Author Topic: Boost Circuit (Read 36895 times)

Breaktru · « **on:** October 23, 2011, 03:05:11 PM »

500kHz High Efficiency 6A Switching Regulator

I modded the Datasheet's Single Li-Ion Cell, 5 volt output circuit to a Variable Output: 3.88v to 6.04v as shown in the Diagrams.
After much tweaking, I got it stable with little voltage drop when a load of 1.5 ohms was added. Unfortunately I can not get the amperage up to more than 1.87 amps before the chip cuts out.
CraigHB sent me a link (thank you Craig) to a simulator which I used to tune the circuit to a max of 4.3amps. However my real world circuit does not perform as the simulation. Must be doing something wrong.

timesarerough · « **Reply #1 on:** October 23, 2011, 03:31:55 PM »

Maybe the chip is crap?

http://www.ebay.com/itm/500kHz-High-Efficiency-6A-Switching-Regulator-LT1370-/110643798967?pt=LH_DefaultDomain_2&hash=item19c2e263b7

Breaktru · « **Reply #2 on:** October 23, 2011, 03:36:53 PM »

Probably me... I'm sure an EE (Electrical Engineer) can figure it out.
I'm just a EKH (Electrical Knuckle Head)

timesarerough · « **Reply #3 on:** October 23, 2011, 04:04:09 PM »

Probably right....

timesarerough · « **Reply #4 on:** October 23, 2011, 04:12:11 PM »

Did you test the chip using the 5V schematic before making alterations to see if it would produce the 3+A output?

Breaktru · « **Reply #5 on:** October 23, 2011, 04:23:10 PM »

Quote from: timesarerough on October 23, 2011, 04:12:11 PM

Did you test the chip using the 5V schematic before making alterations to see if it would produce the 3+A output?

Yes, of course. I got slightly over 2 amps

mjmetz · « **Reply #6 on:** October 23, 2011, 04:27:34 PM »

Inductor?

timesarerough · « **Reply #7 on:** October 23, 2011, 04:29:50 PM »

Quote from: Breaktru Admin on October 23, 2011, 04:23:10 PM

Yes, of course. I got slightly over 2 amps

Shouldn't you be getting a lot more than 2 amps?

It says 3.3amps....

Maybe its not the EKH and its just a POS!

Pantera · « **Reply #8 on:** October 28, 2011, 10:40:41 AM »

I've seen your thread over at ECF and don't understand why following the schematic with the recommended parts should not yield the specified amperage of 3.3 amps

Breaktru · « **Reply #9 on:** October 28, 2011, 10:52:08 AM »

Exactly my point Pantera, I used the exact part numbers and kept critical length of parts as close as possible. Even made a copper ground plane for under the controller chip. I understand what the E.E.'s are saying about an external Mosfet and different inductor for more amperage but I should get at least what the spec's says using their parts.

mjmetz · « **Reply #10 on:** October 28, 2011, 12:15:07 PM »

I guess I didn't catch that. You made their exact test circuit for 5V and couldn't get the spec'd amps out? Yes, you did, I went back and reread it. Hmmm.

CraigHB · « **Reply #11 on:** November 20, 2011, 04:52:48 PM »

Hey Breaktru and other folks, just signed up here, checking out interesting threads.

It's not so unusual to find your real world circuit does not perform as well as the simulation. I just did one that said I could get much more power out of a booster design than I can actually achieve on the bench. Though the design is still good since predicted was 50W and actual is 30W. Still plenty of power for an e-cig mod.

The LTSpice simulator is actually pretty good as far as simulators go (and it's free to boot), but no simulator can account for all factors. There are parasitic characteristics that come into play due to PCB layout and component compositions that can have surprising effects in the real world. Sometimes you can incorporate little models of these effects into your simulation to more accurately predict the behavior of a real world circuit, but it's not always a trivial thing to do. There are also manufacturing tolerances in components that are not always easy to account for.

LTSpcie does provide a panel for each component to enter some of its parasitic characteristics. One of them that is very important in DC-DC converter design is the ESR for your output capacitor(s). Other ones that are important are battery and inductor DCR. If you can accurately specify those in your model, you may be able to successfully tune a higher output.

Time domain analysis is usually the first place you start with any circuit simulation. It just means you're observing the response to an input over time. When using a simulator to model a DC-DC converter, it's quite tricky to set up an accurate closed loop time domain analysis. That can be a big help, but not something I can explain in a paragraph. At the least, you can get a good idea of your circuit's response with an open loop frequency domain analysis which is much easier to set up in a simulation.

Any DC-DC converter that utilizes external feedback compensation requires tuning through a complicated simulation or with special test equipment on the bench, however, there are some converter chips that do not utilize external compensation. They are very easy to use.

I highly recommend the MAX1709 for a high output booster that's super easy to build. No simulation or tuning is required. Just strictly follow the component specifications provided by the data sheet. I built a booster with this chip simply by following the instructions in the data sheet and it it did in fact put out as advertised. Though, it does have its disadvantages. It's limited to 5.5V output and is asynchronous which means it uses a diode rectifier instead of a MOSFET. That results in an efficiency hit 5-10% depending on rectifier selection.

I have not yet come across a high output synchronous controller that does not utilize external feedback compensation and does not require tuning. They're probably out there, I just haven't come across any yet.

Breaktru · « **Reply #12 on:** November 20, 2011, 05:23:11 PM »

Thank you Craig for your expert details on the Boost circuit.

The Max1709 looks like a fairly easy build. Although the LT1370 was also fairly easy to build, the Max is probably more my speed regarding my capability.

CraigHB · « **Reply #13 on:** November 20, 2011, 05:24:51 PM »

Welcome, definitly like to hear if you have a go with it. It really is a great part for simplicity and peformance.

Breaktru · « **Reply #14 on:** November 20, 2011, 05:30:50 PM »

Will do Craig, I'm going to wait to buy the parts for now. In my next parts order I'll include the Max circuit parts.
Just received several orders from various vendors. I got to slow down... I keep ordering stuff. I don't even use half of it

Breaktru · « **Reply #15 on:** November 29, 2011, 10:21:21 PM »

Craig, got my free samples of the Max1709 in a SOIC package. Dam this thing is small. I don't think I can solder to it.
Do you use the TSSOP package? The spec's show it's bigger.

Update:

I realize now that I have a couple (3) of varieties of the 1709 and one is the TSSOP which is also tiny.

I see the datasheet says the TSSOP can provide a 4A output and the SOIC has a 2.4A output.

CraigHB · « **Reply #16 on:** November 30, 2011, 05:46:24 PM »

I'm not sure which variant you plan to use, but I built a couple booster mods with the MAX1709EUI+ which is the package that has the solder pad for heatsinking underneath. I believe that's the TSSOP. You may be able to get away without heatsinking since e-cigs are only on for short periods at a time, but I would heat sink anyway. Just put two larger vias (conductive holes) underneath the IC pad that connect to the ground plane. Solder fill the holes to the chip's pad. The 1709 will put out 22W all day with that.

No one will ever accuse me of being curt in my replies, but here's a bunch of information on soldering if you decide to use that part.

I would suggest you look into doing reflow. I still have not attempted reflow, but I've looked into it quite a bit. You can try the cheap 63/37 Chinese paste you find on eBay or DealExtreme, but there's also some good name brand pastes you can find on Mouser, DigiKey, and Newark. Name brand is much more expensive.

I hand solder all my stuff, but mainly because reflow wasn't around when I first started soldering electronics (that's a long time ago). I'm in that rut. If you want to have a go with hand soldering, I use the ETS tip which is the .015" conical one. Your station uses the same pencil as mine.

I use .015" 63/37 wire solder with a mildly activated rosin core (e.g. Kester #24-6337-9703), but you can use an activated rosin core (e.g. Kester #24-6337-0007) if you clean the board with 99% isopropyl alcohol and remove all the flux afterwards. A flux brush and some canned air works well for cleaning. I would recommend an activated rosin flux. The mildly activated stuff is a little weak sometimes, but I use an activated flux pen (Kester #186) to glue ICs down before soldering. The flux comes out liquid than hardens to a thick paste after a few minutes. Works well for holding ICs in place.

Use a magnifier like watchmakers and jewelers use. I use a 10x stereo microsope myself, but the field of view is limited with that. It has excellent clarity though.

Get a 10' roll of desoldering braid. You will invariably need to sop up overflow on occasion. You want the fine #1 braid, but I also find uses for the wider #2 and #3 braid so I keep plenty of that on-hand as well.

I can't really recommend which soldering method you should use since I haven't tried reflow. Most people would probably say reflow. May have to try both yourself and see which method works best for you. In any case, even with reflow, you'll still need to do some hand soldering here and there.

I posted this before in the ECF forum and you've already seen it, but for the benefit of this thread, here's a picture of what you can do with an iron when you have the right tools and materials. This is the mod I'm currently working on.

Breaktru · « **Reply #17 on:** November 30, 2011, 07:10:34 PM »

Thanks very much Craig for your insight.
I basically have everything except a reflow oven. Have the desolder braid, two kinds of flux (paste & liquid), alcohol, .015" conical tip,
Bausch & Lomb Magnifying glasses:

, 0.015 solder, a couple of MAX1709EUI+, etc.....

This all depends on if I can get the PCB down correctly with the software. If not, I'll scrap the whole idea of building it.

Got to get back to my work shop, I'm fixing an old Guitar Amp I bought back in 1967. My son has been using it. The young-ins real love the old tube amps. They prefer them over the solid state amps and they are a premium to buy. Man it's a pleasure working on electronic parts that you can see. Had a 60 cycle filter hum, replaced the 200uf, 200v Electrolytic. Now that's a part. The cap is 3¼ x 1 inch. A real manly size.
Thanks again Craig.

CraigHB · « **Reply #18 on:** November 30, 2011, 07:21:28 PM »

Welcome, enjoy discussing this stuff.

Yes, artists just love those old tube amps. In theory, modern solid state amplifiers should perform with vast superiority, but I think musicians like the effect tubes have on fidelity, kind of a pure analog reproduction over a digitized one. There are actually quite a few designs for tube audio amps floating around the net and there are still a few suppliers of tubes as well. My very first electronic project as a kid was building a 100W tube amplifier (I was heavily into music then). I never would have though they would still be in demand today.

I'll warn you in advance, Eagle has a steep learning curve. I would probably just use ExpressPCB. You can fill each panel with multiple copies of your design. That way, you can get more than 3 boards out of an order. Otherwise, if you want to home grow your boards, go two sided. You pretty much have to have a ground plane with a booster. Use the ExpressPCB software to print out the designs for toner transfer or photo sensitive transfer (you'll have to decide for yourself which method you prefer). You could use the Eagle software and it's a good product, but it takes a time investment to learn.

vpjunkie510 · « **Reply #19 on:** December 11, 2011, 08:15:46 PM »

hey guy's new to the room , I pretty much had a question about Boosting. I have been looking to build a VV mod for a wood box I build and would really like to keep the size as small as possible. as of right now my box measure's in at 2 3/4 L 1 1/2 w 1 T , Now I know I will have to extend it a little to reach what I want but thought I could save spacxe by boosting.

My main issue if I boost is life span of the batterie , I would like to boost a 18350 or 18500 the biggest. would I be settling space for battery life or is the loss harldly noticable
?

being I sand every box by hand I wanted a plug and play but realy haven't found anything in that area. I think it would take to much time for me to build each indivduale vv board , and box's while being able to produce the box at a resonable turn over rate.

any suggestion's?

Breaktru · « **Reply #20 on:** December 11, 2011, 08:40:13 PM »

Quote from: vpjunkie510 on December 11, 2011, 08:15:46 PM

hey guy's new to the room , I pretty much had a question about Boosting. I have been looking to build a VV mod for a wood box I build and would really like to keep the size as small as possible. as of right now my box measure's in at 2 3/4 L 1 1/2 w 1 T , Now I know I will have to extend it a little to reach what I want but thought I could save spacxe by boosting.

My main issue if I boost is life span of the batteries , I would like to boost a 18350 or 18500 the biggest. would I be settling space for battery life or is the loss harldly noticable
?

being I sand every box by hand I wanted a plug and play but realy haven't found anything in that area. I think it would take to much time for me to build each indivduale vv board , and box's while being able to produce the box at a resonable turn over rate.

any suggestion's?

Welcome to the forum,
I don't know what boost circuit you would be using but I can squeeze an 18350, an 18490 or an 18500 in a 2 3/4" x 1 1/4" x 7/8" internal space box using a 04050c.
Haven't counted the hour usage but I use an 18350 - 1200mah and an 18500 - 1500mah. Definitely longer time than my 14500 - 900mah which can run up to 7hrs.

vpjunkie510 · « **Reply #21 on:** December 11, 2011, 10:29:57 PM »

I haven't actually tried boosting anything yet, besides not being very tech savy I try read and ask as many questions as possible before I put something together .

I forgot who gave me the forum address for here but its been a big help just reading. I think i will start looking for parts to play around . Your copper mod looks great by the way .

Breaktru · « **Reply #22 on:** December 18, 2011, 11:17:07 AM »

Found this boost circuit at TI.com. Sounds interesting.
See attached photo and pdf. Pdf has parts and cost (members only)

http://breaktru.com/ecig/webench_design_1202777_4.pdf

CraigHB · « **Reply #23 on:** December 18, 2011, 06:40:37 PM »

That's one of the parts TI inherited from National with the recent acquisition. The LM3488 controller is current mode which makes it easy to stabilize, though you need that current sense resistor which adds a bulky component and reduces efficiency an amount depending on the value of the resistor.

I'm not sure I'd use that particular design with the LM3488 myself. I'd run a higher frequency and lower value inductor (600kHz, 1uH). Also, I don't understand why they're putting a humongous 1500uF capacitor on the input. Normally, with a booster running around half a megahertz, you only need around 150uF on the output and a fraction of that on the input to keep things clean. Maybe that design was for a customer with unusual requirements?

Efficiency is not going to be anything to get excited about. That's a current mode booster with asynchronous rectification (diode rectifier). The best efficiency is obtained with voltage mode and synchronous rectification (MOSFET rectifier). Voltage mode is a RPITA to design though. However for something easy to build, the LM3488 would be a good one to use.

Looks they're using bulk costs on the prices they've listed. You're going to pay a lot more for parts when buying in small quantities.

Pantera · « **Reply #24 on:** April 16, 2012, 02:05:53 PM »

bump......

jomurp · « **Reply #25 on:** May 10, 2012, 11:20:20 AM »

Anyone else use this chip for a boost mod? Looks easy enough.
I know that this chip comes in a much smaller package also. I imagine it would be challenging to solder.

RAYMONDROOT · « **Reply #26 on:** May 29, 2012, 05:40:55 PM »

these Tantalum Capacitors that your getting what voltage rating you getting them in

Breaktru · « **Reply #27 on:** May 29, 2012, 05:46:39 PM »

Quote from: RAYMONDROOT on May 29, 2012, 05:40:55 PM

these Tantalum Capacitors that your getting what voltage rating you getting them in

Ray, are you talking about the caps for this Topic for the LT1370 ? or another boost circuit?

RAYMONDROOT · « **Reply #28 on:** May 29, 2012, 08:18:12 PM »

oh sorry my bad it was in ref to the 04050 i am tooling through the whole mod section and posted in the wrong spot but the ? is still the same
im using the 6v caps and i am getting some odd behavior thats why im asking 100uf 6v "electrolytic" tantalum

Breaktru · « **Reply #29 on:** May 29, 2012, 10:03:32 PM »

I've used 6v, 10v and 16v in my builds.
The rule is 2x the voltage. It would still behave the same. 2 x would compensate for a surge and blowing out a cap. But... we are using batteries so now surge should be expected.
What odd behavior are you experiencing?

RAYMONDROOT · « **Reply #30 on:** May 29, 2012, 10:48:10 PM »

OK well that explains allot
i'm using a transformer dropped from 12v dc to an adjustable output running at 5v dc to run my bread board for testing and i haven't tested for spikes or fluctuation yet
i intend to run 3.7v ly-po 20c for my new toy

RAYMONDROOT · « **Reply #31 on:** May 29, 2012, 11:00:13 PM »

really this was a brain fart moment i should have thrown some caps in my transformer to stabilize it or just used the battery i intend to use that is what im seeing allover this site battery's act differently even identical batts can give slightly different values in the end
byt thanks for the brain duster (i needed it)