Maple with PWM and onboard LED clarification requested

LeafLabs Garden » Support » Maple Hardware Support

Maple with PWM and onboard LED clarification requested

(17 posts) (4 voices)

Started 3 years ago by juzza_s
Latest reply from siy

12 Next »

juzza_s
Member

Hi all,

Before I write any code, I've been looking at the online documentation to try to get my head around PWM outputs, timers and LEDs with my shiny new Maple board and I beleive that I'm misunderstanding something quite fundemental.

I'm currently trying to understand the relationship with PWM and timers, using the libmaple. The documentation for pwmWrite (http://leaflabs.com/docs/lang/api/pwmwrite.html#lang-pwmwrite) gives a code snippet for using the on-board LED (BOARD_LED_PIN) which seems simple enough.

Now, BOARD_LED_PIN points to pin D13.

According to the Timer Pin Map (http://leaflabs.com/docs/hardware/maple.html#timer-pin-map), there doesn't seem to be a Timer that can be attached to pin D13 for PWM.

So, basically, my question is: is the code sample for pwmWrite wrong and that the onboard LED at D13 can't be used for PWM? Or am I being stupid and missing some obvious bit of information with the relationship between the timers and PWM? Does libmaple use a software implementation for PWM with pins that can't use the PWM feature of the STM32 chip?

To be honest, I'm not fussed whether the onboard LED is capable of hardware PWM, as my project doesn't require it. I shall be using another pin, with han IR LED, for that part of the project (a 38khz IR transceiver). I'm just trying to understand how it works. :-)

I'm not going to get the chance to be able to play with the actual hardware for a few days so I thought I'd try clear it all in my head before I do...

Thanks for listening, and best regards,
J.

Posted 3 years ago #
feurig
Member

Bummer. Pin 33 (the led pin on the mini) is perfectly pwm capable.... I havent dug very deep into the underlying code but i I sincerely hope that analog write is left to the pwm capable pins (and god knows there is enough of them). Software pwm is just stupid on a system with perfectly good hardware and better things to do.

Posted 3 years ago #
feurig
Member

on the other hand the use of pin 13 which is the serial clock for the spi used in the arduino's ISP is subsequently also the serial clock on spi1 for the maple and maple ret6
So it makes sense that it would not be pwm capable. Again we suffer the arduino compatibility here.

Posted 3 years ago #
siy
Member

@juzza_s, you're right, LED pin on Maple is not PWM capable (same is true for Maple Native and Maple RET6 edition). But at Maple Mini LED pin supports PWM. Actually this is the main reason why my Mini64 board has two LEDs - one connected like in Maple and other connected as in Mini. IMHO, in list of design flaws of original Maple lack of PWM support at LED pin, probably, should be right after weak voltage regulator.

Posted 3 years ago #
juzza_s
Member

@feirig and @siy: thanks for the response. At least I now know that I'm not going crazy with reading the documentation. :-)

It sounds like the code sample for pwmWrite() is still perfectly valid for the Mini. How do I go about submitting a change request to add a note for this on the pwmWrite)() document page?

Posted 3 years ago #
feurig
Member

Funny I was just revising our maple mini design with a x1117x regulator (900ma) design since the wireless module we are using draws 250 ma when talking.

Posted 3 years ago #
siy
Member

@feurig, yeah, I've also added x1117 to last version of my Mini48U (Maple Mini clone).

Posted 3 years ago #
gbulmer
Moderator

Should we try to use a similar regulator?

Having commonality on parts would help folks to DIY, and we would accumulate experience faster.

Does that also protect from reverse voltage, or is a diode still needed?
(I can't see any mention of reverse voltage protection on an EXAR part, which some regulators have)

Posted 3 years ago #
siy
Member

@gbulmer, if board has enough space, then definitely more powerful regulator is better. I think LM1117 (and several clones) is one of the very common and popular parts, despite the fact that it is not the best one in its class. Many (if not all) versions of Discovery board for different MCU's use it as well.
About diodes. I'm using diodes in all my boards even if regulator has reverse voltage protection (for example, wonderful automotive grade LM2937 can protect not just from reverse voltage but also from +60V/−50V(!) load dump transients). In fact these diodes are necessary not just to protect board and regulator, but also to decouple external power supply and USB VBUS line. I think this function is even more important than protection of board itself, especially taking into account that external supply can have much more than +5V.

Posted 3 years ago #
gbulmer
Moderator

siy - yes, very important point about protecting USB VBUS line.

I have diodes on all (two:-) voltage input lines (USB or header).
I wasn't planning on removing the USB VBUS diode.

I am looking at (low-cost) ways to supply power from the header.
I'd like to run with 4 NiMh AA (or AAA) cells, which is nominally 4.8V, but dips with motor load.
The 1117 has a 'spec' which says drop out is 1.1V (or even 1.2V) drop out.
So 4.8V - 3.3V -1.1V doesn't leave much headroom for a diode and a battery voltage dip under motor load.
Pete (micromouseonline.com) says folks have had problems with micromice where power dips have caused processor reset.

Also, because of the design of my board, I am trying to find a through-hole regulator to put 'on the back'.
I'd like it to be capable of supplying 500mA from any voltage from 4.8V upto 9.6V.
The best idea so far is to either:
a. use two regulators, one to reach 5V, then a second to get down to 3.3V (cheap, simple, and easy for folks to vary), or
b. use a 'switch mode power supply'.

Option a. adds even more voltage drops, or two Vinput pins (confusing).
Option b. seems to require quite a lot more design, which may put folks off using a design and messing with it, and harder to get parts.

Posted 3 years ago #
siy
Member

Seems I was wrong about Discovery boards, only VL uses 1117, others use smaller regulators. Anyway, the 1117 is quite common, although it is far from the best ones. Mentioned above LM2937 is much better, it has typical drop out only 0.5V. In cases when input voltage range is narrow (less than 7V), I prefer LM3940, which especially designed for 5->3.3V regulation and capable to source 1A.

As for switching regulator, I've experimented with LM2734Y (with lower switching frequency) and it performed very well. Entire regulator consumes board space comparable to space consumed by linear regulator and it capable to source 1A in wide range of input voltages. I've calculated schematic for 5-15V range, for example. Also, TI provides online design tool (Webench) which simplifies design a lot.

Posted 3 years ago #
gbulmer
Moderator

siy -

... I've experimented with LM2734Y (with lower switching frequency) and it performed very well. Entire regulator consumes board space comparable to space consumed by linear regulator and it capable to source 1A

What size components are you using to fit it into a small space? You seem to have it smaller than my estimate.

Posted 3 years ago #
siy
Member

To achieve smallest possible consumed space I've used 0402 resistors and smaller value capacitors. Capacitors at input/output were 0805. I don't remember exact part number for inductor, something like CDRH3D28 or CDRH4D28 from Sumida. Using all 0805 resistors/capacitors would not affect consumed space significantly. As I've mentioned design can be done with Webench and it suggests particular part numbers depending on required parameters and preferences (board space/price/efficiency).

Posted 3 years ago #
gbulmer
Moderator

siy - Are you planning to publish a schematic, or is it better to go to TI Webench?

A small 'switcher' may be a handy add-on for existing Maple boards; I think I remember three different people who have blown regulators.

Were you using the LM2734Y with a microcontroller, or was it a different project. If it was STM32, did you use a linear regulator for the analogue power, or take it straight from the LM2734Y?

Posted 3 years ago #
siy
Member

@gbulmer, I've not used this IC beside separate experimental board. I can send you calculated schematic and Eagle layout for that board, if you're interested. Also, while LM2734Y looks very interesting, it is quite expensive (at least in local stores) comparing to other IC which I'm going to try - NCP3170. This IC about 1/4 of the price of LM2734Y in local stores and capable to source up to 3A. It somewhat larger (SO-8 vs SOT-23), but requires smaller inductor (4.7uH vs ~18uH).

Posted 3 years ago #

RSS feed for this topic

12 Next »

Reply »

You must log in to post.