Roast my circuit - Negative rail

[techninja80]

Hello again all. I am from attending the Just Google It University of electrical engineering. So please feel free to roast my circuit.  I do have a reason to be posting actually. I would like advice on two things. How can I load test the circuit properly.  I typically use an electronic load, but I am sure it does not work on a negative rail. If it does please let me know how to set that up. If not should I just throw a resistor on there, and use ohms law to figure out the amp draw. Ideally I would like to specify the amp draw as I can with an electronic load.  Then finally any advice on any thing else I can add to the circuit to improve it. I have no idea what I am doing, and use google, and reverse engineering to teach myself. So any input would be great. 

P.S. I want to put load on the negrail node.

[Infraviolet]

The best testing is to put it in scenarios based on the eventual application it will be used for. You're setting up a circuit to provide a negative voltage rail from a positive power input, but what larger system are you planning to integrate it in to, that should give you the first indication for what sort of tests you'll want to perform on your circuit.

[techninja80]

But what larger system are you planning to integrate it in to, that should give you the first indication for what sort of tests you'll want to perform on your circuit.

Well to be honest I am making to see if I could. I am learning, and trying things just to see how things work.  I had an assortment kit of opamps, and a lot of them needed a negative rail.  I have looked up circuits to give me a negative rail but they all collapsed with any form of load on them.  This one I found by accident trying to fix an old board.  My mind was blown when it showed a test point of negative voltage and I was like how the heck did that happen. Anyways long story short I think I finally found a circuit that I can test out all these op-amps I have sitting around.  So long story short. No real circuit yet. I am in the discovery phase right now, and I want to just see how much load I can put under this before it collapse like all the other ones I tried before.

[PGPG]

In 80s I was using 555 to drive step-up converter in our product and even using its Compensate or Reset PIN you can put there a feedback to regulate output voltage, but in general nowadays are not the times to use 555 to drive DCDC. There are thousands of other DCDC specialized circuits. Just search for them at Digikey or Mouser, collect datasheet and learn how everything works.
First try to understand some step-up converters. In datasheets you can find how to design them so you will be able to calculate output current you can get from them.
Then if from 5V using step-up you get 10V than using the same idea but everything reversed you can assume that your old +5V is your new GND and your old GND is your new -5V so step-up reversed will give you -10V relative to your new GND so it will be -5V relative your old GND.

[Kim Christensen]

P.S. I want to put load on the negrail node.

If the input of your electronic load is floating, then simply connect it's positive input to the DUT's ground (Device Under Test) and it's negative input to the DUT's negative output rail.
Also works if the DUT's ground is floating. One of them needs to be floating.
"Floating" means not connected to building or AC ground.

[Xena E]

Just an observation, or two.

Using the 555 is perfectly valid for this if its what you have to hand to experiment with. There is better stuff out there, but it has to be bought!

In using the 555, the minimum supply voltage to run them on is around 5 volts, if youre geting a realworld standard 555 circuit to run at 3V, thats lucky.

How had you intended to regulate the negative rail?

[Kim Christensen]

It might be the CMOS LMC555 IC rather than the old school variant.

[Xena E]

It might be the CMOS LMC555 IC rather than the old school variant.

Good point, I always note that on schematics though. (7555).

[Harry_22]

Hi!
Could you attach .cir file here.

[PGPG]

Using the 555 is perfectly valid for this
....
How had you intended to regulate the negative rail?

This is the reason I suggested to not use 555 and look into DCDC control ICs. At original OP schematic 555 just generates square wave without any pulse time control. In such case without load output voltage will theoretically rise and rise (here in negative direction) as you just continuously pump energy into output capacitor.
If I were using DCDC to convert + into - I would write the IC I suggest to start from, but I just don't use such ICs.
I'm using only step-down ICs and when in 2014 I started to read datasheets it ended with TI issuing new revisions of 9 datasheets.

[Ian.M]

An old but readily available hobbyist friendly DC-DC converter IC is the MC34063A.
Also see OnSemi's AN920 which goes into its theory of operation and design calculations in far more detail, and has many more example circuits.

It supports switching currents of up to 1.5A so depending on the duty cycle you can easily get several hundred mA of output current.   Its cheap and readily available in a DIP-8 through hole package.  Its low switching frequency (<100 kHz, typ ~25 kHz, but varies with load as it cycle skips to regulate) means it tolerates poor layout and even solderless breadboarding with few issues.  It can be used in boost, buck and inverting buck-boost applications with or without an external switching transistor.  Its switching transistor and supply voltage has an abs. max rating of 40V, so it can boost to over 30V without a transformer or voltage doubler, and work with over 30V DC in.

The downsides are the low switching frequency which requires far bigger and more expensive inductors and capacitors vs a modern switching DC-DC IC operating at high hundreds of kHz to low MHz, and the high current sense threshold (typ. 300 mV), so high I²R losses in the sense resistor, and bipolar output stage which inevitably has much higher losses (so lower efficiency) than a MOSFET.  Also its 'low' standby current is max. 4mA and typically in the 2.5 - 2.8 mA region so its unsuitable for most battery powered applications without an upstream power switch.

[techninja80]

Hi!
Could you attach .cir file here.

Gladly. I imported the LM317TI from a regulator library. I am not sure if it will store that component in the cir file. I you need the regulator lib let me know.

[techninja80]

Thank you everyone for all the great suggestions. Please keep them coming as one of them will be helpful to someone. I however know there are existing circuits to do the whole thing. I am not trying to get a negative volt rail, I am trying to design and build my own circuit for it.  I also am trying to use only the components I have thought my self about as I am self thought. I like to do projects that force me to demonstrate what I learned. So please once again to everyone continue to suggest these ready made IC's as someone will find it useful  .


Now for me and my design. Again please feel free to roast. I built it, and drum roll please............... It EXPLODED!!!!!!!!!!!!!

Literally C1 exploded. This happened because under no load, the voltage just kept getting higher, and higher. I left it like that for a bit, and was shocked by the explosion of the cap. Picture for awesomeness electroboom style.  Anyways this is awesome to me because I feel like its a wonderful journey. Anyways. My personal solution is use a comparator to open a transistor to a resistor if the voltage to C1 gets too high. I plan to design a circuit to do that later, hope to post it soon.

Thanks for following the adventures of a noob.

[techninja80]

Using the 555 is perfectly valid for this
....
How had you intended to regulate the negative rail?

This is the reason I suggested to not use 555 and look into DCDC control ICs. At original OP schematic 555 just generates square wave without any pulse time control. In such case without load output voltage will theoretically rise and rise (here in negative direction) as you just continuously pump energy into output capacitor.
If I were using DCDC to convert + into - I would write the IC I suggest to start from, but I just don't use such ICs.
I'm using only step-down ICs and when in 2014 I started to read datasheets it ended with TI issuing new revisions of 9 datasheets.

Haha this is literally exactly what happened  see my post above.  You get the nailed it of the day award in my books.  I will look into the IC. But I am on a build it from the workshop project right now. Maybe if my plan does not work. I will go to those.  Thanks again for you on point comments, and perspective.

[Ian.M]

So if you are staying with the 555, to regulate the negative rail you need to pull the RESET pin low once the output voltage reaches your desired negative voltage, and let it go high again when the output voltage starts to droop.  You could do that with a comparator, but the precise implementation varies depending on what reference voltages or regulated positive supply rails you have available.

[Kim Christensen]

If you want only a few milliamps and very rough "regulation", then a simple capacitive inverter using the 555 can work.
Not very effective at low input voltages because of the x2 diode drops, but the unregulated & unloaded output voltage won't spike like the inductor version does. Current is quite limited.

[Xena E]

Using the 555 is perfectly valid for this
....
How had you intended to regulate the negative rail?

This is the reason I suggested to not use 555 and look into DCDC control ICs. At original OP schematic 555 just generates square wave without any pulse time control. In such case without load output voltage will theoretically rise and rise (here in negative direction) as you just continuously pump energy into output capacitor.
If I were using DCDC to convert + into - I would write the IC I suggest to start from, but I just don't use such ICs.
I'm using only step-down ICs and when in 2014 I started to read datasheets it ended with TI issuing new revisions of 9 datasheets.

Haha this is literally exactly what happened  see my post above.  You get the nailed it of the day award in my books.  I will look into the IC. But I am on a build it from the workshop project right now. Maybe if my plan does not work. I will go to those.  Thanks again for you on point comments, and perspective.

A DC/DC PSU chip is going to be the easiest for sure, however this can/and has be done with the 555.

It can be done with a single chip, bang/bang regulation as Ian.M suggests. Or using two 555's one configured astable, driving one as monostable with pulse position, or pulse width modulation, it's more instructive than using off the shelf components... ... ... or more destructive if you get it wrong as you've found out.

[schmitt trigger]

I also vouch for the MC34063.
Its major advantage is that being a hysteretic regulator, the newbie avoids a major problem facing a novice electronics hobbyist: compensating the closed loop response.

It is dirt cheap and widely available, all of the three basic regulator topologies are extensively documented, and without an external boost transistor you may supply almost 100 milliamps of load current.