Power Source Voltage Ripple Regulation

For Short Term Learning Discussions ONLY. This area is for CURSORY questions and connecting with other users ONLY. ALL technical contributions need to be made in the appropriate forums and NOT HERE. All posts are temporary and will be deleted within weeks or months. You should have already search the extensive FAQs in each of the forums before posting here as your question may already be answered.
The Super Nerds Team
Posts: 11
Joined: Sat Jul 14, 2018 8:13 pm
Real name: The Super Nerds Team

Power Source Voltage Ripple Regulation

Post by The Super Nerds Team » Mon Sep 10, 2018 7:34 pm

Greetings,

After getting to meet with and talk to Joshua Guertler - another Fusioneer-to-be - the rest of the Super Nerds Team and I became interested in using an electrostatic precipitator power supply as the main power supply for our small-scale fusor. Due to our relatively large budget, the team was able to afford a relatively expensive power supply that has been claimed to be able to supply 600 Watts and -60 kV DC at the pressures which the fusor operates at, whether or not this is true will be tested. I have included a link to the product for those who are interested to see the eBay product that we are referencing: https://www.ebay.com/itm/High-Voltage-E ... 2405889314.

With the design, there have been a couple of issues that we want to make sure that we solve before getting the supply to power the fusor. First, we decided to remove the four flyback transformers in the power supply and submerse them in mineral oil to act as De Facto transformer oil. Each transformer and its connection will be given silicone insulation to minimize any damage on exposed electronics from mineral oil. While Joshua Guertler told me that this approach should work, I would like to confirm with the fusor community that this approach for a flyback transformer would be safe.

The other potential issue that the team has - and the cause for positing this question - is that people on this forum in the past have noted that the output of these sorts of electrostatic precipitator power supplies produce a relatively ripple-heavy DC output, as the power source relies on flyback transformers and following diodes for rectification. This has been explained in one of Finn Hammer's write-ups about his assessment of a 200 watt and 30 kV DC precipitator power supply seen here: viewtopic.php?f=11&t=11992&p=78210&hili ... ese#p77912.

To combat the problem, I have been considering the apparatus set up shown as an attachment to this post. Since I do not know the frequency or the peak voltage of the ripples from the team's power supply I did not assign a capacitance to the capacitor. The apparatus shows an input voltage of -40 kV DC, as I am not certain that the device can supply the advertised -60 kV DC safely (just a suspicion from a long-time buyer of Chinese electronics). Having experimented around with similar power supplies from the same seller, I have seen that you can exceed the 10 mA output when operating below -60 kV as long as you turn off the current control mechanism which will effectively act as an adjustable ballast resistor to prevent current cascades when you turn up the power. Hopefully, this should put me at around -40 kV DC and 10 mA with ripples. Again, I am not certain about the frequency of the ripples, but I would assume that they are proportional to the ones seen in Finn Hammer's experiment once the fusor's plasma stabilizes. If fusor would work well regardless of the voltage ripples, please make sure to address this to my team and I, as we have generally seen from the forum that such ripples must be dealt with.

On a similar note, my team and I have seen quite a few remarks from others that such a power supply might not be able to function under the conditions provided by a fusor's plasma due to the strong swings in a fusor's internal resistance. Thus, we were curious if there was a relatively inexpensive way that we could fix this problem with out power supply by making certain that a constant and stable current output could be provided. Please keep in mind that the current-moderating capability of the power supply can be turned off completely and gradually decreased, so there is no concern of the power supply shutting off upon establishing plasma.

Also, we were curious if there was a way for us to inexpensively establish an environment where the establishment of plasma would not cause such a strain on the power supply's current output. This could be something like incorporating an accelerator-type design or ion gun to prevent a massive strain on the power supply.

Thank you.

Sincerely, Zach from the Super Nerds Team
Attachments
4bc.jpg

User avatar
Dennis P Brown
Posts: 1521
Joined: Sun May 20, 2012 2:46 pm
Real name: Dennis P Brown
Location: Glen Arm, MD

Re: Power Source Voltage Ripple Regulation

Post by Dennis P Brown » Tue Sep 11, 2018 10:09 am

Someone did, if I recall, not too long ago try one of these exact units and they haven't posted since. There may be issues that cause problems - placing the unit under oil can't hurt but you might consider placing the unit under vacuum first, then adding oil so there is no trapped air within the cores. I plan on doing that with my new 60 kV xformer when I have time later this winter. That shouldn't be very difficult to setup and do.

I am suspicious of transformer systems that are placed in parallel for extra current and I seriously wonder if they can maintain the load in any real balanced manner - an expert EE once warned me that in most resistive circuits that are parallel, 99% of the current goes through the device with the very slightest lower resistance and the others carry little - if they are using a phase system, maybe it will work but again, requires a good design - maybe it is and maybe not.

If one has the funds, why not just obtain a proper power supply or at least a proper transformer and build a supply?

If you already have it, I guess it can't hurt to try but a good ballast resistor would appear essential - that is the only "inexpensive" method I know that will help control the swings within the fusor. Not sure that issue can be overcome if the unit shuts off during swings - that just makes matters more difficult, from what I've experienced dealing with swings. Having a power supply that can handle the current run-aways allowed me to establish a steady plasma by careful deuterium, and voltage control during the swings. I quickly and easily established a good, steady plasma. If my unit was cutting out, not sure that would have been so easy.

Well, hope you succeed because that would be a very useful development for fusors and people wanting to try their hand at this - since the cost of a power supply is, without a doubt, one major drawback.

The Super Nerds Team
Posts: 11
Joined: Sat Jul 14, 2018 8:13 pm
Real name: The Super Nerds Team

Re: Power Source Voltage Ripple Regulation

Post by The Super Nerds Team » Wed Sep 12, 2018 2:00 am

As a quick question to the general community, would this power source even need a capacitor set or device of the sort to smooth down the current ripples, or would a ballast resistor followed by a HV diode do the trick?

Chris Giles
Posts: 16
Joined: Mon May 14, 2018 10:46 pm
Real name: Chris Giles

Re: Power Source Voltage Ripple Regulation

Post by Chris Giles » Sun Sep 23, 2018 8:28 pm

I am suspicious of transformer systems that are placed in parallel for extra current and I seriously wonder if they can maintain the load in any real balanced manner
Absolutely correct, however, I think the Chinese unit referred to has 4 in series. The problem it has is that if you ground one end, then the other end has a potential difference from primary to secondary seemingly greater than the transformers are rated for. The 60kV model PSU could be wired as 30kV using a full-wave rectifier, grounding the connection between transformers 2, 3. That would give 30kV @ 20mA. Perhaps a good solution?
Someone did, if I recall, not too long ago try one of these exact units and they haven't posted since.
Yes, that was in this thread . It looks like the same brand.

So, PSUs suggested could have the following types of output:
  1. Half-wave AC. This is what you get if you ground one side a transformer and put a diode on the other. I see this often suggested, as above, and for NSTs. It gives much heavier ripple as one half of every AC cycle is unused.
  2. Full-wave AC. This is what you would usually find in any linear PSU. This is either a centre-tapped transformer and 2 diodes or 4 diodes in a bridge-rectifier.
  3. DC. From a full wave rectifier plus capacitor(s) and perhaps further regulation circuitry depending on the quality of the PSU. Some of the professional HV supplies units can give a very accurate DC output.
1 & 2 could be either line frequency (50-60Hz) for a linear PSU or high-frequency (30-60kHz) in an SMPS. Smoothing HF needs a much smaller capacitance (good news) but the higher frequency also puts the cap under a lot of stress and requires careful selection of a type graded to handle the required frequency (bad news). Standard ceramic caps for instance, would not to have the rated capacitance at these frequencies. This is even before adding the requirement of them operating at extremely high voltage. We may also need be careful with HF as the Fusor may have quite a reactive impedance at these frequencies? Anyone know?

I wonder how the Fusor behaves electrically as a load and what impact it has on the fusion process if the cathode voltage is not stable (#1 and #2)? Does it matter? I expect that if people have success with 50Hz half-wave (NST) then it can't matter much.

There also may be a disadvantage of using a proper lab grade DC PSU as these have over-current protection. Does anyone experienced have a view on the issues above and whether a professional regulated DC supply which has a current rating of say 20-30mA is good in practice or would keep cutting out?

Thanks

User avatar
Rich Feldman
Posts: 993
Joined: Mon Dec 21, 2009 11:59 pm
Real name: Rich Feldman
Location: Santa Clara County, CA, USA

Re: Power Source Voltage Ripple Regulation

Post by Rich Feldman » Sun Sep 23, 2018 10:16 pm

What's wrong with voltage ripple?
I would think that for the same average voltage, the more ripple the better.
Based on people saying that increased voltage brings a more-than-proportional increase in fusion rate, and applying that instant by instant.

Look up Greg Courville's (sp?) fusion reports here. He observed the HV waveform while his fusor was running. Said the voltage ripple was much less than expected. Not because of capacitance, but because (we think) the plasma was behaving like a voltage regulator tube.
Richard Feldman

Chris Giles
Posts: 16
Joined: Mon May 14, 2018 10:46 pm
Real name: Chris Giles

Re: Power Source Voltage Ripple Regulation

Post by Chris Giles » Mon Sep 24, 2018 9:00 pm

Thanks Rich.
Rich Feldman wrote:
Sun Sep 23, 2018 10:16 pm
What's wrong with voltage ripple?
Maybe it doesn't matter for the Fusor - I was wondering if anyone had any observations as there is a mixture of people using NSTs and professional DC supplies.
Rich Feldman wrote:
Sun Sep 23, 2018 10:16 pm
I would think that for the same average voltage, the more ripple the better.
not so sure about that. One important consideration (perhaps the only one?) is that if a half-wave rectifier is used then half of the power from the transformer is not being used, so for instance the above 600W PSU can only supply 300W max with a half-wave rectifier.

I do expect that using unsmoothed high-frequency output from an SMPS (these Chinese PSUs) will bring its own problems though. For example the inductance and capacitance of the cables may become important. The whole system could resonate, the amount of EMI produced would be incredible... Personally i would suggest a full wave rectifier to get the full output power and a smoothing cap just to avoid any RF effects.

I think the professional PSUs tend to use diode charge pumps which naturally rectify the AC in process of multiplying it.
From Spellman HV Manual:
Typically, high voltage is created by controlling an inverter that feeds a step up transformer which is connected to a voltage multiplier circuit
Best regards
Chris

User avatar
Rich Feldman
Posts: 993
Joined: Mon Dec 21, 2009 11:59 pm
Real name: Rich Feldman
Location: Santa Clara County, CA, USA

Re: Power Source Voltage Ripple Regulation

Post by Rich Feldman » Mon Sep 24, 2018 10:15 pm

Rich Feldman wrote:
Sun Sep 23, 2018 10:16 pm
I would think that for the same average voltage, the more ripple the better.
I'll stand by that statement, and clarify it with a picture.
fwhw.JPG
But you're right about half wave rectification (single diode) being avoided when the power demand is non-trivial.
Other things being equal, it's also hard on the transformer.
For same average current,the RMS current in winding is 41% higher, so heating from copper loss is double. Same kind of arithmetic that, I bet, makes its average fusion rate higher. :-)
And the average current flows as DC in the secondary winding, unproductively magnetizing it in one direction.

You _will_ find single diode recifiers in professionally designed flyback converters. Those circuits use coupled inductors, with parameters that make them a poor choice for real transformer circuits. And flyback circuits are generally not chosen for high power applications, for other reasons.
Richard Feldman

Chris Giles
Posts: 16
Joined: Mon May 14, 2018 10:46 pm
Real name: Chris Giles

Re: Power Source Voltage Ripple Regulation

Post by Chris Giles » Mon Sep 24, 2018 11:09 pm

Thanks Rich
Rich Feldman wrote:
Mon Sep 24, 2018 10:15 pm
I'll stand by that statement, and clarify it with a picture.fwhw.JPG
Yes, if i read it correctly, the pic shows equivalent power full-wave and half-wave PSUs; the half-wave one has double the peak voltage.

However, if we compared the same AC input (transformer) with half-wave vs full-wave or bridge rectifier then you would have the same peak voltage but half the average current due to missing half the waves - like removing 2 diodes from a bridge; it wouldn't increase the peak voltage just halve the power. Constructors need to bear in mind that the design power may be half what they thought in this case. Further info here .

Personally I'm waiting around for a nice lab PSU and if that doesn't magically appear by the time i need it then I think converting one of these "60kV" Chinese units with a full-wave rectifier may be a cheap workable option. Also waiting to see if anyone else does so and survives to describe it!

Best regards
Chris

User avatar
Rich Feldman
Posts: 993
Joined: Mon Dec 21, 2009 11:59 pm
Real name: Rich Feldman
Location: Santa Clara County, CA, USA

Re: Power Source Voltage Ripple Regulation

Post by Rich Feldman » Tue Sep 25, 2018 2:33 am

Obviously you're in the club, Chris. And you posted a nice reference for newbies.
Let's just hope its home document teaches the conventional direction of electric current. Some noobs, and even teachers, have trouble letting go of negative electrons drifting counter to the arrows in schematics. Electrical circuit theory, sign conventions, and units of measurement were well established before the discovery of electrons, and electron theory of conduction in metals, around the end of the 19th century. Forget about the electrons, except when the internal workings of rectifiers, active devices, or plasmas matter. Telephones, electric lights, and three-phase AC power were industrialized without knowledge of electrons. [\soapbox]

My chart does _not_ show equivalent power DC, full wave, and half wave rectified waveforms.
Its title is sloppy and ambiguous, but the three waveforms have the same average voltage.
Trying to shed light on why voltage ripple might be not a bad thing in fusors.

When scaled that way, and if we keep it simple by using a resistive load, all three have the same average current.
But the average full wave _power_ is 23% higher than the DC case, and the average half wave power is higher by a factor of 2 on top of that.
fwhw2.JPG
In this sense, ripple voltage of _any_ shape and amplitude will increase the mean square voltage (thus power, and RMS voltage)
above that of DC component alone. MS(dc+ac) = MS(dc) + MS(ac) = dc^2 + MS(ac).

That would be bad if we were comparing copper losses or power-per-weight, and didn't get to use different transformer designs.

I bet it would be good in a fusor. It's anybody's guess about how good, with these grotesque examples of voltage ripple. Plasma does not behave like a pressure-controlled resistor, and I bet you'd need different pressures to operate at the same V and I point.
Richard Feldman

User avatar
Richard Hull
Site Admin
Posts: 10904
Joined: Fri Jun 15, 2001 1:44 pm
Real name: Richard Hull

Re: Power Source Voltage Ripple Regulation

Post by Richard Hull » Tue Sep 25, 2018 5:07 am

To a stiff, (low Z) power supply, the fusor is a purely resistive load at all current draws. It is a dynamic resistance, (variable), based on pressures and voltage applied. Reactions between the gas load and any reactive circuit components can generate, a lot of RF radio noise. RF wise, the fusor is a great RF noise source. This is why it is so important to use good grounding technique and shielding on all cables and instrumentation such as neutron detectors and amplifiers.

Richard Hull
Progress may have been a good thing once, but it just went on too long. - Yogi Berra
Fusion is the energy of the future....and it always will be
Retired now...Doing only what I want and not what I should...every day is a saturday.

Post Reply