# Boost converter 12V to 36V (25 mA load) using 555

Discussion in 'General Electronics Discussion' started by abuhafss, Jul 18, 2014.

1. ### abuhafss

348
11
Aug 3, 2010
Hi

I have 3 x 12V SLA batteries connected in series as shown.

The current drawn by Circuit A is 25mA and the Circuit B draws about 100mA.

The batteries are to be fixed inside the casing. The battery A will have different discharge status than B and C.

So, how should the battery charging circuit be designed?

1) A+B+C => 36V charger circuit
2) B+C separately and A separately => 24V charger circuit and 12V charger circuit
3) All three batteries separately => 3 x 12V charger circuits

Thanks for the help.

2. ### (*steve*)¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥdModerator

25,449
2,809
Jan 21, 2010
It is problematic to charge all three together because one battery (A) is going to be discharged more than the others and so the others will tend to get overcharged.

0ne 24V and one 12V charger would suffice, however you'd have to ensure that they don't have a grounded output or you'll short something out (probably battery A).

Three 12V chargers would also work, but with the same caveat as expressed above.

3. ### Gryd3

4,098
875
Jun 25, 2014
There are load balancing circuits available to handle the kind of situation where the cells become unbalanced... but the circuit operating at 12V midway between the stack of cells is gong to complicate things.

This is the reference I found that may help, but may need to be adapted.
http://www.digikey.com/en/articles/...n-evs---part-i-passive-balancing-technologies

4. ### KrisBlueNZSadly passed away in 2015

8,393
1,270
Nov 28, 2011
Another option is to use a voltage regulator (switching, preferably, for best efficiency) to convert the 36V down to 12V for circuit B so both circuits draw from the three batteries combined.

Edit: Or use only one battery and step it up to 36V for circuit A.

abuhafss, shumifan50 and (*steve*) like this.
5. ### EeLove

3
0
Jul 19, 2014
Ideally you would use a single 24v DC battery. Buy or design a step up converter to transform 24vDC to 36vDC and also a buck converter to transform 24vDC to 12vDC.

That way you only have to charge one battery and the output voltages 12v and 36v will be far more stable than using your design idea since the output voltage is not dependent on the depletion level of the batteries.

6. ### abuhafss

348
11
Aug 3, 2010
Thanks all of you for your input.

Kris' suggestion for using only one 12V battery is the best choice. It will not only reduce the size and weight but also the make the charging circuit more simple.

This leads me to a new requirement of 12V to 36V (1W) converter. I need to discuss a couple of circuits in this regard. Should I start a new thread or continue here? Or if anybody has some appropriate schematic, shall be thankful for sharing.

7. ### (*steve*)¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥdModerator

25,449
2,809
Jan 21, 2010
I'd keep going here

8. ### abuhafss

348
11
Aug 3, 2010
Okay. A simple 555 voltage tripler gives about 34V (acceptable) but, the power is not enough to provide 25mA.

Alternatively, I have another circuit which simulates ok in LTspice. Please check it and help me optimizing it.

The spice file is attached, please change the extension to .asc

File size:
1.8 KB
Views:
176
9. ### KrisBlueNZSadly passed away in 2015

8,393
1,270
Nov 28, 2011
It would be easier if you posted a screen capture of the schematic, along with simulation waveforms.

Your best option is a boost converter - a switching supply that uses an inductor in flyback mode to produce a higher output voltage. These are available preassembled on eBay with adjustable output voltages, or you can make one yourself using various switching ICs from many companies, including Linear Technology and Texas Instruments. Just search for boost converter.

348
11
Aug 3, 2010

11. ### KrisBlueNZSadly passed away in 2015

8,393
1,270
Nov 28, 2011
OK, that's a boost converter based on a relaxation oscillator. It will waste significant power because the regulation consists of a zener clamp across the output, instead of a circuit that reduces the drive to the inductor. Also you're running the 1N5819 pretty close to its voltage limit. But it will work.

Personally I think I would buy a preassembled module on eBay. I don't normally recommend doing that, but you only need 25 mA so performance isn't a big issue, and it's a very convenient and cheap option.

I'm going to do some messing around with that design to see if I can improve the efficiency, but mainly for my own personal interest. I don't think you should use it unless you're obsessed with doing it yourself, or using a discrete solution.

Edit: That design will also damage the base-emitter junction of the BC557. You should insert a 1N914/4148 diode in series with the base, with its anode to the base.

(*steve*) likes this.
12. ### (*steve*)¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥdModerator

25,449
2,809
Jan 21, 2010
Thanks Kris, I was going to analyse that and comment because it didn't look too convincing when I first glanced at it. I was going to investigate if it could degenerate into a linear mode, but with the other things you spotted I think that's now moot.

KrisBlueNZ likes this.

348
11
Aug 3, 2010

14. ### KrisBlueNZSadly passed away in 2015

8,393
1,270
Nov 28, 2011
R1 is much too high to keep Q1 saturated. You would be better to use a MOSFET. And there's no control loop for voltage regulation. Reduce the load current, and the output voltage will increase! D1's voltage rating is still too close - at least 30% safety margin is recommended (36 * 130% = 47V so use an MBR150. Actually you don't need a Schottky diode at that frequency; a normal fast-recovery diode like a UF4002 would be fine.

If you don't want to buy a preassembled unit, you should look into ICs designed for this purpose. Here are four that I found at Digi-Key, in order of increasing price.

Texas Instruments LMR64010 SOT23-5 (SMT); current mode; 1.6 MHz
http://www.digikey.com/product-detail/en/LMR64010XMFE/NOPB/LMR64010XMFE/NOPBCT-ND/2754198 USD 1.98

Texas Instruments LM2733Y
SOT23-5 (SMT); current mode; 600 kHz
http://www.digikey.com/product-detail/en/LM2733YMF/NOPB/LM2733YMF/NOPBCT-ND/565265 USD 2.44

Texas Instruments LM5001 SO-8 (SMT); current mode; 260/780 kHz
http://www.digikey.com/product-detail/en/LM5001MAX/NOPB/296-35285-1-ND/3738993 USD 4.02

Linear Technology LT1107 SOIC-8 (SMT); 63 kHz
http://www.digikey.com/product-detail/en/LT1107IS8#PBF/LT1107IS8#PBF-ND/889893 USD 5.40
This one is a bit different from the current mode controllers - it is a "gated oscillator" type and the output may be more noisy than current mode types, but it's simple to use - three resistors, inductor, diode, input and output caps.

shumifan50 likes this.
15. ### abuhafss

348
11
Aug 3, 2010
Thanks. Would appreciate if you suggest me some in DIL8 package.

8,393
1,270
Nov 28, 2011

25,449
2,809
Jan 21, 2010
18. ### abuhafss

348
11
Aug 3, 2010
That is really a valuable thread for a hobbyist like me. I have read the whole thread once and got some idea of the working. Unfortunately, failed to simulate it in LTSpice. Maybe I need to adjust the setting of LTSpice. I will do my best to adjust components value in Steve's design to get my required output.

In the meantime, based on Steve's idea of feedback loop, I modified the 555 based boost circuit. And here are the results with and without load. It is quite stable upto 60 or 65mA. The output is always 8V + Zener's voltage. Frankly, I have no exact idea of the working. I shall appreciate if you guys can shed some light on it. I have not yet tested it physically, will do it after getting your valuable comments.

19. ### (*steve*)¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥdModerator

25,449
2,809
Jan 21, 2010
Take a look at the dedicated smps ICs.

The intent of that link was to show how complex it can be to design a "simple" DC-DC converter.

Sure, I could offer improvements that could boost its performance, but all that would do is add more complexity. And that's fine if your aim is to learn more about techniques and pitfalls, but less so if you want something that just works.

The dedicated solutions will give you a guaranteed performance and their datasheets will contain equations or tables to guide you to appropriate value components. There are some that require in the order of 3 additional components.

KrisBlueNZ likes this.
20. ### (*steve*)¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥdModerator

25,449
2,809
Jan 21, 2010
I've never tried simulating that circuit. I did my tests in real life.