Dear All, We are discharging a High Voltage 26 kV 40 nF Capacitor through a Spark Gap. The device is used for medical purposes. The discharge is causing transients (ringing) on the supply voltage and this affects all of our Circuitry. ( Resets on Microcontrollers, sometimes eploding components). Any Idea how we can reduce the voltage transients? Also is there any way to prevent the transient reaching our circuitry. We tried all kinds of optical isolation. Unfortunately, as the transient is on the supply line all circuits are affected. Please please please help. Thanks in advance.
There are perhaps, three main coupling paths. There is the magnetic field produced by the discharge current. There is the electric field change that takes place around the exposed voltage changes and there is a conducted path back through whatever supply charges the capacitor.
You can minimize the magnetic field magnitude by keeping as much of the current coaxial as possible (current going one way concentric with current going the other way. If this is not possible, a twisted pair of paths is almost as good. Avoid any unnecessary open area loops that the discharge current must encircle.
You minimize the electric field effects by enclosing as much of the high voltage system as possible with a shield that returns its capacitive current back to the source, the capacitor. If the case of the capacitor is grounded, the shield should also be grounded at the case.
The best way to keep the discharge pulse from heading back down the supply wiring depends on the supply details, so I will have to let that go for now.
So you developed an EMP machine and wonder why it eats electronics?
A lot will depend on the physical layout.
Exploding components sounds like the brunt of the problem is in the supply electronics - you don't give any information on how the 26 KV is supplied, AC or DC etc..?????????
In Tesla coils we deliberately discharge large HV caps into inductors to produce high frequency ringing in the inductor. To keep it out of the mains supply (where it will destroy other electronics) there's a pair of air core inductors and a "safety gap" just before the source.
When the Tesla coil is in tune - no problem most of the energy is transferred into the load (sparks and heat in the air and on the output terminal or corona). When out of tune, the standing waves are reflected back to the source and that's when the damage occurs, or the safety gap starts firing.
I'm guessing that you have the same system with no load - maximum energy reflected back to the source. A few chokes and filters and a safety gap should get it to the point where it won't eat anything you want to keep. You may still have EMI problems but less destructive ones.
In Tesla coils I'd use an air core choke in preference to ferrite - because the ferrite ones would arc across the windings (through Teflon insulation) but your problem may be different in that respect.
What would be nice to know is AC or DC? what is the source? what power level? How often does the gap fire? What is the goal? Diathermy, lasers, induction heating?
You may not be intending to build an RF transmitter, but anytime you discharge high voltage from a capacitor into wires you have an RF transmitter. You have to start thinking like an RF engineer and concentrate on getting the power to dissipate in the load and not back at the source.
----== Posted via Newsfeeds.Com - Unlimited-Unrestricted-Secure Usenet News==----
formatting link
The #1 Newsgroup Service in the World! 120,000+ Newsgroups
----= East and West-Coast Server Farms - Total Privacy via Encryption =----
The capacitors are charged through a HV Power Supply, its Lambda EMI's
102A. From looking to the schematics I see diode bridges on the TR's secondary, so its DC. An Inverter is present at the The primary. The AC Input is 220V AC. The Caps are discharged every 0.5 s. through a spark gap. All the HV section and the cable (which has to be 2 mts) are shielded. The application will be ESWL (Targeting Kidney stones.). Yes we do have EMI problems, but the shielding reduced them a lot, but the transients are really killing us. Any but any help will be appreciated...
The ideal situation for a run of 2 meters would probably be a shield over a twisted pair of wires carrying the HV and return out to the cap. The cap and spark gap should be close to the point of use (ideally) - that prevents that whole wire from becoming an EMI radiator. Is that practical?
Just behind the cap the twisted pair (shield and all) should wind around a ferrite core a turn or two.
If the cap has to be located at the machine - a ferrite or two at either end of the cable may help the EMI problem but may hurt the production of high energy shock waves.
The shield should be grounded at one point only - at the machine - otherwise insulated from everything else
The patient, tables, ultrasound wiring, etc. should be at ground (from a radiation point of view - safety is your problem)
Another consideration with DC is that there be no exposed wires with sharp edges that can float ions into the surrounding air. Any conductive surface that is insulated from ground can build up quite a static charge - and that includes people. Build up a charge and walk over to a terminal or ground and you may damage some sensitive electronic gear. If the whole thing is shielded that shouldn't be a concern.
So best idea is twisted pair with a shield and cap at the point of use and ferrite core just behind the supply cable as close to the cap and gap as possible. The twisted pair alone should make some difference.
What does the spark gap look like? Is it also shielded?
And for my own edification: How is the sound coupled into the patient? Plastic window? Is the sound focused?
Can something like a barium titanate (ceramic piezoelectric) type transducer serve the same purpose? (at lower voltage with a SCR switch to fire it?)
----== Posted via Newsfeeds.Com - Unlimited-Unrestricted-Secure Usenet News==----
formatting link
The #1 Newsgroup Service in the World! 120,000+ Newsgroups
----= East and West-Coast Server Farms - Total Privacy via Encryption =----
One path will be the switch that isolates the capacitor from the charge unit. There you have a parasitic capacitor even if it is switched off. only a few pF will be enough to transfer a lot of charge through it. Try to make
2 switches with an aditional cap vs 1nF between:cap 1nF between them like this:
The switches in the ground may help too. All your wires have inductance. And if the capacitor is empty the current try to find a way to flow... may be not only through your 40 nF cap. If you use f.ex. only a diode as a switch to charge, this switch will be wide open for the negative swing of each ringing.
Just for everybody's information. I measured (or rather estimated) dv/dt of the spark discharge. ~1us wide 10kV (measured by Tek P6015 probe) pulse is generated by small trigger transformer. The breakdown takes place somewhere around
5kV (half way to the top). The falling edge was picked up by Tek P6139A probe sitting ~1" away. TDS3054B was used to record the data. dv/dt was ~2-3kV/ns (yes, kilovolt per nanosecond!). This will send 2-3A through 1pF capacitor!! I have a sneaky suspicion that actual dv/dt is higher as my measurements were almost at the limit of the instrumentation I used.
Dear All, Thank you very much for all the help. Apart from shielding, I wonder if I can Isolate the system from the HV Supply. What if I place a relay at the input of the HV Supply (220v) and break the phase and neutral just miliseconds before the HV Capacitor discharge happens. Do you believe this can prevent the transient to reach the system?
I doubt it. It will probably just cause an arc across the open contacts. I think you need to draw a full schematic of your system, with all wiring inductance and capacitance to all surrounding conductors added as explicit coupling components, so you can begin to unravel where energy is going and how you might detour it around the sensitive parts. Stabs in the dark are very unlikely to arrive at an optimal solution.
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.