12v=19v 4,7A dcdc

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
N-Channel said:

....As for my humble little setup, depending on the load being drawn, if it draws, say 500W continuously (w/o the engine), maybe it would last say, six hours....

I must've been on drugs or something. :xeye: :xeye: At 500W, it will last 4.08 hours, under IDEAL conditions, with the extra batteries (1.92 with only the two on-board 40Ah batteries) which we rarely, if ever, encounter, not the 6 hours originally predicted. I would say more like 3.5 hours, tops, with the extra batteries, and 1.5hours w/o them, then much longer with the engine running.
 
i have built this DC/DC convertor and it does work very well
it powers my 6 amp laptop with ease and has no problems at all with overheating e.t.c even inside a inclosed metal box. i used a smaller core for the coil feeding the MOSFET's, then the one sugested and i found that it works just as well, but the same colour code white/yellow
this design is well worth saving. and still very price competitive against the comercial units, worked first time as well only thing i needed to adjust was the feedback resistors the ones originally
gave me 21v maybe because of slightly different variations of components. but now it stays at a nice 19.3 v.
thanks for sharing your desgin.
:)
 
See Post #74, this thread. As the opposite F takes over, any remaining flux is automatically cancelled out. Energy is stored during the MOSFET's ON time, and released during it's OFF time. Once the period is complete, when the opposite MOSFET turns on, that is when any remaining flux in the core is cancelled out.
 
We know that ordinary boost regulator is not efficient in ferrite size, since it only use 1 quadrant for its operation.

With your topology (twin boost on common choke), I think that this problem is solved, the core is now more efficient (there's resetting mechanism like in push-pull).

The core size VS power handling is now improved, compared to ordinary boost regulator. Am I right here?
 
You are indeed correct. Thanks for the compliment on it being my topology, however I'm sure proper credit is deserved elsewhere. :)

Anyway, with two of the four quadrants now being used, power density is effectively doubled, and efficiency goes up.

I originally got this idea from the many application notes that come in the monthly issue of Electronic Design. They go up to five Fs, but the majority I have seen are two or three Fs. I think two is sufficient, especially since that's the number that will work on a common core.
 
New 12-19V Boost

All:

Here are a few pics of my new 12V to 19V DC-DC Uppie I am doing for a laptop. I gutted several old Dell PA-1 & PA-2 AC-DC Laptop powerr supplies and re-did them with the aforementioned 2-F topology. Since making my own pc Boards is a royal pain in the a$$, I decided to do them in Express PCB, and have Them professionally done. Double-sided, plated through, but no silk-screening, and no solder mask (who needs 'em anyway?). 6 boards for a little over US $110.00. Can't wait to finish them in time for Field Day.

Comments, Please.
 

Attachments

  • case with lid (676 x 507).jpg
    case with lid (676 x 507).jpg
    77.7 KB · Views: 473
Actually, this won't work as a boost converter if your earlier explanation still is true (wound like a push pull transformer). It is actually a forward converter.

Think of what happens with the voltage on the other side of the primary when one side switches to ground. For 12V input the other side tries to go to 24V. But still - there is leakage inductance so regulation will still work just as in a half-bridge or push-pull without output inductor. The ripple current will be quite large though and it won't scale well. The core used should be ferrite for best performance. And you could add a filter inductor after the diodes if you want to be able to drive it in continous mode. Output voltage is limited between Vin and Vin*2.
 
Actually, it DOES work. This is the second such supply. The best description of the circuit would be a 2-F interleaved flyback. It is ~90-95% efficient, and is based on several related designs already available. Original idea came from N8XJK's booster appearing in Nov '04 QST. Other idea came from ST App Note for a new 2-channel PWM chip for just this purpose. My twist on this was to wind both inductors on the same toroid core in anti-parallel. Their common point goes to +Vin (~10-16VDC), and each end goes to the Source-Cathode junction of a pair of MTP75N06s and an MBR2545CT Schottky diode. V(flyback) is limited to 19V by clamping action.

Here is a pic of the original working model.
 

Attachments

  • sg3525 dc-dc uppie 1 smaller.jpg
    sg3525 dc-dc uppie 1 smaller.jpg
    87.2 KB · Views: 451
I didn't say it doesn't work - just that it isn't a flyback any more. :)

"My twist on this was to wind both inductors on the same toroid core in anti-parallel."
That is a big difference! When one transistor is on the other end will spring up to 24V. Leakage inductance limits the voltage/current though. So it does work, but not much as a boost converter. Stored energy in the core will go to the output but most will be delivered in forward mode.

If the cores are separate it will work as a boost though.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.