First attempt at designing a low input voltage boost converter

[Gasboss775]

I do it all the time.

I had a stroke in 2008 and along with various other anomalies my hands are pretty shaky. Soldering through hole is challenging enough. I have built a few projects using mostly terminal blocks and minimal soldering.

 

[large_ghostman]

When I was young lol (11) and I first tried soldering smt parts I struggled, I could get the part to stay in place. In the end I found if I dabbed either a small bead of glue like Bostik or aradite on the bottom of the chip I could place it in the right place, it would stay movable plenty long enough to mess about positioning the chip until it was correctly placed. Then I waited for it to dry and found I could then easily solder the legs without moving the chips (plenty liquid flux helps alot).
I also had success using small beads of caulkin on the bottom of the chip.

My dad also had a major stroke and he managed to solder smt chip again by using that method. give it a try it might help , also a ring magnifier with a light around the rim helps, as would tinning the board first, but tinning the board first makes it harder to use the glue method as you can get a gap.

I hope that helps, the only other method that worked really well for my dad was to get me to do it :S

 

[Gasboss775]

My dad also had a major stroke and he managed to solder smt chip again by using that method. give it a try it might help , also a ring magnifier with a light around the rim helps, as would tinning the board first, but tinning the board first makes it harder to use the glue method as you can get a gap.

That is certainly encouraging. Thanks for your advice and encouragement.

 

[Flyback]

ref your schem of post 4 or 5 back, you have no current sense filter.
I don't see output voltage feedback divider.
You know the TL431 needs to drink a good milliamp just as bias current
Better to have at least some fet GS resistance, eg 22k or something
Surely your current sense resistor voltage amp has to be non inverting.
You can simulate that in ltspice, do you want me to send you the logic blocks? (inverting Schmitt triggers)

 

[spec]

I wanted to have a go at designing a boost converter that could power say a 9volt circuit from 2 x NiMH cells. The circuit works quite well in fact with as much as 85% efficiency. My problem is that at certain combinations of input voltage and / or load it becomes unstable with the output voltage cycling up and down over about a 300mV range. R5 & C3 were a crude attempt at compensation ( a subject that I can't really claim to understand very well ) from the little I do know I suspect that the problem is related to compensation or lack there of.

Hello Gordon,

Is that your first switch mode power supply design? If so it's a damn good circuit. You certainly jump in at the deep end- not many SMPSUs will work off 2.5V.

I can see a couple of areas that might be giving you problems. As I don't have a workshop I would be very interested in developing your circuit because it would suit an application I have in mind. If so, could you give a datasheet for the inductor or its part number?

spec

 

[spec]

Another great new product spoiled by only being available in SMT. Bah!

Hi again Gordon,
I am the same about surface mount. But you can get adapter PCBs to convert surface mount to thu hole. OK you do need to solder the SM component to the adapter board but provided you get one of the bigger surface mount paks it should not be too bad provided you have a good magnifier and small tip soldering iron. You also need some solder paste. On the other hand you can glue the SM pack in place, having previously solder pasted the pads and then blast the assembly with a heat gun. Even a standard DeWalt paint stripper gun will do if you are careful

 

[Flyback]

GORDON ROBERTSON
LTspice is free and helps you to get a circuit right before making it.
Here is some logic blocks.
You can set the "trip " point of the inverting not gates to be the same as those Schmitt triggers that you have, then they act the same as them

its massively simpler to simulate first, then make later

 

[Gasboss775]

Hello Gordon,

Is that your first switch mode power supply design? If so it's a damn good circuit. You certainly jump in at the deep end- not many SMPSUs will work off 2.5V.

I can see a couple of areas that might be giving you problems. As I don't have a workshop I would be very interested in developing your circuit because it would suit an application I have in mind. If so, could you give a datasheet for the inductor or its part number?

spec

Thanks Spec. I have been hard at it since about the end of October experimenting with DC to DC converters. Started with simple mc34063 based stuff, see this thread

I done some stuff with the LM2575 boosting its output current with an external P MOSFET. More recently I've been trying to design my own control circuits.

The inductor is a bobbin inductor 330uH rated 2.8A I think R is < 0R1, Iam unable to measure lower than one tenth of an ohm. I have a few if you want one? I bought a large lot of them on eBay a few years ago, sold some and kept the rest, I forget the manufacturer but there is a part number on top of them which I can check if you want to try looking them up.

I also built a rig for testing power inductors with the addition of a 'scope. I will attach a copy of its schematic. If you think you could help with the STEPUP idea that would be good. My motivation is purely the joy of learning about a new type of circuit. My workshop is well stocked so Iam able to do a lot of practical experiments and enjoy doing so!

Which version of the circuit do you think holds the most promise? Post #1, #4 or #36?
__________________________________

This is where I got the idea for the Inductor Test Rig: Flyback Converters for Dummies

 

[spec]

Thanks Spec. I have been hard at it since about the end of October experimenting with DC to DC converters. Started with simple mc34063 based stuff, see this thread

I done some stuff with the LM2575 boosting its output current with an external P MOSFET. More recently I've been trying to design my own control circuits.

The inductor is a bobbin inductor 330uH rated 2.8A I think R is < 0R1, Iam unable to measure lower than one tenth of an ohm. I have a few if you want one? I bought a large lot of them on eBay a few years ago, sold some and kept the rest, I forget the manufacturer but there is a part number on top of them which I can check if you want to try looking them up.

I also built a rig for testing power inductors with the addition of a 'scope. I will attach a copy of its schematic. If you think you could help with the STEPUP idea that would be good. My motivation is purely the joy of learning about a new type of circuit. My workshop is well stocked so Iam able to do a lot of practical experiments and enjoy doing so!

Which version of the circuit do you think holds the most promise? Post #1, #4 or #36?
__________________________________

This is where I got the idea for the Inductor Test Rig: Flyback Converters for Dummies

Hi Gordon,

I would like to work on the first circuit if that is OK with you. Please don't think I am criticizing your work though- the first switch mode PSU I did was just a transistor killer with no output at all. What I propose is is two modification paths resulting in two final circuits, hopefully - have got pretty bad flue at the moment so am not firing on all two cylinders.

 

[Gasboss775]

Hi Gordon,

I would like to work on the first circuit if that is OK with you. Please don't think I am criticizing your work though- the first switch mode PSU I did was just a transistor killer with no output at all. What I propose is is two modification paths resulting in two final circuits, hopefully - have got pretty bad flue at the moment so am not firing on all two cylinders.

I've had a sore throat on and off since late November. I should perhaps go to the docs, but expect he will just say it's a virus and will go away units own time.

Glad you are interested in working on the circuit, I'm eager to learn and at the moment I've just been making educated ( hopefully ) guesses. Still not fully understanding why certain things do / do not happen.

 

[Gasboss775]

Apparently the Schmitt trigger type regulator that I came up with is called a hysteretic converter. I can honestly say that I've never heard of this topology and briefly wondered if I'd actually came up with something original! I will admit that I may have come across something similar, but without actual reference to that name.

 

[spec]

In view of your positive response I will go through some of the design processes as we go along- please don't think I am being patronizing. Would you be happy to go for another switching power transistor. The BD139 family are rather difficult in this application.

 

[spec]

Apparently the Schmitt trigger type regulator that I came up with is called a hysteretic converter. I can honestly say that I've never heard of this topology and briefly wondered if I'd actually came up with something original! I will admit that I may have come across something similar, but without actual reference to that name.

If you evolved that circuit yourself. you did invent it. The fact that it is already widely used shows that you are a good inventor. I have invented hundreds of things only to find that they have already been invented. It is a good process to get you thinking in a creative mode.

(It's 2.45am and I can't sleep for coughing. Like you, I have had a cough since November 2015, but day before yesterday it developed into something really nasty- afraid my brain is not working at the moment. 71 today as well!)

 

[spec]

I had a stroke in 2008 and along with various other anomalies my hands are pretty shaky. Soldering through hole is challenging enough. I have built a few projects using mostly terminal blocks and minimal soldering.

That is a drag- glad to see that you are not letting anything get in the way of your endeavors though. It really pees me off that my eyesight has deteriorated- makes surface mount work difficult, but I have a microscope in mind which should help a lot. I just can't decide on approach - digital or optical.

 

[spec]

A switch mode power supply relies on one fundamental principal- energy transfer. It achieves this in two parts of a cycle:
(1) charge up an inductor
(2) dump the charge stored in the inductor into a capacitor.

The formula for the charge stored in an inductor is:
Q= LI= VT .....(f1)
Where
Q= charge in Coulombs
L= inductance in Henrys
I= current in Amps
T= time in seconds

For example you place 2V across a 1H inductor for 4 seconds.
What is the current flowing in the inductor at the end of the 4 seconds:
Transposing from (f1) you get:
I=(VT)/L .....(f2)
V= 2V
T= 4 seconds
L= 1H
Thus I = (2*4)/1 =8A

The parameters to consider for the switching bipolar transistor are:
(1) hfe (ic/ib) at maximum inductor current
(2) vce (sat) at the maximum inductor current
(3) vbe at the maximum inductor current

(1) tells you the minimum current you need to drive the transistor
(2) tells you how much voltage will be lost across the CE of the transistor. This voltage must be subtracted from the supply line to establish how much voltage the inductor actually sees
(3) vbe tells you the minimum voltage needed to drive the transistor sufficiently (there is a complication here in that you rarely drive a bipolar transistor base with a voltage but for the time being ignore it)

A typical power transistor to do the 8A job outlined above would have:
(1) hfe = 20
(2) vce(sat)= 1V
(3) vbe = 1.2V

Thus, the minimum base current drive required will be 8/20 = 0.4A (400ma) at 1.2V

Suppose the transistor were driven by a voltage source swinging between OV and 5V. The value of the base ballast resistor would be (5V-1.2V)/0.4 = 9.5 Ohms

The fundamental procedure for designing the essence of a switch mode power supply has now been defined.

There are all sorts of secondary issues to consider but don't let anyone cloud the issue with those for the time being.

From the above you should be able to calculate the characteristics of the original switch mode power supply and identify two major problems that need to be sorted. Don't worry if not because I will explain later.

 

[Gasboss775]

OK, think I got that, but how do you determine the amount of charge that needs to be stored in the inductor given a particular load current / voltage at the output?

 

[spec]

Assuming 100% efficiency and a 1:1 mark to space ratio of the switching waveform.

V1*I1=V2*I2
Where
V1= charging voltage across inductor
I1 = final charging current through inductor
V2= output voltage required
I2= output current required divided by 4

The reason why I2 is only a quarter of the required output current is that the input current is a ramp and the RMS (average) value of a ramp is Ipeak/2. The current is further divided by 2 because the switching waveform has a mark to space ratio of 1:1 so the output circuit is only receiving current for half of the time.

Note that I have used this rather odd form to make clear what is going on and to simplify design calculations.

The energy stored in an inductor = (I*I*L)/2 Jules
Where:
I = the final current flowing through the inductor in Amps
L= the inductance of the inductor in Henrys