Simplified AC PWM Controller

Thread Starter

nekojita

Joined Nov 19, 2010
170
Hi,

In an earlier post for an AC PWM controller, there was a lot of great commentary and feedback. As I showed in one of my last posts, I did create a working prototype.

In moving this into KiCAD to begin a layout, I realize that there are simply too many components for the space that I have to work with.

Looking back at some of the articles that I found while researching the AC PWM controller, I found one that has a similar circuit to the final circuit that the thread mentioned above but with many fewer components.

The circuit and output are shown below:
1610393168229.png
1610393220175.png
Simulating this circuit in LTSPICE does not give the same results. Attached is my LTSPICE version and the complete article.

If anyone in the Forum has any thoughts or comments, they would be greatly appreciated.

Thanks,

Neko
 

Attachments

Danko

Joined Nov 22, 2017
1,829
Your schematic and my posted schematic look identical. What am I missing?
Hi Neko,
Added V3 20V.
VCC of U1 = 20V.
L1 = 47m instead 47u.
L2 and L3 = 100m instead 100u.
C4 = 10u instead 1u.
Added ammeter "Motor".
 
Last edited:

Thread Starter

nekojita

Joined Nov 19, 2010
170
Hi Danko,

Thanks for pointing out those changes. My wife says that there is a Cantonese saying that goes, "Some people have eyes but they can't see." :)
A couple of comments and questions on your mods to the original circuit:

• The transformer inductances are an order of magnitude higher than what I had. Do they really need to that high?
• I just grabbed the LTC4442 gate driver from the LTSPICE Power Library because it was available. May use some other gate driver. Thoughts?
• A VDD of 20V is high. Could this be reduced by increasing transformer secondary inductance? (or reducing primary inductance)
• Did you test the circuit over a range of PWM duty cycles? The reference paper suggests that 10-90% DC can be achieved with transformer coupling.

Please let me know your thoughts on these points.

Thanks,

Neko
 

Danko

Joined Nov 22, 2017
1,829
Hi Danko,

Thanks for pointing out those changes. My wife says that there is a Cantonese saying that goes, "Some people have eyes but they can't see." :)
A couple of comments and questions on your mods to the original circuit:

• The transformer inductances are an order of magnitude higher than what I had. Do they really need to that high?
• I just grabbed the LTC4442 gate driver from the LTSPICE Power Library because it was available. May use some other gate driver. Thoughts?
• A VDD of 20V is high. Could this be reduced by increasing transformer secondary inductance? (or reducing primary inductance)
• Did you test the circuit over a range of PWM duty cycles? The reference paper suggests that 10-90% DC can be achieved with transformer coupling.

Please let me know your thoughts on these points.

Thanks,

Neko
Hi Neko,
I think, you have only one big problem: to design wideband (0.1 Hz - 10 MHz) transformer.
Or you can use miniature HF transformer, as @Bordodynov suggested in this post.
You are designer, so it should be your choice.
 

Thread Starter

nekojita

Joined Nov 19, 2010
170
Hi Danko,

Transformers are always a tough one. I haven't tried to dial in a wideband transformer for this application. As my colleague always reminds me, "There's no free lunch."

Due to the transformer design limitations, my thought is to go to the circuit in Figure 6 of the research paper. Using your simple DC-DC converter circuits previously posted but in a single positive voltage as shown in Figure 6, this comes in between the last Danko circuit (prototype tested) and the very simplified circuit last posted.

What do you think? What are the perils and pitfalls?

Thanks,

Neko
 

Thread Starter

nekojita

Joined Nov 19, 2010
170
Hi Danko,

This version starts adding parts back in, but still probably easier than dealing with the wideband transformer.

I think I hve it drawn correctly but there is only a very small output. Do you see what I am doing wrong?

Thanks,

Neko


1610662787844.png
 

Attachments

Danko

Joined Nov 22, 2017
1,829
This version starts adding parts back in, but still probably easier than dealing with the wideband transformer.
I think I hve it drawn correctly but there is only a very small output. Do you see what I am doing wrong?
Hi Neko,
You can start with this circuit:
1610709687521.png
1610709800761.png
ADDED:
Pay attention to short circuit current through chain of M1, M2, M3, M4.
In real circuit this current may be much bigger:
1610735360227.png
 

Attachments

Last edited:

Thread Starter

nekojita

Joined Nov 19, 2010
170
Hi Danko,

Thanks for pointing this out. Transient conditions with high voltages can blow out even the sturdiest MOSFET. Digikey shows this spec for M1-M4:

Detailed Description

N-Channel 250V 64A (Tc) 300W (Tc) Surface Mount D²PAK (TO-263AB)

To further protect M1-M2, I added the delay back in so that M1-M2 can't turn before the +5V is stable. See below:

1610840837216.png

Also, it turns out that I don't need the LTC1693 driver since the LTC6992 can drive the optos.

Neko
 

Thread Starter

nekojita

Joined Nov 19, 2010
170
Hi Danko,

Another consideration to put into the mix. The LTC6992-1 provides a duty cycle of 0% to 100%. Ceiling fans in poice articular, have a minimum stall torque. Running these fans without enough drive to overcome the stall torque will burn out the motor winding. Maybe a better choice of PWM chip would be the LTC6992-2, DC= 5% to 95%. Putting in a trim based on a Vref would allow the fan user to setup a minimum duty cycle if the 5% is still too low.

Commercial fan speed controls have a trim pot for this.

Make sense?

Thanks,

Neko
 

Thread Starter

nekojita

Joined Nov 19, 2010
170
Hi KeepitSimpleStupid,

The operation of the circuit is explained well in the attached paper at the beginning of this thread.

PWMACChopperControlofSingle-Phase-2.pdf

Neko
 
Top