Seeking: ~120V DC 20A supply

[David Lesher]

Sure it does. You have a big fat switcher that can be
precisely controlled and you don't have to stack several
telco switchers. Stacking switchers is always a white-knuckle
scenario.

The switcher in the VFD's I know are switching rectified line
voltage to create 3 phase; there are thus six sets of power
semis. The control function is to vary the frequency while
firing the semis in sequence. The output is 3phase AC, but
hardly sine waves.

I can't see what that has in common with a high power multi-tens
of KHz switcher, with a transformer suitable at the frequency &
power level.

The control circuitry is easy, what's difficult are the ferrite
transformer, the big semiconductors and the layout (for good EMC). The
latter are all done. I don't think you will find the perfect solution
with your desired control gear on a silver platter for this job.

While I've not repaired any VFD's, I've read a bunch of spec sheets
while seeking one. None I saw alluded to a HF transformer.

UPS and VFDs _are_ switching supplies. For the UPS kind you just have to
make sure that it is a true power conditioner type and can work sans
battery.

The UPS's I have scraped have real, 60 Hz, transformer iron.

 

[David Lesher]

It's not isolated. Why not just do a triac phase control, if isolation
isn't needed?

We want isolation. That means a transformer. At 60 Hz, that is huge,
but at 100 KHz, far less so.

 

[Joerg]

The switcher in the VFD's I know are switching rectified line
voltage to create 3 phase; there are thus six sets of power
semis. The control function is to vary the frequency while
firing the semis in sequence. The output is 3phase AC, but
hardly sine waves.

Modified sine or the ancient six-step? That must have been a real cheapo
VFD. The ones I dealt with had a real clean sine wave coming out. You do
not have to buy three-phase, they also come single-phase.

I can't see what that has in common with a high power multi-tens
of KHz switcher, with a transformer suitable at the frequency &
power level.

Well, it's been a while but the ones I saw had that in there. Plus a
bridge which you'd have to disable or bypass.

While I've not repaired any VFD's, I've read a bunch of spec sheets
while seeking one. None I saw alluded to a HF transformer.

If you mean a big ferrite transformer, all the ones I looked at had those.

The UPS's I have scraped have real, 60 Hz, transformer iron.

Wow, those must be from the days of Fred Flintstone

Fast forward 40 years and check out a modern "real" UPS in detail:

http://www.freescale.com/files/dsp/doc/app_note/AN3113.pdf

 

[P E Schoen]

"David Lesher" wrote in message

We want isolation. That means a transformer. At 60 Hz, that is huge,
but at 100 KHz, far less so.

There is another thread where someone showed a 3kW charger. Here are the
specs:
http://www.ev-power.cz/docs/GWL-Power-POW-PFC_1-8kW-Spec.pdf

The thread is:
http://www.diyelectriccar.com/forum...ion-vw-polo-ac-industrial-motor-78701p12.html

Meanwell makes battery chargers up to 1000W standard, and possibly up to
3000W special:
http://www.meanwell.com/product/inverter/inverter02_2.html

Maybe you can use three chargers.

Paul

 

[Joerg]

Still, kilowatts of drivers and transformers and rectifiers is
non-trivial. There will be hundreds of watts of heat to get rid of.

Not necessarily. I just sim'd another converter, a step-up to 28V.
Including magnetics losses it clocks in at slightly above 97%. When the
boards come back they are usually about where the sim said.

I still think Mike Engelhardt should get a Nobel prize for LTSpice.

 

[[email protected]]

As you should. You should have said "Selenium bridge rectifier..."`

I've actually done this to charge a traction battery pack. The only
difference was that instead of a stepper motor, a student kept an eye on
the multimeter and cranked the Variac down as the battery charged up.
Works fine, if a bit labor-intensive.

It sounds like no one has seen what we need;

Zivan makes exactly what you need, if you don't mind buying the charge
controller and power supply all packaged into one unit. If you want to
roll your own controller, Sorensen probably makes what you need for the
power supply.

and/or the 750W 48V max supplies are not that bad after all.

I just remembered another manufacturer of telecom-type supplies, Valere.
Same problem as you have now, though; for most of their products, the
world ends at 54 V. They do have a few that are available in 125 V DC
output; they are the style where you buy a shelf and then buy modules
to plug into it. Usually you can get some kind of data bus to the
shelf.

There is a separate charging controller processor. It has lots of AD
inputs and measures every cell while charging.

Watchdog timer driving the I/O pin that goes to the big contactor?

If any cell gets too high in voltage, or too warm; charging terminates
with an error report. At $135/cell, you do not want to cook any.

Yow! At cell voltages of popular chemistries, and a 120 V pack, you've
got somewhere between about 33 and 100 cells, or $4,455 to $13,500 in
all.

Matt Roberds

 

[David Lesher]

I've actually done this to charge a traction battery pack. The only
difference was that instead of a stepper motor, a student kept an eye on
the multimeter and cranked the Variac down as the battery charged up.
Works fine, if a bit labor-intensive.

With a selenium stack? I hope not...

Zivan makes exactly what you need, if you don't mind buying the charge
controller and power supply all packaged into one unit. If you want to
roll your own controller, Sorensen probably makes what you need for the
power supply.

But we don't want to blow oogobs of $$ on fullpriced new...

I just remembered another manufacturer of telecom-type supplies, Valere.
Same problem as you have now, though; for most of their products, the
world ends at 54 V. They do have a few that are available in 125 V DC
output; they are the style where you buy a shelf and then buy modules
to plug into it. Usually you can get some kind of data bus to the
shelf.

The 750W ones are $30 each, so we can tolerate some design pain
at that delta-$

Watchdog timer driving the I/O pin that goes to the big contactor?

Not the BIG contactor; charge isolation one.

Yow! At cell voltages of popular chemistries, and a 120 V pack, you've
got somewhere between about 33 and 100 cells, or $4,455 to $13,500 in
all.

It's on the lower end but still the biggest investment...

 

[[email protected]]

With a selenium stack? I hope not...

No, just a plain old square brick rectifier bridge with Si diodes,
something like 600 V, 35 A. I found a piece of random sheet metal lying
around the shop and bolted the bridge to that for a heatsink. Heat sink
grease was in short supply, so we used a little wheel bearing grease
instead. Worked fine, if a bit greasy-smelling when warm.

But we don't want to blow oogobs of $$ on fullpriced new...

Understood. I was trying to supply some brand names that might be good
keywords for Ebay, Craigslist, etc.

The 750W ones are $30 each, so we can tolerate some design pain
at that delta-$

Are the outputs full floating? You could *try* stacking three of them,
but as mentioned elsethread, that's generally a bit of a gamble with
switchers.

Matt Roberds

 

[Jasen Betts]

Modified sine or the ancient six-step? That must have been a real cheapo
VFD. The ones I dealt with had a real clean sine wave coming out. You do
not have to buy three-phase, they also come single-phase.



Well, it's been a while but the ones I saw had that in there. Plus a
bridge which you'd have to disable or bypass.



If you mean a big ferrite transformer, all the ones I looked at had those.




Wow, those must be from the days of Fred Flintstone

I've got an 3 year old one here with a 50Hz transformer.
the device weighs about as much in iron and copper as it does
in lead. the badge says "Dyn-a-mix" which is the importer's
mark.

It's just a little one. 1200VA output, 230Wh of battery.