Buck/boost converter

[LarsT]

I've been looking around for a while trying to find simple buck/boost converters based on either 555-timers or relaxation oscillators (preferably lm339 which I happen to have) with some kind of feedback loop to keep the output voltage stable. I want to make a 5V to 12V, 0.5A boost converter and a 5V to 3V, 1A buck converter.

I've tried making several boost converters but the output voltage goes to approx. 50-60V, even when using some of the calculators out there to choose the appropriate duty cycle/inductor value. However when I put even the smallest load on ~10mA, the voltage drops to maybe 10V.

[Yansi]

Bad circuit design.

Use appropriate IC for that. No reason to build that around 555 in the year of 2015. O_o

There's not ehough information to guess anything else.

[janoc]

Ehm, and why would you want to build this with a 555 timer? (apart from it being a learning challenge, of course). There are specialized chips and modules for this. Designing a DC-DC converter with good efficiency from scratch is a fairly nontrivial problem.

The 60+V voltage unloaded and then drop to 10V when loaded that you are seeing means that you have  no feedback system. DC-DC converters use a feedback loop to control the switching frequency - usually by PWM. The more current you draw, the lower the output drops and the faster the converter should switch to compensate (by "pumping" more energy into the circuit). And vice versa - if the output is too high, the converter frequency drops or it even stops switching until the output voltage drops again.

[madires]

A view into my glass ball tells me that you don't got any feedback loop to control the output voltage. The most simple way is to buy some of those cheap modules via ebay. If you want to DIY you could look for LM2575 (buck) or LM2577 (boost), just to name two. There are tons of buck/boost converter ICs.

[LarsT]

A view into my glass ball tells me that you don't got any feedback loop to control the output voltage. The most simple way is to buy some of those cheap modules via ebay. If you want to DIY you could look for LM2575 (buck) or LM2577 (boost), just to name two. There are tons of buck/boost converter ICs.

Ehm, and why would you want to build this with a 555 timer? (apart from it being a learning challenge, of course). There are specialized chips and modules for this. Designing a DC-DC converter with good efficiency from scratch is a fairly nontrivial problem.

The 60+V voltage unloaded and then drop to 10V when loaded that you are seeing means that you have  no feedback system. DC-DC converters use a feedback loop to control the switching frequency - usually by PWM. The more current you draw, the lower the output drops and the faster the converter should switch to compensate (by "pumping" more energy into the circuit). And vice versa - if the output is too high, the converter frequency drops or it even stops switching until the output voltage drops again.

It's both for the learning challenge as well as the practical need. I guess I'll just order a couple of buck/boost-ICs, but it would be fun to be able to make something that works in the meantime.

So do these dedicated ICs regulate both the frequency and the duty cycle or just the duty cycle? By looking at examples on buck/boost-converters it seems that the output voltage is affected by the duty cycle alone? E.g V_out = V_in*1/(1-D), given that the inductor corresponds to the set switching frequency. Is it that simple?

[madires]

So do these dedicated ICs regulate both the frequency and the duty cycle or just the duty cycle? By looking at examples on buck/boost-converters it seems that the output voltage is affected by the duty cycle alone? E.g V_out = V_in*1/(1-D), given that the inductor corresponds to the set switching frequency. Is it that simple?

Just the duty cycle. And the output voltage is affected by the duty cycle, inductor, output cap, operation mode and so on. Most datasheets include a recipe and examples to calculate the proper values for inductor, caps, feedback or what have you.

[janoc]

It's both for the learning challenge as well as the practical need. I guess I'll just order a couple of buck/boost-ICs, but it would be fun to be able to make something that works in the meantime.

I think you will have the chips home faster than you manage to get something semi-working from the naked 555. Just sayin' ....

So do these dedicated ICs regulate both the frequency and the duty cycle or just the duty cycle?

They regulate duty cycle, the frequency tends to be constant. The amount of energy passed to the inductor for each cycle depends on how long was the output switch open/closed = duty cycle (the area under the voltage curve, etc.). So you don't really need to change the frequency too. The ICs don't do it, because the switching frequency has a large impact on the inductor and capacitor choices in the circuit.

By looking at examples on buck/boost-converters it seems that the output voltage is affected by the duty cycle alone? E.g V_out = V_in*1/(1-D), given that the inductor corresponds to the set switching frequency. Is it that simple?

The voltage depends on the duty cycle, but also the inductor choice, output load, etc. The choice of the inductor is quite complex affair - it doesn't depend only on frequency but also saturation current. Ideally you want an inductor that is driven almost to saturation so that it is not unnecessarily large. However that is no good if your load isn't constant because you don't want to actually drive the inductor into saturation = losses and the circuit is unable to deliver more energy demanded by the load, so poor regulation. So this is quite a complex dance how to pick the right part. Datasheets for the converter ICs will tell you how to choose the inductors and what has to be considered.

[janoc]

Here is a good primer explaining the various types of SMPS:

http://www.learnabout-electronics.org/PSU/psu30.php

And an app note from Maxim is useful too:
http://www.maximintegrated.com/en/app-notes/index.mvp/id/660

[kerrsmith]

I found the following post which may be of interest:

https://www.eevblog.com/forum/beginners/voltage-regulator-circuit-for-smd-leds/msg688619/#msg688619

If you click the circuit image on the post there is a constant current boost converter based on a 555 timer.

It seems quite similar to what you are looking to achieve so though I would post a link.

[janoc]

I found the following post which may be of interest:

https://www.eevblog.com/forum/beginners/voltage-regulator-circuit-for-smd-leds/msg688619/#msg688619

If you click the circuit image on the post there is a constant current boost converter based on a 555 timer.

It seems quite similar to what you are looking to achieve so though I would post a link.

Haven't seen that one before, that's a seriously neat circuit.

[LarsT]

I found the following post which may be of interest:

https://www.eevblog.com/forum/beginners/voltage-regulator-circuit-for-smd-leds/msg688619/#msg688619

If you click the circuit image on the post there is a constant current boost converter based on a 555 timer.

It seems quite similar to what you are looking to achieve so though I would post a link.

Haven't seen that one before, that's a seriously neat circuit.

I found the following post which may be of interest:

https://www.eevblog.com/forum/beginners/voltage-regulator-circuit-for-smd-leds/msg688619/#msg688619

If you click the circuit image on the post there is a constant current boost converter based on a 555 timer.

It seems quite similar to what you are looking to achieve so though I would post a link.

Nice, I will try that one. Is there a way to make this work for higher currents, e.g 500mA? I guess the transistor at the control pin will have problems handling that much current to the base.