SEPIC converter MOD

[CapeCAD]

2 Months, 2 PCB spins, 100s of simulations, and many hours of testing later I was able to put together this converter with the help of my toaster oven.

Output voltage can be greater than, equal to, or less than input voltage.

Vin: 4.5-18V
Vout: 1.24-6V

This mod is wired with batteries in series (5.5-8.4V) and will operate across the full range with no dropout.

 

[Emris]

So you hae a Booster and a regulator! Nice!!!

 

[CapeCAD]

Additional information and questions:
All switching power supplies in Boost mode require higher current in vs out, Buck mode uses lower current in vs out.

I am seeing 5Amp at 5V input for 5.5V/3.3Amp output. For this reason I am considering either reducing current limit or adding UVLO (Under Voltage Lock Out) set to 5.5V minimum input to prevent excessive current drain off a single battery. The converter continues to operate down to 2.5V input on the bench (Both protected batteries are completely drained in my mod). I am worried this could be a problem with unprotected batteries, or use with a single/dead battery.

I would appreciate any comments/suggestions on which way to go with this. Reducing current limit would allow me to increase the maximum output voltage or use cheaper components. I use a 2.8 Ohm atty and have designed this so I can go down to 1.5 Ohm at 6V. Is there any advantage to it? Seems like it may be too much since I'm only running my 2.8 Ohm at 4.5 Volts.

• Current limit currently set to 4A Max Output current (1.5Ohm load @6V).
• ~82%-84% efficiency. New components are on order to increase this a few percent.
• .1 pin spacing was used to simplify my modding, I used the same pinout as Industry standard Buck converters (Evercool and Okami).
• Converter is on by default, enable pin to Ground shuts it off. Next revision will be off by default and enable to ground will turn it on.
• Size: .5" X .825" x .5"

Once I get this circuit nailed down, I already have a 8.4V charger circuit (USB or 12V) completed that I'll be combining to a single PC Board so I can complete my mechanical packaging design with a juice feeder.

View attachment Efficiency.pdf
View attachment 5Vin5.5Vout.pdf

 

[SurvivorMcGyver]

Additional information and questions:
All switching power supplies in Boost mode require higher current in vs out, Buck mode uses lower current in vs out.

I am seeing 5Amp at 5V input for 5.5V/3.3Amp output. For this reason I am considering either reducing current limit or adding UVLO (Under Voltage Lock Out) set to 5.5V minimum input to prevent excessive current drain off a single battery. The converter continues to operate down to 2.5V input on the bench (Both protected batteries are completely drained in my mod). I am worried this could be a problem with unprotected batteries, or use with a single/dead battery.

I would appreciate any comments/suggestions on which way to go with this. Reducing current limit would allow me to increase the maximum output voltage or use cheaper components. I use a 2.8 Ohm atty and have designed this so I can go down to 1.5 Ohm at 6V. Is there any advantage to it? Seems like it may be too much since I'm only running my 2.8 Ohm at 4.5 Volts.

• Current limit currently set to 4A Max Output current (1.5Ohm load @6V).
• ~82%-84% efficiency. New components are on order to increase this a few percent.
• .1 pin spacing was used to simplify my modding, I used the same pinout as Industry standard Buck converters (Evercool and Okami).
• Converter is on by default, enable pin to Ground shuts it off. Next revision will be off by default and enable to ground will turn it on.
• Size: .5" X .825" x .5"

Once I get this circuit nailed down, I already have a 8.4V charger circuit (USB or 12V) completed that I'll be combining to a single PC Board so I can complete my mechanical packaging design with a juice feeder.

View attachment 25814
View attachment 25813

Great work CapeCad...

The UVLO is a smart way to go -- I would bump that up to 5.6 or even a tad more for safety.

Personally, I would reduce the current limit out to around 3 as the give more than enough to satify even the Most hardcore LR atty user -- especially if it will allow a boost in Voltage range. 1.5 ohms at 6V ?? 23 WATTS ? way overkill in my book. Plus The higher voltages would then allow for the use of the HV atties currently available. With the lower amp draw of these HV atties you'd be hard pressed to even pull a 2 AMP load -- I rairly go above 1.5 A on the output -- Much less taxing on the batteries as well

I hope your charger is monitoring both batts. Some say the stacked charger method for protected batteries will present a few problems - especially when one cell reaches cut off before the other. Perhaps including a failsafe batteriy protection circuit on the board might be a good approach -with that - use unprotected batteries -- the Tekkeon does it quite nicely.

Mind if I ask what chip your using? (PM on this issue is fine). I've been playing in spice with well over a dozen Sepic's over the past year and simply can't find one to do my ditty without problems. (my ditty = operate accross all V ranges and all atty configurations known to be vapable) . The closest I've come to success is with the LTC3872. It's current mode and not sepic. Seems there may be some promise with inverting controllers but I haven't gone there yet.

Anyways -- Good luck and keep up the good work!

Happy Holidays everyone!

 

[SurvivorMcGyver]

PS. Mind if I ask what ripple you get with the 1.5 Ohm Load at say...... 4.5Vo?

 

[CapeCAD]

Great work CapeCad...

The UVLO is a smart way to go -- I would bump that up to 5.6 or even a tad more for safety.

Personally, I would reduce the current limit out to around 3 as the give more than enough to satify even the Most hardcore LR atty user -- especially if it will allow a boost in Voltage range. 1.5 ohms at 6V ?? 23 WATTS ? way overkill in my book. Plus The higher voltages would then allow for the use of the HV atties currently available. With the lower amp draw of these HV atties you'd be hard pressed to even pull a 2 AMP load -- I rairly go above 1.5 A on the output -- Much less taxing on the batteries as well

Thanks, great feedback. This is exactly the type of information I need.

I'm running at ~8 watts myself, I didn't think > 20 Watts would be needed. I've burned out a few 2.8 Ohm atomizers at 5.5V (Pot changed accidentally) in a matter of seconds.

I hope your charger is monitoring both batts. Some say the stacked charger method for protected batteries will present a few problems - especially when one cell reaches cut off before the other. Perhaps including a failsafe batteriy protection circuit on the board might be a good approach -with that - use unprotected batteries -- the Tekkeon does it quite nicely.

Yes, I have 3 levels of protection, including thermal. Not sure if all 3 will make it to the final board, but I will not rely on one. It all depends on how small of a component I can assemble in my toaster oven while hand dispensing solder paste with a syringe, smallest I've tried so far is 0603 and MSOP8.

Mind if I ask what chip your using? (PM on this issue is fine). I've been playing in spice with well over a dozen Sepic's over the past year and simply can't find one to do my ditty without problems. (my ditty = operate accross all V ranges and all atty configurations known to be vapable) . The closest I've come to success is with the LTC3872. It's current mode and not sepic. Seems there may be some promise with inverting controllers but I haven't gone there yet.

PM sent.

I spoke to Linear Tech: Buck-Boost in these ranges would be more efficient but circuit becomes more complicated, Cuk (Inverting) would work, but Sepic seemed to be the easiest for my first Converter circuit design.

Note: I wouldn't have even attempted this if not for all the questions WillyB posted to my Buck converter mod thread, I can't begin to explain all that I've learned from this experience. Thanks WillyB.

 

[CapeCAD]

PS. Mind if I ask what ripple you get with the 1.5 Ohm Load at say...... 4.5Vo?

Circuit was designed around 25%, I'll have to measure and get back to you.

 

[o4_srt]

you say enable pin to ground shuts it off.

Is it an inhibit, or a full off?

The evercool still draws 1 mA of quiescent current. To increase battery life, I've been playing around with a P channel MOSFET on the VCC of the evercool, after the bypass cap, keeping the cap charged to further increase battery life and tying the inhibit pin to ground to keep it on.

What is the quiescent current of your device?

 

[CapeCAD]

you say enable pin to ground shuts it off.

Is it an inhibit, or a full off?

The evercool still draws 1 mA of quiescent current. To increase battery life, I've been playing around with a P channel MOSFET on the VCC of the evercool, after the bypass cap, keeping the cap charged to further increase battery life and tying the inhibit pin to ground to keep it on.

What is the quiescent current of your device?

Inhibit, my ammeter minimum is 200 microAmps and it is less than that. PWM draws 15 microAmps in shutdown mode.

 

[CapeCAD]

PS. Mind if I ask what ripple you get with the 1.5 Ohm Load at say...... 4.5Vo?

I measured 50mV Pk-Pk
ScopeBW= 100MHz
Vin = 7.2V
Vout = 4.5V
Iout=3A

I'd send screen shots, but my 20Amp electronic load puffed after running 30 minutes at 3Amps - I'ts hot.

 

[SurvivorMcGyver]

I measured 50mV Pk-Pk
ScopeBW= 100MHz
Vin = 7.2V
Vout = 4.5V
Iout=3A


I'd send screen shots, but my 20Amp electronic load puffed after running 30 minutes at 3Amps - I'ts hot.

VERY sorry to hear about the "puff" -- we ALL should know what that means by now......

Thanks for the feed back. Spot on. Great ripple! And great design. A few of my simulations (with other designs and chips) were giving me up to a 300 mV ripple (which is unaceptable to me). My endeavor is to break to 6V barrier and get at least 6.3-6.5Vo , perferably 7.2ish and and still maintain the adjustability down to 3.4 without horendous spikes on the Vi on start-up across all voltage ranges with the smallest Vo ripple as possible. Still haven't hit it. Are you for hire {chuckling} I need a consultant (or at least your notes for a cheat sheet)