Hotwire Regulator Design Collaboration Project part 3

[mpteach]

Some users want to use 52V (12S Li-Ion). The present parts are rated at 40V so we need to stay below that.

For most users the 40V parts are fine.

Diodes were investigated to reduce the inductive overshoot, they did not appear to help.

12S, thats a very long and skinny tube!

did you try a schottky diode? http://en.wikipedia.org/wiki/Flyback_diode#Design

 

[Alan B]

12S, thats a very long and skinny tube!

did you try a schottky diode? http://en.wikipedia.org/wiki/Flyback_diode#Design

wquiles did that testing, and possibly JimmyM as well. I believe a schottky diode was tried. It had zero effect as I recall. You can find the discussion and measurements earlier in one of the design threads.

 

[JimmyM]

Yeah. I tried a Shottkey during development with no effect. The inductive kick is only about 1 uS and our design is very low frequency. The Avalanche energy capability of the FETs we're using is sufficient to handle the "kick".

 

[Alan B]

12S, thats a very long and skinny tube!

Actually not. Folks using 12S are generally using elephant type lights that have 3 or 4 cells alongside, so a 12S stack is only 4 or 3 cells in length. Four 18650 cells alongside makes a thick handle but three is not much thicker than a D M@g. These are custom builds but available from several sources here on CPF.

 

[sylathnie]

12S Body + Light

Hopefully it will be hosting a high voltage regulator soon! :twothumbs

 

[mpteach]

What kind of lithium cells do you have it there?

 

[sylathnie]

IMR 18650s from Ryobi One+ battery packs.

 

[mpteach]

are the 18650 electrically in series but physcially parrelel? thats long but it doesnt look 650*12 long to me.

 

[KiwiMark]

are the 18650 electrically in series but physcially parrelel? thats long but it doesnt look 650*12 long to me.

An Elephant II body with 12 cells would be around 3 cells long inside the battery compartment. The cells are arranged in a 2x2 arrangement - as you say electrically in series but physically parallel. I have an Elephant II with 8 cells, same deal just 2/3rds as long. The 12 cell version can be reduced to 8 cells or 4 cells very easily with a little unscrewing of the body extensions and the battery holder extensions.

 

[mpteach]

Which is exactly why I didnt understand the need for lithium 12S or high voltages since i doubt anyone will ever stack 12 in a single row.

 

[Alan B]

are the 18650 electrically in series but physcially parrelel? thats long but it doesnt look 650*12 long to me.

Yes, they are electrically in series which is useful for some of the high powered filaments. Several folks on this list have 12S voltages to deal with (about 52 volts). The four-alongside packaging may only be 3S in physical length but it is still 52V.

There are three parts that need to see this voltage. The FET, the voltage regulator and the input capacitor. The 50V capacitor is probably ok, but the 40V rating on the FET and voltage regulator are too overstressed.

I found some 60V FETs, and the voltage regulator is made in a 60V model, but finding it is not easy. Newark shows some stock in the UK but it is costly to ship from there unless you live there.

The regulator and FET are the same on both the through-hole version and the hybrid SMD version of the PCB. The capacitor did change to a SMD. It is a ceramic capacitor and the rating is so close that I don't expect it will have problems, but it could be changed to a higher voltage unit as well.

 

[petrev]

Yes, they are electrically in series which is useful for some of the high powered filaments. Several folks on this list have 12S voltages to deal with (about 52 volts). The four-alongside packaging may only be 3S in physical length but it is still 52V.

There are three parts that need to see this voltage. The FET, the voltage regulator and the input capacitor. The 50V capacitor is probably ok, but the 40V rating on the FET and voltage regulator are too overstressed.

I found some 60V FETs, and the voltage regulator is made in a 60V model, but finding it is not easy. Newark shows some stock in the UK but it is costly to ship from there unless you live there.

The regulator and FET are the same on both the through-hole version and the hybrid SMD version of the PCB. The capacitor did change to a SMD. It is a ceramic capacitor and the rating is so close that I don't expect it will have problems, but it could be changed to a higher voltage unit as well.

Hi Alan

I order from Farnell (which is Newark in the UK) - Let me know the part codes and I can look them up and check availability. I can post whatever you need.

Cheers
Pete

 

[Alan B]

Hi Alan

I order from Farnell (which is Newark in the UK) - Let me know the part codes and I can look them up and check availability. I can post whatever you need.

Cheers
Pete

The part we're looking for is LM2936HV-5.0 in TO92 package. Looking at a data sheet again just now I'm not sure they even make that combination. I've seen the HV model at Farnells, but the surface mount package is too large for the regulator board.

What we need is a TO92 regulator, low quiescent current, 5.0V output, 60V input. Load current is small, 50mA is fine.

 

[petrev]

The part we're looking for is LM2936HV-5.0 in TO92 package. Looking at a data sheet again just now I'm not sure they even make that combination. I've seen the HV model at Farnells, but the surface mount package is too large for the regulator board.

What we need is a TO92 regulator, low quiescent current, 5.0V output, 60V input. Load current is small, 50mA is fine.

Hi

Only one in UK Farnell is
LM2936HVBMA-5.0/NOPBand that is 8-pin package and it is only available from US stock ? ? ?

Data sheet suggests HV-5.0 not available in TO92
Cheers
Pete

 

[Alan B]

I just did another Digikey search for any TO92 regulator over 40V and found NONE. So the High Voltage version of the regulator is apparently not going to be as easy as changing a couple of parts. It will probably require a new PCB layout to accept some other package, and that part of the board is pretty crowded, though the SMD relayout did free up some space. But realistically I can't work on that now with all the other things in my que.

I was misled by the data sheet which indicates a high voltage version. What they didn't make clear is that this high voltage version is apparently only available in a fairly large surface mount package that I can't easily use.

So what High Power is possible with the "Standard 40V Regulator":

40V Divided by 12 is 3.33V, so A123 18650's that are charged to 3.3V will likely do quite well. They pack quite a punch, too. Easily handle 10-12 amps. Put out close to 400-480 watts (40V 10A).

Alternately 11S A123 cells can be fully charged to 3.6V. They will produce 35V at 350 to 400 watts.

Next down is 10S LiMN charged to 4.0V. Again an easily 400-480W power source. Charging to 4.0V will increase the lifetime of the cells tremendously at the cost of some fully charged capacity loss.

At the full 4.2V the limit is 9S. With LiMN you would get about 32V under load for 300-380W.

 

[mpteach]

What voltage do the high wattage bulbs need?

 

[Alan B]

What voltage do the high wattage bulbs need?

About 36 volts. Here's a 400W setup:

https://www.candlepowerforums.com/threads/205841

 

[sylathnie]

So the High Voltage version of the regulator is apparently not going to be as easy as changing a couple of parts.

:mecry::mecry::mecry:

Ouch. This is crippling to my plans for world domination via flashlight....

So anyone else have a high voltage soft starter in the works? :naughty:

 

[JimmyM]

Are the sleep_cpu(), sleep_enable(), and sei() your defined functions?
If so, can you post the the details of those functions?
Is your code posted anywhere so I don't have to bother with these questions?
Thanks, Alan.

I've gotten the sleep to work.

OCR1A = 0;  // Sets PWM Duty to 0 (Zero)
ADCSRA = 0; // disable ADC to save power
            
set_sleep_mode(SLEEP_MODE_PWR_DOWN); // Sets sleep mode to Power-Down.
sleep_mode(); // Puts the AVR to sleep using above set mode.

Of course this won't work in a pushbutton interface since I haven't enabled any interrupts. But for the current mode of operation. When operating it uses 4.4mA. In power-down mode, it uses 0.12 mA

 

[Alan B]

:mecry::mecry::mecry:

Ouch. This is crippling to my plans for world domination via flashlight....

So anyone else have a high voltage soft starter in the works? :naughty:

I'm using the same FET that you are already using, and the regulator is spec'd to survive 60V. It might work now.

The specs on the regulator seem to say that it will not fail at 60V, but it is not spec'd to operate over 40V. It is rated at 50ma and we're using about 6ma so it may well be the case that it will work fine above 40V. I can't predict that.

You might decide to try it and let us know how it works.