Helpful ReplyProblems with MCP2200

Author
hanzibal
Starting Member
  • Total Posts : 47
  • Reward points : 0
  • Joined: 2012/06/11 00:40:35
  • Location: Sweden
  • Status: offline
2012/06/26 15:10:03 (permalink)
0

Problems with MCP2200

Hi all!
 
I've been having some real trouble getting the MCP2200 chip to work for more than a week now. I recently started another thread about this but thought I had it solved and because of that, the thread is now closed so I started this new thread instead.
 
Anyway, I'm working on a prototype in a breadboard and using a DIP20 adapter for the MCP2200. I got the crystal (a regular hc49/s) soldered between pins 2 and 3 and then the two load caps (smd 1206) soldered from there (i.e. again from pins 2 and 3) to ground on pin 20 via a small wire.  Like this:

The dapter is somwhat brused from soldering different caps on and off. The MCP2200 is sitting on the bottom side.
 
I'm using these schematics: EDIT: Currently using other schematics, see further down this thread.

 
Symptoms are that the MCP2200 sometimes runs for a couple of minutes and then dies while sometimes it doesn't run at all. By "running" I mean that it shows up in Windows (XP SP3) for a couple of minutes and sometimes not at all. I've tried 22pF, 33pF and 47pF load caps but it seem not to make a difference. I have read the excellent app notes from M on how to get an oscillator to work.
 
So far I've trashed 7 x MCP2200 (yes, seven!), an equal amount of DIP20 adapters and about 20 crystals...is it really supposed to be this hard or am I doing something else wrong?
 
Please help so that I can sleep again!
post edited by hanzibal - 2012/07/01 07:51:49

Attached Image(s)

#1
hanzibal
Starting Member
  • Total Posts : 47
  • Reward points : 0
  • Joined: 2012/06/11 00:40:35
  • Location: Sweden
  • Status: offline
Re:Problems with MCP2200 2012/06/28 14:22:51 (permalink)
0
Having not received any suggestions, I tried a few more ideas in pure desperation and discovered that if I disconnect from USB host and wait long enough the board works but only for a minute or so. Immediately after that it doesn't work at all but after being disconnected again for a couple of hours, it works for about 20 seconds.

Also, I discovered that if I short Vdd and Vss for a minute while disconnected to USB host, the board works again for a little while longer compared to simply waiting while disconnected from USB host without shorting Vdd to Vss.

In conclusion, the duration time of correct functioning between USB connection cycles is clearly proportional to the duration of time being disconnected from USB host. I.e the longer I wait before re-connecting, the longer the board works.

This gives me the distinct feeling that there is something involving capacitance going on here. By shorting Vdd and Vss I probably discharge some voltage accumulated in the device.

Does this ring any bells anyone?
 
I can't understand why I have these problems while nobody else seem to. I've seen people gluing these chips directly onto USB connectors and soldering tiny wires between pins here and there, and it works for them. In short, I've seen the chip working well under much worse conditions.
post edited by hanzibal - 2012/06/28 14:31:56
#2
hanzibal
Starting Member
  • Total Posts : 47
  • Reward points : 0
  • Joined: 2012/06/11 00:40:35
  • Location: Sweden
  • Status: offline
Re:Problems with MCP2200 2012/06/28 17:03:15 (permalink)
0
Struggling on, my last post rang a bell to me myself...of course, floating pins! Smile
 
So I pulled down my unused GPIOs (only those configured as inputs) to ground via a 10k resistor respectively and the device runs for yet some more time but still no more than about 10 minutes.
 
Ok, so pulling down floating pins helped somewhat but not all the way...
 
There are lots of breakout boards for the MCP2200, most of which have different kinds of oscillators (i.e. resonators or crystals) and load caps (all between 10 and 47pF) but none of them (including MCP breakout module) pull down (or up) any unconnected pins.
 
So, why do I need the pull-downs when nobody else does and why doesn't it work all the way when I actually do so?
 
Questions:
1. Do I really need to pull down (or up) unused pins - inputs/outputs or all?
2. Is the MCP2200 (a.k.a. PIC18F14K50) extremely sensitive to ESD and/or EMF?
3. At what maximum temperature can I solder the MCP2200 and for how long?
4. Can I use a HC49/S crystal or do I need a HC49/U or ceramic resonator instead?
5. What load caps are optimal given the design depicted above,(SOIC20 adapter) under normal working conditions (room temperature, etc)?
6. Is a ferrite bead on USB 5V required?
7. Am I loosing my mind?
 
Thanks and best wishes!
 
EDIT: Another thing - I got the MCP2200 configured to output the USBCFG signal on GPIO1 and got a LED connected to that. Thing is, the LED doesn't go off when the device stops working. I find this somewhat strange.
post edited by hanzibal - 2012/06/28 17:08:02
#3
hanzibal
Starting Member
  • Total Posts : 47
  • Reward points : 0
  • Joined: 2012/06/11 00:40:35
  • Location: Sweden
  • Status: offline
Re:Problems with MCP2200 2012/06/29 04:29:33 (permalink)
0
Turned out it wasn't the oscillator since my scope shows a nice 12Mhz sine regardless of whether the chip is running or not.
 
Doesn't seem to be heat either, I brought out a fan och blew on the chip but didn't help in getting it going again.
 
USB current consumption goes down from ~20mA to ~12mA when the chip stops, dunno what this means, maybe the chip halts execution?
 
Still I got these questions:

1. Do I really need to pull down (or up) unused pins - inputs/outputs or all? 
2. Is the MCP2200 (a.k.a. PIC18F14K50) extremely sensitive to ESD and/or EMI? 
3. At what maximum temperature can I solder the MCP2200 and for how long? Can't find this info in the d/s. 
4. Can I use a HC49/S crystal or should I be using a HC49/U or ceramic resonator instead? 
5. What load caps are optimal given the design depicted above (SOIC20 adapter) under normal working conditions (room temperature, etc)? 
6. Is a ferrite bead on USB 5V required?
 
Again thanks!
 
EDIT: It's not heat, the chip is about 26C when it stops.
 
post edited by hanzibal - 2012/06/29 06:52:50
#4
newfound
Super Member
  • Total Posts : 1814
  • Reward points : 0
  • Joined: 2003/11/07 12:35:49
  • Status: offline
Re:Problems with MCP2200 2012/06/29 09:09:32 (permalink)
0
I don't see a cap on Vusb anywhere or what you have done with the reset pin.
#5
hanzibal
Starting Member
  • Total Posts : 47
  • Reward points : 0
  • Joined: 2012/06/11 00:40:35
  • Location: Sweden
  • Status: offline
Re:Problems with MCP2200 2012/06/29 14:25:16 (permalink)
0
Sorry for being unclear, the image of the adapter is taken before put into the breadboard and just to show the oscillator circuit itself. While in the breadboard, these schematics are now used (taken from the MCHP demo board) so there's a little reset-circuitry and a 470nF cap on Vusb:

 
post edited by hanzibal - 2012/06/29 22:15:53

Attached Image(s)

#6
hanzibal
Starting Member
  • Total Posts : 47
  • Reward points : 0
  • Joined: 2012/06/11 00:40:35
  • Location: Sweden
  • Status: offline
Re:Problems with MCP2200 2012/06/30 18:14:05 (permalink)
0
Anyone please help!
I'm about to loose my mind over this...and it's such a simple circuit too. I'm currently using the schematics above but with a HC49/S crystal and a pair of 22pF load caps instead of the resonator depicted.
 
#7
hanzibal
Starting Member
  • Total Posts : 47
  • Reward points : 0
  • Joined: 2012/06/11 00:40:35
  • Location: Sweden
  • Status: offline
Re:Problems with MCP2200 2012/07/03 14:58:52 (permalink) ☄ Helpful
0

I've been working on this simple little circuit for about a month now. I had the first 90% done within the first couple of days but since then I've been struggling with the oscillator.

I've been loosing sleep over this and so I was really at the point of giving up, when I (after a too long story) threw out the crystal and fitted one of these simple resonators instead:
http://uk.farnell.com/ael-crystals/c12m000000l003/resonator-ztt-12mhz/dp/1448128?Ntt=144-8128
Funny thing is that now the device works like a charm! Resonators are generally less accurate and even though I feel it's unprofessional at this point, I don't care why the I couldn't get to work with a crystal.

The d/s also suggests a resonator but I thought I'd be a good boy and fit a crystal instead, so much for trying to be clever obviously.

Cheers!

#8
cubicus
New Member
  • Total Posts : 2
  • Reward points : 0
  • Joined: 2012/07/28 10:14:41
  • Location: 0
  • Status: offline
Re:Problems with MCP2200 2012/07/28 11:04:08 (permalink)
0
Hello Hanzibal,
I have build the exact circuit with MCP2200 and its not working. (Nothing happends when i connect it to the USB port).  I think theat is the oscillator problem becose i dont´t get any wave form on my oscilloscope when i connect the probe beetwen OSC2 and the ground. I will try to get the crystal running outside the circuit to see what happends.  
Even if the d/s suggests a resonator, maybe I can learn more about crystals anyway.
#9
cubicus
New Member
  • Total Posts : 2
  • Reward points : 0
  • Joined: 2012/07/28 10:14:41
  • Location: 0
  • Status: offline
Re:Problems with MCP2200 2012/07/31 02:59:22 (permalink)
0
Hello again!
I have managed to make theat crystal oscillator to oscillate by connecting an 22 Ohm resistor between OSC2 pin and the crystal. Microchip did say something about an series resistor with the crystal but they don't say anything about the value, so i had to try myself.
Did anyone else find the same solution or another?
#10
hanzibal
Starting Member
  • Total Posts : 47
  • Reward points : 0
  • Joined: 2012/06/11 00:40:35
  • Location: Sweden
  • Status: offline
Re:Problems with MCP2200 2012/07/31 03:38:09 (permalink)
0
Hi cubicus!
Sorry for not answering sooner, I forgot to check the "notify me of replies" box. About the series resistor, I read an app note (can't find it now) where they mentioned a suitable value to be around 2k.
 
My problems went away after switching to a resonator but the actual problem turned out to be wrong soldering technique - I used a too big of a soldering tip and applied too much heat on the chip. I'm pretty sure I would get the a crystal working well too if I would solder with the same smaller tip that I used for the resonator. I really don't think the series resistor is necessary for ordinary HC49/S crystals which are not the "high drive" kind.
 
Try using a narrower tip @ ~270 degrees C and keep the soldering time to a minimum. Hope that helps you too.
 
Cheers!
#11
geortiz@une.net.co
New Member
  • Total Posts : 1
  • Reward points : 0
  • Joined: 2012/11/19 19:46:53
  • Location: 0
  • Status: offline
Re:Problems with MCP2200 2012/11/19 20:02:13 (permalink)
0
I went through all these cases, and redesign the PCB as 3 or 4 times thinking that the problem was in the hardware, but the problem is really in the drivers.The driver then generate a physical disconnection even, and with a program connected to the serial port is not able to release it and the only solution is to disconnect and connect.
I have a case of systems that never fail and other cases that do not last more than 20min connected, I have no problem with being disconnected, the problem is that the blocks port driver and is unable to reconnects.
How does one build the DLL, or have the driver source and change it?
#12
hanzibal
Starting Member
  • Total Posts : 47
  • Reward points : 0
  • Joined: 2012/06/11 00:40:35
  • Location: Sweden
  • Status: offline
Re:Problems with MCP2200 2012/11/20 09:03:46 (permalink)
0
What driver are you refering to? Windows driver usbser.dll or which one?

For WinXP, Service pack 3 is required since the mcp2200 is a composite device.

I doubt the driver is faulty, rather I think the disconnect event is a direct (and correct) result of the chip failing at which point it probably drops the data line pulldown.

Try configuring the !sspnd and !usbcfg outputs and monitor their state at the point of failure. My guess is that at least one of them turns high.

I now face a problem where the virtual com port fails. Windows says the corresponding com port does not exist but the HID interface continues to function correctly and the com port still shows in device mgr where it is reported to "function correctly" a statement that is obviously false. Also, no disconnect event fires.
#13
AndrewKeatley
New Member
  • Total Posts : 24
  • Reward points : 0
  • Joined: 2007/04/27 04:07:38
  • Location: 0
  • Status: offline
Re:Problems with MCP2200 2013/02/15 05:42:46 (permalink)
0
I have exactly the same problem where an MCP2200 comms channel appears to be there in Control Panel->Device Manager and says working correctly but it doesn't work as a comm port. Trying to open it says it doesn't exist.
 
This can happen after the port has been running perfectly for a while.
 
Did you find a way round it or a way to reset it other than unplugging the USB lead?
 
Thanks
 
#14
hanzibal
Starting Member
  • Total Posts : 47
  • Reward points : 0
  • Joined: 2012/06/11 00:40:35
  • Location: Sweden
  • Status: offline
Re:Problems with MCP2200 2013/02/15 06:48:47 (permalink)
0
Sadly no, it's stilll occurs sometimes :-(
#15
tcremel
New Member
  • Total Posts : 3
  • Reward points : 0
  • Joined: 2013/03/20 03:36:49
  • Location: 0
  • Status: offline
Re:Problems with MCP2200 2013/03/20 03:43:54 (permalink)
0
USB seems to  require a serial resistor on D+ / D-
 
Add a 22Ohm resistor between the D+ (on the mcp2200) and the D+ on USB connector. Same for D-.
 
You can also add a capacitor between the D+ (on MCP2200) and ground (22pFto 68pF) to improve immunity (same for D-)
 
Let me know if it works better.
 
I have an issue on my board, sometime, the communication goes only one direction. (from PC to board but not from board to PC). Could you check that you still have a one direction working when your MCP is hanged
#16
hanzibal
Starting Member
  • Total Posts : 47
  • Reward points : 0
  • Joined: 2012/06/11 00:40:35
  • Location: Sweden
  • Status: offline
Re:Problems with MCP2200 2013/03/20 13:56:49 (permalink)
0
Many thanks, I will also try this for my next design (may be a while though).
#17
Pacer
Super Member
  • Total Posts : 1171
  • Reward points : 0
  • Joined: 2004/12/01 09:29:20
  • Status: offline
Re:Problems with MCP2200 2013/03/20 15:21:20 (permalink)
0
tcremel

USB seems to  require a serial resistor on D+ / D-

Add a 22Ohm resistor between the D+ (on the mcp2200) and the D+ on USB connector. Same for D-.

You can also add a capacitor between the D+ (on MCP2200) and ground (22pFto 68pF) to improve immunity (same for D-)

Let me know if it works better.

I have an issue on my board, sometime, the communication goes only one direction. (from PC to board but not from board to PC). Could you check that you still have a one direction working when your MCP is hanged

tbh - kluges like these are not necessary and just make it sound as though you have some other underlying problem, which needs addressing. I don't think that you should be recommending these measures, particularly if your own device is not fully functional.
 
Just my 2 pence worth.
#18
tcremel
New Member
  • Total Posts : 3
  • Reward points : 0
  • Joined: 2013/03/20 03:36:49
  • Location: 0
  • Status: offline
Re:Problems with MCP2200 2013/03/21 00:53:29 (permalink)
0
well my implementation works.
I have opened a topic for my own issue. ( http://www.microchip.com/...tree=true&m=713067 )
 
In my first steps there were ne resistor on my USB line and some USB hubs were seeing the card, but not all the hubs. I had issues to get the whole thing recognized until I added the resistors.
I can give you back 1 penny.
 
Smile
 
#19
Axxel
New Member
  • Total Posts : 1
  • Reward points : 0
  • Joined: 2017/06/22 05:23:53
  • Location: 0
  • Status: offline
Re:Problems with MCP2200 2017/06/22 20:36:56 (permalink)
0
Dear friend.
I had exactly the same problems as you described.
Apparently, I solved this problem removing the ground plan just under the MCP2200.
Remove all ground plans near the MCP2200 and say goodbye to this problem :)
#20
Jump to:
© 2019 APG vNext Commercial Version 4.5