New Project - 200W, HV SMPS, need advice

[tartan5]

Hello,

I am an electrical engineer, embarking on my first adventure of designing a smps system. Seeing as there is so much info out there on so many different topologies, I was hoping that if I list my parameters I could get some feedback from the community on which topology they feel would be the most suitable.

If I leave any information out, please feel free to let me know, as at this point I am sure that I don't know what I don't know....

1) Use: This supply will be used to create a low pressure (100 mTorr to 900 mTorr) plasma
2) Input power: Primarily 120V, 60 Hz, with the possibility of 240V 50 Hz
3) Required output: -1000V (or so) DC at strike. Once plasma is initiated, the plasma current will be adjusted between 20 to 100 mA of current.
4) Low electrical noise: Unit is to be used in area with high sensitivity to radiated noise.

The low noise requirement has me looking at resonant type topologies with ZVS. If anyone has suggestions as to what would work best for these requirements, I would love to hear them!

Thanks!
Tartan5

 

[Kanwar]

Hello,

I am an electrical engineer, embarking on my first adventure of designing a smps system. Seeing as there is so much info out there on so many different topologies, I was hoping that if I list my parameters I could get some feedback from the community on which topology they feel would be the most suitable.

If I leave any information out, please feel free to let me know, as at this point I am sure that I don't know what I don't know....

1) Use: This supply will be used to create a low pressure (100 mTorr to 900 mTorr) plasma
2) Input power: Primarily 120V, 60 Hz, with the possibility of 240V 50 Hz
3) Required output: -1000V (or so) DC at strike. Once plasma is initiated, the plasma current will be adjusted between 20 to 100 mA of current.
4) Low electrical noise: Unit is to be used in area with high sensitivity to radiated noise.

The low noise requirement has me looking at resonant type topologies with ZVS. If anyone has suggestions as to what would work best for these requirements, I would love to hear them!

Thanks!
Tartan5

What you need is ZVS with fully resonant drive and for such a high voltage you need 3-4 transformers with primary paralleled and secondary in series for safe and reliable operation.

 

[tartan5]

What you need is ZVS with fully resonant drive and for such a high voltage you need 3-4 transformers with primary paralleled and secondary in series for safe and reliable operation.

Is it not possible to get a transformer with the secondary capable of handling 1000V? I am assuming this would be because of dielectric breakdown? Or is there some other reason to do it this way?

 

[wally7856]

Standard double insulated magnet wire does not even have a rated working voltage. 60V is probably the max you would ever push it to. I am talking about turn to turn voltage, as in turns next to each other. That is one of the reasons commercial transformers varnish there winding’s, with the little nicks during winding the 60V goes even lower.

For higher voltage outputs you have to figure out a winding scheme for a core you can actually buy that will not exceed maybe 40V per turn.

Some methods are sectional bobbins were the 1000V would be broken up to 2 or 3 sections were the turn to turn voltage does not exceed 40V or so.

Another method is to wind it like a television flyback transformer. A few turns wide and then another layer, making a tall skinny winding.

I think for you, were you only want 1 piece and you will have trouble buying transformer cores and bobbins, the best answer is to through some money at it. Buy some Teflon wire with a known working voltage and a working temperature rating of 250 Deg C or so. You want the thinnest Teflon that will give you the rating you need. The thicker the insulation the less room you will have for winding’s. And i think the leakage inductance goes up also. My guess is that with 100 or 200 volts turn to turn rating you can just wind this as a conventional transformer. But you will have to play with the numbers to be sure.

 

[tartan5]

Thanks for the responses........Eventually this will be a low volume commercial product. Does anyone have a list of sources for buying transformers and/or bobbins and cores?

 

[wally7856]

The major manufactures of transformer cores will have a distributor list on there web site. The distributer site will have a list of all transformer manufactures they represent.

Big manufactures are TDK, Ferroxcube, and Magnetics.

Typical minimum orders will be around $100.00 USD.

If you go with the resonant design you will have very few choices of cores and bobbins. I only remember seeing them in the ETD series. Only 3 or 4 sizes had bobbins for resonant design.

I did a quick search and found these resonant transformers with an ap note.

http://www.tdk.co.jp/tefe02/e636_srx.pdf

Do not be to surprised if you can not buy them. Many of the transformers listed in catalogs are not available, or only available in very small samples. If you try to buy any kind of quantity you are told 50 weeks or so.

Buy the way. It is the sectional bobbin that makes a transformer suitable for resonant use, any power ferrite transformer core will work, and there are very few sectional bobbins made.

 

[wally7856]

One other thing. A sectional bobbin for resonance transformer use is just a few millimeters of plastic to separate the primary and secondary winding’s to get a large value of leakage inductance. The manufacture of the bobbin just guesses were to place the divider for the most common designs. If you use an ETD core with a round middle leg you can machine up your own bobbins with just a lathe to fit your design. ETD cores are the easiest to buy, even Farnel and Newerk has them.

 

[tartan5]

Thanks for all the great info!
In your opinion, what topologies are easiest to get running initially, and are there topologies other than resonant that utilize ZVS? It seems like the resonant design might be a bit more difficult to get running correctly?

Thanks again for your input!

 

[wally7856]

“In your opinion, what topologies are easiest to get running initially”

The easiest may be a low voltage half bridge like 12vdc input. Then the control chip and transformer can run off of the same power supply.

“are there topologies other than resonant that utilize ZVS?”

As i was writing those messages i kept thinking i heard of other ZVS SMPS but i could not remember what.

“It seems like the resonant design might be a bit more difficult to get running correctly?”

They look a little complicated to me to. In this case though, because the primary is isolated from the secondary, i thought that it would be a plus for a higher voltage supply. Also, there are many ap notes out there on making a resonant supply and there are single chip solutions.

Look at.
FAN7631

And this one.
http://www.fairchildsemi.com/ds/FS/FSFR2100U.pdf

Full design ap note here.
http://www.fairchildsemi.com/an/AN/AN-4151.pdf

 

[Redwire]

This sound very similar to a gas-laser SMPS. I built one (although push-pull square wave) and used teflon stranded wire for the secondary and ran it into a voltage-tripler to get DC.

I would look at the Dual Resonant Solid State Tesla Coils as they are similar, you just need much lower secondary voltage.

 

[szhighstar]

Dear friend, if you need help about magnetic components, including transformer and choke and filte and SMD power inductor and bobbin and core, I maybe can help, thanks!

 

[wally7856]

Dear friend, if you need help about magnetic components, including transformer and choke and filte and SMD power inductor and bobbin and core, I maybe can help, thanks!

Go ahead and help. Pick the components for and show your calculations for the transformer, choke, filter and SMD power inductor.

 

[szhighstar]

To design transformer, need know some necessary information:
1.Topology circuit, forward or flyback or other circuit.
2.Work frequency
3.Input voltage and current/ouput voltage and current
4.Diemsion requested
5.Safety requirement.

 

[szhighstar]

PFC inductor design:

1, selection core:
We should be know below necessary information:
(a) Minimum Inductance based on DC basic, e.g. 1mH
(b) IDC current, e.g. 3A
According to formula LI*I= 1.0*3.0*3.0=9.0
We can find 60u which is in range from curve which is LI*I vs core part number, and find core
55586 is suitable .
Inductance factor of 55586 is 38uH/N*N+/-8%, so lower limit is 34.96uH/N*N
For meeting 1.0mH, it needs to wind 169Ts, according to formula IN=HL, get H is 56.6A/cm, we
continue to review curve which is DC vs initial permeability, can find initial permeability at DC
current is 68% than initial permeability at 0A, so we adjust turns from 169Ts to 249Ts.

2. wire selection
We select AWG#20 when current is 3.0A, due to big current and high temperature raise, we use
PEW or EIW wire, which temperature rate is 180 degree and reinfore insulation, ensure not short
between wire and wire.

3. Verify design result
249Ts AWG#20 is equivalent 157.9mm*mm, winding factor is about 39%, total area of window
is 401mm*mm.
Minimum inductance at 3.0DCA is 1.04mH.

4.Configuration design
Because PFC inductor is heavy and coil is big, we advise to add base, and full fix coil and base
using epoxy glue, design pin pitch according to PCB of customer.

5.Reliability testing
We run high temperature and low temperature testing, ensure the inductor can bear high and low
temperature.

6. Electrical parameter testing
L at oA
L at DC current
DCR for coil
Impulse voltage testing for winding.

 

[szhighstar]

For Cmmon mode choke and difference mode choke, need know below information:
1.Input voltage and current
2.Inductance
3.Dimension requested

 

[szhighstar]

Fror SMD power inductor design, need know below basic information:
1.DC current, (Isat and Irms)
2.Inductance
3.Diemsion requested
4.DCR

 

[tartan5]

Wally, some great info there.....I'm going to study the app note for a bit....

Redwire, looking at the DRSSTC design,while the results are impressive, it seems that it is geared toward much higher power than I would be looking for, and I'm not sure how clean the output voltage would be? However, the push-pull circuit sounds interesting, does that hard switch?

 

[tartan5]

Hi szhighstar,

I don't have information about what the topology is yet, that is what I am working on determining. The basic requirements are as listed in my first post....

 

[szhighstar]

Hi tartan5:

1) Use: This supply will be used to create a low pressure (100 mTorr to 900 mTorr) plasma
2) Input power: Primarily 120V, 60 Hz, with the possibility of 240V 50 Hz
3) Required output: -1000V (or so) DC at strike. Once plasma is initiated, the plasma current will be adjusted between 20 to 100 mA of current.
4) Low electrical noise: Unit is to be used in area with high sensitivity to radiated noise.

According to above information, my comment is below:
1.If work frequency is only 60Hz or 50Hz,sould be select lamination configuration transformer.
2.Input is 120V and 240V, there are two input windings in the transformer, with center tap.
3.Output is 1000V, it is step up transformer, note short between output winding, bacause there are many turns in secondary winding and high voltage output.
Please comment,thanks!

 

[tartan5]

szhighstar, I don't think that would work, as I need to be able to regulate the output transformer, thus need to regulate the input power. My existing system utilizes a step-up transformer with a variac, but I am looking to be able to digitally control the current. Yes, I could use a motor on the variac, but I'd rather not....