View previous topic :: View next topic |
Author |
Message |
pmuldoon
Joined: 26 Sep 2003 Posts: 218 Location: Northern Indiana
|
can anybody select a 16F1779 part? |
Posted: Tue Apr 05, 2016 12:01 pm |
|
|
I have the latest PCM and PCH ver 5.056 and it isn't in the PICC\devices directory. It does show up as supported on the CCS web page. Am I missing something? (besides the device file, I mean)
It didn't show up in MPLAB 8 but it did in MPX so I will finally be forced to become an X man. Will I have to use the Microchip complier as well? Argh...more retraining! |
|
![](templates/subSilver/images/spacer.gif) |
Ttelmah
Joined: 11 Mar 2010 Posts: 19378
|
|
Posted: Tue Apr 05, 2016 12:24 pm |
|
|
Talk to CCS. They have probably just forgotten to include the file in the download. |
|
![](templates/subSilver/images/spacer.gif) |
pmuldoon
Joined: 26 Sep 2003 Posts: 218 Location: Northern Indiana
|
|
Posted: Tue Apr 05, 2016 12:48 pm |
|
|
I've done that, too.
Thanks, |
|
![](templates/subSilver/images/spacer.gif) |
asmboy
Joined: 20 Nov 2007 Posts: 2128 Location: albany ny
|
|
Posted: Tue Apr 05, 2016 1:44 pm |
|
|
Just wondering:
Its a very new part - prelim datasheet only released in December '15
---
The 40 pin 18f46k22 has way more processing power , memory and prog flash for less $$ but lacks some of the function features of the 1779......
So what are the unique features of the 1779 that made it your choice for a project ?? |
|
![](templates/subSilver/images/spacer.gif) |
Ttelmah
Joined: 11 Mar 2010 Posts: 19378
|
|
Posted: Tue Apr 05, 2016 2:41 pm |
|
|
The most interesting 'non standard' feature, is the high current I/O.
Obviously you have to be super careful to avoid the power and total current limits, but it does potentially mean quite a few more things that can be done without needing external drivers. The PWM using CCP8 for example could directly drive a MOSFET and give quite reasonable transition times.
There are rather a lot of errata as well, so 'beware'..... |
|
![](templates/subSilver/images/spacer.gif) |
asmboy
Joined: 20 Nov 2007 Posts: 2128 Location: albany ny
|
|
Posted: Tue Apr 05, 2016 3:51 pm |
|
|
"high current I/O. "
It would be GREAT if one pair of the Vss Vdd pins were devoted solely to Digital I/O source and sinking ........ and the other was designated as analog "safe" uses etc
I'm still under the impression that migrating to 18F for most of my work was the right thing to do.
I'd be over the moon if there was an 18F?6K?? part with DUAL NCO's that could be mapped to the same pins as the dual PWM is now -- wishful thinking i guess. |
|
![](templates/subSilver/images/spacer.gif) |
pmuldoon
Joined: 26 Sep 2003 Posts: 218 Location: Northern Indiana
|
|
Posted: Wed Apr 06, 2016 5:06 am |
|
|
I needed 3-4 channels of 16-bit pwm. This one came up on microchips search thingy. Don't really need anything beyond that other than the I2C. I hope that much of it works. The configurable pins seem a bit vague in the datasheet, but I'll get there. The pcb is already done, so I might be stuck with it. Rush deadline, of course. |
|
![](templates/subSilver/images/spacer.gif) |
Ttelmah
Joined: 11 Mar 2010 Posts: 19378
|
|
Posted: Wed Apr 06, 2016 7:32 am |
|
|
I'd have gone for something that was readily available now, like the PIC16F1788 or 9. 3 separate PSMC modules.
Or (of course) almost any PIC24.
Also, depending on the frequency all the CCP's can give this, if running in compare mode, but at the cost of having to reload on each half cycle.
I find it hard to actually visualise many PWM applications that actually need 16bit. I spend most of my time turning the resolution down to give faster responses!.... |
|
![](templates/subSilver/images/spacer.gif) |
pmuldoon
Joined: 26 Sep 2003 Posts: 218 Location: Northern Indiana
|
|
Posted: Wed Apr 06, 2016 8:12 am |
|
|
The chip is available, already on order.
The errata seems minimal compared to most other PICs I've used - maybe it's still early in the discovery phase?
I don't have a CCS compiler for the 24's, only PCM and PCH. I tried the PCD years ago on the 30's & 33's but it was too buggy to use at the time so I quit renewing it.
Looks like I'm stuck with MPlabX and the mchip compiler. MPlab 8 doesn't even support that part, so now I get to dive into X.
MPLabX make doesn't like filepaths that contain spaces - still stuck in the 1980's era for some reason. And even though I updated my dll files per what CCS sent me for the chip, MPlabX doesn't see CCS as a compiler option for that part. It does for the old 18F6722 part I use, but not the 16F1779.
Arg! This dog's too old to learn new tricks, especially when the tricks don't really give you anything new. Oh well, just keep moving forward. These will be last weeks problems next week. I always seem to manage to resolve them somehow tomorrow even when I don't have a clue how today. Hmmm, I think that could be a definition of engineering, LOL! |
|
![](templates/subSilver/images/spacer.gif) |
temtronic
Joined: 01 Jul 2010 Posts: 9177 Location: Greensville,Ontario
|
|
Posted: Wed Apr 06, 2016 8:21 am |
|
|
This 'old dog' doesn't WANT to learn 'new PICs, er tricks'. Just because something new comes out doesn't mean it's better, which you probably already know. Have to ask why 16bit PWM? Unless really ,really, really needed couldn't 12bits work ??
If speed isn't a factor did you think about 3 I2C>16bit PWM ?
Cost HAS to include R&D time and hassles ! Sometimes a proven PIC is cheaper than time wasted on new compiler, new PIC 'oddities', etc. Though I figure you already know this stuff...
I like 'old parts', ones of a proven 'pedigree' though I do like the 18F46K22 as my goto PIC.
Jay |
|
![](templates/subSilver/images/spacer.gif) |
pmuldoon
Joined: 26 Sep 2003 Posts: 218 Location: Northern Indiana
|
|
Posted: Wed Apr 06, 2016 8:42 am |
|
|
All of the PIC's I've used in the past have only had 10-bit resolution on the PWM. I need precision control of a light dimmer and knew 10-bit wouldn't hack it. Small steps at the high end, but big steps at the low end. 12-bit might have been okay - didn't see that one. I just didn't see 16-bit pwm being a difficult implementation on a device that has 16-bit timers. I didn't think about the part being new and MPlab8 no longer being supported so the library became stagnant.
Most other PICs have very extensive errata sheets. To the point that I wish Microchip would re-format page 1 of there datasheet to show the features that work separate from the features that work work. I've yet to find a PIC with a 12-bit adc that works correctly and have no clue how to search for the ones that work work using their selection tool.
definitions:
work - saw it function properly once
work work - can be expected to function properly all the time |
|
![](templates/subSilver/images/spacer.gif) |
Ttelmah
Joined: 11 Mar 2010 Posts: 19378
|
|
Posted: Wed Apr 06, 2016 9:04 am |
|
|
If you are talking about light dimming for things like stage/film, you don't even need close to 10bit accuracy.
You are very unlikely to require 16bit, unless you are using regulated power. Problem is that mains itself fluctuates by several percent in amplitude, and the waveform is far enough from a sinusoid that this too leads to variations in the actual output power that has to be corrected for in the dimmer.
There are specialist IC's on the market that implement a feedback loop round the dimmer control, measuring the power actually being delivered and adjusting the dimming to to give the required power. If you need good accuracy, you are going to have to think about adding something like this.
I do some power LED lights that are designed to give high accuracies, and on these we have a light meter measuring the actual output, and adjust to set this to the requested level. This is needed to compensate for ageing, temperature, etc.. |
|
![](templates/subSilver/images/spacer.gif) |
pmuldoon
Joined: 26 Sep 2003 Posts: 218 Location: Northern Indiana
|
|
Posted: Wed Apr 06, 2016 10:14 am |
|
|
You're probably right, I probably do have more pwm resolution than I need, but we're getting a bit off-topic. The point is, I found a 16-bit pwm that I thought should be an easy implementation and would provide more than enough resolution. I can always make a 16-bit pwm act like a 10-bit, etc.
CCS just provided me with the missing link - an update to the CCS plugin for MPlabX. The installation instructions were impeccable. Even I couldn't screw it up! That knocks out this weeks problem and now I can get a head-start on next weeks problems. |
|
![](templates/subSilver/images/spacer.gif) |
temtronic
Joined: 01 Jul 2010 Posts: 9177 Location: Greensville,Ontario
|
|
Posted: Wed Apr 06, 2016 11:26 am |
|
|
hmm stage lighting...BTDT, 20 year ago, used the 'new' 16F877 and 8 bit PWM for the then 'new' DMX type. Found out fast you have to 'preheat' the stage lamps, else they go POOF. There's also an 'eyeball vs PWM' chart one has to consult to get a 'linear ramp' effect. Nowadays all this stuff is on the Web. DALI is the other standard.
I know 16 bit changes are far too subtle for my aging eyeballs.
But, if that's what the client wants....that's what he gets !!
Jay |
|
![](templates/subSilver/images/spacer.gif) |
pmuldoon
Joined: 26 Sep 2003 Posts: 218 Location: Northern Indiana
|
|
Posted: Wed Apr 06, 2016 11:44 am |
|
|
I did a project years ago with an led and, i think 8 or 10 bit pwm. The problem was they wanted it to turn from off to on slowly and the low end pwm steps were very discernible. I was trying not to fall into that trap again. I was also thinking that to make a noticeable 'step' of light output you about have to double or half the pwm. Kind of what T said about the eyeball vs PWM chart. In which case you run through the pwm count very quickly.
You guys sure ask a lot of questions, |
|
![](templates/subSilver/images/spacer.gif) |
|