# AC-AC step down voltage + frequency converter simulation (LT Spice), file provided

#### Mr Su

Joined Sep 2, 2021
2
Hi, below is a AC-AC frequency converter with a step down voltage from 200 to 20 Volts

Question: As you can see that the efficiency is 1.55/2.251 which is about 66%. May i know how to improve the efficiency of this circuit. I have tried to change the capacitance of capacitor since it absorbs a lot of power, however, if you reduce it too much, a lot of ripple will form which causes the frequency converter at the next circuit to be inefficient. Would be glad if anyone can offer me some guidance. I have post my LT Spice file below, try to simulate it and maybe offer some correction. I know that the voltage are too high for capacitor can handle but let's just forget about that now because I can adjust it. Thank you

#### Attachments

• 4 KB Views: 8

#### Papabravo

Joined Feb 24, 2006
21,225
I'm going to hazard a guess that switching losses are your main problem. Take a look at the waveform on the gate of M1 and ask yourself how much time the MOSFET is spending in the linear region. Your driving method has an active pulldown (good) with Q1, and a passive pullup (bad) with R1. There is probably a plateau on the gate voltage where time goes by, the gate voltage is not changing, and the MOSFET is in the linear region, wasting tons of power.

#### crutschow

Joined Mar 14, 2008
34,431
One problem is that you are applying 200V to the MOSFET gate, and it's maximum is likely no more than 20V.

And you cannot pull 2kW from an AC line with a half-wave rectifier as it would likely saturate the line transformer.
You should be using a diode bridge (and don't use the generic diode model).

Where will you get a 20V, 200A transformer?

Last edited:

#### ronsimpson

Joined Oct 7, 2019
3,037
One problem is that you are applying 200V to the MOSFET gate, and it's maximum is likely no more than 20V.
The 2N2222s are also broken.
The entire oscillator needs to be on a 15 volt supply.
The frequency is very slow. Is that by design?

#### Mr Su

Joined Sep 2, 2021
2
One problem is that you are applying 200V to the MOSFET gate, and it's maximum is likely no more than 20V.

And you cannot pull 2kW from an AC line with a half-wave rectifier as it would likely saturate the line transformer.
You should be using a diode bridge (and don't use the generic diode model).

Where will you get a 20V, 200A transformer?
This is like a charger. Frequency is set very high so that trabsformer used cab be small.

#### Papabravo

Joined Feb 24, 2006
21,225
This is like a charger. Frequency is set very high so that trabsformer used cab be small.
I see problems getting this to work the way you want it to.

#### ronsimpson

Joined Oct 7, 2019
3,037
Frequency is set very high so that trabsformer used cab be small.
I asked about frequency because my SPICE shows it running very low.

#### Irving

Joined Jan 30, 2016
3,885
I get the feeling there's some fundamental issues here. Why are you using discrete parts rather than a proper SMPS device, or is this some theoretical design? How will you accurately control the duty cycle and/or frequency to regulate the output? What do you plan to do with 20V @ 200A??? Have you considered the size of the secondary winding and the core material to handle that current? Do you know about 'skin effect' and how that's going to make such a transformer impractical?

#### Papabravo

Joined Feb 24, 2006
21,225
I get the feeling there's some fundamental issues here. Why are you using discrete parts rather than a proper SMPS device, or is this some theoretical design? How will you accurately control the duty cycle and/or frequency to regulate the output? What do you plan to do with 20V @ 200A??? Have you considered the size of the secondary winding and the core material to handle that current? Do you know about 'skin effect' and how that's going to make such a transformer impractical?
I think that 'some fundamental issues' was manifestly illustrated by @crutschow in post #3. It might be useful if simulators would pump out warnings, the way a compiler would, but sadly that technology appears to be uncommon. Maybe we could have the equivalent of a #pragma where we pass information to the simulator about absolute maximum ratings. In the meanwhile we have to rely on 'friends and relations' to cool our wilder fantasies.

#### ronsimpson

Joined Oct 7, 2019
3,037
if simulators would pump out warnings,
If SPICE could blow flames and smoke out at your face. I long for the sound of exploding capacitors. When SPICE = "10,000,000 volts" I wish we could send some of that to the mouse pad.
I would be happy for a simple "2N222 is dead from too much Vce".