- Joined
- Jan 21, 2010
- Messages
- 25,510
You should be able to purchase access to a single Silicon Chip issue. Access to errata in their magazines is free, so if there are any updates you'll be able to read them too.
The main thing you'll get from it is the method of operation.
For 1kHz, the mosfet needs to be on for 50uS every 1ms. The inductors are 1mH and 220uH and the capacitor is 100uF.
Some rough "back of a napkin" calculations are that for 3000 times the frequency, you would require the mosfet to be on for 17ns each 330ns, appropriate inductors would be 330nH and 70nH, and the capacitor around 0.033uF
In the SC article, they use a IRF540N. This has a gate charge of 71nC. By my estimation, if you want to turn the device on and off in (say) 4ns (which for this circuit is probably too slow) then you need a gate current of around 18 amps. That is way too high (the specs suggest that you should expect a turn on time of 11ns and a rise time of 35ns, so lclearly this device is not appropriate) . A mosfet with higher Rds(on) will have lower gate capacitance.
Another random mosfet, ST 5NE10L-1 has a lower gate charge (around 10nC, but a higher Rds(on) of around 0.27 ohms). This *might* be able to be switched on and off fast enough. You would require a gate current of 4A probably.
So, So all you require is a driver capable of delivering a 4A gate current to turn it on and off. This is also a non-trivial exercise (the gate current probably approaches the current you're switching with the mosfet!
You probably need one of the other moderators to assist you with this.
Once you have figures all of that out, you need a way to prevent massive RFI as you will have created a very powerful RF transmitter. You will also need to minimise inductance in the leads to the battery or all that power will go precisely nowhere.
Many of the components (the mosfet and the diode at a minimum) will require heatsinking because there will be significant switching losses.
The main thing you'll get from it is the method of operation.
For 1kHz, the mosfet needs to be on for 50uS every 1ms. The inductors are 1mH and 220uH and the capacitor is 100uF.
Some rough "back of a napkin" calculations are that for 3000 times the frequency, you would require the mosfet to be on for 17ns each 330ns, appropriate inductors would be 330nH and 70nH, and the capacitor around 0.033uF
In the SC article, they use a IRF540N. This has a gate charge of 71nC. By my estimation, if you want to turn the device on and off in (say) 4ns (which for this circuit is probably too slow) then you need a gate current of around 18 amps. That is way too high (the specs suggest that you should expect a turn on time of 11ns and a rise time of 35ns, so lclearly this device is not appropriate) . A mosfet with higher Rds(on) will have lower gate capacitance.
Another random mosfet, ST 5NE10L-1 has a lower gate charge (around 10nC, but a higher Rds(on) of around 0.27 ohms). This *might* be able to be switched on and off fast enough. You would require a gate current of 4A probably.
So, So all you require is a driver capable of delivering a 4A gate current to turn it on and off. This is also a non-trivial exercise (the gate current probably approaches the current you're switching with the mosfet!
You probably need one of the other moderators to assist you with this.
Once you have figures all of that out, you need a way to prevent massive RFI as you will have created a very powerful RF transmitter. You will also need to minimise inductance in the leads to the battery or all that power will go precisely nowhere.
Many of the components (the mosfet and the diode at a minimum) will require heatsinking because there will be significant switching losses.