• Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

RT8059 step-down DC/DC converter - can't get it to be stable, annyone with experience?

Status
Not open for further replies.

Andy_C_

New Member
Hi all,

I have an RT8059 step-down DC/DC converter. I want to use it to regulate the voltage from a li-ion battery (seeing as the battery voltage drops from 4.2V to ~3.4V as its charge level decreases). I've constructed the circuit from the IC's datasheet, using suggested component values and component types. However, the circuit doesn't provide a stable output - the output voltage wanders and doesn't stay at a fixed level. I've tried several identical IC's, different components around the IC, and tried constructing the circuit on a breadboard, and copper stripboard, but still no stability.

Does anyone have any experience of this regulator, or any suggestions for how to get it to perform in a stable fashion?

Many thanks.
 

Tony Stewart

Well-Known Member
Most Helpful Member
If feedback R is too low in values, phase margin can be poor due to excess bandwidth. We assumed you used components recommended and same layout.
 
Last edited:

spec

Well-Known Member
Most Helpful Member
The RT8059 is a commonly used part, and being simple should give no trouble.

But the RT8059 operates at a high frequency (1.5MHz) and also has three modes of operation: normal pulse width modulation, relaxation, and linear.

Because of this, the choice of components is critical and the physical layout is also critical.

As Ron says in post #2, please post a schematic of your actual circuit and also an image of your layout and any other information, especially the power source you are using to test the circuit. This will help us to sort the problem.

One thing you can try right away is to connect a 220 Ohm, or so, load on the output and see if that stabilizes the RT8059.

Also, long leads from the power source to RT8059 can cause problems.

spec
 
Last edited:

Andy_C_

New Member
Hi Guys, thanks for your replies.

Here's the link to the RT8059 datasheet
http://www.richtek.com/assets/product_file/RT8059/DS8059-05.pdf

Here's the schematic of the constructed circuit


Here's a picture of the breadboard circuit


Here's a picture of the copper stripboard circuit


The circuit feedback network resistors are R1 = 1M ohm; and R2 = 200k ohm.
From the formula in the datasheet, this should give an output voltage of 3.6V.
I have an LED and 100 ohm resistor in series as the load (basic 5mm indicator LED, forward voltage ~2.4V)

The circuit is powered by an 18650 li-ion battery, which I've measured regularly and it has a stable output.
I've been using a digital multimeter to measure the output voltage from the circuit. The output voltage typically starts at 1.4V and drops steadily at a rate of approximately 100mV per second. The voltage sporadically rises and falls, without settling at a stable level. I've noticed that touching the ground wire with my finger causes the output voltage to rise significantly.

Spec, I suspect as you said, that the type of components and their layout are the important factor for stability - the datasheet discusses this, however, I don't have the ability to make a PCB with the appropriate layout! Any suggestions about how to construct the circuit with the components in the required layout but without a custom made PCB?

Ronsimpson, I am in Dublin, Ireland.

Thanks again guys.
 

spec

Well-Known Member
Most Helpful Member
Hy Andy,

My initial reaction is that R1 and R2 are too higher resistance and causing the circuit to go unstable in the frequency domain. It is bed time now, but I will get back to this in about 10 hours.

spec
 

ronsimpson

Well-Known Member
Most Helpful Member
but I will get back to this in about 10 hours
spec is down for nightly maintenance. Defrag his hard drive. lol

Look at post #6. R1, R2, C1
There is a time constant. R1 X C1 110k X 10p. I think your circuit should have the same time constant. (or some where near)
 

Tony Stewart

Well-Known Member
Most Helpful Member
I doubt that you have any great success unless you can create a layout etched, milled or similar that resembles figure 3 with SMT parts. The reason is all your loops and this chip switching at 1.5MHz with bandwidth over 50MHz, you are looking at a digital RF circuit and without good ground layer underneath and very short current loops, it will behave unstable.

Unless you find someone who can make it for you locally, it may be pure frustration and a waste of time.
 

spec

Well-Known Member
Most Helpful Member
Hi Guys, thanks for your replies.

Here's the link to the RT8059 datasheet
http://www.richtek.com/assets/product_file/RT8059/DS8059-05.pdf

Here's the schematic of the constructed circuit


Here's a picture of the breadboard circuit


Here's a picture of the copper stripboard circuit


The circuit feedback network resistors are R1 = 1M ohm; and R2 = 200k ohm.
From the formula in the datasheet, this should give an output voltage of 3.6V.
I have an LED and 100 ohm resistor in series as the load (basic 5mm indicator LED, forward voltage ~2.4V)

The circuit is powered by an 18650 li-ion battery, which I've measured regularly and it has a stable output.
I've been using a digital multimeter to measure the output voltage from the circuit. The output voltage typically starts at 1.4V and drops steadily at a rate of approximately 100mV per second. The voltage sporadically rises and falls, without settling at a stable level. I've noticed that touching the ground wire with my finger causes the output voltage to rise significantly.

Spec, I suspect as you said, that the type of components and their layout are the important factor for stability - the datasheet discusses this, however, I don't have the ability to make a PCB with the appropriate layout! Any suggestions about how to construct the circuit with the components in the required layout but without a custom made PCB?

Ronsimpson, I am in Dublin, Ireland.

Thanks again guys.
Hello Andy,

I have never been to Ireland but have always meant to.

Sorry to say this but, to be blunt, both of your circuits have no chance of working.:sorry:

The reason is as Tony says, but you should be able to get the RT8059 chip working as a power supply using strip board. This is an interesting challenge and I will give the details some thought.:)

But in the meantime (before I go back to bed) just a bit of theory:

As some of us on ETO have said many times before, a schematic is an illusion; it is an idealized representation of the actual circuit. For example, even a simple piece of wire is comprised of resistors, capacitors and inductors along its length, and a resistor can turn into an inductor at a high frequency. Similarly, an inductor can turn into a capacitor. Then there are all the capacitive and inductive couplings so that a wire carrying a current can induce a current into another wire. These effects increase dramatically as the frequency increases, and 1.5 MHz is relatively high, especially for a switching power supply which generates a square wave which has harmonics going up to much higher frequencies.

The upshot of all this is that component choice and physical layout is critical for some circuits. The real fun starts at microwave frequencies.:D

spec
 
Last edited:

Les Jones

Well-Known Member
Most Helpful Member
Hi Andy,
Yet another clue to why it does not work is in the title of the regulator "step-down DC/DC converter" From your first post I am assuming you are powering it with a single lithium ion cell. Which will have a terminal voltage from 4.2 volts when just fully charged dropping down to 3.0 volts As you seem to want 3.6 volts output it MIGHT just work with 4.2 volts in (Even that only gives 0.6 volts for the regulator to work with.) The inductor that you are using is not suitable for this application. (It probably has too high a resistance and will probably saturate.) To achieve what you are trying to achieve you will need both a step up and step down regulator as the input voltage swings above and below the required output voltage.
 

spec

Well-Known Member
Most Helpful Member
Hi Andy,
Yet another clue to why it does not work is in the title of the regulator "step-down DC/DC converter" From your first post I am assuming you are powering it with a single lithium ion cell. Which will have a terminal voltage from 4.2 volts when just fully charged dropping down to 3.0 volts ...
Hy Les,

You may not believe this, but I was just about to post the same thing!:p

Andy,

Can you change R1 to 390K and R2 to 100K to give an output voltage of near 3V and also put a 100 Ohm resistor across the output to provide a load of 30mA?

Let us know what happens.

spec
 

Andy_C_

New Member
Hi guys, many thanks again for all your help so far - lots of useful info there...

So, the long and the short of it seems to be that I need to use the proper PCB layout, in order to minimise RF issues, stray signals, etc.
I will look into making a PCB...

Spec and Les, the point you made about needing a step-down and step-up converter - I have a MAX1912, which is exactly that. However, it is highly inefficient, I've measured it at ~45% efficiency (230mW at 4.2V input vs 100mW at 3.4V output). So it's not good for this battery powered application unfortunately!
https://www.maximintegrated.com/en/products/power/led-drivers/MAX1912.html

My aim with this project is to power some LEDs, with somewhere around 100mA total current being supplied to them - while trying to get the highest level of efficiency because the circuit is battery powered.

This is a pretty light load (100mA) for a li-ion cell (~2,000mAH capacity), and Richtek (manufacturer of RT8059) has some info about li-ion discharge.
http://www.richtek.com/battery-management/en/designing-liion.html
Richtek has a discharge chart which shows the battery voltage being >3.6V for >90% of battery charge (under light load conditions). Also, the RT8059 datasheet shows ('Electrical Characteristics' table, page 3) that Vout = Vin - 0.2V (I'm assuming that the 0.2V drop is relatively constant).
So, this 0.2V regulator drop, with a minimum 3.6V supply from the battery, means that I should be able to set the output voltage to around 3.2V - and power the LEDs with a small series current-limiting resistor. Or am I being a little optimistic, any thoughts?

Spec, I'll try your suggestion for the values of R1, R2, and a load resistor and get back to you.

Thanks again guys, much appreciated.
 

spec

Well-Known Member
Most Helpful Member
Hy Andy,

Care to specify exactly what your objective is with the LEDs?

An etched PCB would be ideal, but you can make a small simple PCB from a piece of copper clad fiberglass board, a scalpel, a 0.8mm drill (for through holes), and a fine soldering iron. Just mark out the traces, cut along the lines and heat and peel off the copper you do not want.

I suspect though that you could make a reasonable RT8059 power supply using strip board (with the right components that is).

spec
 
Last edited:

Andy_C_

New Member
Spec,

Yup, my objective with the LEDs is a solar light project. I'm planning to have 3 clusters of LEDs, each cluster containing 3 LEDs, with 10-15mA per LED (all tentative figures at this stage).

I had been thinking the same thing as you suggested, about taking a fine knife to a copper clad board, I'll need to be accurate though!

I have SMD versions of the capacitors and inductor - all ceramic. Regardless of these being surface mount, are they suitable component types (ceramic) for this high frequency application?

Thanks.
 

Les Jones

Well-Known Member
Most Helpful Member
Hi spec,
Yes I can believe that you would post similar comments once some proper information was posted about the problem rather than "the circuit is as in the application notes" and the "voltage is varying". Once we knew the output voltage was so close to the input voltage and a picture of the actual construction was posted then we both had proper information on which to base our conclusions. A method I used in the past to make printed circuit boards was to paint the circuit on the board with cellulose car paint then etch them. The first boards I made around 1962 when I was about 19 were made useing thin phosphor bronze sheet (My father used to get offcuts from work.) glued to paxolin (SRBP) with contact adhesive. They were then etched using nitric acid as I did not know about ferric chloride at the time.

Les.
 

spec

Well-Known Member
Most Helpful Member
Hy Andy,

Have you thought about using high brightness LEDs which are bright with just 2mA, which would give a total current consumption of only 18mA. You could then probably forget the voltage converter.

You do not have to be that accurate with the PCB cutting!:)

Surface mount ceramic capacitors should be OK but I have never heard of a ceramic inductor. As Les says the inductor is a critical item and needs to be a high frequency low loss type.

spec
 

spec

Well-Known Member
Most Helpful Member
Hi spec,
Yes I can believe that you would post similar comments once some proper information was posted about the problem rather than "the circuit is as in the application notes" and the "voltage is varying". Once we knew the output voltage was so close to the input voltage and a picture of the actual construction was posted then we both had proper information on which to base our conclusions. A method I used in the past to make printed circuit boards was to paint the circuit on the board with cellulose car paint then etch them. The first boards I made around 1962 when I was about 19 were made useing thin phosphor bronze sheet (My father used to get offcuts from work.) glued to paxolin (SRBP) with contact adhesive. They were then etched using nitric acid as I did not know about ferric chloride at the time.

Les.
Seems like we have done similar things.:)

In the beginning there was tag strips. Then, before DIL integrated circuits became common, my standard board technique was to drill 0.8mm holes in fiberglass and hammer pins into the holes. The circuit was then made by wiring the pins as necessary with Tinned Copper Wire (TCW). Later I moved over to painting and etching with ferric chloride, then taping and etching. Then the normal UV resist and etch. Finally just sending Gerbers to a PCCT house.

Never heard of using phosphor bronze before- quite a good idea.:cool: I bet the etching didn't take long with nitric acid. When I was a kid I went through a chemistry phase and had nitric, sulphuric, and hdrochloric acid in the shed. Health and safety- what is that.:D

spec
 
Last edited:

Andy_C_

New Member
Hi Spec,

What type/model of LEDs can you suggest with those parameters... I have been experimenting with some warm white 5050 types, they have three LEDs per chip, approx 10mA for each LED, high brightness (19 lumens per LED)
https://www.superbrightleds.com/mor...ewing-angle-6000-mcd/317/#/tab/Specifications

The reason that I want the regulator is to keep the supply to the LEDs constant... if I forget the regulator and simply go with series resistors to limit current, then the current (and LED brightness) with change significantly as the li-ion battery voltage drops over its range from 4.2V to ~3.4V. Also, I need to be able to switch the LED on and off... using a transistor to do this will lose 0.6V, but the RT8059 would only lose 0.2V (and the RT8059 has a handy enable pin to switch the device on and off).

Essentially what I'm trying to do here, is drive a LED(s) from a consistent source that is powered by a li-ion battery (whose voltage isn't consistent).

I hadn't heard of a ceramic inductor before either! But here they are... (this is just an example, not where I bought them)
http://www.coilcraft.com/0603cs.cfm


Andy.
 
Status
Not open for further replies.

Latest threads

EE World Online Articles

Loading
Top