AC to DC 65+ volts 60 amp Power Supply Plans?

[Ulysses]

Greetings. I hope I'm in the right place.

I want to build or buy a power supply which will in effect mimic the output
of solarvoltaic panels. I want to feed the output into an OutBack MX60
charge controller. The maximum input is 141 VDC and 60 amps. I can use
any voltage above about 63 VDC (up to 141) to charge my 48 volt battery bank
by programming the MX60. This charge controller will automatically adjust
the absorb, float, and equalize charge rate voltages and times. I plan to
power it with a 120 VAC 60 Hz generator.

What I had in mind is a step-down transformer, a couple of rectifying
diodes, and a smoothing capacitor. The problem that I expect to encounter
is ripple. The MX60 is only designed for solar panels so it may behave
incorrectly (or worse) if there is any ripple or pulsing present. Any
suggestions? I don't want to do anything that might damage my rather
expensive charge controller. I'd be happy to buy a power supply if it's not
too expensive but so far all of my searches have not turned up anything.
Are there any web sites that have high current power supply plans? I
couldn't find any.

Thanks.

 

[Paul Mathews]

Greetings. I hope I'm in the right place.

I want to build or buy a power supply which will in effect mimic the output
of solarvoltaic panels. I want to feed the output into an OutBack MX60
charge controller. The maximum input is 141 VDC and 60 amps. I can use
any voltage above about 63 VDC (up to 141) to charge my 48 volt battery bank
by programming the MX60. This charge controller will automatically adjust
the absorb, float, and equalize charge rate voltages and times. I plan to
power it with a 120 VAC 60 Hz generator.

What I had in mind is a step-down transformer, a couple of rectifying
diodes, and a smoothing capacitor. The problem that I expect to encounter
is ripple. The MX60 is only designed for solar panels so it may behave
incorrectly (or worse) if there is any ripple or pulsing present. Any
suggestions? I don't want to do anything that might damage my rather
expensive charge controller. I'd be happy to buy a power supply if it's not
too expensive but so far all of my searches have not turned up anything.
Are there any web sites that have high current power supply plans? I
couldn't find any.

Thanks.

You might save yourself a lot of trouble by locating a couple of 48 V
ACDC supplies that can be put in series to provide 96V. Or, you might
find a 72V supply. Search for something like a Vicor MegaPac. There's
not much competition for the 48V units, so you'll probably find some
kind of a deal.
If your nominal output voltage is 72V for a transformer/rectifier
running on single-phase 60 Hz power, at 60 Amps output, you'd need
about 60,000 microFarads capacitance to keep ripple down to about 8
volts peak to peak (11%). This is not out of the question, but you'll
end up spending more than the cost of some surplus switchmode power
units.
Paul Mathews

 

[Tom Bruhns]

Greetings. I hope I'm in the right place.

I want to build or buy a power supply which will in effect mimic the output
of solarvoltaic panels. I want to feed the output into an OutBack MX60
charge controller. The maximum input is 141 VDC and 60 amps. I can use
any voltage above about 63 VDC (up to 141) to charge my 48 volt battery bank
by programming the MX60. This charge controller will automatically adjust
the absorb, float, and equalize charge rate voltages and times. I plan to
power it with a 120 VAC 60 Hz generator.

What I had in mind is a step-down transformer, a couple of rectifying
diodes, and a smoothing capacitor. The problem that I expect to encounter
is ripple. The MX60 is only designed for solar panels so it may behave
incorrectly (or worse) if there is any ripple or pulsing present. Any
suggestions? I don't want to do anything that might damage my rather
expensive charge controller. I'd be happy to buy a power supply if it's not
too expensive but so far all of my searches have not turned up anything.
Are there any web sites that have high current power supply plans? I
couldn't find any.

Thanks.

That's a lot of solar panels you're wanting to mimic. At 10%
efficiency, I suppose it's about 25 to 30 square meters of surface.
Do you really need that much current? If you really want your source
to look like solar panels, you should have it look like a moderately
high impedance. Into a short circuit, solar panels look about like
constant current sources. Also, if a shadow passes over the panels
(even the right small portion of them) the output drops drastically.
Controllers have to be able to handle that sort of thing.

A mains-frequency transformer that size won't be a small piece of iron
either. You'd probably be a lot better off with switching supplies,
and you can filter the output to get the ripple to a low value with a
lot smaller components than it will take to do it with a mains-
frequency supply. Now if you have three phase mains available, the
ripple can be much more tolerable, but it's still a lot of capacitance
(or a little less capacitance and some inductance; a little inductance
goes a long ways in reducing ripple especially if you want to get to a
very low ripple percentage.

There are surplus houses that deal in power supplies...though 4kW
supplies aren't all that common anyway. Also, I'm not too surprised
you're not finding plans for them (switching type, at least) on the
web. That's a lot of power, potentially kinda dangerous to be playing
with, and because of the currents involved, the physical construction
details and proper selection of components is critical.

Cheers,
Tom

 

[Ecnerwal]

I'd be happy to buy a power supply if it's not
too expensive but so far all of my searches have not turned up anything.
Are there any web sites that have high current power supply plans? I
couldn't find any.

Hi Uly,

Knowing a fair bit about why you are trying to do it this way, several
things come to mind. The obvious-but-it-hurts-the-wallet method to get
an off the shelf charger is to drop bucks on another inverter (which
contains a nice high current switched-mode supply, filter capacitors,
charge control smarts so the MX60 can stay put on the solar panels and
not get involved), and set it up such that you are not using the
inverter section (where all that irritating pass-through-the-generator
stuff comes up when feeding the house power). Could be another Outback
(which would give you a spare to move around in case of any issues),
could be a less spendy brand.

Another off-the-shelf option is to visit the local golf cart dealer, as
48V golf carts need 48V (really 60+V, same as your "48V" bank) chargers,
but I don't know that they pay much attention to efficiency when
designing those.

The best method for fuel to DC is fuel to DC, not fuel to 60 Hz AC to
DC. In reality, what's commonly available (other than esoteric and
overpriced "48V" DC gensets for the cell tower market) is fuel driven
welders, most of which are doing some form of 3 phase on the way to DC,
though there may be "true" direct DC versions (sounds like a lot of
brush wear issues, though). One with a decent OCV (in the 80 V range)
should work for you, though the main output power is typically intended
to be in the 24V range; but if you get one without too many new-fangled
smarts, it should be possible (perhaps with some taming of the MX60's
MPPT settings) to operate on the higher-voltage part of the curve. Might
take some fiddling with (or at least monitoring) the control
arrangement, I'd think.

Another approach is to grab a high-amperage car/truck alternator that
uses an off-board regulator, drive it with a fuel engine, and crank up
the voltage. These are also typically 3-phase AC, with a 6-diode bridge
rectifier. So long as the diodes are not overly low voltage, and you
keep the current within/well below the current rating, they are not
overly picky about what voltage they are putting out. The insulation
should be good for it, and the heat generated is a function of current.

Perhaps a heavy-duty truck alternator that's already intended for 24V
operation would be a good starting point, but I recall people talking
about ramping 12V units up to 70 volts for various purposes with either
no trouble, or frying low-spec diodes and replacing them with
higher-voltage units when they die. In the case of feeding your MX60,
I'd replace them at the start. 100-200 amp units are rather common, and
should not cost much at the junkyard.

 

[Ulysses]

You might save yourself a lot of trouble by locating a couple of 48 V
ACDC supplies that can be put in series to provide 96V. Or, you might
find a 72V supply. Search for something like a Vicor MegaPac. There's
not much competition for the 48V units, so you'll probably find some
kind of a deal.
If your nominal output voltage is 72V for a transformer/rectifier
running on single-phase 60 Hz power, at 60 Amps output, you'd need
about 60,000 microFarads capacitance to keep ripple down to about 8
volts peak to peak (11%). This is not out of the question, but you'll
end up spending more than the cost of some surplus switchmode power
units.
Paul Mathews

Thanks for the pointers. Having a name (Victor MegaPac) to search for
should help a lot. I've tried searching with every combination of words I
could think of and I usually end up finding stuff around 1-2 amps.

I have something like a 24,000 uF capacitor in my junk box but I didn't
think it would be enough. 60,000 eh? Time to start digging through ALL of
my boxes of stuff in the garage. I have an old NCR cash register somewhere
that has a lot of good stuff in it if I can only find it...

 

[Ulysses]

That's a lot of solar panels you're wanting to mimic. At 10%
efficiency, I suppose it's about 25 to 30 square meters of surface.
Do you really need that much current? If you really want your source
to look like solar panels, you should have it look like a moderately
high impedance. Into a short circuit, solar panels look about like
constant current sources. Also, if a shadow passes over the panels
(even the right small portion of them) the output drops drastically.
Controllers have to be able to handle that sort of thing.

If I was going to be using that many solar panels, then yes, it would be
quite a few. That just happens to be the maximum amount that the charge
controller will accept. I could go down to, say, 40 amps and still ge
reasonably content. I'm just trying to cut down on generator run-time and
hopefully gasoline consumption.

A mains-frequency transformer that size won't be a small piece of iron
either. You'd probably be a lot better off with switching supplies,
and you can filter the output to get the ripple to a low value with a
lot smaller components than it will take to do it with a mains-
frequency supply. Now if you have three phase mains available, the
ripple can be much more tolerable, but it's still a lot of capacitance
(or a little less capacitance and some inductance; a little inductance
goes a long ways in reducing ripple especially if you want to get to a
very low ripple percentage.

I just have single-phase AC. Someone suggested a surplus line transformer
(has a catchy name but I can't think of it right now) but there don't seem
to be any such things available near where I live. The closest town is
basically a huge trailer park and I can't find any junk yards or places that
sell surplus electrical stuff. No scrap metal places either. Before I buy
a huge transformer I want to be reasonable sure I can deal with the
frequency/ripple problem effectively. But then I could always find a use
for a big 2:1 transformer...

 

[Ulysses]

"Ulysses" <[email protected]/> wrote:

Ah, the upiquitous Ecnerwal ;-) Someone uses your "cats" signature on the
OB forum so I assumed it was you.

Hi Uly,

Knowing a fair bit about why you are trying to do it this way, several
things come to mind. The obvious-but-it-hurts-the-wallet method to get
an off the shelf charger is to drop bucks on another inverter (which
contains a nice high current switched-mode supply, filter capacitors,
charge control smarts so the MX60 can stay put on the solar panels and
not get involved), and set it up such that you are not using the
inverter section (where all that irritating pass-through-the-generator
stuff comes up when feeding the house power). Could be another Outback
(which would give you a spare to move around in case of any issues),
could be a less spendy brand.

I'm using Inverter #2 only to charge the batteries right now. Yes, it was
an expensive battery charger and not all that great (as far as charging
currnet goes) for the money. I just completed my gravity-fed water system
with float switch and I want my other inverter back so I can have my water
on automatic. With only one 3600 watt inverter (OB VFX 3648) no doubt some
bad stuff will happen if someone is using the vacuum cleaner or washing
machine when the well pump comes on. I don't have any solar panels yet
(other dreams of mine have come true) so in the meantime I'm looking for a
better way the charge the batteries. Of all the things I've thought of so
far using the MX60 seems to make the most sense as far as ease of use and
properly charging the batteries. It has all the smart stuff built into it.
OutBack says just get a bigger (12,000 watts or more) generator. Well, like
you said, running the house from a generator stinks unless I spend $4000 on
a 5000 watt inverter generator and I'd still have the "where the hell is all
the power going?" issue. Right now I'm charging my batteries with a $200
2000 watt generator via Inverter #2. That seems to be about as small as I
can get away with. I usually have to run it about 3-4 hours a day (evening)
to get the batteries up to 100%. Sometimes a little longer.

The good news is that since I switched to propane refrigeration and have my
well pump on it's own generator I could probably get by nicely with only six
175 watt panels. Hopefully I could even squeeze in the well pump as I only
run it about 15 minutes a day.

Another off-the-shelf option is to visit the local golf cart dealer, as
48V golf carts need 48V (really 60+V, same as your "48V" bank) chargers,
but I don't know that they pay much attention to efficiency when
designing those.

I found something like what you described but it didn't seem to have a float
or equalize setting. It's in my referece file if all else fails.

The best method for fuel to DC is fuel to DC, not fuel to 60 Hz AC to
DC. In reality, what's commonly available (other than esoteric and
overpriced "48V" DC gensets for the cell tower market) is fuel driven
welders, most of which are doing some form of 3 phase on the way to DC,
though there may be "true" direct DC versions (sounds like a lot of
brush wear issues, though). One with a decent OCV (in the 80 V range)
should work for you, though the main output power is typically intended
to be in the 24V range; but if you get one without too many new-fangled
smarts, it should be possible (perhaps with some taming of the MX60's
MPPT settings) to operate on the higher-voltage part of the curve. Might
take some fiddling with (or at least monitoring) the control
arrangement, I'd think.

Welders eh? I was looking at welders a while back and I don't even remember
why now. They seem to be fairly inexpensive compared to a lot of things, at
least the stuff Harbor Freight sells. I'll take a look.

Another approach is to grab a high-amperage car/truck alternator that
uses an off-board regulator, drive it with a fuel engine, and crank up
the voltage. These are also typically 3-phase AC, with a 6-diode bridge
rectifier. So long as the diodes are not overly low voltage, and you
keep the current within/well below the current rating, they are not
overly picky about what voltage they are putting out. The insulation
should be good for it, and the heat generated is a function of current.

I have a 12 volt 63 amp alternator that I was using to charge a 12 volt
battery bank. A couple of people on the homepower NG told me I could use it
for 48 volt charging if I replace the diode packs with heftier diodes and
heat sinks and somehow modify the voltage regulator. Well, I'm almost smart
enough to do it, but not quite. I tried putting a VR in series with the
field control and managed to get about 30 volts. My understanding is that I
have to remove the voltage regulator and bypass the circuit to get the full
output. Then I have to control the voltage one way or another. Controlling
the field would seem to be the way to go, but I'm not sure. If I could just
find the instructions somewhere I could do it. I've looked at some
permanent magnet alternators and that seems like it would work just fine.
The voltage could be controlled by engine speed and I could probably adjust
it at no load for equalizing too. The problem is that I've only found one
source and a couple of people told me "don't buy anything from them." I
have the booklet that I downloaded from the wind generator guy (can't think
of his name) and building a PMA and driving it directly from a small engine
is yet another option. It's not yet clear to me just what determines how
much current you will get--the voltage part is apparent.

In any case I'm not going to connect anything to my MX60 unless I'm pretty
darn sure it won't hurt it. Some kind of alternator seems like a good
option and could charge the batteries directly, with some futzing.

 

[Rich Grise]

I just have single-phase AC. Someone suggested a surplus line
transformer (has a catchy name but I can't think of it right now)

....

"Pole Pig"?

Cheers!
Rich

 

[Tom Bruhns]

If I was going to be using that many solar panels, then yes, it would be
quite a few. That just happens to be the maximum amount that the charge
controller will accept. I could go down to, say, 40 amps and still ge
reasonably content. I'm just trying to cut down on generator run-time and
hopefully gasoline consumption.




I just have single-phase AC. Someone suggested a surplus line transformer
(has a catchy name but I can't think of it right now) but there don't seem
to be any such things available near where I live. The closest town is
basically a huge trailer park and I can't find any junk yards or places that
sell surplus electrical stuff. No scrap metal places either. Before I buy
a huge transformer I want to be reasonable sure I can deal with the
frequency/ripple problem effectively. But then I could always find a use
for a big 2:1 transformer...

Sounds like the safest way for you to do this, given the lack of
resources, is find a big 2:1 transformer and rectify and filter it.
Unfortunately, a "pole pig" is very unlikely to be 2:1, though you
could use the secondary as a center-tapped winding that you put 120V
across and get 60V from. Does the output have to be ground-
referenced? If so, things get ugly again because you don't have a
grounded center-tap you can work from -- unless you can float the 120V
input. You'll need to float either the input or output, to make a
center-tapped full wave rectifier. A half-wave rectifier would be
just too awful to try to filter.

For the filter, I'd highly recommend using a choke in your filter, if
you want to get to seriously low ripple. A quick simulation using an
0.1 farad filter cap with 1 milliohm effective series resistance and a
40 amp load shows about 1/4 volt ripple, which is probably plenty low
enough for you. But if you can get a 20 millihenry 60 amp choke, you
can use just a couple 4700uF caps (less than 1/10 as much total
capacitance as in the first case) and have about 1/10 volt ripple.
Not that the choke would be either cheap or easy to find, but it can
help a lot in the filtering.

Cheers,
Tom

 

[Paul Mathews]

Thanks for the pointers. Having a name (Victor MegaPac) to search for
should help a lot. I've tried searching with every combination of words I
could think of and I usually end up finding stuff around 1-2 amps.

I have something like a 24,000 uF capacitor in my junk box but I didn't
think it would be enough. 60,000 eh? Time to start digging through ALL of
my boxes of stuff in the garage. I have an old NCR cash register somewhere
that has a lot of good stuff in it if I can only find it...- Hide quoted text -

- Show quoted text -

It's VICOR, not Victor. A couple of 2 kW 48V supplies in series
should do it, and most work fine in series. (Not so with parallel
operation.) By the way, if you do use transformer/rectifier/capacitor,
I doubt that the ripple needs to be much below 10% at 120 Hz. Some
posters have suggested ways to get down to under 1% ripple.
Paul Mathews

 

[Ulysses]

It's VICOR, not Victor. A couple of 2 kW 48V supplies in series
should do it, and most work fine in series. (Not so with parallel
operation.) By the way, if you do use transformer/rectifier/capacitor,
I doubt that the ripple needs to be much below 10% at 120 Hz. Some
posters have suggested ways to get down to under 1% ripple.
Paul Mathews

Oops. And I was being careful NOT to write Victor. Anyway, I found their
web site and they have a model that sounds like it's in the right power
range. It's just kinda scary because they don't have any prices posted.
Thanks for the info!

 

[Ulysses]

Sounds like the safest way for you to do this, given the lack of
resources, is find a big 2:1 transformer and rectify and filter it.
Unfortunately, a "pole pig" is very unlikely to be 2:1, though you
could use the secondary as a center-tapped winding that you put 120V
across and get 60V from. Does the output have to be ground-
referenced? If so, things get ugly again because you don't have a
grounded center-tap you can work from -- unless you can float the 120V
input. You'll need to float either the input or output, to make a
center-tapped full wave rectifier. A half-wave rectifier would be
just too awful to try to filter.

For the filter, I'd highly recommend using a choke in your filter, if
you want to get to seriously low ripple. A quick simulation using an
0.1 farad filter cap with 1 milliohm effective series resistance and a
40 amp load shows about 1/4 volt ripple, which is probably plenty low
enough for you. But if you can get a 20 millihenry 60 amp choke, you
can use just a couple 4700uF caps (less than 1/10 as much total
capacitance as in the first case) and have about 1/10 volt ripple.
Not that the choke would be either cheap or easy to find, but it can
help a lot in the filtering.

Cheers,
Tom

Wow. Millihenrys. Sounds big. Makes Faraday sound like a midget. I asked
about this on another NG a while back and a couple of people told me not to
use full-wave rectification but to just use two diodes. What you are saying
makes more sense to me. I've built several small (1-2 amp) power supplies
but never anything this big. My problem seems to be that I have no formal
education in electronics and it has always just been a hobby for me.
Because if this I don't know what it is that I don't know.

By "choke" do you mean an AC capacitor connected across the outputs before
it is rectified? And then the 4700uF caps after it's rectified? I think
it's time to call my father-in-law. He just might have something like that.

From what everyone has said it sounds like the ripple can be dealt with
satisfactorily. It also sounds like I'm going to have to break down someday
and get a better oscilloscope. The one I have is a Heathkit from the 60s
and the tubes are so weak I can't even get a good sine wave image any more.

 

[Tim Shoppa]

Greetings. I hope I'm in the right place.

I want to build or buy a power supply which will in effect mimic the output
of solarvoltaic panels. I want to feed the output into an OutBack MX60
charge controller.

This may sound too circular... but probably the most economical and
versatile way of getting what you want will be a bunch of car/golf
cart/etc. batteries connected in series, with some series resistance
(paralleled car auto headlamps to get your desired current) to switch
in/out as series drop elements if you want to mimic internal impedance
of a solar panel.

Ironic, huh? Using batteries to power a battery charger!

If there is no need for isolation you could also do this from AC power
with no transformer, and SCR's for voltage control. The output
waveform will look like that from a giant light dimmer and I don't
know if this can be adequately filtered to do what you want to test.

Tim.

 

[Tom Bruhns]

Wow. Millihenrys. Sounds big. Makes Faraday sound like a midget. I asked
about this on another NG a while back and a couple of people told me not to
use full-wave rectification but to just use two diodes. What you are saying
makes more sense to me. I've built several small (1-2 amp) power supplies
but never anything this big. My problem seems to be that I have no formal
education in electronics and it has always just been a hobby for me.
Because if this I don't know what it is that I don't know.

By "choke" do you mean an AC capacitor connected across the outputs before
it is rectified? And then the 4700uF caps after it's rectified? I think
it's time to call my father-in-law. He just might have something like that.

From what everyone has said it sounds like the ripple can be dealt with
satisfactorily. It also sounds like I'm going to have to break down someday
and get a better oscilloscope. The one I have is a Heathkit from the 60s
and the tubes are so weak I can't even get a good sine wave image any more.

No, the choke is used as follows: diodes to a capacitor. The choke
is an inductor (coil of wire, generally on an iron core for stuff like
this), and it goes from that first capacitor to a second capacitor.
The other side of both capacitors are connected together, and to the
transformer center tap. If the diodes are connected anode to the
transformer, and cathodes to the capacitor terminal, then that
terminal should be the POSITIVE one (or you get exploding
capacitors!).

You can do the filter with one big capacitance (maybe made up of
several capacitors connected in parallel), or with a couple much
smaller capacitors with a choke (inductor) in between them. What's
most economical probably depends on what you can pick up surplus at a
reasonable price. Maybe I'm biased toward using the choke because I
have some kicking around. ;-)

Be careful with this stuff...it's not terribly high voltage, but the
currents can melt things pretty quickly. I've heard of people doing
serious damage to fingers when they get a ring across something that
delivers enough current to make the ring really hot. You really don't
want to be trying to hold liquid gold in your hand.

Cheers,
Tom

 

[Paul Mathews]

text -



Oops. And I was being careful NOT to write Victor. Anyway, I found their
web site and they have a model that sounds like it's in the right power
range. It's just kinda scary because they don't have any prices posted.
Thanks for the info!- Hide quoted text -

- Show quoted text -

Hint: ebay
Paul Mathews

 

[Ulysses]

more.

No, the choke is used as follows: diodes to a capacitor. The choke
is an inductor (coil of wire, generally on an iron core for stuff like
this), and it goes from that first capacitor to a second capacitor.
The other side of both capacitors are connected together, and to the
transformer center tap. If the diodes are connected anode to the
transformer, and cathodes to the capacitor terminal, then that
terminal should be the POSITIVE one (or you get exploding
capacitors!).

OK, thanks!

You can do the filter with one big capacitance (maybe made up of
several capacitors connected in parallel), or with a couple much
smaller capacitors with a choke (inductor) in between them. What's
most economical probably depends on what you can pick up surplus at a
reasonable price. Maybe I'm biased toward using the choke because I
have some kicking around. ;-)

Well, I generally estimate how much it would cost to build something and
then figure out how to do it with the stuff I already have. If you have
enough boxes of junk....

Be careful with this stuff...it's not terribly high voltage, but the
currents can melt things pretty quickly. I've heard of people doing
serious damage to fingers when they get a ring across something that
delivers enough current to make the ring really hot. You really don't
want to be trying to hold liquid gold in your hand.

This is one reason I don't wear any jewelry. Power saws are other reasons.

 

[Ulysses]

This may sound too circular... but probably the most economical and
versatile way of getting what you want will be a bunch of car/golf
cart/etc. batteries connected in series, with some series resistance
(paralleled car auto headlamps to get your desired current) to switch
in/out as series drop elements if you want to mimic internal impedance
of a solar panel.

Ironic, huh? Using batteries to power a battery charger!

At least I'm not the only one. I've done something similar by using an
alternator, inverter, and 12 volt battery to run a battery charger to charge
12 volt batteries. It also made for a cheap inverter generator (sorta).

If there is no need for isolation you could also do this from AC power
with no transformer, and SCR's for voltage control. The output
waveform will look like that from a giant light dimmer and I don't
know if this can be adequately filtered to do what you want to test.

That's about what I was thinking when I first got the idea. With a maximum
input of 141 volts it seems like it the 120V AC from a generator could be
rectified and filtered then it might work.