Body
Hi Sven,
"ok, twist Gate Source lines and I always thought to twist both gate lines, then I can save the one line from the capacitor."
When a loop is formed this is a coil with one turn! When two such loops shares the area as in the picture they form a transformer. I.e. one loop is the primary and the other loop that receives the magnetic field of the first loop is the secondary.
Be very careful with the ground wires as they may share current from different circuits. A normal way to avoid this is to make a star connection. That is a central point where different ground wires from different circuits meet. This prevents sharing the current from a different circuit along part of a ground wire that is shared by the other circuit which then sees the voltage drop of the first circuit. The voltage along the ground wire can be the flyback voltage from an abrupt turn off of a MOSFET. This voltage can be enough to charge the gate-source voltage of another MOSFET to cause it to turn on when it actually had to be in the off state.
Put decoupling capacitors (local energy reservoirs) between the power rails close to semiconductor switches and their loads.
For even better decoupling each of the three circuits (a switch with load and decoupling capacitor) each has its own ground wire directly to the main capacitor of the voltage supply. This is star coupled ground which avoids voltage drops to the other parts of a circuit. In the above image the two first switch circuits will see the voltage drop along the ground wire (or the +V wire if this is used for reference). This may cause false gate signals in the first two switches if the gate drivers are referenced to the left end of the GND wire. The left end has GND potential while the voltage increases when moving to the right and the right hand switch is closed. Remember every component has the three fundamental parasites of resistance capacitance and inductance. Thus the ground wire gets a voltage drop along its length when traversed by a current. It induces voltages during changing currents.
Regards
Ole
---In [email protected], <s.friedrich@...> wrote :
Hi Ole, ok, twist Gate Source lines and I always thought to twist both gate lines, then I can save the one line from the capacitor. The source line from the condenser to the Mosfet then just run and bridge or grind to the next mosfet!?!. No matter what I do a driver turns off at a certain pulse width uncontrolled from the image of the right TLP250 imme if the voltage is about 100V is about. I will revise it and take a new picture, first thank you. Greetings Sven
Von Samsung Mobile gesendet
-------- Ursprüngliche Nachricht --------
Von: "onielsen@... [EVGRAY]"
Datum:08.07.2018 17:01 (GMT+01:00)
An: [email protected]
Betreff: Re: Aw: [EVGRAY] Re: Push pull generator. [1 Attachment]
Re-post with the picture attached.
Regards
Ole
---In [email protected], <onielsen@...> wrote :
Hi Sven,
The capacitor in parallel to the parasitic gate-source capacitor can be omitted if using a low impedance gate driver placed close to the MOSFET for minimizing inductive loops. Twisting the gate source wires forms a transmission line of less impedance than the big loops on the picture. Twisting the wires of different gate circuits together couples them together. This can be seen in the scope shot where spikes are transferred between the two gate circuits. This also happens between the big loops that work as loop antennas which both send and receive the signals. See the picture below if it makes it through Yahoo.
"For me, C1 is 1μF MKP and C2 is 68μF Elko.
R2 2.2 ohms and R3 now 10K."
This doesn't tell anything without knowing how or what they're connected to.
Regards
Ole