topology suggestions for 3.5kW batt charger?

For a low power case, this would be fairly easy to get working. At many KW, all designs will be hard to get working. Remember that I did say I posted it "just for fun". It, I think is a cute idea because it is so well suited to doubling the voltage which it just happens is about what you need.

The output diodes will take it on the chin if there is a short. The circuit will be quite broken as a result.

I think it would take more parts to do it with one of them than without. The output drive would not be enough to drive the needed MOSFETs, there is no good way to make the overlapping drive and it needs a forest of parts to make it happy.

The LT3827 looks easier to talk into making the needed waveforms.

I'm sure someone will point out that a PIC can do it much easier.

Well, my attempt to dismantle the transformer from a cheap-o CompUSA power supply last night met with total disaster. The, one detail we both seem to have forgotten is that commercial switchers tend to have their transformers dipped and/or vacuum impregnated. Needless to say, when I tried to separate the core halves I ended up breaking the ferrite into lots of little pieces.... This does not bode well for re-winding the transformer on this beast:

An otherwise excellent choice! Only $100 and you get yourself an honest 3kW of 48V output. I'll probably buy one anyway and hope it uses a PQ44040 core (as I have those in stock). The education one can glean just from taking something like this apart is worth the price of admission!

-Jeff

I took a look at the LT part and agree it would be ideal for producing the right gate drive signals, however, it is expecting to control a buck regulator and the KenSmith idea certainly ain't one of those. Closing the loop, then, might be difficult with this chip. Also, is that a right half plane zero I see lurking in this creation of yours???

Don't even say the P-word ("Phil" or "PIC", now that I think about it)!#$^&

Yeah if theyr glued or dipped or worse potted, then its hard, sometimes soaking them in something realy disolving ie acetone and sometimes the glue gives up if you get it hot enough, but then that melts the plastic core wich is usefull to keep intact, but I have used paper soaked with epoxy for a core on many occasion.

fortunatly the supply I had just had a clip holding it together, no glue or varnish. ive glued broken cores back together before, ofc this is only realy ok if its a gapped core anyway. but ive got quite a few bits of broken cores lol.

but also I managed to get my full voltage out of the 5v winding by putting in a much higher voltage switching transistor. you maye be able to use the

you can of course buy quite big ferrite core/transformers.

Colin =^.^=

A pity. Some years ago I designed something similar using a motorised Variac, with a current-sensing relay for control. It charged a large bank of NiFe batteries.

The relay armature hovered between two contacts when the current was correct. If the current fell or rose, the armature touched one of the contacts, pulled-in an auxliary relay and drove the motor to wind the Variac up or down by means of sprockets and a chain.

There was an interlock that ensured that the Variac always wound down to zero before the battery circuit was closed. The main battery circuit breaker was magneto-mechanical with a magnetic arc blow-out coil.

There was also a timing cam that pulsed a uniselector around at hourly intervals until the voltage rose above a pre-set threshold value, then a second uniselector was pulsed at half-hourly intervals until it caught up with the first one and terminated the charge. That ensured that the battery received the final third of its charge after it reached the gassing point.

"Adrian Tuddenham"

Banks of large capacity Nickel Iron cells ?

Inside a Wehrmacht Tiger Tank maybe ??

Or a British Navy diesel submarine ?

Or the radio in a Ferret ?

....... Phil

What the designers intended the chip for is just a limition of their imaginations not ours. In fact, the KenSmith circuit sometimes looks a little like a bucker. Right at the peak of the waveform, the transistors aren't overlapping. It looks a little like a bucker feeding a 2:1 step up squarewave.

We are only charging a battery. The transient responce doesn't matter too much. We can slow the circuit way down. There can be a very slow soft start to ease it up to the regulated charging current and voltage.

You really need a diode ORed control signal because you want the charger to pour in the current right up to the charged state and then back off to perhaps a holding voltage.

The pole in the RHP isn't lurking. It is jumping up and down, waving and yelling. It has a couple of high frequency friends hidden out in the weeds too.