Ili9341 power and restock available

Status
Not open for further replies.
Duhjoker

Duhjoker

Well-known member
powering options and general power consumption questions ili9341 and teensy3.6

Hi guys

I just had a couple questuons about the ILi9341 tft. In most of the ads for the same type on other sites, it says that they can be powered by 3.3v or 5v. Does that mean i dont have to use 5v to power the tft, use 3.3v for both the LED pin and the vcc. Or does the vcc have to 5v?

Also when can we expect a restock of the tft's?
 
Last edited:
Ok I do have the 2.8 LCD ili9341. But had to enlarge my handheld case from 132mm to 163. Which is long and not very pocket sized.

So the main point of my handheld DIY gaming project that I've been trying to develop is that the hardware must be easy to find and affordable and have an easy to use library to use for gaming. The work you and Kurt have done with the ili9341 libraries has very rapidly boosted my work with new programming that pushes the TFT to its limit. Screen buffer and some stuff as an example.

What would you reccomend for a TFT that should be around for a while? I can go up the next resolution at 480 X 320 and use the 3.2 to 3.5 inch screens which I can do but I don't know if the teensy could process the graphics for the upgrade. Of course I could go esp and utilize the 4mb of flash which would vastly improve how much bitmaps can be used but that's whole new set problems to learn.

Any way I what would you reccomend I do at this point. I need a lil guidance
 
ok can you tell me the current draw in mah that both the screen and the teensy take. I'm trying to calculate battery size and draw. Right now I'm using a 3.7v lipo stepped up to 5v and I get 40minutes of battery life.

Also is there a way to use a couple pins for battery level indication?

Edit::::

Why do the screens get hot in one corner? Today it melted my plastic a lil.
 
Last edited:
Ok quick catch up..... I'm trying to develop an easy to use gaming library and set of hardware to allow any one, even if you are new to programming to easily make games for a handheld console you build your self.

I managed to get the console working minus a couple buttons but I am only getting 40-45 minutes from a 3.7v lipo at 1500mah stepped up to 5v powering both a teensy3.6 and an ili9341 2.2tft. I've been looking around trying to figure out how much current draw both items use but I seem to be coming up short.



I did see that the LCD can be powered by 3.3v only so that will make a difference but I also would like to use the 3.3v input voltage instead of 5v but kinda can't figure out How to supply 3.3v since most 3v lipos start at 3.7v and higher. Kind of odd.

Edit:::

Ok I see on the card that it can be 3.6 to 6.0v's. So does that mean if I want to use the 3.3v logic I can just plug in my 3.7v lipo straight up?

Also just to make sure is the 3.3v (250 max) above pin 23 an out put or an input?


What other ways can I get more battery life?

Edit::: i ask about a battery monitor pin because the iriginal gamebuino code utilizes a battery monitor code and uses a pin i found in the schematics here.....

http://gamebuino.com/wiki/images/6/6f/Gamebuino_r2_schematics.pdf

It also uses a crystal but i dont know exactly for whar or how to utilize it on the teensy3.6
 
Last edited:
3.7V lip are going to be tricky, since they will cross through the 3.3V target voltage as they discharge. So you can't just plug them into the 5V output and let the regs deal with it and you can't just plug it into the 3.3V directly.

At the moment about 1/3 of your power is being burnt as heat by the regulators on the board, the display and the LED (5->3.3 = 1.7V drop across regulator that's turned into heat).

With the current draw you really need to mod things to measure that yourself, since it's heavily hardware dependant and you will be makeing tweaks to improve things and watching the results.

The ref card 250mA @ 3.3V talks about output, where the onboard reg is driving the Teensy+your hardware.

At the moment there are three places to tighten things up:
Optimising the display LED so it's just as bright as it needs to be, and as much of the power going to it is actually turning into light as possible. If it's on 5V, try moving to 3.3V, and since that'll probably make it too dim try tracing the tracks and putting a lower value R on it. Make sure there aren't any other components on the LED chain sucking power

Swap to using all 3.3V power from a buck/boost converter, with the highest effciency you can find at your operating current (check, since many quote a best case value that me not be near your design current). Then use that 3.3V to power the Teensy and LCD. To do this you may need to hack the LCD to remove/bypass an onboard 5.5->3.3V reg.

Get your Teensy sleeping - see the snooze library and look at slotting in sleep time whenever your code is not preping the next frame, The sleep current for a Teensy is close to zero for your purposes, so from there you try to strip out instructions that you don't actually need and let it sleep .

Which one you use depends on what you find when you put a meter on your system, go for the part that's pulling the most load to start with.
 
Duhjoker said:
I managed to get the console working minus a couple buttons but I am only getting 40-45 minutes from a 3.7v lipo at 1500mah stepped up to 5v powering both a teensy3.6 and an ili9341 2.2tft.
Click to expand...
That sounds way too low (maybe your display is defective or your wiring wrong). The Teensy @ 180mHz should draw around 90mA, the screen from the numbers I have seen less than 50mA. With a 75% efficient boost converter, that should be 5-6 hours battery life.

Some Lipo batteries are mostly empty at 3.5V, so depending on the battery, powering Teensy + Display directly may be an option.
 
lol yea it's low. I run my pi's on the same batteries and get at least 2 1/2 to 3hours. But I do use an adafruit powerboost 500 for them.

I dont know about hacking the screen. My skills arent totally there. I cant even solder an smt component to a board. I just want to give back to a large community of electronics enthusiasts and makers for all the help ive been given over the years.

I don't think it's the wiring or at least shorting. I try to set things up where its modular and uses jumper wire ends as connectors with shrink tube around the ends.

The charger for the battery is a TP4056 and the step up can be found here......

http://pages.ebay.com/link/?nav=item.view&alt=web&id=172420743783&globalID=EBAY-US

For the circuit I attached two sets of wires to the charger, two for the battery and two that lead to the step up then the step up to the t3.6 with a switch Inbetween the step up and the charger/battery so it's not expelling energy while it's not in use.

The SPI wires match the scheme from the ili9341 that's being sold at the pjrc store. The button wires I try to set where there is a pin space in between each wire.

Maybe I'm not supplying enough amperage from the step up and it's sucking the power down to compensate for what it's not being given? I do have a DROK 3A step up i can use if that's the case.
 
Last edited:
Simple check if your not up for putting a meter in series is to check for anything running hot. My math says 1.5 Amp hours in 45 minutes is 2 amps draw, at 3.7v is 7.4 watts. Given likely consumption for the Teensy+ screen is 0.1amps finding the part that gets the hotest may be a low tech way to do this.

Also potentially useful is a lipo fuel gauge that tracks actual current vs current out, at least for your prototype
https://www.sparkfun.com/products/10617
Though I think I've seen better ones that actually track current. Useful since if you start doing sleep and similar things current draw will pulse and be hard to track on a meter.

Given the limited information you've given (no actual measurments on your system) my guess is that your boost converter isn't doing a very good job, and most likely neither is the LED power circuit in the display. The LED test could be as simple as loading code that writes to EEprom once a minute to increment a counter and leaving it to run with and without the LCD backlight connected.

With the power supply that last paragraph needs some answers to get your mental model in a better position.

If you can't supply enough current to the boost converter it can't make more current, the output voltage will droop and/or the PSU shuts down.

If your battery can't provide the current that the boost converter needs, it's voltage will droop and may hit the point where it's protection circuit trips and the whole thing dies (so possible you have a surge load that's dropping out a battery still at half charge, though your design should have no surge loads at the moment).

Boost converters can vary widely in efficiency, and that efficiency only applies at one current, and that current may not be Imax. If you can't get a I vs effeciency graph then that's a warning sign.

You don't care what the total capacity of the boost converter is, you want the one that is operating at 70-80% while there, hopefully with some head room. So dropping a 3A unit in there may not help, unless it's good at feeding <300mA loads.

And if you boost to 5V you will still be wasteing 1/3 of the battery capacity in the step down regulators on the Teensy and LCD, though at the moment that 60 or so mA is not a big part of the loses.
 
quick check on polulu got the graphs down near the bottom of
https://www.pololu.com/product/2565
So for those devices peak efficiency is well below Imax, and for that randomly chosen model boosting 5V from a ~3.7 supply it'll be above 90% efficiency between 100mA and 1 Amp. So not bad, but good design will hopefully take your operating current below 100mA where a different device might be better. Edit - these graphs should in no way be seen as representing what you have in front of you, you need to find or measure it to make decisions.

If all this is seeming complex remember a number of major corporations specilising in this area have managed to get this wrong in recent years, to the point of things exploding. And you haven't set anything on fire yet, so you are doing better than them.
 
Last edited:
One corner of the tft does get hot. Yesterday when i left it on to see the actual time till it died and there was enough heat that it had melted and warped the case it was imbedded in.

I found this lil doohickey here........

https://www.ebay.com/itm/121868205078

Should drop my voltage from 3.7 to 3.3v
 
Nearly all the display's power is for the LED backlight. If you can PWM control the backlight, you can reduce power (and heat), but of course the result will be lower brightness. Maybe that's worthwhile?

On the availability of parts, the reality is we're all pretty much at the mercy of whatever a small number of huge factories in China decide to do. For years it seemed we would always have 2.2 inch displays, then suddenly the supply completely dried up. Nobody can say how much longer the 2.8 inch size will last. Robin and I have considered buying a large quantity, but that's quite risky for our tiny little company. Right now we tend to keep about 100 to 200 in stock.

I'm also considering making a 480x320 display, similar to the one Adafruit sells, with SPI interface on the bottom side and 8 bit parallel on the top. But even if we get a custom made 3.5 inch TFT from China and assemble onto a PCB here in the USA (which is probably what Adafruit is doing), long-term the supply could still dry up if the upstream companies decide to stop making the 3.5 inch glass or other materials.

The only way to be 100% sure of a long-term supply is to buy all the inventory you'll ever need up front. Who wants to take that kind of crazy financial risk?!
 
One corner getting hot enough to melt isn't right, Since there isn't any photo's would guess that the backlight is intended for 3.3V and has been connected to 5V.
 
Yea I do have the Vcc and LED pins connected and supplied at 5v. Guess that was a no no. I don't mind the contrast being a lil lower and I think I have the contrast as pmw and set at 60%. I think I have to look at the code again.

All right I had dig out my meter but I tested the out put of my stepup and it shot up to 7.04 and 7.08v!!!! How has that not killed my teensy. I can test other stuff but you will have to instruct me on how to to do each test.

Thanks Paul for your explanation on the restock.

Hold on for pics.........
 
better get another psu, i use a 5v 30a dc-dc psu that runs teensy, servo, 2 lcds, and quite a few i2c/spi devices 24/7, the psu is epoxy sealed and waterproof with input voltage between 7.8 to 40v and output is always solid 5v
 
Ok here are some pics

This shows the screen in its case for orientation.



If you look at the left side of the screen you can see how the plastic warped up a lil. You can also see where the hot glue melted the case when installed the screen.



Can you link me to a psu you use or would reccomend?
 
If you measured unloaded the 7V is not surprising because of the way the boost circuits work. if it was 7V with the LCD etc connected that would be very much a bin job. Useful tests at the moment are the voltage on your rig while loaded, and with meter set to current (will need to move leads and a 2Amp scale) measure in series with the output of your boost converter (red or black doesn't matter, whichever is easiest to access).

Swapping your LED input to 3.3V will also be interesting. Different boards are around but the cheap way to build it would be white LEDs in parallel with a single series resistor. Given the ~3V drop of white LEDs that R would be quite small giving a fair heat dump when run on 5V
 
Ok so when I test the step up at full battery connected to the teensy and at the module the voltage goes 1045. I measured several times and those are the numbers that came up. No decimal either.
 
Would need some information on how you have you meter setup, since thats not a voltage measurement. Normal method when testing something unknown and especially if a measurement doesn't make sense is to test something known (like USB 5V or such) first and sort out what your meter is up.
 
Heres a pic of the multimeter.......



It was a gift and i have never really used one. I have to turn the knob to 20v to get a reading, if i put it on 2v i get the low voltage signal error or what ever. 0L.

It looks like it can measure Mah but i havent figured that part out yet.
 
Last edited:
Should have come with a manual which might be worth a read, or doing some googling. Happy to help but I shouldn't be having to tell you the 'OL' means 'overload'. Fortunately with current generation meters that doesn't normally result in a bang.

Your next step is to find some known voltages to measure (Teensy has 5V and 3.3), and find out if it's actually accurate if not not which way. With cheap meters watch for leads that only work at a specific angle due to dodgy wiring.
 
Ok so with the 5v going in to the t3.6 vcc input i get 4.58v out through the 3.3v pins.

Today i connected the tft to screen using all 3.3v power and everything comes on but no picture. Maybe i need to power the VCC with 5v and the reset and LED need 3.3v?

The brightness is fine though and not to dim when it all comes on.

Edit::: oh yea the sleep programming funtions are already included in the library

View attachment TFT_ILI93XX-master2.0.zip
 
Last edited:
Ok so now i cant get a picture at all. I switched to the bigger 2.8tft and hooked all the wires up then connected like i always do but no picture. Both the teensy and the screen come on though.

I took some pics. The first is the screen wiring.



Bottom to top,

Red-----vcc 5v
Black ------gnd
Orange-----CS------10
Yellow ------reset----3.3v
Green ------D/C -----9
Blue--------mosi-----11
Purple------SCK -----13
Grey--------LED-----vcc 3.3v to 5V
Brown------miso -----12

Here is the the ternsy. The red and white pins are connected to a small board i made with pins to test and use to power so it acts like bridge. The red is ground and the white is 5v.



Here is a poc showing the bridge...


Here is a pic of the step up. Green amd grey is ground, red and white are positive.

 
Something is pretty wrong with with the meter, the teensy or where things are connected to have 4.58 on the teensy 3.3V pin. If real the teensy is really dead. Since it seems to be still working means you can't trust your measurments. You really want to sort that out since pretty much all chasing of your power from here will involve measurments.

The LCD should be able to run on 3.3V, but try the 5V on VCC to start with, your LCD is different to mine but if you get that far check the solder jumper just above the little black object in the photo above, on the edge of the tape. Most likely the three pin device is the onboard regulator and the jumper is how you bypass it for 3.3V use. Don't try that until you have a much better handle on your wiring though.

Other than sorting out the meter next step may be to see if the display still gets hot now the LED power is 3.3V. And getting it to communicate again.
 
Status
Not open for further replies.
Back
Top