High voltage power supply / High Voltage / Forums

Forums

4hv.org :: Forums :: High Voltage

« Previous topic | Next topic »

High voltage power supply

1 2 next

Move Thread

LAN_403

darthtony

Sat Jan 19 2013, 02:14PM

Registered Member #8464 Joined: Mon Dec 03 2012, 02:24PM
Location:
Posts: 6

Hi, i am a beginner in electronic and HV stuff, and i would like to build a bench top variable high voltage powers supply using a flyback transformer. I have played around with the simple 2n3055 driver and an arduni/mosfet driver, but what circuit would be suitable for my application? By varying the input voltage to the flyback, will the output voltage vary linearly?(the power supply isnt for something specific,just for playing around and maybe some physics experiments)

Ash Small

Sat Jan 19 2013, 03:18PM

Registered Member #3414 Joined: Sun Nov 14 2010, 05:05PM
Location: UK
Posts: 4245

It will (The flyback topology can also be used to reduce voltage), another way to regulate voltage is to use a fixed voltage, but vary the number of primary turns. This way you will get more current at lower voltage with more primary turns, however, other factors come into play, so you might not get a truly linear response-flybacks tend to have a 'sweet spot', due to all the various factors involved.

Also remember you have a maximum volts per turn (usualy ~3-4V) before the core saturates (I think this applies to flybacks-it certainly applies to other topologies-This is one of the reasons why you should try different numbers of primary turns, and different numbers of feedback turns, when using the single 2N3055 driver. Frequency also seems to play a part here.

The best thing to do is experiment. No two flybacks ever seem to have identical characteristics, in my experience.

Another thing you can try is paralleling two or more flybacks (preferably identical ones), using one primary around two (or more) flybacks. This effectively increases the cross sectional area of the core, allowing a greated voltage per turn (hope that makes sense). You can then parallel the secondaries for more output current (connecting the secondaries in series for more voltage usually leads to failure).

darthtony

Sat Jan 19 2013, 04:57PM

Registered Member #8464 Joined: Mon Dec 03 2012, 02:24PM
Location:
Posts: 6

So if i understand you correctly, the number of turns i should wind depends on the voltage i am using. Varying the number of turns is not really practical beacuse i want it to be a bench power supply. So i am thinking of using a simple LM317/LM338 variable power supply, which i will use to control the primary voltage, and for a maximum voltage of, lets say 12 Volts i should wind 3/4 turns is that correct?

Now about the driver cicuit, i am leaning more towards a arduino/mosfet or 555/mosfet design, mainly beacause i do not want to have to experiment with different feedback windings. Also instead of using a variabe power supply, can i use PWM?

What about the ZVS circuit?I was watching some videos, and i noticed that the sparks were very differnet from what i was making using a arduin/555 configuration.They were fatter, but you had to pull the electroded close to eatch other to start the spark.In my circuit, the were less fat and right, but could jump longer distances. why does this happen?

Sulaiman

Sat Jan 19 2013, 05:16PM

Registered Member #162 Joined: Mon Feb 13 2006, 10:25AM
Location: United Kingdom
Posts: 3141

Output POWER varies linearly with supply voltage and/or duty cycle,
NOT output voltage.
The voltage output mainly depends upon the load.

Only if you have a constant resistance load will the output voltage will be proportional to supply voltage and/or duty cycle.

Ash Small

Sat Jan 19 2013, 06:52PM

Registered Member #3414 Joined: Sun Nov 14 2010, 05:05PM
Location: UK
Posts: 4245

darthtony wrote ...

What about the ZVS circuit?I was watching some videos, and i noticed that the sparks were very differnet from what i was making using a arduin/555 configuration.They were fatter, but you had to pull the electroded close to eatch other to start the spark.In my circuit, the were less fat and right, but could jump longer distances. why does this happen?

While a lot of people use what is generally called a 'TV flyback' with a Mazilli ZVS circuit, it doesn't operate in 'flyback mode' when used this way. It operates in 'push-pull' mode (like a half bridge or full bridge). For this reason, the airgap is not required, and should be removed. Also, because it is not using the flyback topology in this case, the output voltage is lower (it follows the 'usual' primary: secondary ratio), however, quite a lot of power can be pushed through them, and the output current is higher, which results in fat, juicy arcs, which can be pulled to quite a distance once they have been struck.

darthtony

Sat Jan 19 2013, 07:31PM

Registered Member #8464 Joined: Mon Dec 03 2012, 02:24PM
Location:
Posts: 6

Thank you very much for your responces, they have been very helpfull, but i have some more questions:
1)What sort of power supply would i need to use for the primary, to make a deent lab high voltage power supply?

2)Since i plan it to be a variable power supply, i was intending to use a regular panel voltmeter, which would measure the primary voltage, and then extrapolate the output voltage, in order to avoid having to build a HV meter, but from what you said , I understand that this extrapolation will work without load, but it will be off with a load?
3)any suggestions on the type of driver circuit i should be looking at?

4)Do you have any links with similar projects?

Ash Small

Sat Jan 19 2013, 08:58PM

Registered Member #3414 Joined: Sun Nov 14 2010, 05:05PM
Location: UK
Posts: 4245

Well, you'll only get a milliamp or so from most flybacks, assuming you are using a DC one (old AC ones may be a bit higher), genarally speaking.

It depends what you want to use it for really.

We'd really need to know what you have in mind before advising you further.

I think it would be reasonably straightforward, though, to have a fixed frequency, fixed voltage driver circuit, which is switched by a feedback circuit from a voltage divider on the output, and use the flyback to charge a capacitor (or bank of capacitors) to a voltage selected by you (could be implemented with a potentiometer, maybe at the bottom of the divider). You could then draw power from the capacitor bank, and the flyback would keep the capacitor bank charged to your selected voltage, switching off when that voltage is reached, and switching on again as current is drawn to maintain the voltage on the capacitor bank. This would only work for resistive loads though, or with a resistor on the output of the capacitor, which is why I ask what you intend to use it for.

There are several circuits on this site, and elsewhere, for charging capacitors to a preset voltage which could easily be adapted for something like this.

The disadvantage of this circuit is that if you try to draw too much current, the charger circuit won't be able to maintain your selected voltage.

The above is just one example of a possible solution. I'm sure others will suggest alternatives.

EDIT: the part of the circuit in the bottom left hand corner in this link shows a similar system, but you'd want to adapt it to control a flyback driver, rather than the Mazilli ZVS circuit illustrated, for reasons I gave earlier in this thread.

darthtony

Sat Jan 19 2013, 11:19PM

Registered Member #8464 Joined: Mon Dec 03 2012, 02:24PM
Location:
Posts: 6

Well i dont have a specific use for it. I just have a flyback transformer laying around(DC flycback) and i thought of making a power supply, just for playing around and maybe some physics experiments.(suggestions?) Well 1mA maybe low, but at least it will be reasonably safe, so i am thinking of not using a capacitor bank what do you think?(or if i need more current for a short amount of time, i could connect one to the output). Also can you explain why oyur approach will only work with resistive loads?

Ash Small

Sun Jan 20 2013, 07:05AM

Registered Member #3414 Joined: Sun Nov 14 2010, 05:05PM
Location: UK
Posts: 4245

It will only work with loads that drain the capacitor at a slower rate than the capacitor charges, otherwise the voltage on the capacitor will drop.

You may be able to get a bit more than one mA from a DC flyback, but not more than a few at most. For a bit more power you could use two or three in parallel to charge a capacitor bank, or whatever.

You needn't use a capacitor bank, but it would be a fairly simple way to maintain a pre-selected voltage. Obviously there are risks involved when charging capacitors to these voltages.

If you just wanted to do experiments with corona discharge you could manage without it, for example.

You could still draw arcs, for example, with the capacitor, but you'd need a resistor in series with the capacitor and arc. It would probably need to be in the range of tens of Mega Ohms, though, and rated for a few watts (the maths is fairly simple).

There are obviously more safety precautions required when charging capacitors, though.

I believe most commercial HV DC power supplies have capacitors and resistors before the output terminals. Certainly in the circuit diagrams I've seen. The resistor would limit the output current to 'safe' levels.

Erlend^SE

Wed Jan 23 2013, 10:31PM

Registered Member #1565 Joined: Wed Jun 25 2008, 09:08PM
Location: Norway
Posts: 159

Exactly what are you seeking?

Constant voltage (use internal resistors, or voltage on other winding as feedback to PWM controller(TL494, UC3842 e.t.c.). Add current limit if used for arcing.
(I have a HV supply like that extracted from a computer monitor: 30 kV, 0.5 mA)

Constant current: connect a resistor in series with the HV return and use that for feedback to PWM controller.
Limit voltage so transistors/FET doesn't blow.

Constant power: PWM/555 controller with no feedback.
Or multiply "opamp" that combine current and voltage.

Constant voltage is neat for electrostatic stuff, but I haven't seen too much interest about it around here.
(And my HV supply refuse to do arcs, but corona and electrostatic stuff, and CRT stuff works nicely)

I guess a constant voltage one with current limit would be most familar since it behave like most other lab-supply's.

1 2 next

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob