Post by thread:[PIC] MCP2200

MCP2200 looks like a valiant contender to the FT232R series, both permit GPIO pin programming from the host which is interesting for a number of reasons. The price is also very interesting for MCP2200 imho. Unlike the FT232R the MCP2200 seems to provide GPIO access from Linux drivers too. Has anyone done something *other* than serial RS232 to USB conversion using these chips? The way back to the easy peasy IO interfacing, lost when parallel ports disappeared, seems to have reappeared. It took a while but, as they say, better later than never. -- Peter

On Fri, Jan 7, 2011 at 6:00 PM, Peter <spam_OUTplpeter2006TakeThisOuTyahoo.com> wrote: > MCP2200 looks like a valiant contender to the FT232R series, both permit GPIO > pin programming from the host which is interesting for a number of reasons. The > price is also very interesting for MCP2200 imho. Unlike the FT232R the MCP2200 > seems to provide GPIO access from Linux drivers too. > > Has anyone done something *other* than serial RS232 to USB conversion using > these chips? > It is said to be a preprogrammed PIC18F14K50 after all and it is no contender to FT232R at all, especially under Windows due to the driver support issues.. I tend to think you may have better luck with this one. http://www.schmalzhaus.com/UBW/ http://www.schmalzhaus.com/UBW32/ -- Xiaofa

> Has anyone done something *other* than serial RS232 to USB > conversion using these chips? Peter, I'm waiting on samples. AIUI, it is for USB to serial only, there is no chip control from the UART end. My plan is to use it to dump emailed data from a PC down to a PIC + RAM board The MCP2200 datasheet USB 2.0 to UART Protocol Converter with GPIO ww1.microchip.com/downloads/en/DeviceDoc/22228A.pdf http://ww1.microchip.com/downloads/en/DeviceDoc/80499A.pdf and MCP2200EV-VCP USB to RS232 development and evaluation board http://ww1.microchip.com/downloads/en/DeviceDoc/51901A.pdf Also has its own page Jo

> MCP2200 looks like a valiant contender to the FT232R series, both permit GPIO > pin programming from the host which is interesting for a number of reasons. The > price is also very interesting for MCP2200 imho. Unlike the FT232R the MCP2200 > seems to provide GPIO access from Linux drivers too. > > Has anyone done something *other* than serial RS232 to USB conversion using > these chips? > > The way back to the easy peasy IO interfacing, lost when parallel ports > disappeared, seems to have reappeared. It took a while but, as they say, better > later than never. The MCP2200 is a pre-programmed 18F14K50 chip. I have seen reports that they don't seem to have it ideally set up, but don't have any experience with it myself. -- Scanned by iCritical.

One significant missing feature when I took a close look at it 4 months ago was the inability to configure how much current you want it to ask for from the USB host. For USB-powered designs this can obviously be a showstopper. It is a shame because that seems like something easily added and not having it makes a whole class of designs not possible to do. I got the confirmation of the inability directly from someone @ microchip. This made it not possible to replace an expensive and inconvenient to use for various reasons FTDI part with microchip in an evolving high volume design I'm deeply involved in. Alas. Rolling our own USB is not an option. J .....alan.b.pearceKILLspam@spam@stfc.ac.uk wrote: {Quote hidden}>> MCP2200 looks like a valiant contender to the FT232R series, both permit GPIO >> pin programming from the host which is interesting for a number of reasons. The >> price is also very interesting for MCP2200 imho. Unlike the FT232R the MCP2200 >> seems to provide GPIO access from Linux drivers too. >> >> Has anyone done something *other* than serial RS232 to USB conversion using >> these chips? >> >> The way back to the easy peasy IO interfacing, lost when parallel ports >> disappeared, seems to have reappeared. It took a while but, as they say, better >> later than never. > > The MCP2200 is a pre-programmed 18F14K50 chip. I have seen reports that they don't seem to have it ideally set up, but don't have any experience with it myself

Jesse Lackey wrote: > This made it not possible to replace an expensive and inconvenient to > use for various reasons FTDI part with microchip in an evolving high > volume design I'm deeply involved in. Alas. Rolling our own USB is > not an option. What about having someone outside with a existing USB stack and good knowledge of PICs do for you what Microchip did (black box a PIC running USB code as a USB to serial converter), but with the features you actually want? If this is indeed a high volume design and the alternative is "expensive and incovenient" this might be worth a look. ******************************************************************** Embed Inc, Littleton Massachusetts, http://www.embedinc.com/products (978) 742-9014. Gold level PIC consultants since 2000

I went to search the PIClist for the thread about the MCP2200 and was told that searching does not function.... I apologize that I am asking a question I believe was answered earlier. What is a good choice for a chip that allows USB to bit twiddling .... that has driver support for Windows, OS X and Linux ??? Gus in Denver 99gus

YES NOPE9 wrote: >I went to search the PIClist for the > thread about the MCP2200 and was told > that searching does not function.... > > I apologize that I am asking a question > I believe was answered earlier. > > What is a good choice for a chip that > allows USB to bit twiddling .... > that has driver support for > Windows, OS X and Linux ??? FT232RL? It's got several I/O lines, and IIRC it can even do SPI. Vitaliy

William "Chops" Westfield wrote: >> MCP2200 looks like a valiant contender to the FT232R series > > I was disappointed that it needs an external crystal. There went the > size and cost savings (over ftdi) I was hoping to achieve. We originally designed it into OBDLink SX, in our quantities it was cheaper -- even with the crystal. However, after playing with the prototypes, we switched back to FTDI. If Microchip is serious about competing with FTDI, they need to improve their USB drivers. Vitaliy

On Jan 7, 2011, at 7:53 PM, Vitaliy wrote: >> What is a good choice for a chip that allows USB to bit >> twiddling .... >> that has driver support for Windows, OS X and Linux ??? > > FT232RL? It's got several I/O lines, and IIRC it can even do SPI. I think the bit-banging support for the extra IO lines is only supported by the windows driver. This is unfortunately typical :-( Many so-called standard USB/Serial chips have trouble supporting even the serial function on any OS other than windows... BillW

On Fri, Jan 7, 2011 at 11:28 PM, William "Chops" Westfield <westfwKILLspammac.com> wrote: > > On Jan 7, 2011, at 7:53 PM, Vitaliy wrote: > >>> What is a good choice for a chip that allows USB to bit >>> twiddling .... >>> that has driver support for Windows, OS X and Linux ??? >> >> FT232RL? It's got several I/O lines, and IIRC it can even do SPI. > > I think the bit-banging support for the extra IO lines is only > supported by the windows driver. > This is unfortunately typical :-( Many so-called standard USB/Serial > chips have trouble supporting even the serial function on any OS other > than windows... > > BillW > Bit-bang on FT232RL works fine on Linux with libftdi (entirely userland, uses libusb). And FTDI provides the proprietary D2XX drivers for the big three operating systems as well. Regards, Mark markrages@gmail -- Mark Rages, Engineer Midwest Telecine LLC .....markragesKILLspam.....midwesttelecine.com

> One significant missing feature when I took a close look at it 4 months > ago was the inability to configure how much current you want it to ask > for from the USB host. For USB-powered designs this can obviously be a > showstopper. It is a shame because that seems like something easily > added and not having it makes a whole class of designs not possible to do.. A "Heath-Robinson" solution that may make the difference in some cases would be to explicitly control the maximum current that you draw - . which is easily and cheaply done. R

<alan.b.pearce <at> stfc.ac.uk> writes: > The MCP2200 is a pre-programmed > 18F14K50 chip. I have seen reports that they don't seem to have it ideally > set up, but don't have any experience with it myself. Thanks for all the answers. Even if the MCP2200 is just a custom programmed 18F14K50 it is still interesting to design in, I think. For example, a tiny breakout board designed for MCP2200 (USB to TTL RS232) could easily become something else later by simply mounting another 18F14K50 on it. We will see where this is going later. I also assume that if the MCP2200 is indeed a 18F14K50 then Mchip can update the firmware (and maybe users can do that too?), fix bugs after the fact, and maybe build in more capabilites. Bit banging GPIO pins does indeed seem to be more difficult than on a FTDI chip which is now the de facto standard for such work. But that may change in the future, especially if people start writing code for PIC18F14K50, that being drop-in compatible with any MCP2200 USB to RS232 board, however tiny. -- Peter

> > MCP2200 looks like a valiant contender to the FT232R series > > I was disappointed that it needs an external crystal. There went the > size and cost savings (over ftdi) I was hoping to achieve. Yes, unfortunately the internal oscillator is not stable enough to do USB - even though they do have other chips with an internal oscillator that is good enough, but I think they are all PIC24 chips. One would have thought they would have tried very hard to get that technology into this chip, as it does seem to be such an essential cost saving for this chip and its smaller brother, the 18F13K50. -- Scanned by iCritical.

Peter <plpeter2006 <at> yahoo.com> writes: > <alan.b.pearce <at> stfc.ac.uk> writes: > > The MCP2200 is a pre-programmed > > 18F14K50 chip. I have seen reports that they don't seem to have it ideally > > set up, but don't have any experience with it myself. I just got the whole 100+MB Mchip development libraries and focused on the CDC serial RS232 interface whose hardware option 4 is the "low pin count USB development board", a straightforward PIC18F14K50 application. When I will find the time I will test this deeper. I understand that the USB example code has a restrictive license but I do not understand what that restriction may be. In a HW device, the situation is clear, you buy the device (f.ex. from FTDI), use it, done. No open ends. (I am not covering Vinculum chips here, just FTxxx). In a MCP2200 the situation should be the same. In a PIC18F14K50 programmed with the firmware from the examples, what *exactly* do the limitations on licensing/use? Lawyerspeak is inscrutable. Is one free to modify the code in the examples and use it as is in a product? If not, where does the buck stop? The USB stack? main()? The whole project? -- Peter

On Jan 10, 2011, at 1:52 PM, Peter wrote: > In a PIC18F14K50 programmed with the firmware from the examples, what > *exactly* do the limitations on licensing/use? Lawyerspeak is > inscrutable. > > Is one free to modify the code in the examples and use it as is in a > product? If > not, where does the buck stop? The USB stack? main()? The whole > project? I believe you can use the code in a product as long as the product uses a microchip microprocessor. Adapting the Microchip USB stack to an Atmel USB controller would get someone annoyed. This of course means that the CDC implementations for PIC, Atmel, TI, Silicon Labs, Cypress, etc all have different bugs. Which is why I was thinking that the world would be ripe for a common implementation. (The Microchip CDC Serial port seems to work reasonably well, FWIW. The EggBot uses it and AFAIK hasn't had any USB-related problems on any of windows/linux/mac...) BillW

Again, I'm digging up an old thread but thought this was worth mentioning: http://dangerousprototypes.com/2011/01/18/hack-open-source-usb-stack-on-mcp2200/ As of right now: MCP2200 in qty 100 (SSOP) @ digikey = $1.47 w/ 1,722 available PIC18F14K50 in qty 100 (SSOP) @ digikey = $1.74 w/ 18,001 available $0.50 difference for qty 1. Any chance the MCP2200s are 14K50s that didn't fully pass QC, or is the lower price due simply to marketing? -k

On Tue, January 18, 2011 10:18 pm, Nathan Nottingham wrote: > > Again, I'm digging up an old thread but thought this was worth mentioning: > > http://dangerousprototypes.com/2011/01/18/hack-open-source-usb-stack-on-mcp2200/ > > As of right now: > > MCP2200 in qty 100 (SSOP) @ digikey = $1.47 w/ 1,722 available > PIC18F14K50 in qty 100 (SSOP) @ digikey = $1.74 w/ 18,001 available > > $0.50 difference for qty 1. > > Any chance the MCP2200s are 14K50s that didn't fully pass QC, or is the > lower price due simply to marketing? It's a marketing decision, but it also leaves room for migrating the part to a PIC18F13K50, or even a version that removes the MCU completely and makes it a non-reprogrammable part. Don't forget that specs that aren't in the datasheet can change. Maybe you can reprogram what you ordered today, but there is no guarantee that you can reprogram what you order tomorrow. Not much of a loss if you order 1, but if you order 100,000, you could be out big bucks, with zero recourse. There is also the support cost built in- if you're in production and go to Microchip with MCP2200 re-flashed to a general purpose MCU, you're not going to get much (if any) support. Matt Bennett Just outside of Austin, TX 30.510843,-97.919286 The views I express are my own, not that of my employer, a large multinational corporation that you are familiar with