverter so that it peaks at 24V pos and neg...then you wont have a dangerously high peak voltage.

Even like a square wave (with bit of dead time) pos and neg +/-24V would be good....send that down the line......obviosuly it needs to be AC because of the rectifier at the other end. and if you send dc then only one half of the rect will be conducting and may overheat.

Be careful that what ever electronics at "the other end" can handle the raised voltage.

If EMC is an issue, then smooth the square wave AC with a filter before sending it down the line.

This sounds like a security alarm system?

 

[FvM]

12AWG has 1.7 to 1.8 ohms/kft, depending on the stranding. Presume you mean 2x250 ft total wire length, 0.9 ohm total resistance.

Maximal available power is achieved with impedance matching, 50 % cable voltage drop, 12V^2 / (4*0.9 ohm) = 40 W. Means you can't get more than 3.33A @ 12 V through this cable, even with lossless conversion. Not completely clear what's the required output, is it 5A@12V DC?

Please correct if I read your specification wrong.

 

[riwalker]

we need 5A@12VDC from a 12VAC line.
we cannot change the line voltage, the market uses 12VAC, and cannot change to 12VDC
its not a line 'physics problem', or drop problem per se,
12AWG 12/2 resistance approx. 1.6ohm/1000ft, or for 250ft = 0.4 ohm, so loss I2R=10W, or 2V
a BOOST convertor can easily handle 8V to 12V

On measuring, we do see good voltage pre-bridge.
The issue is for the buck-boost, to get 60W out, we have the efficiency of the regulator, but more importantly is the voltage drop after the bridge, thus for the same given power, the voltage drops exponentially, thus the current draw rises exponentially
Having said that, even with 5A draw, we still see the line voltage > +-8V (worst case, empirically)
So line volts seems fine.

the issue is the drop after a bridge with any kind of rectifier then DCDC
hence asking for AC/DC alternative topologies that have been tested. i am reliably told BCM has been proven to work.
also interested in anyone who have a PFC circuit working for 12VAC input

 

[FvM]

O.k. if it's 0.45 ohm total resistance (125 ft cable, 250 ft wire length), a 8.5V input PFC boost converter with 90% efficiency can do. Presume bridgeless PFC topology or active rectifier bridge, custom designed controller. Definitely feasible.

 

[riwalker]

yes possible, but so far empirically a dream. hence need actual circuits with this condition.

We are limited to 3900uF caps, maybe 2x at a push

TI & Analog ‘active rectifier bridge ‘ fail because the voltage drops too low to fire the NFET (even changing to logic FET)

Voltage doubler fail also

I am asking for that, actual tested circuits as have tested a bunch (DC/DC) and all fail due to the bridge drop

Hence

• AC/DC topology
• ‘bridgeless PFC’ topology
• BCM topology

The only example 'Low-Voltage PFC' I can see is a Microchip science park, and no longer supported, and no production silicon.

 

[24vtingle]

Its not the real mains system, so any thing that will draw current and not give high reactive current will do...what about a COT booster with a divider on its vin to give your sine current reference...then your vout error amplifier should just clip the sine downward for when the load gets lighter.
You can do this with a cheap hacked up ucc28c43 circuit...ill send you a LTspice simulation of it if you want?

Yes you can do BCM PFC but the difference in efficiency would be insignificant.
Are you using Schottky diodes in the bridge rectifier.?.. (best put transient protectors (TVS) with them as Schottky = sensitive).
I woudl doubt TotemPole is worth while.

And If you really want to get rid of some of the bridge rectifier drop you could always just make the rectifiers synchronous...with your low voltage that would be easy to do....but i doubt even that would be necessary.
If you just make the two ground connected rectifier diodes synchronous then thats really easy, as theres no hi side drive.
Is it a security alarm system or fire alarm system you are doing?

 

[riwalker]

Thank you, that (a simulation of possible solution) would be ideal.
also, 1 detail I missed is that we also have a 150uH on + and - at the end of the line to our power circuit (12AWG x 250ft) + 150uH) in order to filter 'noise' from our power circuit going back onto the line as we have PLC communication. We need to avoind anythign in the G3 PLC band 154.6875-487.5KHz range. of course, we can get the Power circuit to work first, then work on the filtering....

 

[24vtingle]

Here is the "Jack" PFC sim in LTspice...this one is for 400VDC output and mains 230vac input, so you will need to adjust it.....give me a shout if you need
help.
Becuase of your input inductors, you will need some bit of capacitance at the input to the attached...but not too much.

 

[riwalker]

Here is the "Jack" PFC sim in LTspice...this one is for 400VDC output and mains 230vac input, so you will need to adjust it.....give me a shout if you need
help.
Becuase of your input inductors, you will need some bit of capacitance at the input to the attached...but not too much.

Need help please, i have pspice, but havent used ltspice since college days !! and find many circuits and magnetics for mains AC do not resize or make sense (windings) for 12VAC
I see, RFN5TF8S is Vf 2.1V – will need to change to Schottky Rectifier, to example ST STPS3045DJF (I used RBR10T30ANZ).
also did not have model for fast recovery D1, so used VS-E5PH3006
changed load ? R7 to be 2.4ohm (for 5A @ 12V) ?
changed VIN to be 12VAC, so is this correct ?

 

[24vtingle]

Hi, Here it is adjusted a bit for a quick first pass adjustment
You can tweak it a bit nicer if you want.
Check for any resonance with the input LC filter...this shouldnt be same as Fsw.
Some of the nodes could do with a bit of filtering.

12vac is 17vpk, so the boost should have output volts set a bit above 17v..so i made it 25v

--- Updated May 10, 2025 ---

Woops i did the limit command a bit wrong , so corrected it here attached.

Simple Constant Power Loads in LTspice

I recently needed to model active and passive power factor correction techniques as part of the design process for a 100 watt offline LED ba...

russellsrandomthoughts.blogspot.com

Also, there is some LC resonance so the input filter LC needs adjusting to stop that

 

[riwalker]

Thank you. much appreciated.
1) Undefined model "rb10l-40" - suggest alternative
2) does this circuit exist such that we can evaluate ?
3) we need to model the 12AWG 250ft for 12VAC 60Hz, 12/2 cable (i.e Southwire's Low-Voltage (Low Energy) Underground Landscape Lighting Cable, 12AWG), Spec is here (pdf or excel chart)

 

[barry]

Maybe I'm missing something, but you're making this way harder than it is. I can't believe you've spent "MONTHS" on this. You can either put your boost converter at the source end of the wire and set the output voltage high enough to compensate for the IR drop (bad idea if the load current varies) or you can put the boost converter at the load end.

 

[riwalker]

Barry, not sure you read through, i explained several times. 1) The line voltage is set, 12VAC, 2) the Boost convertor IS AT THE END, and the bridge voltage drops as current is drawn.

 

[24vtingle]

1) Undefined model "rb10l-40" - suggest alternative

Yes it sounds like you have a different LTspice library than me...just use a Schottky rated to the voltage.......use "m=2" etc to put schottkys in parallel if the one you find isnt rated for the
current on its own.....

Find the resistance of the cable, and add that resistance in the LTspice to get a check on your calcs.

Attached i changed the schottky and paralleled them as i could only find 1A schottkys

does this circuit exist such that we can evaluate ?

If you have PCB layout prog then just do a quick PCB and solder i up...use the pin4pin UCC28C43 in stead of the expensive LT1243,
Are you "au fait" with PCB layout of switchers,? or would you like a doc showing how to do it.?

Its essentially just a constant off time boost converter with a variable reference and a slow feedback loop so the current goes kind of sine-ish.
UCC28C43 is probz the most well known current mode control chip in the world...but i dont find any demo boards of it on digikey etc.

I think a bcm circuit will be unecessarily more than needed as it needs a bias coil on the boost inductor to indicate inductor "discharge".

So you cant run the sim?
If you are running it and see it working but you still dont think it works "in real life"?

 

[barry]

Barry, not sure you read through, i explained several times. 1) The line voltage is set, 12VAC, 2) the Boost convertor IS AT THE END, and the bridge voltage drops as current is drawn.

I've read it, YOU haven't explained it.

1) There are TWO ends of a wire. Which end?
2) Your rectifier drops voltage. So what? Your boost converter is AFTER the rectifier.

 

[riwalker]

Yes it sounds like you have a different LTspice library than me...just use a Schottky rated to the voltage.......use "m=2" etc to put schottkys in parallel if the one you find isnt rated for the
current on its own.....

My LTSpice is fresh today, but managed to find the model for the RB10L-40 (and add as a spice directive on the sheet (
.model RB10L-40 D(Is=2.6u Rs=.0674 N=1.163 Cjo=180p M=.6 Eg=.69 Xti=2 Iave=1 Vpk=40 mfg=Rohm type=Schottky)

Find the resistance of the cable, and add that resistance in the LTspice to get a check on your calcs.

Really need a LCR model for the cable

Attached i changed the schottky and paralleled them as i could only find 1A schottkys

If you have PCB layout prog then just do a quick PCB and solder i up...use the pin4pin UCC28C43 in stead of the expensive LT1243,
Are you "au fait" with PCB layout of switchers,? or would you like a doc showing how to do it.?

yes, experienced PCB layout, just a lot of hasstle to build this circuit to proove it works. will need to finish this model firsh, complete with DCDC (i.e MAX20048, or TI LM51772)

Its essentially just a constant off time boost converter with a variable reference and a slow feedback loop so the current goes kind of sine-ish.
UCC28C43 is probz the most well known current mode control chip in the world...but i dont find any demo boards of it on digikey etc

For UCC28C43, says UCC28C53 is the replacement, The eval module is UCC28C56EVM-066 but its a special eval for 800V, and not helpful

 

[24vtingle]

LM51772 is 4 switch .....if you use the cct i showed you only need a sync buck...ti.com do some really good cheap sync bucks which will take 25v down to 12v 5a.
Attached is a good list of synch bucks.
If you want nice and easy you can do sepic also with ucc28c43 (or .......53), but it likely wont be quite as efficient as synch buck.

You could always buy in a buck module ...loads on the shelf that do say 18-30V to 12V at 5a. You often buy open frame and just solder it in to your setup.

12VAC, 60Hz used to be very very common ayk, but not any more, as such, i doubt you will find any demo board that does anything like that.
Whats the secondary winding res of the 12vac transformer.?

Really need a LCR model for the cable

...i would just put enough capacitance at the cable end such that you can forget about any LC effects of the cable. Though be wary of turn on ringing to high voltage due to the cable stray L...but most likely that would get quenched in the cout of the boost converter as shown.

Also, your input inductors will "drown out" any cable stray inductance.

I dare say i would add an input common mode choke for emc...or at least a balanced diff mode filter with equal L in go and return.

 

[dick_freebird]

Why not get a transformer and scale up to (say) 18 or 24V.
Fix the UVLO problem, if that's the problem.

Then you could use a simple buck to knock it back down
relatively efficiently.

You do not mention any limitation to 12VAC current
/ any source impedance. So why not?