Pierre's 170W in 1600W out Looped Very impressive Build continued & moderated

[listener192]

Latest result with 10 coils in series North and 10 coils in series South.An issue in code was giving a bad result for this configuration previously.
If I break the series loop of coils, output falls a little, and waveform distorts.

You can see the input current is 3A, so if this can be increased  9A, there would be 230V RMS across this load.
It will maintain this output to over 70Hz.
I tried Gotoluc's 3 coil scheme however that had a square waveform and lower output.
Perhaps each of the 5 coils have to be on an isolated H bridge.

Note we have a bit of phase shift between voltage and current due to the small clean up cap on the output.
Hardly any recovery current present.
L192

[pmgr]

Hi pmgr
Can you please say some more on this? Do you mean just using the body diodes without driving those two mosfets? And if yes, is there any particular reason for choosing this method instead of two simple diodes?

Thanks

Jeg, you can check this TI note on back to back FETs:

http://www.ti.com/lit/an/slva948/slva948.pdf

Compared to using simple diodes, this method won't have the voltage drop of the diode when turned on as the current will flow through the FET. When turned off, the body diodes back-to-back will block any current in any direction.

PmgR

[Jeg]

Jeg, you can check this TI note on back to back FETs:

http://www.ti.com/lit/an/slva948/slva948.pdf

Compared to using simple diodes, this method won't have the voltage drop of the diode when turned on as the current will flow through the FET. When turned off, the body diodes back-to-back will block any current in any direction.

PmgR

Nice gift pmgr! Thank you

[listener192]

This is a  scope shot that accompanies the last set.
It shows an average of 0.5A recovery current into every bridge board rail cap.
This for a forward pulse current of approx 4A.
This can only be seen by extending the cap off board with wires, so a current clamp be be applied.
This shows that most of the recovery current  is absorbed locally and not on the main cap, bank even though external diodes are provided.
L192

[listener192]

To obtain lower input voltage high current operation, each H Bridge could control 1 or 2 coils. In the second coil arrangement the second coil would be 180 deg opposite and would be connected in reverse parallel to form the opposite pole.

As then we would not need to reverse operate the H bridge, in this configuration we could use one Half bridge as a boost converter, which I could never see working with Pierre's series coil connection.
In the figure attached,  Q1 and Q3 are switched on to charge the coil, Q2 is not used and remains open.When the coil turns off, Q3 is opened and Q4 is closed and the inductor charge is pushed by the rail (input) voltage into C. Q1 remains closed.

You may see from the above that Q1 & Q2 are not required, so just a half bridge is used. C = the big cap bank .
The input voltage push ensures all of the coil charge is transferred to C.
In Pierre's series coil arrangement, for the boost to work, the HHS are all connected to the + rail so they cannot have body diodes if they are MOSFETS. Relays would not have that problem. As Pierre has shown, the recovery diode goes from the coil to the super cap.   
The LSS's also don't need a body diode, although the presence of one is not detrimental to operation.So when discharging the coil the HSS charging the coil needs to stay closed. The LSS that was on for charging, then switches off. The recovery diode connected to the LSS and the coil conducts the coil charge pushed by the rail in the cap bank.

So the first thing we are doing wrong is having HSS body diodes in circuit (in the series configuration). These body diodes  left in place, stop the current flow as they are connected to the same DC rail as the input, so no push. There would be some recovery via the current flow through the other LSS body diode but not a boost function. One way of eliminating the action of the body diodes would be series diode in the supply side of the HSS's. This would stop the body diodes conducting during coil discharge and external diodes connected to the coil can need perform a boost recovery

The second problem is the overlap.. not the second HSS switch being turned on, but the second LSS being turned on before the first LSS is turned off. This will stop the boost function by not allowing the end of the coil chain to float, keeping the coil chain in charge. I cannot see anyway of have an overlap and a boost function together in the series configuration.

So using the half bridge on a parallel coil connection, the HHS body diode (Q4 in this example) is fine, because the it is not connected to the input rail, only the cap bank, so the boost push will function.
Overlap is not a problem, as the adjacent coil has its own Half bridge.
If using the BTS7960B bridge board, the rail connection would have to be split into two i.e. separate + inputs for each half bridge.

You can use the figure attached to visualize this.
Taking the concept further, as Q4 in our half bridge is not connected to rail , we could replace the HSS with a diode, as current only needs to flow in one direction.This leaves us with one LSS Q3, which could have isolated drive, but otherwise simplifies control and reduces cost.


L192