Varying output voltage for SPI pins?

LeafLabs Garden » Support » Maple Hardware Support

Varying output voltage for SPI pins?

(28 posts) (4 voices)

Started 2 years ago by LostLess
Latest reply from crenn

12 Next »

LostLess
Member

Hello! I am currently building a project that involves a Maple Mini, a BlueSmirf Silver bluetooth, and six segments of an LPD8806 LED strip. The Maple Mini will consume a maximum of 30mA, the BlueSmirf will consume 25mA, and the six segments should consume about 24mA. When I have the system powered via PC USB (that is, the Maple Mini is powered by my PC's USB port), everything works. When I swap to use my 4.8V 700mAH battery or 5VDC @ 2A power supply (driving power to the Maple Mini via its Vin pin), only the Maple Mini and the BlueSmirf seem to receive power-the LEDs do not light up as they should at all, despite the fact that they are getting power from the same location as the Maple Mini and the modem (and not from the Maple Mini directly).

What might be the problem? The LPD8806 requires 5V for power and communication, and I suspect that when I am not using USB, the SPI pins are operating at 3.3V, which is too low, but I am unsure why this would be.

Here are descriptions of each wiring diagram.

Working configuration
Maple GND -> negative column on breadboard
Maple Rx -> BlueSmirf Tx
Maple Tx -> BlueSmirf Rx
Segment GND -> negative column on breadboard
Segment Vin -> Maple Mini Vin
Segment Clock Input -> Maple SCLK (SPI1)
Segment Data Input -> Maple MOSI (SPI1)
BlueSmirf Vin -> Maple Mini Vin
BlueSmirf GND -> negative column on breadboard
Power is driven from Maple Mini USB port.

What should also work (but does not)
Maple GND -> negative column on breadboard
Maple Rx -> BlueSmirf Tx
Maple Tx -> BlueSmirf Rx
Segment GND -> negative column on breadboard
Segment Vin ->positive column on breadboard
Segment Clock Input -> Maple SCLK (SPI1)
Segment Data Input -> Maple MOSI (SPI1)
BlueSmirf Vin -> positive column on breadboard
BlueSmirf GND -> negative column on breadboard
Power is driven from either 4.8V 700mAh battery or 5V @ 2A power supply whose leads are connected to the negative and positive columns on breadboard.

Thanks for your help!

Posted 2 years ago #
gbulmer
Moderator

LostLess - the SPI pins of the Maple Mini only output between 0-3.3V unless you provide external electronics (at least a pull-up resistor, or some sort of voltage-level shift).

My reading of the LPD8806 datasheet at http://adafruit.com/products/306
suggests that it needs 3.8V for a data high (http://www.adafruit.com/datasheets/HL1606E.pdf). It may work at 3.3V, but it might mean the circuit is teetering on the brink of not working. Please note, adafruit say that that datasheet might not be accurate.

So, one explanation might be the circuit needs pull-up resistors or voltage level shifting on SPI, and you have been a bit lucky.

If you have an Arduino, which runs at 5V, it should be easy to test everything works okay at 5V. You could then simulate driving the LPD8806 at 3.3V (or anything else you choose) from the Arduino using voltage dividers.

A good way to debug hardware is to simplify it, and remove anything that is not essential or might add confusion. Then, once the 'simplest system' works, start adding back one piece at a time.

Try only:
- single LED strip
- Maple Mini + pull-ups or voltage level shifters
- 5V 2A power supply
nothing else.

This might be simple enough for you to see what is happening.

Compare that with the same two components powered from PC USB. I assume you have some way to measure voltage as you say the SPI is running at 3.3V, what is it? An ordinary multi-meter might not read the true peak voltage, even on AC, if the voltage is not a sine-wave.

Posted 2 years ago #
LostLess
Member

Thank you for the informative and helpful reply gbulmer. I will give that a shot when I get the chance. In the meantime, I have a Logic Level Converter from Sparkfun that supposedly can bring the voltage up to 5 from 3.3V. Would this work for a voltage level shifter?

Posted 2 years ago #
gbulmer
Moderator

LostLess - "I have a Logic Level Converter from Sparkfun that supposedly can bring the voltage up to 5 from 3.3V. Would this work for a voltage level shifter?"

Please post a link to the product so that we can be sure. However the last time I looked, Sparkfun did have something with that sort of name that is for exactly this job.

You can actually do level-shifting for less than $0.20/signal with a MOSFET, but if you have something ready to go, use that. You can always circle round later, when everything is stable, and improve it.

Posted 2 years ago #
crenn
Moderator

He's possibly talking about these:
https://www.sparkfun.com/products/8745?

If so, you can use them fine for SPI or I2C, I've used them before. It is also possible to use pull up resistors and put the pins into open drain (OUTPUT_DRAIN if I recall) meaning the pull up resistors will pull up the signal to 5V with minimal hassle. However ensure the pins you do this to are 5V tolerant.

Posted 2 years ago #
LostLess
Member

Hello crenn,

Yes, that is what I was talking about. The pins for SPI1 are not 5V tolerant according to the specs of the Maple Mini.

@gbulmer:

I am not all that familiar with MOSFETs...I think I will just run with the logic level converter for now. Thanks!

I will keep you all posted.

Posted 2 years ago #
crenn
Moderator

If you're curious, that Logic Level Converter is actually using MOSFETs, you can take a look into the schematic, there is a application data sheet somewhere which explains more. Published by Phillips (NXP now) I believe, I can dig it up since I have it saved somewhere.

If you can, you could use SPI2 which is 5V tolerant.

Posted 2 years ago #
gbulmer
Moderator

The NXP application note on I2C level shifter is at http://ics.nxp.com/support/documents/interface/pdf/an97055.pdf
a subset of the same technique is in http://www.nxp.com/documents/application_note/AN10441.pdf

The usual SPI pins are not 5V tolerant, however I glanced at the STM32F103xB datasheet, and SPI2 and the alternative pins for SPI1 are 5V tolerant. It'd be easier to use the level shifter.

Posted 2 years ago #
LostLess
Member

I have no problem re-writing my code to use SPI2; I will do that next. Unfortunately, I can't find the logic level converter, so it looks like I have to order a new one.

In the meantime, I feel confused about this I2C level shifter (pardon my ignorance-I am an experienced software person, but a novice electronics person). Are we talking about a component on the logic level converter or something else that can be used in place of it to get the 5V? The logic shifter would be more appealing to me if it did not require two power sources like the logic level converter does.

Still, I wonder-is it possible that the Mini is outputting a slightly higher voltage (~3.8V) when powered by USB? I can't believe I have been so lucky all this time.

Posted 2 years ago #
ala42
Member

>The logic shifter would be more appealing to me if it did not require two power sources like the logic level converter does.
The level shifter can not magically pull the signals to the desired levels if you do not connect these levels to the shifter. It can not guess the needed voltage levels and make up the unknown target voltage out of fresh air. This is more about logic than hardware :)

Posted 2 years ago #
LostLess
Member

Update, 2/9/13:

I ordered a logic level converter that does 3.3 to 5V (here are the specs: https://www.sparkfun.com/products/8745). The logic level converter requires two power sources: one at the lower voltage, the other at the higher voltage. For the higher voltage source, I am using the 5V adapter. For the lower voltage, I am using a 3V CR2032 battery. I connected the SCLK and MOSI pins to the TX lower voltage lines, and I connected the LPD8806 CI and DI pins to the TX lines of the higher voltage. It sounds contradictory to do so, but according to the description of the converter:

Pins are labeled as Inputs and Outputs. These are relative to the board. A digital one going into the RXI pin on the 5V side will show up on the RXO pin on the 3.3V side as 3.3V. A digital one going into the TXI pin on the 3.3V side will show up on the TXO pin on the 5V side as 5V.

Given that, still no joy. However, I think there may be a newer problem. The CR2032 is outputting closer to 4V (the last readout was something like 3.89V). Might it be that the power supply for the lower voltage is too high for the output of the Maple Mini?

Posted 2 years ago #
gbulmer
Moderator

LostLess - could you just use the 3.3V output by the Maple Mini instead of a battery?

Posted 2 years ago #
crenn
Moderator

The text of that page is a little misleading. If you look at the data sheet, you'll see a voltage divider used for the 'RX' pins. So if 5V is put into the HV side, you'll actually get 2.5V out, not 3.3V. This doesn't matter since for the STM32F103, the lowest voltage to be considered a HIGH state is 1.546V for 5V tolerant pins and 1.833V for all other pins. This information can be found in the Datasheet for the STM32F103 in section 5.3.14 I/O Port characteristics. Not really related to your problem, but something you should know.

Can you use pastebin so we can look at your code? Are you using Adafruit's library?

Posted 2 years ago #
LostLess
Member

LostLess - could you just use the 3.3V output by the Maple Mini instead of a battery?

Not long after posting my update, I realized that Vcc could work. Is that what you are referring to? I could use it, and I will try later today.

Can you use pastebin so we can look at your code? Are you using Adafruit's library?

Here's the pastebin to my code: http://pastebin.com/4Xt5gL2W. I'm not using Adafruit's library because it looked like it needed translation (isn't the language used in the Maple IDE a subset of C/C++?); I may reconsider using it at a later date.

As for the information about the Rx's, thanks crenn. However, yes, I don't think that this is an issue here as the communication is one way between the Maple Mini and the LPD8806.

Posted 2 years ago #
gbulmer
Moderator

LostLess - yes Vcc.

The Maple IDE language is close to Arduino IDE language, but it isn't identical. LeafLabs aimed to make some of the extra features of the STM32F available, so it wasn't going to match exactly.

Arduino and Maple both use C++ (not a subset). A source of confusion is they use a bit of pre-processing to remove the need to #include a header file, and it auto-magically creates function declarations at the top of the source file (which does make it easier for newbies to get started, and oldies to hack something together quickly). There is a bit of error message processing too. Otherwise, it's C++

Posted 2 years ago #

RSS feed for this topic

12 Next »

Reply »

You must log in to post.