30kV isolation in 24V DC-DC converter 100W

Have you considered a motor generator?

A motor and the alternator from an automobile should be within your budget.

A insulating rod from the motor to the generator could provide the required isolation.

Carl,

Yes, I think that is feasible. However I am looking for smart solution using electronics and transformers such that is can fit into compact space of electronics.

I found this agency making somewhat similar to what I want. May be they can develop for me a custom solution. Only they are going to be a bit expensive.

http://www.absopulse.com/news_May%2008-DIO122-24.html

Mr. Shyam

10E+12 isolation will be very difficult to achieve.

PTFE is great insulation, however there are very few things that will stick to it, unless it is etched. Also if the bobbin is make from a block of Teflon, it may be expensive.

5 stages is an interesting approach. The problems may be that if transformers are used, there may be excessive leakage inductance, If each stage is rectified and then converted to AC, there may be an efficiency problem. There may also be the problem that if the output shorts to ground, (due to stray capacitance) the full 30kV may be across the last stage.

How much time do you have to get the isolated 24 VDC working. A motor generator could be working in a matter of weeks.

It will probably take months to build or have made a 10E+12 ohm transformer. Then probably several months to build the switching circuit.

Is there a problem with vibration. If there is then a motor generator may be a problem.

A switching power supply will put out signals that are very similar to corona. The power supply should be very well shielded.

This sounds like a fun project.

Good Luck

Carl

Carl Pugh,

Some people do use piezo electric vibrator to generate isolated power but that has little power. Other means are photoelectric power using LASER and photocell, which again limits power to about 2W. My requirement is about 60W/120W. It is a compact portable instrument and not enough place for motor or generator and also instruments operating time is unpredictable so batteries with limited life can not be used.

I thought about charging batteries and then switching its circuit to high isolation which is possible but batteries will run out in few hours and will need recharging and interrupt the experiment. Perhaps for lower power need under 4W I can try Lithium cells.

I still need this DC-DC converter and is a serious requirement. The reference I have given was for 10kV isolation and it is affordable to use their products but they are not rated for 30kV which is my actual need.

Instrument casing is 2U 19" standard rack module.

I have other things to fit in and looking for isolated DC-DC converter now. I have 2-3 months to complete this. This a part of the research and development project I am currently involved. It is not a commercial requirement.

Shyam: Is the power supply 24VDC input and output?

Bill ML,

Yes, the power supply we often use is 24V DC 60W and 120W ranges. This is derived through AC-DC converters with 6kV isolation from mains Ground. The same to be further isolated by 30000V between input and output. Output is same 24V with some loses in conversion.

I am thinking to use a high frequency oscillator coupled with magnetic core with lots of electrical isolation using PTFE wire and less turns and high current per turn in the range of 5A to 10A. For 100W, transformer core size 1000cc or 10cmX10cmx10cm with potting material should be OK and it can also house fast high current rectifier Schottky diodes, 1000uF/50V low ESR capacitor and 120W power voltage regulator.

Vacuum potting of high insulation is essential else nasty corona discharge will take place, which is very bad for my experiments. My circuits are designed to trip on 10uA corona discharge.

Any one having experience in TV high voltage transformer will be able to do this economically and also very effectively. I think Silicone potting material is often used in those transformers. They also have nearly 25kV isolation between primary and secondary.

Bill ML,

Yes. It is a DC-DC requirement to isolate 24V 60W/120W power supply by 30kV.This 24V will again generate isolated 6kV 60W DC-DC and this 6kV will be biased to another 20kV referenced to Ground. In all 26kV will get formed and this 20kV is intermediate reference point at which most of the control electronics will also operate. Hence, this point to be isolated from real ground by 20kV operating and to be safe side 30kV is taken as working point for isolation with no corona discharge to Ground. This entire electronics will be placed in an insulated box with finally metal shielding. All internal parts are also having their own shielding as they generate high frequency switching noise.

The digital signals are easy to isolate using short fibre optic cables or line of sight opto- couplers. Your timing is very demanding though. You might want to look at a serial bus instead of parallel signals. It might be cost effective if you have processors on both sides.

The power is harder to do. Your transformer will need thick isolation layers between the coils which will result in large leakage inductance. Thick wire isolation is a bad choice as it will make the coils too large. You want small coils separated by a thick insulation. Don't forget to isolate the core. A shield between the coils and core might reduce the leakage but will increase the parasitic capacitance.

To deal with the large leakage inductance, choose an inductive kick back topology to take advantage of the leakage inductance. A direct transformer topology will have poor regulation.

There are various capacitive coupling that could be tested but I don't have any experience with those.

I would say that this is something of a toughie, where the 30KV isolation and the given other parameters tends to be a real challenge.

Flyback concepts I think are generally limited to the milliampere range. Nevertheless, I'm not really sure the 30KV isolation could be achieved with the use of a transformer. Or perhaps even a MG.

Silicone potting otherwise is commonly used to prevent corona up to 25KV or more.

Maybe I'm missing something; and other can come up with ideas.

In so far as the TTL interface, I've used these in past, at 10MB/s:

http://www.avagotech.com/pages/en/optocouplers_plastic/common_mode_rejection_motor_drive_optical_isolator_optocoupler_optoisolator_photo_ic_photo_isolator_photocoupler/?gclid=CImHs_K3tZ0CFSBN5QodDXYriA

Bill ML,

For 30kV these ICs will not work so fiber optic link is considered for TTL signals.

Within the link I suggested are both fiber optic transmitters and receivers.

Mr. Shyam

Is it possible to arrange your circuit so that 10E+12 ohm is not required?

If you would post your circuit, someone may have some suggestions.

Good Luck

Carl

Hi,

I am quoting such an isolation power supply right now but it is not an off the shelf type of product. Though after the design is compled $500 each would the be price range.

For the TTL signals use fibre optic isolators, for off the shelf products see Newark Electronics distributor but use the glass fibre not plastic as it is more relaible.

Watch your level of corona leakage.

Dave K.

Dear Dave K,

Isolated power sources of voltage and current have a lot of application and hence your efforts will not go in waste if you cn make these.

For example 0-5A isolated power supplies are used for filament power. 24V DC are used to provide isolated power source in Trains. These are also used in Ion beam electronics.

Keep in touch.

Hi, The most challenging one was for a 180kV X-ray tube. The heater was +/- 0.1Vdc accuracy, 70watt, dynamic and no active circuits on the High voltage side. The high voltage side was only a rectifier and filter making the design very reliable.

I can be reached at david.k@1-LTL.com

An affiliated company link for capability is

http://www.stonepathpower.com/downloads/stonepath_design_capability.pdf

I do not include isolated heater supplies on this list.

Dave

Simply Great!

I'm about to need something similar - to get 24DVC floating on 30kV. Thread has been idle for some time - seems Shiam found a solution. Could you share what solution has worked best?

Would it work to do a transformer with thick isolation around sec winding? Usual HV potting epoxies has ~15..20kV/mm dielectric strength. So making say 3mm thick walls would be enough to isolate HV from primary and from core.

Has somebody tried that?

15kV isolation was easier than 20kV to 30kV ranges. Charge leakage is higher and chances of breakdown higher unless dielectric is good quality. PE dielectric is OK as long as it has no voids or cracks or air gaps. Just filling separation dielectric alone does not work well. Switching frequencies of 50kHz to 100kHz range are ideal.

we have 9kV isolated toroidal transformers commercial version and 15kV isolated is constructed locally on ETD59/N97 core and bobbin of EPCOS. We will manufacture 15kV isolation version soon for volume use. Our cost is about US$30000 for 100 units. and have procured all material for the first batch.

20kV and 30kV version require potting or oil filling due to dielectric breakdown possibilities in air. This development is slightly delayed and to be taken up in next stage.