Fusor Forums

I have found these little high frequency transformers very handy for supplies up to 2kv. They can supply a decent amount of current of 1ma if driven hard. Most really large screen TV's that have twenty or more CCFL tubes used 4 or more of these on their driver boards. Such boards feed these with 24-48 volts DC. Yes, they are current hogs, but do supply decent current out of up to several mA needed to light the long tubes. Each transformer might supply 5 or more long tubes in banks. I have collected about 100 various types of these from scrapped flat screens. I originally hoped to use them as bias supplies for GM, PM and neutron detectors, but you have too many varieties in such a pile that each one or set of 4 would require to hand trim each group of supplies to spec. For one offs, they are great, however.

They can be made portable for battery power if under driven and only a few microamps are needed. I find the trend was to use 4-6 small transformers on the driver boards on big screens, but now they tend to use only two much larger, weighty transformers to reduce the component count. Battery, portable operation would tend to make these larger ones problematic, but mains operation could see these used for rather stiff high frequency, high voltage supplies.

Richard Hull

Hi .
Isn't easier to buy one?
Like this one it tried it works just fine
https://a.aliexpress.com/_dXvy7er
Good luck

The board says 200 volts. Is that enough? Can you blast in more than the stated 5 volts to get more voltage. If so, what is the current drain from the low voltage supply. Will the pot on the board allow for more than 200 volt adjustment. Not many nuclear detectors will work at 200 volts. what is your detector? Ion chambers can work as low as 150 volts.

Richard Hull

Dear Richard Hull
No its not 200v !
On 5 volts maximum you will get 700v
But on 9 v you will get 1000v
There is a soldiering points fixed voltage 200v, 50v , 170v
The green terminal is the main out voltage
And you can adjust the voltage with the variable resistor
Have a good day

And the easiest way to find a high voltage Module is from
Rechargeable Electric Mosquito Zapper it gives 2000 volts
And some adjusting with variable resistor you can use it
Its called the poor mans Hv you can get it for 10$

Hi:

I was aware of the mosquitoe killing HV source. It is a "Joule Thief" like the one I previously discussed (i.e a blocking oscillator) with added voltage multiplication at the output. Thus, it's noise makes it unsuitable to my needs.
What is the noise in the commercial module you described before?
EDIT: It seems like the different voltages are taking out at different stages of the output multiplier. Can it sustain the current claimed at the latter stages of multiplication?

Cai
most of these can do up to 10 watts of output some a little more
from that you can work it out

John:

That seems way too high. Taking the output directly the output at the transformer (thus avoiding any losses in the multiplier) we get a power 1.5 W (the 75V claimed times the *optimistic* 20mA). At 400V that is around 4 mA, which is respectable.
Still, the noise is what worries me the most... It also seems like there is no feedback. The circuit is just a 555 square wave generator and a N-Channel MOSFET! A quick homemade blocking oscillator is cheaper, doesn't have shipping time and should be stiffer. Furthermore, by using a small capacitor between between base and colector (multiplied by Miller Effect) one compromises on the rise time (and thus output voltage) but can significantly reduce noise.

Cai
Both Doug Coulter and myself have used similar bought from Mouser and I got 15 watts max out of mine they are meant to run one or two 5-8 watt flouro tubes as used in laptop LCDs

John:

I think we are talking about different things. You probably reference something like this: https://www.mouser.es/ProductDetail/JKL ... qjc3YDs%3D. Which I have read about in Doug post "Standard HV Supply for detectors". I was talking about the cheapo module referenced above from Aliexpress.

yes I was talking about reputable supplies not aliexpress or banggood both of the later sell knockoffs that generally take their specs from the original but the product is nowhere near the original
this also appies to components like gate driver ics mosfets, IGBTs, RF transistors, ZVS driver boards, just see what the tesla coiling people are saying ---STAY AWAY from the chinese junk you get exactly what you pay for --high priced rubbish

John,

Any chance you could find and share a link (Mouser or etc.) to the kind of supply you are talking about?

Or details of the manufacturer/model.

Rex
I'll try
I'm in lockdown still but there are people at work who know what we bought i'll see

After the final exams and some unexpected health issues, I'm finally able to continue working. I just got the inverters for a power supply like that demonstrated by Nathan Marshall, but before that I'm (again) revising the HV supply.

More specifically, I've found a way to use the blocking oscillator effectively: by connecting an appropiate sized capacitor across the primary (the winding between Vcc and the collector) the charging and discharging happens in a resonant way, thus the output waveform is really similar to a sinewave and the noise (using the same filter as before: another thing I have to optimize) is way down: I cant distinguish it from background (which is around 100mV). This is promising, because it's a stellar performance (with a relatively stiff output too) for very few components and a very low cost.

I'm also using this cheap chinese flyback transformer (https://www.ebay.es/itm/15KV-High-Volta ... Sw-89ZRPh7) (I say flyback because despite not having any specifications, it has two small pieces of paper separating the yoke). The turns-ratio is also really high so I can operate at really low primary voltages.
The only thing I have to figure out is how to regulate the output without including excessive noise. In the internet I've seen a transistor stealing current from the base of the original oscillator, but without lineal control that seems noisy. Any ideas?

As soon as I have everything figured out, I will post en extensive explanation (LTSpice simulations, schematics, etc)

After becoming tired of feedback compensation shanaenigans, poor regulation and spurious oscillations, i'm now looking to linear high voltage, maida style regulators.
I've even found the fascinating LR8, a one-package, 400V, cheap, linear regulator! Wow! Is there any problems or quirks (aside from power regulation) I should be aware of?

This linear regulator are toys, with terrible regulation under load and tremendously easy to pop. I have finished the supply now, using a classic series regulator with a HV transistor from a CRT TV.
If anyone is interested in pursuing low noise power supplies, GET A DIFFERENTIAL PROBE! I can not tell you how many times I've being fooled by common mode noise and how many hours wasted tracking inexistent noise supplies. If you don't have it, you are handicapped (like me). Short ground clips (use the tiny ground springs) and always triggering the scope on the input waveform are a must. If the "noise" also does not trigger, or/and changes amplitude changing the lenght or shape of the ground clip, then you should most likely ignore it.
And do some serious indepth reading about noise before going down the rabbit hole.

Sounds like you have a good grip on the effort related to noise. I bought a Tektronix differential probe system at a hamfest at a "silent key" sale of $10.00 (the sellers did not know what it was. It was listed at over a thousand dollars by Tek!!). It is a high voltage differential probe and can handle 1kv. Sweet! I have only needed it a couple of times since I bought it. You are right about triggering and grounding issues went poking about for noise. Did I mention, I hate noise!! I hate chasing it down.... Fortunately, noise is not an issue in most of my current work.

Richard Hull

Why not filter with a 2-stage RC filter, like what every preamp uses? Since the bias current is low the resistor values can be MOhms.

http://sim.okawa-denshi.jp/en/CRCRtool.php

Joe:

If you use a fluorescent lamp inverter and this filter (just like Jim Williams used) the noise will be indistinguishable from background. However I wanted to add feedback as to make the supply regulated, and it always seemed to introduce noise. As of right now, and trying to learn as much as possible about feedback compensation and such, and are experimenting with a simple circuit that seems to work, more or less.
One of my main learning points have been that you can not say simply "add feedback to make it regulated" to a noiseless supply and expect it to stay that way without additional care.

Feedback will cause instability if the gain loop crosses 0dB at more than -20db/decade

The problem is that I dont know the characteristic function of the CCFL-Inverter and bridge-filter themselves, so planning for compensation is, well, impossible.
The best I'm doing is inserting various capacitance values between the feedback and the output pin of the TL072 until the noise (which very clearly is caused by oscillations since by killing the bandwidht by putting 0.1uF at the output of the opamp eliminates it) goes down to acceptable levels and no oscillation is present.

Not impossible. It can be measured.

How can this be done?

This should get you started:

https://www.ti.com/lit/an/snva364a/snva ... 8718652691

If I understood correctly, my main objective using this technique is to observe the phase shift at 0db and make sure it is high enough (between 45 or 50 degrees).
If I make my function generator float, then the transformer is unnecessary, right?