Fusor Metering Questions

[lanewaddell]

Hey all,

I wanted to bounce some ideas for current and voltage metering off you. My aim is #1 safety and #2 simplicity as long as the measurements are fairly accurate.

I've read through the FAQ posts related to this and had what I'm my mind is a very simple concept and basically want to know if it could work or why it's a bad idea.

My concept is kind of based on "goes ins = goes outs" and IGNORES the efficiency losses of the flyback transformers in my percipitator supply (my understanding is losses should be mere percents but I could be wrong).

Essentially, I want to measure the current using a TC induction sensor on the 240v input line (very safe). Since the voltage is roughly known or can be measured, I'd then know the input power to the supply.

Then I'd like the measure the current only through the case ground. From that information, I could approximate the high voltage since I know the power in and the current out (again, ignoring losses).

For current measuring at the ground, I'm thinking having a split with two equal low resistance resistors in parallel and measuring the voltage drop across one (with the goal of 1 ma equally 1 v).

Using simple V=IR, a 1k ohm resistor will yield a 1 volt drop per a milliamp of current. So I'd have two 1k ohm resistors in parallel and measure voltage across one to find current. The other resistor is the "safety" backup in case one gets disconnected. The resistances have to be equal on both legs or else I wouldn't measure a voltage drop due to zero resistance on the non- metered leg.

The supply can do up to 10 ma at 60kv. Using P=VA, 5 ma PER LEG (because current is split) at 5 volts (1 ma yields a 1 volt drop on a 1kohm resistor) yields only 0.025 watts. So in this case a 1 watt resistor is okay.

This is my conceptual idea, now I'm ready to hear why it's dumb : p

[JoeBallantyne]

The output voltage on the precipitator supply is adjustable. So measuring the return current flow will not tell you what the voltage output level is.

Why don't you just build a proper HV voltage meter for your fusor and use that?

Then you aren't making any assumptions about what is going on, you are just directly measuring the voltage going into the fusor.

You can build a voltage meter for about $50, and a proper current meter for about $25. (In fact I have a kit for each, but you can build your own also.)

Joe.

[lanewaddell]

Joe,

Honestly, a big part of it is wanting to absolutely minimize anything touching HV.

But I'm happy to learn about your kit and how to buy it. I will add In looking to have the signal be digital for labview.


Why wouldn't the return current tell you about the HV at that particular point in time? The resistance is fixed.

[JoeBallantyne]

For the kit, look in the items for sale section of these forums. I posted about the HV meter in the thread on the low cost fusor build.

My HV meter kit is most definitely analog, not digital.

You can't measure the HV, unless you connect to it.

You can't get into even the plasma club, if you aren't measuring the actual voltage you are feeding your fusor, let alone the neutron club.

The resistance of the fusor itself is most definitely not fixed, which is why the return current tells you nothing about voltage level on the cathode.

You can have a current of 1mA flowing through the fusor with many different voltages - pretty much an infinite number of different voltages - because what determines the current that flows is a combination of the voltage on the cathode and the pressure in the fusor.

On my fusor I can get a 1mA current flowing with 10kV on the cathode or 20kV on the cathode. It is just a question of what the pressure is adjusted to. That is true of pretty much any glow discharge fusor.

Joe.

[lanewaddell]

Yea I think I might understand your point.

Processing....

[Rich Gorski]

Lane,

Please take Joe’s advice. It is sound. You are asking for trouble and great inaccuracy if you try to calculate the precipitator output voltage by knowing the input power and the ground current from the fusor. Also know that if your vacuum level is low enough so that the fusor plasma does not ignite you will have zero current on the precipitator output even though there could be tens of kVs there. So measure it directly. If you are going to take on the project of building a fusor you will need to get used to working with (and connecting to) high voltages. So learn and get help from someone who has experience working with high voltages. Then build yourself a high voltage 1000 to 1 divider and measure directly by connecting to the precipitator output. Install your 1000:1 divider with meter in a plexiglass box and use HV cable rated at your max kVs going in so no one can touch anything inside.

Rich G.

[lanewaddell]

Hey Rich,

Thanks for the candid feedback, it is appreciated.

I suppose I will look into voltage dividers, but I'm going to need an optically coupled 12,000:1 divider circuit to do what I want to do with an arduino.

Do you have any practical advice in terms of picking the resistors for the voltage divider circuit? Fundamentally the voltage ratios could be set by any combination, but I am finding conflicting information about how to practically go about picking based on current. Some sources say they should be 10% of the load, while other sources say that is totally wrong and flipped and the load should be only 10% of the divider! Then there are posts saying it is more of an art and you just get good at picking a resistor size based on the load you are dealing with.

Any practical advice here? Reviewing some posts it looks like the 10s of m ohm range for the first resistor is the sweet spot.


Perhaps R1 of 400 m ohm and R2 of 30 k ohm? I'm showing that would yield:

R2 Output (measured) voltage - 4.5 volts @ 60kv
R2 Output current - 0.15 ma
R2 Power Dissipation - 0.675 mW

R1 Output voltage - 59,995 volts @ 60 kv
R1 Output Current- 0.15 ma
R1 Power Dissipation - 9 watts

[Rich Gorski]

Lane,
There’s no magic number to use for a voltage divider. Mostly it comes down to being practical and cost effective. I would consider two things in my choice.

1. How much should the divider load the power supply? For a high voltage supply I wouldn’t go much higher than 1 or 2 percent of the supply rated current. With a 60kV, 10mA supply I would not go to 10% (1mA) unless you want the divider (R1) to dissipate many tens of Watts of power. Your choice of R1 = 400M ohm, 0.15mA (1.5%) seems good dissipating 9 watts.

2. How are you going to measure the voltage on the small side of the divider? DVM, analog meter, Arduino? For a DVM you could nicely set a simple 1000:1 ratio to readout 0-60V when for 0 to 60kV on the supply. The Arduino however allows only 0-5V on the analog input. Thus you need the 12000:1 ratio as you have indicated in your post.

You could even do both the Arduino and a permanently installed DVM in the box with the supply if you use a 1000:1 ratio. From your example a 400M and a 400K. The DVM would read across the 400k ohms to give you 0 to 60V direct readout on the DVM. The DVM will likely have a 10M input impedance so you might want to adjust the 400k a bit to compensate the 10M parallel resistance. For the Arduino which has I believe a 100M input impedance you could break up the 400k (R2) to like 367k and a 33k ohm in series and have the Arduino read across the 33k. If I did the math right the Arduino input would see the 0 - 4.5 volts when the supply goes from 0 to 60kV. I like the idea of having a permanent readout of the high voltage on a simple digital meter that's on all the time and easily visible.

I did a quick search on Amazon for HV resistors and found a pair of 200M ohm, 5 watt, high voltage glass resistors for about $10. Maybe buy 4 sets of these and arrange them as 4 pairs in series to give you a 400M ohm, 40 watt R1. That should easily handle the 9 watts you calculated. Amazon didn’t give the voltage rating but they are 2 inches long so maybe capable of holding off 15kV in air (oil immersion is better).

Good luck with your build.
Rich G.

[lanewaddell]

Thanks Rich!

This is super helpful!

I also appreciate the effort you put into your post to lend some practical advice. I understand that one has to do their own work, but sometimes the community is a little rough on "newbies". Thanks for the useful positivity and especially encouragement that I won't kill myself if I'm careful.

[Daniel Harrer]

Just a few more things from my experience:

(a) I strongly recommend a voltage buffer (i.e. an op-amp at gain 1) after a voltage divider with large ratio. The current will be quite low and even small losses can change the result a lot. Especially relevant if you want to have a capacitor there. You can use a cheap sub-$ op-amp like LM324 in most cases. For the current measurement a zero-drift buffer might be preferable, such as the LTC1051/1053 at around $10.

(b) In addition to the above, a voltage buffer can either double as, or be followed by, an op-amp at low adjustable gain and optionally a capacitor to filter out AC components. This allows for low-level analogue filtering and correction of the signal.

(c) Don't fret too much about resistor precision for both voltage and current. If anything, then thermal stability is the most important factor for accuracy, and low power dissipation and/or an already highly recommended oil immersion should fix that. Any other variance can be adjusted for on the Arduino or with an op-amp.

(d) The ADC in many cheaper Arduinos is not that great, and using code to average many measurements can clog the thing. Instead an affordable and much more precise solution is an ADC such as the ADS1115 run at low sample rate 'continuous mode' and connected via I²C.

(e) If you want like me almost total separation of the HV from the rest of the world, then fiber connections are a viable and affordable option (unlike HV optocouplers which somehow are ridiculously pricey even for non-linear ones) . You can either send UART or USB over fiber with modules around $20 each. I communicate over fiber with an Arduino on the inside to read and adjust voltage and current to have no outside connections but grounding, mains, and the outgoing highly insulated HV line.

[lanewaddell]

Thanks for all these tidbits! Do you have a preferred optical module? I was considering using a cheap laser pointer and solar cell. I was going to experiment with linearity later today

I saw a cool rig where someone was able to transmit music across a room using a cheap cat oy laser laser and a solar panel. If it's linear that could be an option also.

[Daniel Harrer]

The optical modules I've been using are by OPTCORE (https://www.optcore.net/product/0-5mb50 ... ansceiver/) or on PCB as basic module by DFrobot (https://www.dfrobot.com/product-2588.html). At least I think those are the same hardware, they look and work the very same.

Sadly lacking a 20km cable I cannot vouch for their true range, but they worked without issue on a 4m one. OPTCORE also has a few other similarly priced fiber connectors for various protocols, but I haven't used them.