Doubts regarding 100W inverter design

[Tahmid]

Regulation will be a little worse. Although maybe it won't hurt too much.
Taking feedback from AC output will compensate for everything - losses and drops in both the DC-DC converter and DC-AC converter. Taking output from 320VDC will only compensate for losses in the DC-DC converter. Although losses in the DC-AC converter aren't too great.

Hope this helps.
Tahmid.

 

[sriblue7]

Thank you Tahmid..
I have a 12V 135Ah battery with me,thats why i am continuously insisting on 12V.Though i design a 1000W inverter i may use it to provide a max of just 600W.And my aim is to design an inverter with the available battery.

1)I have attached two photos of the circuit designed based on your suggestion.please confirm me whether the design is efficient(except for the 12V battery!) with minimal power loss.Its just a basic diagram.


2)In the above circuit is it necessary to use two IR2110 at stage 2?
3)Should i put MOS drivers at the push-pull stage too?

 

[MAXPAYNE]

Regulation will be a little worse. Although maybe it won't hurt too much.
Taking feedback from AC output will compensate for everything - losses and drops in both the DC-DC converter and DC-AC converter. Taking output from 320VDC will only compensate for losses in the DC-DC converter. Although losses in the DC-AC converter aren't too great.

Hope this helps.
Tahmid.

Dear Tahmid,

I took feedback(voltage divider feedback) from the DC-DC stage and waveshape in the MOSFET drains was severely distorted, due to this MOSFET was getting severely heated. I didn't attached the DC-AC stage in that ckt and put bulb loads in the DC-DC stage output. Do I have to isolate the feedback using optocoupler ?

 

[Tahmid]

Thank you Tahmid..
I have a 12V 135Ah battery with me,thats why i am continuously insisting on 12V.Though i design a 1000W inverter i may use it to provide a max of just 600W.And my aim is to design an inverter with the available battery.

1)I have attached two photos of the circuit designed based on your suggestion.please confirm me whether the design is efficient(except for the 12V battery!) with minimal power loss.Its just a basic diagram.


2)In the above circuit is it necessary to use two IR2110 at stage 2?
3)Should i put MOS drivers at the push-pull stage too?

Connect gate resistors for each MOSFET - between the input signal (from the SG3525 output) and the MOSFET gate.
Connect two pull-down resistors from the SG3525 outputs to ground. This is just a safety precaution and may be left out. But it's better you connect it.
You should connect snubbers across the transformer windings. Refer to the project here: **broken link removed**
You should use a 11V-to-340V transformer or 12V-to-370V transformer. Otherwise headroom for feedback will be very small.
Between the output capacitor and the bridge output, connect an inductor.

Same for 2nd circuit.
Connect gate resistors for each MOSFET.
In the 2nd circuit, connect a 1k resistor from the gate to source of each MOSFET.
Yes, you need two IR2110s.

In the push-pull stage, you should put a MOS driver as you're using 3 MOSFETs in each leg at high frequency. And these MOSFETs have quite large gate charge and capacitance. You can use a BJT-based totem-pole driver as shown in the project: **broken link removed**. Or you may use an IC-based solution.

Hope this helps.
Tahmid.

- - - Updated - - -

Dear Tahmid,

I took feedback(voltage divider feedback) from the DC-DC stage and waveshape in the MOSFET drains was severely distorted, due to this MOSFET was getting severely heated. I didn't attached the DC-AC stage in that ckt and put bulb loads in the DC-DC stage output. Do I have to isolate the feedback using optocoupler ?

Did you use snubbers?
Did you check the drain waveforms without feedback? What were they like?
Did you drive the MOSFETs with a driver or directly from the SG3525? Which MOSFETs did you use and how many?
Try to slow down the feedback a little.

 

[MAXPAYNE]

I used snubber.

without feedback the drain waveform was totally clean with no distortion and mosfet temp. was almost at room temp.
i drove them using the same design as **broken link removed**
I used IRF3205 X 2 per channel. I didn't used any inductor in the output dc. I just put capacitor after the rectification stage.

 

[Tahmid]

Increase the value of the capacitance between pin 9 (COMP) and ground to 100nF. And also use an inductor between the output of the bridge rectifier and the capacitor.

Let me know the result.

Hope this helps.
Tahmid.

 

[MAXPAYNE]

Increase the value of the capacitance between pin 9 (COMP) and ground to 100nF. And also use an inductor between the output of the bridge rectifier and the capacitor.

Let me know the result.

Hope this helps.
Tahmid.

Ok. I will let you know the result..

Can you give a tutorial on feedback loop design and compensation in smps.

-Roomi

 

[sriblue7]

How to calculate the value of the resistor and capacitor in the snubber and also the inductor at the output of the rectifier?
is it ok to use the same BJT (BD139,BD140) in the link that you provided?

 

[Tahmid]

Ok. I will let you know the result..

Can you give a tutorial on feedback loop design and compensation in smps.

-Roomi

I think you can better understand this if you read from a book.

You can read it in "Power Supply Cookbook" by Marty Brown.

And, what is the result?

- - - Updated - - -

How to calculate the value of the resistor and capacitor in the snubber and also the inductor at the output of the rectifier?
is it ok to use the same BJT (BD139,BD140) in the link that you provided?

Which circuit are you referring to? Project 89 - ESP?

 

[MAXPAYNE]

I think you can better understand this if you read from a book.

You can read it in "Power Supply Cookbook" by Marty Brown.

And, what is the result?

- - - Updated - - -



Which circuit are you referring to? Project 89 - ESP?

talking abt this ckt, **broken link removed**

still didn't manage to test the ckt as i don't own an oscilloscope

 

[Tahmid]

Snubber is, I would say, more of a trial and error thing than just calculating from raw numbers. The variables involved really come into play in the real circuit. And these vary from sample to sample, circuit to circuit, especially in the case of hand wound transformers. So, I would say that you start from one value - the one in the link provided - and see if it provides reasonable results. If not, adjust the values and observe the results.

 

[sriblue7]

Thank you Tahmid....

 

[sam781]

For 1000W, you can use an ETD54 core. You can increase the frequency to use the transformer at higher power. But at 50kHz, you should be able to use ETD54 core. Maybe even ETD49 core, depending on the core. But you say that you won't design the transformer. But maybe when purchasing a core, you can take a look at the core used.

Hope this helps.
Tahmid.

@Tahmid: So far I know, the number after ETD (54) is related to the dimension of the core. A core can be made up of different core materials having different magnetic property. Same dimensioned core can have different magnetic property. Is my understanding okay?

How can we choose Ferrite Core w.r.t desired power handling capacity? For example, I need a transformer for full bridge dc-dc converter having 500VA power handling capability. Which core do I need to use here?

 

[Tahmid]

The number is related to the dimension. The power capacity is related to dimension and core material. In general, the greater the dimensions, the greater the power handling capacity.

Yes, same dimensioned cores can have different properties.

You can find a chart in Abraham Pressman's book that you can use as a reference.

You might find this helpful: www.tdk.co.jp/tefe02/e141.pdf

Hope this helps.
Tahmid.

 

[sam781]

Thanks...
Now I need to find shops where ferrite code is available. Could you please help to find such shops here in Bangladesh?