Why is this astable oscillator limited in duty cycle/frequency?

Thread Starter

Allenph

Joined May 27, 2015
76

I have this circuit. (Ignore the MOSFET for now.)

It's an astable multivibrator based on two BJTs. It works...kind of. There's two issues.

This is a sim of the output where R13 and R16 are 5k ohms:



I would like the fall times to be faster (any advice on how to do this would be appreciated), but I digress. It works.

Here's the same exact circuit where R13 and R16 are 10k ohms:



I think that this happens because I'm not limiting current from the voltage divider...I'm just decreasing the output voltage of the voltage divider which isn't what I want. In this case that voltage divider yields about .4V which isn't enough to forward bias the transistors.

In addition, here's what happens if the resistor values are too low: 1k ohms in this case...



In this case, I'm not sure why this is happening.

How can I use the RC time constant to my advantage while not changing the voltage I'm receiving from the voltage divider I want to "isolate" the original voltage divider and treat it like a rail.
 

Papabravo

Joined Feb 24, 2006
21,159
The simple answer is that, the astable is fixed in duty cycle and frequency because all of the components that affect duty cycle and frequency have fixed values. If the components that determine duty cycle and frequency were variable then you could change things.
 

Thread Starter

Allenph

Joined May 27, 2015
76
The simple answer is that, the astable is fixed in duty cycle and frequency because all of the components that affect duty cycle and frequency have fixed values. If the components that determine duty cycle and frequency were variable then you could change things.
Well...aren't the RC time constants the only thing that affect frequency and duty cycle? I would have MUCH wider margins if I didn't have the voltage divider issue.
 

Thread Starter

Allenph

Joined May 27, 2015
76
Can't you ditch the voltage dividers? You may not get enough gate drive voltage for M1 if R8 is used.
I suppose I could. The input will be about 12V DC. 2N3904s can handle that, especially since the switch is voltage controlled and can have a low current.

Really, the reason I haven't ditched them thus far is I was trying to come up with a solution that WOULD work at a voltage that standard logic level BJTs CAN'T handle.
 

Thread Starter

Allenph

Joined May 27, 2015
76
I suppose I could. The input will be about 12V DC. 2N3904s can handle that, especially since the switch is voltage controlled and can have a low current.

Really, the reason I haven't ditched them thus far is I was trying to come up with a solution that WOULD work at a voltage that standard logic level BJTs CAN'T handle.
Perhaps the solution is to drop the BJT oscillator and substitute an Op-amp multivibrator. That wouldn't require an input voltage, and the rails could be directly controlled with a voltage divider.
 

Papabravo

Joined Feb 24, 2006
21,159
Well...aren't the RC time constants the only thing that affect frequency and duty cycle? I would have MUCH wider margins if I didn't have the voltage divider issue.
Yes, and the threshold voltages of the switches for both on to off and off to on.
 

shortbus

Joined Sep 30, 2009
10,045
Again like your power supply thread, why go through all the aggravation of trying to reinvent the wheel? There are chips to do what your asking this discreet circuit to do. Not trying to put you down, but from the questions your asking in all your threads if you just want to build a machine do that. The wheel has been invented and works very well. :)
 

Thread Starter

Allenph

Joined May 27, 2015
76
Again like your power supply thread, why go through all the aggravation of trying to reinvent the wheel? There are chips to do what your asking this discreet circuit to do. Not trying to put you down, but from the questions your asking in all your threads if you just want to build a machine do that. The wheel has been invented and works very well. :)
I suppose that's true...actually I don't suppose that's true I know it's true. It's more of a personal project, a learning experience. I want to make it from scratch BECAUSE it's challenging, and because I think it's probably a pretty good way to learn a lot of useful concepts.

But, if I'm asking annoying time-consumung questions by all means tell me. I appreciate all the effort that's gone into answering my numerous questions, if I'm asking too much, I'm asking too much.
 

crutschow

Joined Mar 14, 2008
34,282
........................
Really, the reason I haven't ditched them thus far is I was trying to come up with a solution that WOULD work at a voltage that standard logic level BJTs CAN'T handle.
What do you mean "logic level BJTs"? :confused:
Most common BJT's are rated at 30V or more and you can buy some that are rated at over a 100V (the 2N5550 is rated at 150V, for example).
At how high a voltage do you want the circuit to work?
 

Thread Starter

Allenph

Joined May 27, 2015
76
What do you mean "logic level BJTs"? :confused:
Most common BJT's are rated at 30V or more and you can buy some that are rated at over a 100V (the 2N5550 is rated at 150V, for example).
At how high a voltage do you want the circuit to work?
Well, I honestly didn't know that. Although, using high voltage like that as an input would lead to some problems on the gate side of the MOSFET in this case, and anything you needed to connect the output signal to. In this case since we're dealing with voltage controlled components, you could use a very high voltage output signal and then break it with a voltage divider at a very low current. In fact, thinking about it you could use that same voltage divider tactic to switch the base of a very sensitive BJT or whatever. Right?
 

Thread Starter

Allenph

Joined May 27, 2015
76
Well, I honestly didn't know that. Although, using high voltage like that as an input would lead to some problems on the gate side of the MOSFET in this case, and anything you needed to connect the output signal to. In this case since we're dealing with voltage controlled components, you could use a very high voltage output signal and then break it with a voltage divider at a very low current. In fact, thinking about it you could use that same voltage divider tactic to switch the base of a very sensitive BJT or whatever. Right?
Actually, I take that back.

The problem using transistors with a high voltage rating in this case would be the base voltage and current, which is coincidentally what I believe the problem to be here.

I'm using a voltage divider to get to a safe base voltage/current, yet I also need to set the RC time constant of each capacitor through the same channel without altering the voltage I want to get out of the divider.
 

shortbus

Joined Sep 30, 2009
10,045
I suppose that's true...actually I don't suppose that's true I know it's true. It's more of a personal project, a learning experience. I want to make it from scratch BECAUSE it's challenging, and because I think it's probably a pretty good way to learn a lot of useful concepts.

But, if I'm asking annoying time-consumung questions by all means tell me. I appreciate all the effort that's gone into answering my numerous questions, if I'm asking too much, I'm asking too much.
While I can relate to what your saying, I'm just trying to save you some time and heart ache/head ache. I'm now 67yrs old, my project that I started learning electronics for was first started when I was ~45. I came into it with the idea, "how hard can it be?" I'll build every electronic thing needed in discrete circuits, I'll be learning and saving money doing it.:) Well I'm just now getting able to see the light at end of the tunnel! Thanks to the good people on this website, especially ronv. But I have more time and money in failed ideas and a pile of electronic parts both that probably won't be used in my lifetime and ones that I 'smoked' trying to make work.

And when it comes to annoying questions I'm the king. Most of the regulars in this group can tell you that! But if you stick around they will help you.

Believe me when I say, even using existing chips and modules, you'll learn, learn more then you ever thought. And it will take longer than you thought.
 

crutschow

Joined Mar 14, 2008
34,282
Of course this begs the question, why do you want the circuit to operate at a high voltage?
I see no good reason for that. You can always reduce the voltage to a reasonable level with a voltage regulator.
 

Thread Starter

Allenph

Joined May 27, 2015
76
Of course this begs the question, why do you want the circuit to operate at a high voltage?
I see no good reason for that. You can always reduce the voltage to a reasonable level with a voltage regulator.
I think that's the only possibility. I had hoped I wouldn't have to do that, but in practice it will probably waste less energy in heat than carefully chosen resistors, and will offer me more flexibility in the oscillator. I'm just going to have to be really careful what I connect to ground to avoid catching things on fire.

I want to be able to supply a low current to a SMPS at a high voltage that will power some sensitive electronics in another project, but I've heard SMPS don't play nice with big inductive loads and I have 3-4 of them. The idea is to make two channels from the 12V transformer.

Power channel one goes to a big capacitor bank to supply a bunch of big motors.

Power channel two goes to a SMPS which draws negligible current and has it's own independent smoothing cap. That buck converter SMPS takes down the 12V to ~5V with a varying load. That load will be anywhere from 10mA to ~120mA at any given time.
 

Thread Starter

Allenph

Joined May 27, 2015
76
temp.png

Okay, so here we are. I'm kind of back at square one though. (Because it's a square wave...ha...haha...ha)

The output of this needs to change its duty cycle based on feedback. I plan on doing this by using a transistor as a variable resistor and replacing R3.

That's all fine and dandy, but when I do that it will change the duty cycle AND the frequency. We want a steady frequency of ~100kHz for the buck but with a modulating duty cycle right?

Thinking about this, I'm not sure how it's even possible to change duty cycle without effecting frequency using this, a timer, a MCU, or what have you. Changing the duty cycle by definition changes the frequency, but modulating frequency does not neccesarily change the duty cycle. Am I missing something here?
 

Attachments

Alec_t

Joined Sep 17, 2013
14,280
Since the LT1236-5 already provides a 5V regulated output able to handle varying loads, what is the point of the multivibrator and SMPS?
 

Thread Starter

Allenph

Joined May 27, 2015
76
Since the LT1236-5 already provides a 5V regulated output able to handle varying loads, what is the point of the multivibrator and SMPS?
Really there isn't. Now I'm just kind of asking theoretical questions. I'm just going to go completely linear on this one. When I'm just pulling a bit of current and I want a steady voltage for my control ectronics it's the best option, and my motors I can't use a SMPS for.

Linear regulators, even ones with eroor amplifiers and dynamic current shunts and adders and all that crap shouldn't be that bad to design around some LT3080s.
 

ericgibbs

Joined Jan 29, 2010
18,766
It's been a long time since I did anything with transformers, but doesn't the inductance go as the turns ratio, not the square of the turns ratio?

@WBahn,

I would use 10H for the transformer primary and 100mH for the secondary, thats a 10:1 step down, also set the voltage to 169V for the mains.
I should have said more precisely, thats a 10:1 step down voltage ratio.
Example of LTSpice 10H:100mH transformer plot.
http://www.linear.com/solutions/5092

E.
The other Thread has been closed.
 

Attachments

Last edited:
Top