Precipitator Failure

[Alan Sailer]

My attempts to operate a fusor in a neutron generating regime are on hold due to a puzzling issue.

I have been using the newly popular precipitator power supply and have been frustrated trying to easily adjust the voltage.
My first attempt used a plastic tube which I press fit to the 10kohm adjustment pot. The pot fell apart after a few runs.
I replaced the pot with a better quality pot but this pot also had no easy way to adjust the voltage.

Thinking that a high quality Bourns 10 turn pot would do the trick I replaced the small screw adjust pot. When I turned on
the supply the thing exploded. Both output MOSFETs in the half bridge are gone, one has a lead evaporate and the output
traces leading to one of the flyback primaries disappeared.

I know enough electronics to be dangerous but I am puzzled as to why this happened. The ten turn pot is a wirewound unit
but I can't see why the inductance would matter in this case.

Another thing I noticed when I powered up the original (unmodified) unit was that using a metal screwdriver to trim the
voltage would cause the output voltage to fluctuate. I though that this might be because the pot was picking up feedback
voltage from the flybacks. If this is true then a wire-wound pot would do the same with possibly horrible results.

Any comments?

Cheers.

[Mark Rowley]

Hi Alan
I’m sorry to hear about the failure. Months ago Bob and I attempted the extension / potentiometer upgrade with similar results. In my case the driver IC failed, I don’t recall what components failed in Bobs. Once I replaced the chip and reverted back to an oem type pot, all was good again.

As far as the reason for this issue, I can’t say for sure. Others have suspected capacitance in the extension leads or inductive currents forming in the pot windings.

The oem pots are definitely susceptible to mechanical failure if they are overtaxed. A couple months ago I had one becoming slightly loose which prompted me to replace it. I then adjusted the plastic rod extension to minimize lateral and downward force on the control stem. Unless I went into long term daily use of the fusor for months, I don’t see any further risk of mechanical breakdown.

A reminder for anyone else reading the thread, precipitator supplies provide enough power for decent neutron production in a 2.75” system. I highly doubt their suitability for larger cylindrical or spherical systems as those demand current flow a precip supply can’t provide.

Mark Rowley

[Alan Sailer]

Mark,

It's great to know that this is a known problem. If I was planning to use the precipitator supply over the long run I'd
scrap the whole circuit except for the flybacks and build my own half bridge circuit.

By the way what was the replacement self oscillating driver chip you used?

I have some IRS2153 and 21531 on order.

Cheers.

[Mark Rowley]

The IR2153 is the oem replacement.
Not sure I’d scrap the original driver circuit. The 60kv unit for my Fusor is working fine and it easily provided enough power to activate indium a few weeks ago. The supply has quite a few hours of operation under its belt now and I don’t see any threat of failure. It’s just a matter of not playing brinkmanship with its output or employing risky modifications.

Speaking of modifications, I do plan on adding some capacitance to the output to smooth out some voltage fluctuations when operating over 30kV.

Mark Rowley

[Richard Hull]

Why did the ten turn precision pot cause the system to explode?

Now...........before you took out the old 270 degree pot, did you put a dot of paint on the board at one of the outside arms on the board and on the pot. Upon removal did you take an ohm reading from the dot to the center lug, and write it down. Then mark the precision pot at one outside end and preset it against the center lug to the reading on the old pot and make sure to install it dot to dot?

Were you sure of the new precision pot's value to be exactly that of the old pot? Did you exercise the precision pot to make sure there were no dead spots or opens center lug to outside lugs?

It is obvious that if all the above were adhered to then it would not or should not have blown up.

In the rush to enhance with a precision pot one of the above may have been overlooked. I have never bought a brand new potentiometer in my life! I have hundreds of used, removed pots of all types, sizes and ranges in many boxes in my lab. Before I use a surplus pot, I perform all of the above QC checks first. Stock older pots may be noisy (micro open spots as it is exercised), if so, I spray the opening with a cleaner made for the job. Sealed pots may need to be heavily exercised over many sweeps.

1. Measure replacement pot outside ear to outside ear. Is it what you need?
2. Center the pot wiper. Measure center lug to each side lug. Are they equal? yes...The pot is a linear pot....No, the pot is a log taper.
3. Place the ohmmeter center ear to any outside ear. Slowly move the arm over the range (turn the knob)... Notice any slight interruptions? yes... clean the pot internally.... Exercise it... Repeat until OK or toss it.

Richard Hull

[Alan Sailer]

Mark,

I did notice that the output gets more ripple at higher voltages. Does that cause plasma issues?

Cheers.

[Cristiano_Machado]

Hello Alan,

I am sorry about the failure. I use a 30kv precipitator supply without problems. I put all the precipitator under oil and to adjust the voltage I use a small screwdriver. I didnt change the trimpot because of the problems reported before.

Regards,

Cristiano

[Mark Rowley]

Alan,
Not ripple but fluctuations with the output that are most likely caused by small vacuum inconsistencies (virtually leaks, outgassing of tungsten, etc). In my case, added capacitance to the output will be closer to how a stiffening capacitor works in a mobile audio amplifier setting. It won’t be specifically for filtering ripple although it most likely will assist with that as well.

Mark Rowley

[Alan Sailer]

Christiano,

I wish I had known about the problem before I made the deadly change.

I made some progress with the supply last night. I had changed the obviously damaged parts with no effect.
After a half hour checking parts with an ohmmeter I found two additional parts, both zener diodes, that had failed.
The supply is at least alive now. I'll test it more extensively later. I'm trying to decide if I want to re-attach the
current limit circuit

Mark,

Had never heard the turn stiffener capacitor. It seems that it means adding a larger capacitor to the output to allow for
variable loads. Kind of a scary addition. Planning on a bleeder across the cap?

Cheers.

[Mark Rowley]

Of course....basic 101

Mark Rowley

[Mark Rowley]

I'd reattach the current limiter until you're familiar with it's operation during neutron production. Operating a precipitator psu is somewhat different from a iron-core transformer so it's best to become familiar with its differences before attempting any modifications.

Fwiw, I've never altered the limiter on the 30kV unit so you're firmly in uncharted territory with that one.

Lot's of posts about the precipitator including the potentiometer issue. Search "precipitator" and you'll get all the old posts. Rex put together a stellar report on how they work including schematics. Well worth reading.

Mark Rowley

[Alan Sailer]

Richard,

The question you ask are good questions. I also have never bought a new wire-wound ten turn pot.
Not that it makes any difference, but I have used them in around twenty projects without any problem.

It is also a 10K pot just like the OEM unit. The pot was applied with teh correct polarity. Checked twice before soldering.

As far as linearity or skips that is not an issue in this case. The pot was turned to minimum resistance before
I turned on the supply. There was no adjustment made, the supply simply arced and died. Both output transistors dead.
Traces leading from the transistors vaporized.

If you could turn on the unit slowly with a Variac I'd try more experiments. But it's designed to boom! on.

Mark,

Good on the bleeder.

As part of my cynicism about Chinese electrical design I will point out that the 220uF input capacitor has no bleeder*.
Guess China has not reached basic EE 101 yet...

I'm not going to stay on my soap box long but I can name a half dozen quality stories about Chinese high voltage design.
The most painful was the death of a trusty HP power supply to a 5 dollar piece of Chinese crap. It was supposed to input
12 volts and turn it into a few thousand. It also put some of that few thousand back on the 12 volt line and popped my supply.

Cheers.

* It does now.

[Richard Hull]

Sorry about the rebound into your 12 volt supply. I am leary of most any switcher by the chi-coms. High voltage ones are no real gems, as a rule. Switchers are all about a snap-on of the drive voltage. Variacs are not on the most favored list for start up. These supplies work off of inrush current to get moving, as a rule. Good switchers are designed to take it and not pass dangerous over-volting to sensitive parts of the switcher circuitry.

Richard Hull