ZVS Driver not working when fixed with 50N321 IGBTs / High Voltage / Forums

Forums

4hv.org :: Forums :: High Voltage

« Previous topic | Next topic »

ZVS Driver not working when fixed with 50N321 IGBTs

first previous 2 3 4 5 next

Move Thread

LAN_403

Kiwihvguy

Thu Sept 29 2011, 10:01AM

Registered Member #3395 Joined: Thu Nov 04 2010, 08:42AM
Location: Christchurch, New Zealand
Posts: 193

By the way, how do people get more stars (left side under pic or name)?

astralhighway

Thu Sept 29 2011, 11:50AM

Registered Member #4107 Joined: Sun Sept 25 2011, 07:30PM
Location: London
Posts: 53

Hi Jozef, when I say 'ground', I mean that you need a conductive surface that is relatively large compared to the rest of your circuit (imagine an unetched printed circuit board), to which you return all relevant connections. Your lash-up doesn't have a ground at all. It's hard to grasp this concept at first - it doesn't mean the same as something at earth potential. But you do need it.

I'll try to answer your other questions later this afternoon. But in outline, the stray inductance in the leads to the gate can be bad because it can induce a high back-emf that is conducted through the MOSFET's gate, causing the device to break down. Gate leads are always very short in commercial circuits, with the resistor directly soldered to the device via as short a lead as possible. Also, imagine the powerful electro-magnetic fields generated by the choke and by the transformer. These could easily induce a signal directly into the long leads, which you don't want.

Meanwhile, I think your ultra-fast diode is nowhere near juicy enough. I haven't worked it out exactly but I'll bet it handles peak currents way in excess of the rating of the component that you've got there.

Don't be demoralised. Some crucial design adjustments will reward you in the long term with a circuit that is reliable and much less prone to component breakdown .

astralhighway

Thu Sept 29 2011, 11:55AM

Registered Member #4107 Joined: Sun Sept 25 2011, 07:30PM
Location: London
Posts: 53

[quote]
I've never completely understood the role of ground in a circuit. Is it a safety thing? Is it there to act as a reservoir of electrons? Is it a consistent reference voltage in a circuit?

What does its large area actually do? Does it act as one half of a capacitor with the rest of the circuit? Is it for shielding? Or is it simply so that it has a very low impedance?
[/quote1317297086]

Hello, yes, it is a consistent reference voltage. I can't better the Wikipedia entry, which is under a creative commons license, so permissible to replicate here:

'An electrical ground system should have an appropriate current-carrying capability in order to serve as an adequate zero-voltage reference level. In electronic circuit theory, a "ground" is usually idealized as an infinite source or sink for charge, which can absorb an unlimited amount of current without changing its potential. Where a real ground connection has a significant resistance, the approximation of zero potential is no longer valid. Stray voltages or earth potential rise effects will occur, which may create noise in signals or if large enough will produce an electric shock hazard.

The use of the term ground (or earth) is so common in electrical and electronics applications that circuits in portable electronic devices such as cell phones and media players as well as circuits in vehicles such as ships, aircraft, and spacecraft may be spoken of as having a "ground" connection without any actual connection to the Earth. This is usually a large conductor attached to one side of the power supply (such as the "ground plane" on a printed circuit board) which serves as the common return path for current from many different components in the circuit.'

Hope that helps? My bold for emphasis...

astralhighway

Thu Sept 29 2011, 01:52PM

Registered Member #4107 Joined: Sun Sept 25 2011, 07:30PM
Location: London
Posts: 53

Jozef, you may be overcooking your FETS, because as you say, your supply voltage is 48volts. Let's say your operating frequency is 30Kz. Then your time period is 33us. And let's say your choke is 47uH. Then peak current = voltage times time period, divided by inductance.

So Ipk=48*33E-6/47E-6=33Amps. This is the maximum rating of your MOSFETS. Looks like you could use MOSFETs with a higher current rating...

astralhighway

Thu Sept 29 2011, 04:52PM

Registered Member #4107 Joined: Sun Sept 25 2011, 07:30PM
Location: London
Posts: 53

Hi again Jozef, what is the current rating of your toroidal choke? All of the current of your circuit passes through this component. Are you able to check? If it is under-rated ithe core will become completely saturated at a certain point each switching cycle and it will cease to behave as a choke.

Kiwihvguy

Thu Sept 29 2011, 11:03PM

Registered Member #3395 Joined: Thu Nov 04 2010, 08:42AM
Location: Christchurch, New Zealand
Posts: 193

astralhighway wrote ...

Jozef, you may be overcooking your FETS, because as you say, your supply voltage is 48volts. Let's say your operating frequency is 30Kz. Then your time period is 33us. And let's say your choke is 47uH. Then peak current = voltage times time period, divided by inductance.

So Ipk=48*33E-6/47E-6=33Amps. This is the maximum rating of your MOSFETS. Looks like you could use MOSFETs with a higher current rating...

astralhighway wrote ...

Hi Jozef, when I say 'ground', I mean that you need a conductive surface that is relatively large compared to the rest of your circuit (imagine an unetched printed circuit board), to which you return all relevant connections. Your lash-up doesn't have a ground at all. It's hard to grasp this concept at first - it doesn't mean the same as something at earth potential. But you do need it.

I'll try to answer your other questions later this afternoon. But in outline, the stray inductance in the leads to the gate can be bad because it can induce a high back-emf that is conducted through the MOSFET's gate, causing the device to break down. Gate leads are always very short in commercial circuits, with the resistor directly soldered to the device via as short a lead as possible. Also, imagine the powerful electro-magnetic fields generated by the choke and by the transformer. These could easily induce a signal directly into the long leads, which you don't want.

Meanwhile, I think your ultra-fast diode is nowhere near juicy enough. I haven't worked it out exactly but I'll bet it handles peak currents way in excess of the rating of the component that you've got there.

Don't be demoralised. Some crucial design adjustments will reward you in the long term with a circuit that is reliable and much less prone to component breakdown .

astralhighway wrote ...

Hi again Jozef, what is the current rating of your toroidal choke? All of the current of your circuit passes through this component. Are you able to check? If it is under-rated ithe core will become completely saturated at a certain point each switching cycle and it will cease to behave as a choke.

I am planning to use some IGBTs with a higher current and voltage rating so that's why I'm buying some soon.

My inductor is rated at 5a BUT the input is 6Ah. It barely gets warm.

I don't understand your explanation of ground. To me, I do have one. It's where the zeners and 10k resistors connect and it's where the black negative wire is connected to.

Kiwihvguy

Fri Sept 30 2011, 03:49AM

Registered Member #3395 Joined: Thu Nov 04 2010, 08:42AM
Location: Christchurch, New Zealand
Posts: 193

I just fixed the ZVS with some new 18v zeners. They were the convict, everything else is fine.

Here are some pics:

astralhighway, this is the best I could do to minimize stray inductance. Hope it's alright by your terms:-)

Kiwihvguy

Fri Sept 30 2011, 05:04AM

Registered Member #3395 Joined: Thu Nov 04 2010, 08:42AM
Location: Christchurch, New Zealand
Posts: 193

Here the ZVS Drivers setup. in the pics.

NOTES: added 2x 10w 100r resistors to the primary (see here:

)

48v SLA bank

Flyback cold is grounded to eliminate floating potentials.

Batteries aren't connected up to the ZVS yet.

Kiwihvguy

Fri Sept 30 2011, 05:05AM

Registered Member #3395 Joined: Thu Nov 04 2010, 08:42AM
Location: Christchurch, New Zealand
Posts: 193

Tell me what you think.

astralhighway

Fri Sept 30 2011, 07:54AM

Registered Member #4107 Joined: Sun Sept 25 2011, 07:30PM
Location: London
Posts: 53

Jozef, well done for shortening those lead lengths. It's also a smart precaution to have your circuit in the platic container there - good call. I can't really add to that explanation of a ground. Which bit of it is causing you problems?

Translated to your circuit, it would mean that your battery -ve and all the other points at the same potential are returne to a large conductive surface, such as a circuit board. The other components would be in physical close proximity to this large surface nominally at 0v potential, without being electrically in contact with it.

Have you ever looked inside a vintage amplifier or radio with electron tubes (uk, valves)? The chassis is the ground here. If you can, get hold of a junk radio from pre-1965 and take a look. (Ebay for a few dollars if it doesn't work).

It is best practice to have a ground in all electronic circuits and apart from the reasons listed aobve in the Wikipeia article, it gives benefits such as screening other parts of the circuit from stray signals.

I think you do need that beefier choke. You are operating way above its saturation point. Good luck with refining your circuit and well done on your problem solving so far.

first previous 2 3 4 5 next

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob