Continue to Site

Welcome to our site!

Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

  • Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

small HV AC power source needed

Status
Not open for further replies.

Ropa II

New Member
I need a physically small and moderately priced high voltage generator. High voltage is a relative term...so….

VIN = 5Vdc, up to 500 mA, the 5V supply is external (not located inside the specified enclosure)
VOUT = AC or pulsating DC, 1500V peak to peak, more is better, but not necessary
Physical size (MAX) 7cM Length, 4cM Wide and .6cM height
My load will range from 10uA up to 2mA
Regulation is not needed (or desirable)
The output is only needed for 5 to 15 seconds at a time with a rest time of at least 3 hours between uses so the need for heatsinks is not a significant issue

It seems a step up transformer is needed, but where to get a smaller transformer that allows electronics and the transformer to fit within the enclosures limited volume? The obvious answer to me is to use an SMPS with a 50/50 on/off driver.

Step up transformers/LCD inverter/backlight chip seems like a viable answer, but the transformer needed at 20 to 40 kHz switching frequency gets to be very large. I’m thinking I need to use an HF frequency transformer (in the 2 to 5 MHz range), which significantly reduces the physical size of the step up transformer. There are no LCD inverters/backlight chips that function in the MHz range however.

I have reverse engineered some electronic arc lighters that are popular these days, but the transformer for 20 to 30 kHz operating frequency are much to large for my enclosure.

I have thought of using an rf generator with a transformer (sinewave/non-smps), but I’m having trouble finding the needed ferrite to work between 2 and 5 MHz. Common toroids for that frequency range are almost impossible to wind, so an E core (or similar) ferrite core with a bobbin is probably needed.

The use of a Planar transformer a possibility, but no one makes a commercial unit that is small enough for my use. The smallest planar transformer I could locate is a 3 watt core (Coilcraft), which is much to large for my enclosure.

I cannot use energy storage capacitors or a cockcroft-walton voltage multiplier because they will only produce DC.

Any suggestions?
 
Thanks Ron,

The planars can be used at high step up ratios. The primary is 1/4 turn on each primary layer, so they can be wired together. So 4 layers makes 1 turn on the primary. And, 20 turns on each secondary (x4) does give an 80 to 1 turns ratio.

Planars at high voltages work ok with (extra layer of) polystyrene coatings on each of the layers. There is an awesome technical paper describing the pitfalls of a less than perfect planar encapsulation. And supplementary insulation made out of .003 inch thick teflon can improve the isolation. The paper is at https://www.magazines007.com/pdf/High-Voltage-PCDesign.pdf.

I mentioned planars only because there was a possibility because they are generally low profile, so they might be able to be shoe horned into my enclosure.

I knew this project would not be easy::>

Regards
 
I have a lash up that uses an old tv lopty, regulation and switching is done with a Uc3843, the primary is just wound over the core.
It makes from a couple of hundred volts up to a few Kv.
 
In round numbers, 5 V to 1500 V is a 300:1 ratio, and 500 mA / 300 = 1.67 mA. Or, the source is good for 2.5 W, and 1500 V at 2 mA is 3 W. So even at 100% efficiency, your source cannot power your load.

Relative to an AC output, the 5 V source can make only a 5 V p-p AC input (again, 100% efficiency) to the converter, or 1.77 Vrms. This leads to a turns ratio of 850:1.

A photoflash trigger transformer comes to mind, but that 1/4" height will be a problem there, too. Would a non-isolated flyback circuit without an output rectifier work for your (secret) load?

ak
 
Relative to an AC output, the 5 V source can make only a 5 V p-p AC input (again, 100% efficiency) to the converter, or 1.77 Vrms. This leads to a turns ratio of 850:1.

A photoflash trigger transformer comes to mind, but that 1/4" height will be a problem there, too. Would a non-isolated flyback circuit without an output rectifier work for your (secret) load?

ak


First, I need to thank all who replied. Unfortunately (for me), I thought the forum would advise me if there was a new reply to my original post. So, I just saw the replies, please believe me...I didn't intentionally ignore those who replied to the initial post!

Lots 'o water been undah dah bridge since my original post. Here's an update with relaxed design criteria and additional information.

Since the original post, I did additional research and found that I could use a higher input current supply. So, I have 4 volts (LIPO) at 2 amps of input power available....2 amps max at startup and ~1A actual operating current. I have a larger enclosure also, a cylinder at .9 inches wide by 2 inches long is suitable. I can get by with 500 volts peak to peak output but it must be AC or a modified square wave output. The supply DOES NOT need to be constant duty, 3 or 4 seconds of output with 5 or 10 minutes of off time is ok. I can tolerate some ringing at the transitions.

The closest off the shelf solution is a LCD inverter. I ordered a surplus unit that is small and compact and that produces up to 3KV at up to 4 mA. It works fine, except the driver chip doesn't have technical information available. It operates fine at approximately 25 kHz, but it is to bulky. And, it needs a minimum of 1 mA of load current in order to operate. If I could make a higher frequency transformer (which is much smaller physically), I might be able to build my own driver circuit that is far smaller than the LCD inverter unit that I am experimenting with. However, since the specs for the Chinese made LCD inverter aren't available, there is little chance that I can increase it's operating frequency significantly. I would like to have a low profile transformer that operates at 1 to 2 MHz rather than 25 KHz. I am very familiar with high voltage (ham radio operator) and microelectronics soldering/rework and with custom transformer winding.

My load will vary from an arc down to 1 megohm or so, my preference is not to use a royer or similar type of power oscillator that utilizes a feedback winding on the transformer since these types of drivers don't do well with a load that varies as much as mine does. Regulation is not critical, but it is desirable to have some form of crude/minimal regulation.

I'm currently looking for a ferrite that is suitable for a 1 to 2 MHz transformer which is easier said than done::>

Any suggestions are appreciated.
 
1 - 2 mhz would require some special engineering as thats a high freq, smps's commonly work around 100khz, though some of the smaller phone chargers I believe now run close to 1mhz, however there would have been quite a bit of research & destruction getting that to work.
Reading your last post a simple blocking oscillator would be really simple and might do the job.
 
Hi Doctor P et al,

I think I need the higher frequency due to my enclosure size limitations-I just don't have the space for typical 20 to 50kHz based magnetics. If HF is used, the magnetics shrinks by quite a large margin. Further, the wire for the magnetics becomes much smaller due to the skin effect improvement. Yes, there will be some increased losses because the leakage inductance/parasitic capacitance becomes much greater as the frequency of operation increases.

I will look into blocking oscillators again, but I think they are not good performers as any kind of oscillator doesn't do well with a widely variable load...such as mine. This is a well known problem for royer oscillators and other oscillator types that use a feedback winding on the magnetics. When an oscillator is used for communications purposes, it is common practice to use a buffer stage between the oscillator and the load. Even the well accepted cmos inverter gated oscillator doesn't start all the time, except when a 3 stage oscillator is used. A three stage cmos inverter oscillator is the standard (2 stage) oscillator, except that the third stage functions as a buffer.

Somehow I gotta make the unit physically smaller without sacrificing the power delivered to the load...smaller high frequency magnetics seems like a contender. Right now, I do have a bobbin that is small enough to do the job...but I can't find ferrite that is square and that is suitable for operation at 2 MHz. I am considering buying some ferrite power and doing the sintering in a microwave oven, multiple web pages and videos depict melting of metal by using a crucible and a high power microwave unit. I hope I don't have to do that to get a suitable HF ferrite core for my bobbin::>

Planar transformers can operate at 10 (or more) MHz, which MIGHT work for me because I am not trying to charge capacitors or rectify the output in order to make DC. Planar transformers can have 2 mm thicknesses, and the power density is extremely favorable. But, the lowest power planar core I know is is a 3 watt unit and it is still to large for my enclosure.

Regards,

BB
 
If I could make a higher frequency transformer (which is much smaller physically)
Small will probably lower the resonance frequency. Pushing the wires together will increase the capacitance.

Here is a picture of "scatter wound". It is a way of adding space between the wires and allowing the part to work at a higher frequency. I have also added transformer tape between layers of wire to increase the ability to work at high frequency. Both methods do not make the windings smaller.
upload_2017-8-22_7-24-13.png
 
One thing not mentioned so far and will be one of the biggest issues is driving the fets, or whatever switching devices, driving the gate charge at 1mc will take a fair amount.
Ron that looks suspiciously like an old tube radio oscillator coil.
 
Status
Not open for further replies.

Latest threads

New Articles From Microcontroller Tips

Back
Top