Design help for 200kHz 0-(1-2) Linear Ramp Generator

[Owen333]

I am building an analog buck converter switching at 200kHz. I need some help designing a ramp generator at 200kHz (0 to 1-2V) Ramp Generator. Any ideas or resources?

 

[KlausST]

Hi,

internet search, forum search?
There will come up a lot of possible solutions. What´s wrong with them?

Your question gives us no informations about your concerns. Is it the levels, the frequency, the part selection, the power supply, the PCB layout...

Btw: Your "0V" specification ... does not work in single supply situations.
You need a dual supply ... or you need to go on with a more relaxed - say +100mV - limit.

Klaus

 

[Easy peasy]

standard pwm chips give you a ramp at a fixed freq, e.g. UC3842

https://www.google.com/url?sa=t&source=web&rct=j&opi=89978449&url=https://www.mouser.com/datasheet/2/149/uc3843-309322.pdf&ved=2ahUKEwivts-Hg86GAxUC1TQHHTbNKVwQFnoECBUQAQ&usg=AOvVaw3RWIIe8UU3LdRyoUdBfbYK

or the UC3856 for push pull output

--- Updated Jun 9, 2024 ---

also comparator oscillator, ramp comes off the timing cap

top 100k = 22k, cap = 1n0, diode and 150R across the top 100k to ger rapid down part of the ramp.

 

[Owen333]

Hi,

internet search, forum search?
There will come up a lot of possible solutions. What´s wrong with them?

Your question gives us no informations about your concerns. Is it the levels, the frequency, the part selection, the power supply, the PCB layout...

Btw: Your "0V" specification ... does not work in single supply situations.
You need a dual supply ... or you need to go on with a more relaxed - say +100mV - limit.

Klaus

The +- 100mV should be fine, this is just a personal project, I've been doing a lot of research and I can't seem to find one, I want to do the controls by myself so I didn't want to use a DC-DC controller and make a pure analog buck converter. I've seen some ramp generators but with 555 timers I can't seem to get them to 200Khz. I also can't seem to find what I am looking for on here so that's why I decided to make a post. I can't seem to find a general circuit that produces a positive voltage linear ramp generator that can run at 200khz. I was hoping someone could point me in the right direction or post. Thank you for the replies!

 

[KlausST]

Hi,

Any fast enough OPAMP should be able to do what you described so far.
I don´t see any progress in disussion. No reference to what circuits you are "looking for" or what documents you did read so far.

What is a "pure analog buck converter". In my eyes a buck converter is switching .. thus: not analog.

Also with "vague descriptions" like "I want to do controls by myself" tells us nothing. Hardware, timing, control loop? .... or what else?
And why this "mystic control" can not be achieved with a "dedicated SMPS controller". There may be thousands of controllers....

As long as we don´t know what you want to achieve, it´s hard / impossible to help.

Klaus

 

[Owen333]

Hi,

Any fast enough OPAMP should be able to do what you described so far.
I don´t see any progress in disussion. No reference to what circuits you are "looking for" or what documents you did read so far.

What is a "pure analog buck converter". In my eyes a buck converter is switching .. thus: not analog.

Also with "vague descriptions" like "I want to do controls by myself" tells us nothing. Hardware, timing, control loop? .... or what else?
And why this "mystic control" can not be achieved with a "dedicated SMPS controller". There may be thousands of controllers....

As long as we don´t know what you want to achieve, it´s hard / impossible to help.

Klaus

Sorry I am not trying to be difficult this is my first PCB I am designing just to gain exposure for power electronics, Here I will attach a picture of what I am trying to achieve.

I am trying to create that Vosc, I know how to create the rest of the control loop of a buck converter I am just trying to create a circuit or find an IC that creates a ramp oscillator.

 

[Owen333]

I've modeled this circuit in LT spice and got a 200Khz Sawtooth OSC. Is there any concerns or tips for using this for PWM or putting this onto a PCB?

 

[FvM]

Nice design. I wonder if frequency and thresholds are sufficiently stable against component parameter variations expectable in a discrete implementation. E.g. D1/Q3 seems to work as a current mirror, current gain can easily vary in range of 0.5 to 2.

 

[KlausST]

Hi,

so it´s just a "reference" for generating a PWM.
If so:
* why this voltage specification?
* why a ramp? Any symmetric or unsymmetric triangle shape will do (no need for a steep edge, at all)
* in a feedbacked loop even the linearity errors will be regulated out

Klaus

 

[Owen333]

Hi,

so it´s just a "reference" for generating a PWM.
If so:
* why this voltage specification?
* why a ramp? Any symmetric or unsymmetric triangle shape will do (no need for a steep edge, at all)
* in a feedbacked loop even the linearity errors will be regulated out

Klaus

Ok that makes sense, a lot of my research has come from books, mainly Fundamentals of Power Electronics by Robert Erickson and they model the reference signal using ramp. I would be interested if it's easier to make and more predictable using a symmetric triangle wave. If I use the symmetric triangle wave, I don't need any specific voltage specification and I can just use any triangle wave that I can make at 200Khz? Do I model this differently in my control loop then 1/(Vramp)? Again Thank you so much for the responses!

 

[Easy peasy]

... for single-supply - R2 goes to mid rail, or use split supply OP-AMP operation, Rf & C on the left set freq, C on the right and R3 set amplitude.

R1, R2 set the switch points for the osc at left, and so can lower Freq if the set points are wide

for 200kHz - you want a fast opamp - e.g. MC34072

ROM = R3

 

[KlausST]

Hi,

sadly the Cs are not individually marked thus, let´s say
* the left side one is Ct (timing capacitor)
* the right side one is Ci (integrator C)

For simple solutions the voltage across Ct is a (distorted) triangle. Then there is no need ofr the integrator part at all.

For improved triangle waveform you may additionally use the integrator. I recommend Ci to be a film capacitor (for improved waveform) and one needs to add a high ohmic resistor across Ci to avoid runaway. Without it - sooner or later - the integrator output will move towards either rail side ... and become distorted.

Klaus

... for single-supply - R2 goes to mid rail,

as well as R_OM...

 

[Easy peasy]

Oh yes, C(t) left, R2, Rom - can all go to mid-rail - for single supply operation - the 1Meg ( or 2M2 ) or so across C(int) right - is a good idea - often called a leaky integrator - in other ckts the top and bottom of the triangle form the switch points feeding back to the oscillator so the leaky R isn't needed.
Cheers Klaus.

 

[FvM]

Principle OP based triangle generator shown in literature