decipher this german circuit 1

[wingneckedhorse]

http://patentscope.wipo.int/search/en/WO2006015713

What does the circuit in the above link do? We've been asked by a client to look into using this to combine power sources with different voltages into a "halfway voltage" power supply. E.g. Input 5V and 3V supply and out the other side comes a 4V supply. I can't make heads or tails of the circuit in the link. Simulations aren't matching up to what the breadboard proto does either; I can provide more details if anybody is curious, but neither supports a the requested use case. Any ideas?

 

[IRstuff]

?? It looks to be a source follower totem pole, n-channel on top, p-channel on bottom. However, it will have a gigantic dead zone in the middle. On bipolar versions, there would be two diodes between the base connections to remove the dead zone.

TTFN
faq731-376

Need help writing a question or understanding a reply? forum1529

 

[IRstuff]

see:

http://www.datasheetarchive.com/dl/Scans-005/Scans-00107226.pdf

The LH0033 has two diode connected BJTs, while the LH0063 uses a single, biased BJT.

TTFN
faq731-376

Need help writing a question or understanding a reply? forum1529

 

[IRstuff]

Oh, OK, should have run the translator first. So, all it is is a buffered voltage divider, but there doesn't seem to be anything solve the dead zone problem. And, it's not a particularly good buffer, because of that.

TTFN
faq731-376

Need help writing a question or understanding a reply? forum1529

 

[wingneckedhorse]

Ah ok. Thanks. How would it buffer the voltage value of the resistor divider across to the output? I can't visualize how that would happen unless the FETs were operating in their linear regions, but that would be no good for a power source, no? To me, it looks like either the top rail being passed or the bottom rail being passed with a dead zone from -Vfet_threshold to + Vfet_threshold. Is that not the case?

 

[IRstuff]

No, they're source followers, so they'll follow, with a FET threshold drop, the voltage at the tap in the resistor divider. If this is what you want to do, the LH0063 would probably be a better choice, since it's got something like 98% gain, which is better than the circuit in the patent, and it'll supply a reasonably amount of current.

http://www.datasheetarchive.com/dl/Scans-063/DSA2IH00121155.pdf

the original National Semi datasheet. ~1/4 amp from the LH0063.

Alternately, if you want to buffer the voltage divider, a simple power op amp circuit would be less of a bother, since its gain would be in the 99.9% range.

TTFN
faq731-376

Need help writing a question or understanding a reply? forum1529

 

[wingneckedhorse]

Got it now. Thanks!

 

[benta]

A patent on a complementary source-follower charging a couple of caps?

Wow... just wow...

Apart from that, if the FETs are depletion-mode types, the circuit will work. But those are scarce nowadays.

Benta.

 

[IRstuff]

Sure, but depletion mode transistors don't really help the accuracy, and they're typically not intended for high current applications.

TTFN
faq731-376

Need help writing a question or understanding a reply? forum1529

 

[electricuwe]

As a native German speaker I'd like to shed some light into the topic:

on the circuit itslef:

This is not a circuit for signal processing but for power. In power elctronics e.g in Inverters connected to 400 V three-phase mains, electrolytic caps have to be connectec in series due to their Limitation in voltage rating. To balance the voltage parallel resistors have to be added, creating significant no-load losses. The circuit ensures proper voltage sharing with much lower losses.

On patent law in the ROW (Rest of world, not US): The document attached is an international patent application, not a granted patent. According to depatisnet - an internet service of the german patent office - no national patent related to this application was granted anywhere.