Solution Finder Get Tech Support

Using HiperLCS in adjustable wide output voltage range converter.

Posted by: a_funtick on

Hello!

We need high efficiency high voltage output (600V 0.6A) converter. Now we're using FSFR2100 from Fairchild at 100 kHz with 500V 0.5A output at constant voltage.

In order to decrease dimensions and volume, I want to try to build the converter, based on HiperLCS LCS708HG IC runnig at high frequency (200-250 kHz). But, I need wide output voltage range at wide input voltage and wide load variations.

Basic design requirements:

VAC in – 198-242 V (280 – 341 VDC)

V out max – 600 V.

I out max – 0.6A

Load variations 10-100%

V out range - 200 - 600V

Is it possible to build such converter based on LCS708HG?

Thanks!

 

Comments

Submitted by PI-Tucker on 04/02/2015

If you were able to meet the design requirement with the FSFR part, it is likely you can do it with LCS708 at double the frequency.

1) What is the operating frequency of your existing design at the extreme combinations: 

a) minimum input voltage, maximum output voltage and current

b) maximum input voltage, minimum output voltage and current.

If you are having trouble with the above compbinations, will you consider adding a PFC stage?

We're using FSFR2100 at 100 kHz with 500V 0.5A output at constant voltage. Another words- in the old design we don't use adjusting of output voltage. Load Current may vary from 10 mA to 500 mA.

Now we would like to design new converter with adjustable output: Vout - 200-600 V and load current up to 600 mA.

1)      What is the operating frequency of your existing design at the extreme combinations?

a)      Operating frequency – 60 kHz at minimum input voltage 240VDC and maximum load current 550 mA at 500V output.

b)      At the maximum input voltage and minimum load current the FSFR2100 works in the burst mode.

2)      If you are having trouble with the above combinations, will you consider adding a PFC stage?

Yes, we considered using PFC in our design, but we should have max efficiency at minimum dimensions, and another conversion stage was undesirable. That’s the main reason why we decide to use HiperLCS LCS708HG in our new converter. This device will allow us to increase switching frequency and decrease transformer dimensions.

We will prefer to narrow output voltage range to 300-600 V instead of using PFC.

 In worst case, if we cannot reach required parameters without PFC, we will ready to use it.

From the information you shared with us, it is going to be a challenged design, although no direct show stopper per information you shared with us.The challenges could be thermal, core selection to meet your dimention specs as well as operation requirement, and no load specs if there is any.

Since your output is high, 200-600V, you might consider using votlage doubler output. Refer to our AN-55 page 33 for high voltaeg output configurations.

 

 

Submitted by a_funtick on 04/10/2015

Hello.

I'll plan to use ELP 32/6/20 or ELP 38/8/25 N87 cores with PCB windings. In the previous design, I successfully used ELP43/10/28 with good performance. The output rectifier stage is Full-Bridge, based on the Silicon Carbide Schottky Diodes.

It’s not so hard to calculate the LLC converter with exact output voltage, but, for me, it’s difficult to calculate LLC with variable input and output voltage! I don’t know what step should be first. Please, give me some suggestions, how I can calculate the wide output voltage range LLC converter using PIXLs Designer and using my initial parameters.

Thanks!

to use the PIXL for the design, start with following conditions:

1. start with lower freq, say 200kHz ( or below if possible) for your design. It gives wider operation range before hitting burst mode.

2. set burst mode 1 in PIXL so that it enters burst mode at as high a freq as possible. Ideally you don't want to hit burst mode untill you reach a very light load.

3. set Vres relatively high, which gives you wider operation range. This also mean that at highest Vo, power supply is running below resonance. For example, 350Vres is a good starting point.

Other comment: for high freq operation, PCB winding maybe lossy in an integrated transformer. It turns out to be a problem, suggest to use seperate resonating inductor.

 

Submitted by a_funtick on 04/15/2015

Thank you for suggestions.

Another questions:

1)What Vbulk_nom and VO1 should I set in my case for optimum design ?

2)Recommended Lpri & Kratio

Thank you.

Submitted by PI-Neela on 04/15/2015

In reply to by a_funtick

1) Vbulk_nom=250, Vmin 198*1.414*factor of minus ripple as min B+ voltages

2) Vo1=600V, it is the highest power and design has to deliver

3) Lpri&Kration: spreadsheet will give default values depending on your design. Use them to start.

Submitted by a_funtick on 04/16/2015

Thank you very much!

Submitted by a_funtick on 10/30/2015

Hi!

I built up converter based on suggestions posted above and in the topic "LCS708 does not properly" . Now I have good performance and stability. But I have questions about working frequency. As mentioned above, the converter should be runing below resonance at any output load. But, in the new posted design DER-447 "184 W Battery Charger Power Supply", the converter was designed to work above F res at all load conditions. My design has the same requirements - wide output load variations- 270V @ 10 - 700 mA output current. What will you recommend to me: to run converter above or below resonance?

Submitted by PI-Neela on 10/30/2015

Thanks for the comments. The recommended better design will be: design the transformer so that it runs slightly below resoncne at norminal B+ voltage, and full load.   

Submitted by a_funtick on 10/31/2015

Thanks! Another question: What was the main reason to run converter above resonance in the DER-447 "184 W Battery Charger Power Supply"? Why you didn't run it below resonance?

Submitted by PI-Neela on 11/02/2015

we design out transformer based on Vbulk_nom of 220V, which is actually the low end of the input Vdc (after doubler and rectifier). With this input, spread sheet can calculate the line sensing resitor divider values properly. However, it results in operation above resonance for most of conditions as you pointed out. The better way of design it is mentioned in the last post.