Continue to Site

Welcome to our site!

Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

  • Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

LM3410 for LED backlight of LCD help

Status
Not open for further replies.

Micro9900

New Member
I finally got some new lm3410Y and a video. Unfortunately I can't tell if the part is no longer operational. I previously I was able to read 17.5V then it went down to 14.5 and then 4.6V. With these new parts the same thing happens I read 14.5V across the resistor and now I only read 4.6V as shown in the video. Is there a better way to test this? Should I try and make a PCB and forget the breadboard?

Video:
https://www.youtube.com/watch?v=KWL5mLMj0B0&feature=youtu.be

I am testing this using two 1K ohm resistors in parallel to get 500 ohms since ~20V/500 = 40mA.

The end goal is to power this via USB. Here is the display:

https://www.electro-tech-online.com/custompdfs/2013/08/NHD-43-480272EF-ATXL-T.pdf

Here is the datasheet for the lm3410:
https://www.electro-tech-online.com/custompdfs/2013/08/lm3410-2.pdf

I need 19.2V @40mA. So far I am unsuccessful. I'm not sure how to go about this and I could really use some help. Maybe I should use a different step up constant current regulator such as the FAN5333B? Thank you.
 
Last edited:
19.2V*40mA is 768mW. At 80% efficiency, that means the chip is dissipating 156mW. I'm guessing that since it is not soldered to pc traces, it is heating up and going into thermal shutdown...

To get (768 + 156)mW from the USB, you will be drawing 185mA.
 
Thanks for the reply Mike. I had the board soldered for me, I actually have 4 of these in total, so I'm not exactly sure if this is a soldering issue. I'm thinking it may be more with the connections. Could you see it okay? If not I can go a bit slower and upload another video if you need to see it better.
 
Post your schematic as you have it hooked up. Your video is worthless...

Read page 15 of the data sheet. I still think you are letting the chip get hot, and it is shutting down to protect itself.
 
I gave it a look. I'm sorry if I came off as dismissive, I just am a bit confused. I am using this part for the adapter:

https://www.proto-advantage.com/store/product_info.php?products_id=2200093

Are you saying that to begin with that the adapter is the problem? If so, that could explain a lot. If this is the case would the FAN5333B be a better way to go for prototyping with a breadboard?

I am following the schematic on pg.31 for the LM3410Y that includes over voltage protection with a zener diode.
 
Last edited:
I don't see the caps. Are they there? Ceramic? What inductor and sense resistor?

It's a pretty high frequency to try and run on a breadboard.
 
It's a pretty high frequency to try and run on a breadboard.
Nearly every time somebody says a circuit does not work, it was built on a lousy and stupid breadboard.
A breadboard might be OK for a battery, a resistor and an LED. Obviously DC, not high frequency AC.
 
I didn't realize breadboards were that cumbersome to high AC frequency. This is obviously my first time doing something like this and I would definitely like to "unstupify" my circuit :p. Capacitors are ceramic (input C1 10uF rated at 6.3V and output C2 is 2.2uF rated at 25V), inductor is 22uH, sense resistor R1 is a 3W 5.05 ohms. I am following the schematic on pg.31 and some of the parts are very similar in rating.

Should I try and make a PCB and then take the load to the breadboard and into the LED- and LED+ pins of the display or is that still no good?
 
Last edited:
I didn't realize breadboards were that cumbersome to high AC frequency.
A breadboard has hundreds of intermittent contacts that make poor contact. A soldered joint is much better.
It has rows (of contacts) that have fairly high capacitance between them that is bad at high frequencies. The connecting wires also have high capacitance (coupling?) between them (oscillation?).
The rows of contacts and the long connecting wires have series inductance that is bad at high frequencies and causer oscillation.
Use a breadboard only for DC or for low frequencies.

Should I try and make a PCB and then take the load to the breadboard and into the LED- and LED+ pins of the display or is that still no good?
For 51 years I have used Veroboard (stripboard) for my prototypes and nearly every one was sold because they worked perfectly and looked good.
Of course every joint was excellent because it was soldered.

The strips of perforated copper are cut with a drill bit at a hole so that the same strip can be used again and again and the circuit is compact (short wires).
The strips of copper form half of a pcb, and the components and a few very short jumper wires form the other half of the wiring.
The holes are already drilled and the strips of copper are already there so it is very simple.

I have designed, made and sold hundreds of Veroboard prototypes (some very complicated) and every one worked perfectly.
The first breadboard circuit I made did not work but worked perfectly on Veroboard so I never used a lousy old breadboard again.
 
Do you think something like this could work:

Those little pads aren't big enough to get rid of the heat produced by the chip. Did you look at the suggested pcb layout in the data sheet I refered you to? The ground tab in that layout is there to act as a heat sink. Yours, not so much...
 
Hmm, so it says a different package such as WSON or MSOP can dissipate more heat. From what I can gather I would need to form a contact to a piece metal or on a PCB to dissipate some heat like you said, so SOT-23 isn't the way to go.

I would still require a contact to a heatsink/padding for the other packages, correct? If I ensured contact with some of the pads on the board that I have linked to (using WSON or MSOP) could that do the trick? What would you suggest using?

EDIT:

How about using something like this adapter (it actually has the required pad that you were talking about - check the first picture out to the right):

https://www.proto-advantage.com/store/product_info.php?products_id=3100086

with a stripboard?
 
Last edited:
The "surface mount" circuit board you found cannot be used for surface mount parts because the pads are too big and are too far apart. It is made to "tack on" through hole parts and many wires.

The "adapter" you found has a tiny pad to solder the chip to but it has very small surface area (that is covered by the chip anyway) so it is useless as a heatsink. A pcb will have a large copper surface area that can dissipate the heat to the air from a connected pad under the chip.
 
My money is still on the board layout. Even at 450C/watt the temperature should be well below the thermal cutoff.

Get everything as close as you can to the IC & the leads as short as possible - or - better yet use a strip board or their pcb layout.
 
I'll order the MSOP breadboard adapter just to be sure (since the MSOP is a bit more resistant to heat) and I'll try it with a strip board. I'll make video and try to make it clear as possible. It's probably going to take a week to get the new adapter. I'll also try the stipboard with the SOT-23 to be sure. If all of that fails then I'll have to try and make a PCB with surface mount parts, at this step I'm probably going to require a bit of help verifying the layout and parts selected (I'll be using Eagle CAD). As always thanks for the help, I'll get things in order and I'll post back as soon as possible.
 
Its only in recent years I've noticed some serious 'sensitive' boost converters (used for low current 30-50mA with 20V out all from 5v) - but when using adapters for the IC, and SMT inductor - they worked on breadboards, just not super efficiently. That said, inductors were connected as close to the chip as possible - no wires, just both plugged straight in. Input and output caps as close to the chip as possible, and no long feedback paths. Although with the FB pin connected directly to a low value resistors then to ground, it should be pretty immune to noise.

I've even made my own PCB's for testing just by cutting lines out of copper clad (even for SOT-23-5) as often the pinout lends itself nicely to a simple layout. As a half measure between custom PCB, and 'solder-less board', strip-board works. You can slice tracks length ways to give you 1.27mm pitch. It isn't pretty, but works for SMT parts, and keeps traces short. using adapters its even easier.

What inductor are you using? And diode? I highly recommend the 'over voltage protection circuit' (page 13 on datasheet) as it doesn't using up any power, just sets the maximum output voltage should you power up the circuits with no LED's attached (or no load). I wouldn't worry about temperature dissipation just yet. First get it lighting LED's :D
 
Status
Not open for further replies.

Latest threads

New Articles From Microcontroller Tips

Back
Top