PicKit3 still suck?

[mzdenkov]

Greetings!

So here we are in 2016 and I was wondering will I, or will I not buy a PicKit3!
After programming with ATMEL ICE and Atmel chips i would like to start testing
Microchip uCs.

Here comes the problem, I have an old 3.rd party PicKit2 and I would like to start
programming their  "ARM(i know that they have their own architecture)" uCs.
My BIG question is, does PicKit3 still suck or not? Because for me
44€ is a big money, and when the work is easier done that is a BIG "+" for me.


Thanks.
Best Regards.
Matija.

[Throy]

Are you talking about programming the PIC32?  I believe the PicKit2 doesn't support all the PIC32 microcontrollers.  You might want to check which ones it does.  Otherwise the PicKit3 should support the PIC32.

I can't tell you how the PicKit3 compares to the PicKit2 since I've never used the PicKit2.

[mzdenkov]

Yes, i was thinking on PIC32. Can you tell me what is your XP with PicKit3... If you can, write positive and negative arguments(few) about it. Is it hard to use?  PicKit2 gave me good XP, no matter that it was 3.rd Party.

Thanks.

[Karel]

My experience with the PICkit 3 is that it does what it has to do and it's reliable.
Only disadvantage, it's slow, specifically when programming pic32's.
So, if you are a hobbyist with a tight budget, buy the PICkit 3.
If you are a professional, buy the ICD 3, it's much faster.

[Throy]

I'd agree with what Karel said.  The other thing to be aware of with the PicKit3 is you can't set a breakpoint while the program is running and expect it to stop.  You have to pause the program and then resume for it to stop at the breakpoint, if it has been set while it's running.  IIRC the PicKit3 also doesn't support runtime watches.

[Howardlong]

Regarding speed, it's largely to do with the size of flash on the target rather than the device family itself, although I accept that PIC32s are innevitably going to have larger flash the a PIC10!

The low pin count PIC32s like the PIC32MX1xx/2xx program relatively quickly, whereas a PIC32MZ2048xxx will take a lot longer. Same applies to the PIC24/dsPIC33 series where there are some monsters nowadays too. In fact, it's a reason to buy devices with smaller flash if you don't need it!

[mikeselectricstuff]

PK3 works fine - you can get clones  on ebay very cheaply - I've bought a couple of the ones labelled "kit 3.5" which are well made & work as well as the original
http://www.ebay.co.uk/itm/New-PICKIT3-PIC-Emulator-kit3-Programmer-offline-Download-adapter-Seat-/151795631744?hash=item2357b98680:g:GrkAAOSwp5JWYkbZ

As mentioned they are a bit slow on larger parts, for which ICD3 is a better option

[Galenbo]

Did anyone compare pickit3/icd3 himself to the Mikroprog of MikroElektronika?
http://www.mikroe.com/mikroprog/pic-dspic-pic32/
I'm curious about speed and compatibility.

(attention: this company lets you pay 299 dollar for the software)

[mzdenkov]

Thanks for more info.

Im going to save some money for the " original " PicKit3.
The upload speed to the uC isnt a problem for me. I think
ill have to live with the Breakpoint part.

One more thing, do you need a external circuitry for programming while
the chip is in circuit? I saw that they added some tris-buffers to protect
programmer(in the prospect).

[Stonent]

£18.90
(£9.45 / 100g)

Sold by weight?

[Howardlong]

Did anyone compare pickit3/icd3 himself to the Mikroprog of MikroElektronika?
http://www.mikroe.com/mikroprog/pic-dspic-pic32/
I'm curious about speed and compatibility.

(attention: this company lets you pay 299 dollar for the software)

The problem with MikroElektonika stuff, or indeed some other third party tool vendors like CCS, is that although they're technically reasonable solutions, you are heavily locked into them. Of course, that may or may not be a problem for you, but there's a much bigger community using Microchip's own tools and that means there are far more examples to choose from and typically you're more likely to find an answer online to a problem.

[Ian.M]

Why would anyone except an educational establishment want to enter Mikroelektronika's walled garden?

* Their programmers/debuggers are significantly behind Microchip's for device support, aren't compatible with MPLAB or anything except mikroIDE or mikroProg.
* Their compilers only run under mikroIDE.
* mikroIDE doesn't support any 3rd party compilers.  It doesn't even support Microchip MPASM.
* Their libraries are closed source with no sourcecode licence available, so you cant modify or extend them, and if you are bug-hunting, you have to work with the disassembly.

OTOH, they have some quite nice hardware (demo boards and modules) so although I wouldn't touch their toolchain with a bargepole, I would not rule out using their boards with a MPLAB supported toolchain.

[dannyf]

I use pickit2 to program pic32 through pic32prog - also possible with other programmers. Pic32prog does not support all pic32 chips however.

Not a problem for me as it does support all pic32 that I use.

[tszaboo]

I used the PK3 at work to program PIC32 years ago. They dont suck, they have less functionality than the other ICDs from Microchip, or the PK2. I think it is a reasonable compromise for the money, if you need Microchip.
On the other hand PIC32 is not fancy enough anymore to bother buying that. The STM32 discovery boards are better value, learning aspect and hardware wise.

[Howardlong]

One more thing, do you need a external circuitry for programming while
the chip is in circuit? I saw that they added some tris-buffers to protect
programmer(in the prospect).

Nope, just a 10k pullup on MCLR is all you need as a rule of thumb.

[mzdenkov]

I used the PK3 at work to program PIC32 years ago. They dont suck, they have less functionality than the other ICDs from Microchip, or the PK2. I think it is a reasonable compromise for the money, if you need Microchip.
On the other hand PIC32 is not fancy enough anymore to bother buying that. The STM32 discovery boards are better value, learning aspect and hardware wise.

Lovely!  I have few DISCO boards from ST. As it looks they are qute inexpensive way to the 32bit ARM uCs. " + "  ST gives  ST-Link nearly for free on this boards. Im only interested in Microchip for great community that it has like Atmel,  but ST will work too.

[Howardlong]

I can't imagine those discovery boards are anything but a loss leader at $10, but then again you look at a Pi Zero and the head starts really scratching!

Back to the PICs, I am pretty partial to the Microsticks, particularly the DM240013-2 and DM330013-2. The first one has one of my favourite devices on it for peripherals, including two 8 bit DACs as well as all the usual ADCs, PWMs and whatnot. The second comes with four chips, including a PIC32. The main reason I like these boards - and PICs in general for low to mid end projects - is that they take 28 pin DIP packages which makes breadboarding and going from demo board to your own prototype a cinch without lifting a soldering iron. (I'm not averse to soldering, far from it in fact, I do everything from 0201 to BGA, but sometimes I feel I can give self flagellation a miss).

[tszaboo]

I can't imagine those discovery boards are anything but a loss leader at $10, but then again you look at a Pi Zero and the head starts really scratching!

So they sell some 1000 boards losing 10 dollars each? Advertisement on cheap. It is really a genius strategy, better than the "free sample".
They usually give it away free on fairs. A company like ST has probably 9 digit budget for marketing.

[Galenbo]

...MikroElektonika stuff... technically reasonable solutions... heavily locked into them...

...Mikroelektronika's walled garden...behind Microchip...doesn't support any ...some quite nice hardware ...not rule out using their boards with a MPLAB supported toolchain.

You both are surely right about everything you say, but how does their mikroPROG compare to PicKit3/ICD3? Speed, settings,...

[Howardlong]

I can't imagine those discovery boards are anything but a loss leader at $10, but then again you look at a Pi Zero and the head starts really scratching!

So they sell some 1000 boards losing 10 dollars each? Advertisement on cheap. It is really a genius strategy, better than the "free sample".
They usually give it away free on fairs. A company like ST has probably 9 digit budget for marketing.

The problem is who are they selling them to? Self proclaimed "makers" I am sure, but volume? It's a very very hard sell if you're dependent on the maker space for real market penetration and volume to make up the for the loss leader.

I agree though about using it as an advertising strategy, at least I went to look at the chip specs, so I'm now aware of the segment they're targetting, but then I couldn't see a reason to change or use right now. It was, for example, interesting that they're using M4F in a low clocked device. But the ARM Cortex M space is already saturated, and picking one device from another between manufacturers is a significant piece of work in itself, and has inadvertently become a barrier to acceptance. It is much more about the peripherals than it is about the core. It's darned cheap per device for M4F though.

Pretty sure that, just like 95%+ of RPis, they'll end up in a drawer. I doubt that directly the remaining 5% is good enough, but as you say as a marketing tool you're right, it's pushed the device family's head up above the noise.

[LordNobady]

...MikroElektonika stuff... technically reasonable solutions... heavily locked into them...

...Mikroelektronika's walled garden...behind Microchip...doesn't support any ...some quite nice hardware ...not rule out using their boards with a MPLAB supported toolchain.

You both are surely right about everything you say, but how does their mikroPROG compare to PicKit3/ICD3? Speed, settings,...

The mikroPROG is faster then the PicKit3 and slower than the ICD2 and ICD3
It mostly works as just an other programmer, and works well as a debugger with the software. note that the pinout of the mikroPROG is differed.
the PicKit3 can with an external power supply ( usb ) work as an stand alone programmer.

all have about the same functionality.

[mzdenkov]

[/quote]
all have about the same functionality.
[/quote]

They do, but from price aspect they are far away from each other.  +I must add that mikroelektronika has great costumer support.
Few years ago, they helped me with the TIVA C board, because their MikroProg program was crashing.

[Ian.M]

If you've got a failed, currently supported, genuine Microchip programmer or debugger, Microchip will generally replace it even if its out of warranty and they are not contractually obliged to do so.  They tend to send out an advance replacement to minimise your downtime.

[Howardlong]

...MikroElektonika stuff... technically reasonable solutions... heavily locked into them...

...Mikroelektronika's walled garden...behind Microchip...doesn't support any ...some quite nice hardware ...not rule out using their boards with a MPLAB supported toolchain.

You both are surely right about everything you say, but how does their mikroPROG compare to PicKit3/ICD3? Speed, settings,...

The mikroPROG is faster then the PicKit3 and slower than the ICD2 and ICD3
It mostly works as just an other programmer, and works well as a debugger with the software. note that the pinout of the mikroPROG is differed.
the PicKit3 can with an external power supply ( usb ) work as an stand alone programmer.

all have about the same functionality.

Does the mikroPROG work as a debugger with MPLAB or is it only compatible with MikroElektronika toolchains?

The pinout also looks substantially different to the ICD/ICSP standard too.

[dannyf]

Microchips decision to scrap pickit2 in favor of pickit3 always puzzles me. The hardware looks pretty much identical, functions almost identically, with pickit3 missing some features of pickit2. Then why restart with a brand new ecosystem for the pickit3, taking on so much risks for seemingly zero benefits.

On the flip side, it does seem to be consistent with their general behaviors of undermining their own objectives,