Boost converter 12V to 36V (25 mA load) using 555

[KrisBlueNZ]

Steve, are you sure Schottky diodes have lower capacitance than fast recovery diodes?

I suggested back in post #14 that abuhafss use a UF4002 or higher because the 1N5819 has a maximum reverse voltage of 40V which is uncomfortably close to the output voltage. Of course you can get Schottky diodes with reverse voltages higher than 40V but I don't think the slight difference in forward voltage (compared to a fast recovery diode) justifies the cost. The switching frequency is 50 kHz maximum, so I think the UF400x's recovery time (45 ns, I think) is fine. Do you agree Steve?

As I mentioned in post #38, second to last paragraph, all the graphs I've posted up to and including post #38 use a C_OUT with an ESR of zero. Typical 10 µF 50V electrolytics have ESRs around 1.2Ω so that's what I've been using since then.

In the last paragraph of post #38 I explained that a second inductor and capacitor would be needed if you need a clean output. With just the single electrolytic with an ESR of 1.2Ω my simulation gives about 310 mV p-p of ripple, both before and after regulation is established.

With the filter inductor (3.3 mH) and second capacitor (10 µF, 50V, 1.2Ω ESR), the ripple at the output is less than 5 mV p-p at 2.5 mA load, and less than 2.2 mV p-p at 25 mA load. It's higher at lower load because the period between pulses is longer.

As I said in post #38, you can optimise the output components according to your requirements for size, cost, and regulation. DC resistance in the filter inductor compromises the output regulation (see last paragraph of post #38) so a larger inductor (thicker wire) and a lower inductance (fewer turns for a given core) will improve output regulation. Filter capacitors with higher capacitance have lower ESR but are larger and more expensive.

 

[abuhafss]

Now the whole picture is clear. Thanks for the clarification.

Referring to the Inductor, for example there are two coils of 470μH both having same core, the one with thicker wire is more preferable here...right? In other words can we say "the one which can withstand more current" ?

 

[KrisBlueNZ]

Not necessarily. Inductors have a specified maximum operating current and/or a saturation current. These should both be at least 500 mA in this application. The saturation current is determined by the number of turns, the overall permeability of the core, and the core material, but it's simplest to just read it off the data sheet for the particular inductance value in the series.

DC resistance is determined by wire thickness and total wire length, which relates to the number of turns. It's significant for the filter inductor (assuming you use one) because DC resistance here will compromise the regulation. But regulation isn't going to be perfect anyway, and your load can probably tolerate moderate variation, so 10Ω DC resistance in the filter inductor shouldn't be a problem. According to Ohm's Law, V = I × R, for a load change of 0~25 mA a DC resistance of 10Ω will cause a DC voltage change of 0~250 mV, which is only 0.7%.

 

[(*steve*)]

I was kinda conflating issues with junction capacitance and minority carriers. However, Schottky diodes are often used at high frequencies because of their low capacitance. Its obviously not a simple issue, and that's at least part of what I was trying to get across I think.

 

[abuhafss]

Hi

Today, I assembled Kris' design on a breadboard, I used following parts:

D1=U4007
D2=1N4148
D3=27V Zener 0.25W
Q1=2SK3435 (60V 80A)
CT=2.2nF
CIN=100μF
COUT=1.25μF Polyester Bipolar (did not had any low ESR electrolytic)
L1=I had a 500μH coil but, wasn't sure if could withstand 320mA so I made handmade coil build with 23mm x 13.8mm x 10mm toroidal core (Yellow/White) - 80 turns of 0.7mm wire. According to the toroidal coil calculator 82 turns should give 470+ μH.
Dz= Not used (did not had 6.2V 0.5W)

Here are the observations:
Vin=12V
Vout=39.2V constant (without load)
Load=1k (1W resistor)
Vout=19.8V (with load)
Vd across 2.2Ω = 6.4mV

 

[KrisBlueNZ]

That regulation is terrible! Can you capture some oscilloscope traces? The voltage at the right end of RS with no load and with load would be an interesting start.

 

[(*steve*)]

The inductance of the coil will depend largely on the material the core is made from. I don't see that in your calculation. Also the maximum current is determined not only by the thickness of the wire, but (and often mostly by) the saturation of the core.

 

[(*steve*)]

That regulation is terrible! Can you capture some oscilloscope traces? The voltage at the right end of RS with no load and with load would be an interesting start.

Maybe the fact he left out the zener diode?

 

[KrisBlueNZ]

Maybe the fact he left out the zener diode?

No, the zener he left out is for regulation against input voltage variations. D3 is the output voltage regulation zener.

But it's odd that the unloaded output voltage is 39V. With a 27V zener, the output voltage shouldn't go above about 32V.

I think it would be a good idea to buy a known correct inductor.

 

[abuhafss]

That regulation is terrible! Can you capture some oscilloscope traces? The voltage at the right end of RS with no load and with load would be an interesting start.

Sorry, I don't have a scope. However, the DVM reads almost the same, with or without load at the junction of Rs and L1. The voltage difference across Rs is 6.4mV.

I have this inductor lying with me. It reads only 501K so, no idea about the maximum current. Shall I try it?

 

[KrisBlueNZ]

I think that inductor is too small.

What information do you have on the core you used?

Ferrite materials have widely varying permeability figures. 80 turns could produce anything from 10 µH to 100 mH depending on the characteristics of the core.

 

[abuhafss]

I think that inductor is too small.

What information do you have on the core you used?

Ferrite materials have widely varying permeability figures. 80 turns could produce anything from 10 µH to 100 mH depending on the characteristics of the core.

Nothing mentioned on it..............only Yellow/White color, which most of the manufacturers use for 75μ permeability.

 

[abuhafss]

Today, I got my LC-meter and found that the coil I had built is 478μH

And for D1 I had used PR1007L which is equivalent to U4007.

The observations are still the same as mentioned in post #65. I repeat them:

Vin=12V
Vout=39.2V constant (without load)
Load=1k (1W resistor)
Vout=19.8V (with load)
Vd across Rs = 6.4mV

And for the scope, I hope to get it within next week. Till then, any tip to work on ??

 

[KrisBlueNZ]

That's great! I'm looking forward to seeing some scope traces - the waveform across RS, and the drain-source waveform for Q1 would be good places to start.

I can't suggest anything to try until you have the scope.

 

[abuhafss]

Frankly speaking, Kris, acquiring an LC-meter and an oscilloscope is a kind of tribute to you for your devotion in this project, though they would be helpful for me in future. I don't to want to leave this project unresolved.

 

[KrisBlueNZ]

Thanks... I sure hope the design lives up to my expectations! But an oscilloscope is only second to a multimeter in its usefulness and versatility, and I hope you will quickly justify your investment!

 

[abuhafss]

I couldn't wait for the oscilloscope so I reassembled the the components thinking that there might be some problem in connections with the earlier assembly. And my doubt was correct. Now the situation, is that Rs (2.2Ω 1/2W) started frying straightaway. Isn't 1/2W enough?

 

[KrisBlueNZ]

0.5W is plenty. There must be something wrong with your new construction.

Try disconnecting the inductor. If RS still overheats, there must be a short (probably to the 0V rail) around RS or RB. If RS is OK now, put the inductor back and disconnect D1. If RS is OK now, check for a short across the output or COUT.. If RS overheats now, check the connections to Q1. If they're OK, disconnect the gate resistor from pin 3 of the 555 and connect it to 0V. If RS is OK now, there's something wrong with the 555 that is causing pin 3 to be stuck high. This could be a short between pins 2/6 and 0V or perhaps Q2 is not connected properly or is damaged.

 

[abuhafss]

Okay, I shall check all these factors and revert.
In the meantime, I should also add that I tested the circuit without any load.

 

[abuhafss]

Cheers Kris!!!

The Kris Boost Converter is working perfectly.
Although the mosfet 2SK3435 (60V 80A 22mΩ) was not dude, I replaced it with similar STP5NF75 (75V 80A 0.0095Ω).
The output is constant 30.2V with 27V zener, without load. The 555, Rs, mosfet and the inductor; all are super cool.
I am really very excited and have decided to use this Kris Boost Converter for my project.

By the way, what mosfet would be ideal for this converter.

Lastly, what changes should I make to get 220V 50W output (to power up a solder iron from 12V battery).