MPLAB Snap $15 programmer/debugger

[macboy]

I haven't seen this mentioned here yet, I just noticed this new programmer/debugger from Microchip, the "MPLAB Snap". It appears to be a cost- and feature-reduced version of the PICkit 4, built around the same 300MHz SAME70 MCU. Supports PIC, dsPIC, and AVR, but fewer devices currently than the PICkit 4 supports. At less than $15, it is definitely the cheapest programmer I've ever seen from Microchip.

Product page:
http://www.microchip.com/Developmenttools/ProductDetails/PG164100

Comparision to PICkit 4:
http://microchipdeveloper.com/snap:versus-pk4

Any first hand experience?

[tsman]

Not worth it IMO considering the PICkit 4 has much better support for devices and it isn't that expensive. The supported device list is far smaller and mostly consists of a limited number of PIC devices with a handful of AVR parts. Popular chips like the ATmega328* aren't supported.

If budget is a problem and you're only using PIC then buy a clone PICkit 3 for the same price as the MPLAB Snap and has most of what PICkit 4 can do and a few extras. Most people seem to prefer the PICkit 3 anyway.

[JPortici]

not yet, but i'll be looking fo an excuse to buy one..
altough, with the PK4 being cheap-ish i don't think i'll bother.

This should be compared against st-link / popolu avr programmer / onboard debuggers on devboards and the likes, everything it lacks is not present on those debuggers either

part support is of course reduced BECAUSE IT LACKS HIGH VOLTAGE PROGRAMMING! only LVP or simillar modes supported and this excludes ancient parts
part support is of course reduced because it's come out a month ago or so. And it's still a struggle with the ICD4/PK4 because they fundamentally changed how the programmer works
citing from the microchip forum

There were some key changes in how the programmers work when we went to the Gen 4 tools. In essence, the tools actually run scripts that execute primitives on the tools, functions for SWD, DebugWire, ICSP, and other protocols. This prevented the issues with Gen 3, where it had to have something programmed into it every time you changed part families. As someone who used to bounce between PIC32 devices, it was very annoying.
 
The scripts themselves don't exist on the tool, but rather within the IDE, and the IDE downloads each script to execute, depending on what is happening (DeviceID, Erase, Program, Verify, etc.). To support PTG, we have to figure out what has to be placed on the SD Card, what the script execution looks like, and then implement it in the firmware. Also, we're having to backfill support for existing parts, since not all parts are supported yet with the Gen 4 tools, and support new parts. So it's all a balancing act of features we would like to implement, versus features you need to have.

on the product page there should be an email address to request specific part support.
when the PK4 first came out, i noticed it lacked support for ALL codeguard-enabled dsPICs, i sent an email from my provate gmail account asking for a specific dsPIC i use a lot, i was answered the same day and on the following release support for only that part of the family was added.

This means that since they are way behind schedule with the part support, each user asking for a specific device tells them which they need to focus to. Just complaining there is no part support won't benefit anyone

[cv007]

It does show they know they had a gap to fill and are trying to fill it (although a long time coming).

I think I posted at one time they should be handing out programmers like candy-

I would rather see them take the cheapest pic32 w/hs usb, <5x5cm size, simple design, cut the parts count down, open source the hardware/software and let the best ideas float to the top. Sell them like candy, toss them out at parades, put them in vending machines. They could bundle this pk-mini with all future curiosity boards also (I don't see them slapping a $15 retail sam micro on future curiosity boards). 

I guess we are closer,.

[VEGETA]

I have PICKit3 and I don't use it anymore due to using Arduinos. So, I will not buy one unless it has a nice enclosure and supports AVR chips. haaa!

[JPortici]

well they ARE working on a replacement for the PKOB, probably based on this..

[cv007]

well they ARE working on a replacement for the PKOB, probably based on this

but in my mind would be a waste of time to keep redesigning the same thing over and over again (layout at least). They now have a cheap programmer/debugger they can hang on any future board- over, under, vertical, horizontal, whatever. One programmer board, one design, one entry to find in mplabx, one vid/pid, one schematic, one set of problems to deal with, one fixed set of parts, one set of documents, and on and on. The only requirement for a curiosity board would be the simple 8 pin header which I imagine would make cranking them out as simple as could be. The $30 curiosity boards could now be $15, and you could buy the $15 SNAP as needed- or maybe not at all if you have a pickit3/4.

But...since they don't want to let you use the usb power
(for my use, I will probably take some wire, jump from 5v0 / 3v3 to the unused pins 7/8 so I can power from usb)

[JPortici]

(for my use, I will probably take some wire, jump from 5v0 / 3v3 to the unused pins 7/8 so I can power from usb)

yeah, i suggested they update the board with at least a jumper or a solder jumper for the (already on board) 3v3 or the (already on board) 5V/VBUS
Hope they listen, because IMHO the most people going for cheap tools don't care for different voltages (see arduino crowd, but not only)

[NorthGuy]

yeah, i suggested they update the board with at least a jumper or a solder jumper for the (already on board) 3v3 or the (already on board) 5V/VBUS

They have a big chip and USB current is limited.

[JPortici]

most of my boards require less than 30 mA

[AloyseTech]

Maybe this is off topic, but could it be used as a SAME70 devboard?

[malagas_on_fire]

Still prefer also the pickit 4, but using pickit 3 and 2 . I'm not on a rush for a new programmer, but for newer to microchip maybe from a startpoint can be more affordable, e.g. schools .

[ataradov]

Maybe this is off topic, but could it be used as a SAME70 devboard?

There is no way to permanently lock the SAM E70, so you can erase and reprogram the device. But what are you going to do with a board that has 10 I/Os? SAM E70 is cheap and making a breakout board is easy enough. Why bother?

If you are interested, I have an Eagle project for a simple SAM E70 breakout board that brings out all pins.

[AloyseTech]

This would be a candidate for the cheap ARM development board list here : https://www.eevblog.com/forum/microcontrollers/list-of-cheap-arm-development-boards/

10 i/o is sufficient for some (relatively) high speed data acquisition project I have. The high speed USB together with the 300MHz core allow data processing and streaming without going the FPGA way. I agree that this is some rare use case. It is more of a curiosity than an actual useful idea... 

[cv007]

I ordered/received 2 of these snappers.
Need MPLABX 5.05, so headed over to mchp for the download- and discovered the webmasters must have been running out of things to do- they now have an ipod 1.0 interface

initial report-
this thing is a lot faster than pk3/pkob (and it should be- usb hs vs usb fs)
on a pic32mm (only thing I tested so far), debugging is actually now usable where before I would say its doable but not very pleasant
click debug- about 3 seconds later (32k code) I'm compiled, programmed and stopped at main (pkob, same project 17 seconds to get to same point)
you can step over functions, view registers/vars, and do things in a responsive way that before would cause you to wonder if anything was going on after a button was pressed

I would say its $15 well spent.

[JPortici]

^ question: can you add/remove breakpoints without having to pause/resume execution like you can do in the ICD?

[cv007]

Yes.

[JPortici]

Awesome.

[cv007]

(for my use, I will probably take some wire, jump from 5v0 / 3v3 to the unused pins 7/8 so I can power from usb)

Now that I have it, a better idea (it seems to me) is just to jumper 5v0/3v3 to a 3pin header (outside ends), center pin to pin 2 of icsp connector (Vdd), then use a 2pin jumper/shunt to select 5v or 3.3.v (or no jumper for no power).

The enclosure I happened to have (Hammond 1551LTBU, $2.23ea) fits perfectly for width, is a little longer than needed but then I have room for a little perfboard for my little mod. The little rubber/silicone feet on the snap board and the case bottom pc board mounts (2) combine to keep the snap board in place (snap rests on pc board mounts, 1 pad keeps from moving as it is right up against mount).

I had to make due with what I had, so I cut a perfboard to fit the unused space, added a 8pin m-m header (bent to make right angle) so the board can be detached, added a 5pin keyed icsp connector as I will only need the 5pins and wanted a keyed connector as that makes more sense. For the 5v0 and 3v3 source points, I just put in a single pin header for each (bent at right angle) and made a couple wires with the female header connector and soldered them to the add-on board. The add-on board can be easily disconnected if I want to something different or go back to original.

I still need to make a 5pin-5pin jumper (keyed to non-keyed), and cut some openings in the case.

It does work, so will now be able to use usb power. The high side power chip they have seems to have been set to about a ~800-1000ma current limit, and I think the 3v3 regulator can handle at least that much. I don't know what the sam and the rest of the circuit consumes, but I would guess a few hundred ma for my own use should be no problem provided the usb port is willing.

One thing I have noticed- the snap likes to disconnect/reconnect from usb when power is applied to Vdd (icsp pin2), so I'm not sure what is going on. The only thing the Vdd pin does on the snap (as far as I can tell), is simply detect the Vdd voltage of the target device to adjust the other pin levels. The same thing happens to my modded snap- with nothing connected I can connect usb, then I place a jumper on my mod to get power and it will disconnect/reconnect from usb. I guess there is either something going on the Vdd detection- either causing some kind of power problem or maybe the sam is simply doing  disconnect/reconnect when Vdd is detected.


edit- I'm going to also try to make a serial-telnet adapter with an esp32-wroom-32d I have so I can eliminate the usb serial-ttl I always have hooked up for high level debug/info. I ordered the wroom not realizing what I was (not) getting so have to create my own 'dev board' (from remains of my power mod board). Good thing only 6 pins needed.

I'm not sure what the best option is now for esp32, but I downloaded arduino and the arduino core for esp32 and have 'hello world' spitting out the uart.

[cv007]

I now have a serial to telnet adapter using a esp32-wroom-32d. My 'dev board' hack doesn't look very nice but it works (I bought the wroom without enough thought, although it turned out ok). I used arduino to program the esp32. The arduino esp32 core seems to have enough examples, so with about 150 lines of code I have 2 telnet servers- one port for transfers between the esp32 serial port 2 and the telnet port, and the other which will be used for any command/info (change baud rate, etc.).  So far has been running ok, but will see how reliable it is over time.

I also have an Onion Omega2 ($10) that I could use which may possibly be a better idea. I have one that I simply use for mp3 streaming and it runs good- never a problem for over a year and I use it all the time. I think all I would need to get serial-telnet going on it would be a simple socat command (socat is not installed, but I'm sure openwrt has it available).

If this all works out ok, I will probably try to design a board where I can 'snap' a SNAP onto it and fit it in an off-the-shelf case (Hammond 1551L or 1551K if a little more room needed).

<commentary>
my pet peeve about all the various bare boards manufacturers put out- can't anyone just go browse mouser/digikey and find a box to use, then design the board to fit it, and if a user wants to protect it from getting shorted out or having it short out all the other stuff on their bench they can buy a readily available inexpensive box to mount it in. It would at least be better than picking dimensions from thin air as they appear to do now.
</commentary>

[oPossum]

Arrow now has these in stock.  $14.16 + free next day shipping

10% off code THANKYOU2018  expires 2018.11.09

https://www.arrow.com/en/products/pg164100/microchip-technology


3D printed case files by Digi-Key:  https://www.thingiverse.com/thing:3074301

[oPossum]

25% off at Microchip Direct -  Use Coupon Code : TP1927  expires 2018.10.31

$14.95 (before 25% off) + shipping

https://www.microchip.com/Developmenttools/ProductDetails/PG164100

[cv007]

I have been running the SNAP programmer for a while now and it works pretty good. Programming speeds are fast and debugging runs fast. I have not had any fail to connects or any other type of errors so far. I'm also putting a relatively large load on the 3v3 rail.

Now the question(s)-

Using a 3 pin header and a jumper to select 3v3/5v was my original idea for output power selection to power the target if wanted, or a small slide switch could also be used. But since we are in the microcntroller forum I am considering using a small pic to do the selection via mosfets.

The attached picture is my current idea. I have a hard time simulating accurately as mosfet selection is slim so just have to work around it. The following mosfets are just something I somehow ended up selecting as potential mosfets-

FDN340P- P-CH Mosfet
SSOT-3 (SOT-23)
2A Id continuous
Rdson ~100mohm

FDN327N - N-CH Mosfet
SSOT-3 (SOT-23)
2A Id continuous
Rdson ~100mohm


The pic will select the output by pulling a gate low. I'm not sure if the N-Ch side can get away without the P-Ch to drive the N-Ch gate, as ~1.4v may not be enough for the N-CH gate if driven by the pic (if I power the pic with 5v).

Is there something I'm obviously missing, or does this look like it will work ok? Or is there a better idea?

edit- there is a problem. I somehow thought I was escaping the body diodes, but I just re-arranged the problem. I need the 5v/3v3 mosfet outputs tied together, but need to avoid both the 5v feeding back to the 3v and the load using 5v/3v when both are off. My circuit violated the former.

edit- added new circuit (diodes added to visualize mosfet body diodes), not sure if good. I'm going to have to find better simulation software where I can enter part parameters. Maybe I need to learn how to use ltspice.

[cdev]

Please forgive me if this is a really stupid question.

I would like to learn PIC - hopefully to play with time-and frequency stuff on the HW platform.

Not knowing in advance if its going to be something I can learn, I really don't want to invest a lot of money in vendor-specific hardware.  $60-90 was way too much.

This new board, well, $15 is too much, but I will pay it, if it means I can access some support resources a bit better and if it supports the older HW fairly well.

I actually have some specific tasks I want to accomplish.

I would especially like to be able to extract the firmware from existing PIC devices to look at it and hopefully modify it.

(I am thinking about my PIC frequency counter which I would like to add a feature - basically a simple gated zero-crossing counter function, to.)

It does have a six pin in-circuit programming header.

Where does one go to start out on PIC, besides the microchip site?


Also, does their development IDE software (MPLAB) run under Wine on Linux? Or are there  viable alternatives to it?

Because its so useful for timing, I think PIC would be fun learning.

[cv007]

This new board, well, $15 is too much, but I will pay it, if it means I can access some support resources a bit better and if it supports the older HW fairly well.

You can access any support resources available without paying a dime for anything, and buying hardware doesn't change that. The SNAP will not support older parts that need high voltage programming. MPLABX runs on linux, windows, osx.

I would especially like to be able to extract the firmware from existing PIC devices to look at it and hopefully modify it.

That's quite a leap from not knowing anything to being able to read a pic, disassemble the instructions, figure out the asm code without any help (like labels, functions names, var names, etc.), then adding your own special functions.  Without the source code, its a waste of time and it wouldn't take you long to realize that. I say that only to possibly adjust your expectations and alter your purchasing decisions based on the probability you do not need to have something that supports older pic parts. For any new projects, there is no reason to use an old part as newer parts are cheaper and much better in many ways.