Capacitive Voltage Reverser - Create a Negative Power Rail With No Transformer - Simply Put
HTML-код
- Опубликовано: 19 май 2020
- Using just a square wave, two diodes, and two capacitors, in a very similar way to how the square wave fed voltage doubler works, you can take a normal positive rail and flip it around to give you an equivalent negative rail (not counting diode drop losses). Want a 5V Arduino Uno with a +100V and -100V signal supply?
Amplifiers Playlist: • Amplifiers
Signal Processing and Filtering Playlist: • Signal Processing and ...
You can join me on Discord as well! -- / discord 
Наука
Awesome explanation
From Sydney, super thanks! You are one of the best communicators in this subject matter. I truly wish you great success in your channel.
This is a charge pump. In order to cascade it, you need every other stage to have an inverted square wave (as can be seen on Wikipedia: en.wikipedia.org/wiki/Charge_pump). The examples on Wikipedia are for a non-inverting charge pump, but you can just flip the diodes and replace the positive connection with a ground connection (as you have here).
While increasing the capacitance is one way to increase the power output, another way is to increase the clock frequency (the capacitors charge through the diodes very quickly, so you're basically topping them up more frequently), though you end up being limited by inductance and resistance within the traces and capacitors.
For example, the LTC7820 is a charge pump IC that can provide upwards of 5A (it seems more complicated to use, but most of it is for feedback, customizability, and extra features. The core is mostly the same). The example circuits show the need for multiple parallel capacitors at the output to minimize ESR.
Losses can be further mitigated by using low-resistance MOSFETs set up either in a diode configuration or in an analog switch configuration.
Awesome breakdown. Thank you King
another great explanation, thank you.
Awesome thank you 😊
that was me who said i used to think of ground as the end!! you were doing something on signal that had a cap coupled to ground that was helping to feed its amplification through an npn or something.. cool video by the way!! thats another one i know ill use.
good explanation
Excellent video, I have a question, you have technical information if there is any to design a DC DC converter with positive variable input and that has a DUAL symmetrical output (high power approx. 2A) of +15 VDC and -15 VDC. Know of some base design with the theoretical source and some switching regulator.
Great video as always! I presume the next video is going to be a circuit to bring all this back down to a workable 5 volts? or is that just a voltage divider? Can't wait!
Pretty much yes: The idea is to be able to take a small single-sided supply, boost and reverse it to power op amps, take a voltage of any magnitude and polarity within the boosted range and do whatever with it, and then generate readable digital signals plus precision rectify the voltage into something the ADC can handle.
@@simplyput2796 aah makes sense. Seems like a lot of work tho. All the best!
It's all for fun and learning. I'm not trying to start a successful commercial venture with it.
Sir,I am learning electronic basic watching videos like you and reading basic from Google.Can you tell me the input voltage is Ac or dc?
Could this be used with a 555 switching a transistor to power around 150ma?
Just use 2 transistor to increase the current.. around 20mA is ok . 50mA is a almost the limit.... Over 100mA it is better to buy a transformer to get neagtive rail.
For 50mA negative rail, the switching current was 350mA.. at 2.2khz. .. at 150mA the switching current was 2.1A .. and it was hot...
If we increase the frequency the switching current will decrease... But then we can't use cheap electrolytic capacitors.. i used 100uf for the middle capacitor and 470uF for the output capacitor... And another 470uF capacitor across the +ve & -ve supply. Then i connected load accross the positive rail to negetive rail.. i had to add a reverse bias diode accross the output capacitor of -ve supply
.. the input voltage was 16V.. and the negative output voltage was -13V..
I used a 555 timer ic to generate the square wave..and 2 transistors for the switching..
The square wave needs to be inverted for multi stage charge pump
Will it also double the voltage, because of the "reverse" cap?
No, it only reverses it. There may be a way to stack the negative stages like you can stack the doubler stages, but I haven't figured one out yet. Right now, you basically take your voltage, double it until it's as high as you need, then reverse it for the other side.
@@simplyput2796 But when the "reverse" cap is charged and the square wave is high again, the voltage will be doubled if the load is very low right?
The voltage doubler has a capacitor charge to the supply voltage on one half wave, and on the other half wave the capacitor discharges with the supply voltage "pushing behind it", so they add together. The voltage reverser doesn't have the supply acting at the same time as the capacitor, because the supply is positive and the capacitor needs to "pull" to make a negative, so the supply is used to charge the capacitor, and then the supply "gets out of the way" while the capacitor pulls through the load and charges the other capacitor.
Could this be used for an OTA?
OTAs aren't high-current devices, so I would expect so. I would try it and see, and if it's unstable, try adding capacitance.
@@simplyput2796 I'll give it a try this week! I picked up an LT1054 which is a switch capacitor voltage converter. Getting a pretty clean -11.94v with a 12v supply. Any thoughts with using such a device for audio signals?
I have no experience with that chip, but if it works, it works.
Dude! Good content. Definitely a sufficient quantity of hair. :) Cool temperament. I like your style. But... eh... in the audio department, maybe a bit heavy on the compression! Permit me to offer an uninvited perspective, for what it's worth. No offense intended; if this doesn't seem helpful, please forgive and forget!
What I'm hearing is the compressor relaxing to it's idle state during the between-phrase pause, then your initial vocal transient comes passing through at full gain loud because the sorta-slow attack doesn't permit any gain reduction yet, and then it sorta catches up with the attack time and cranks the gain WAY down, and the whole rest of the sentence is suppressed by those gobs of gain reduction. So... your sentences kinda come across like... LOUD soft soft soft soft soft soft. BLAH duh duh duh duh duh. WAP ap ap ap ap. GABBA gah gah gah gah gah gah. It's kinda... oh, maybe... scary! Or startling, you might say. Every sentence starting with a little whap in the face with a wet herring. Esh. Mmm. I don't know. Maybe you like it that way! Could be a thing! Fair enough. Me, I kinda prefer the drama in the content, less so in vocal dynamics. Preference, that's all. Were I your sound engineer, and assuming I had a (real or virtual) compressor with lots of knobs and buttons to play with, I'd ease up on the threshold to keep from pushing everything down quite so much, go full-fast on the attack so there's no delay in the gain reduction, and bring the ratio down a bit, so it's not clamping down quite so hard. I tend to think good compression is close to indetectable; that's when you know you're getting it -- when you're not sure it's there. And it's tricky bidness to set a compressor to do that, for sure. Learning to set compression parameters is easiest, I think, by playing with them in the extreme, so that their effects are obvious, and then dialing back and getting comfortable working with the subtleties.
And of course, none of that applies if you don't have a compressor with lots of controls. Harumph! Still, less can be more.
It's actually an audio normalizer that I wrote myself because I wanted it to work in a very specific way, and what I do is use an EBU R128 library to analyze the audio with a 400-ms sliding window as if it were "live" (because I originally wrote the program to normalize my computer audio, and then used it for livestreaming). I really should work on it to make it better, and to actually analyze the full track since it's not live, but the less it's "broken" the harder it is to get myself to work on it. I have tried to add code to account for loud/quiet transitions to minimize the over-amping at the start of sequences, to some limited success. But as you've said in extreme detail, it's very much "brick wall" configured.
I have a personal requirement in order to put out any sort of content: I have to be willing to watch it myself, and enjoy it myself, first, before I try to please anyone else. And that is why the brick-wall normalization (and the jump cuts, because I also wrote my own video editor, when I did my very first video and realized it was going to take an immense amount of effort to edit by hand).
You are absolutely right, and I have received these sorts of comments before (though, nowhere near as expertly), but I just can't bear to do what's actually proper and professional when it would make me itch to even edit it. I hope you don't read this as dismissive, because I don't mean it that way: I just have to be the way I am, and hope enough people are not too annoyed by it that I continue to grow, even if I do it slowly.
@@simplyput2796 Oh! Thanks for the response; I totally get what you're saying. I'm impressed with what you've designed and implemented; I guess we're talking about tuning, here. One of the notions that came to mind, that you're no doubt also not going to have time for, nor sufficient impetus, was borrowing technology from PID controllers. Then you end up with coefficients for rapid changes, slow changes, accumulations, that sort of thing.
Not pushing, just thinking... I have a feeling that shortening the window by a little or lot might actually help. I often run compression with, say, attack times in the millisecond region, or less, and release times of 100 milliseconds or maybe 50. So it's all working pretty quickly, and with voice, the characteristics are amenable.
I also wonder about lookahead... or half-lookahead, at least in an implementation that's working from a recording. So half-window before the current time, and half-window after, sorta smoothing out the transitions, at least by a fair margin. Full lookahead would just mean you transfer the issue to the noise-to-silence transitions.
But... enough! On to funner things, I guess. Best wishes to you!
but it’s your style. I like it, energetic and and sharp, catches my attention. Keep doing it, please :)
I just found this channel - i want to subscribe but ah - the editing is so relentless.
interresting video but too many movie cutoff make the viewing painfull for eyes and brain... Sorry.
Awesome thank you 😊