Olimex ICD2 Broken!

[3v0]

What text did you sent to Olimex?

It is clear that the body who responded was having a bad day.

At least the msg you got back was not a canned response.

The Junebug works well for me. It is about the same price as the assembled Pickit2. All of the units I built worked the first time. It is good to know that if any of them break I can fix them. Either by myself or with the help of other nice people here.

EDIT: Did you try testing MCLR PGC and PGD between the processor socket pins and the ICSP connector as I suggested earlier ?

 

[HerbertMunch]

All appears to be well with the tests that you suggested.

 

[HerbertMunch]

I dont understand!
It works about 3/10 for programming, and 0/10 for debugging.
Now olimex has replied back saying that the ICD2 is not broken, when it clearly is.

 

[HerbertMunch]

Here is the conversation between olimex and myself:

The very fact that it doesnt work, is grounds enough to blame your ICD2!
----- Original Message -----
From: OLIMEX Support
To: CPB22
Sent: Thursday, January 31, 2008 10:34 AM
Subject: Re: Faulty ICD2


Hi,
these are perfect values and with the self test passing I think you don't have ground to blame ICD2 to be faulty
Best regards
Tsvetan / Olimex
----- Original Message -----
From: CPB22
To: OLIMEX Support
Sent: Thursday, January 31, 2008 12:32 PM
Subject: Re: Faulty ICD2


Target VDD = 5.04
Target VPP = 13.26
MPLAB VPP = 13.26

Thanks.
----- Original Message -----
From: OLIMEX Support
To: CPB22
Sent: Thursday, January 31, 2008 10:26 AM
Subject: Re: Faulty ICD2


and the target voltages Vcc and Vpp during programming and debug are?...
----- Original Message -----
From: CPB22
To: OLIMEX Support
Sent: Thursday, January 31, 2008 12:19 PM
Subject: Re: Faulty ICD2


Thanks very much for the lightening quick reply.

I can appreciate that you have a wide range of customers, all with varying skills.
May i suggest that if you dont know, you shouldn't make assumptions.

The ICD2 is definately at fault.
I usually power the target through the ICD2. I have however tried applying power to the board through a 9v battery and a regulator. Still doesnt work.

The Hex that i obtained from blue room electronics functioned exactly as your hex did.
As stated before, MPLab had no problems talking to the replacement chip and firmware (**broken link removed**) that i installed. Even with the new chip, the problem still remains.

I know that the ICD2 is faulty, because i have been using it for many months, but only now im experiencing problems.

I have tested the ICD2 on both my PC and my laptop. Both fail. I have tried using different versions of MPLAB, all with no success.

There is definately something wrong with the ICD2.


Many thanks,
chris.
----- Original Message -----
From: OLIMEX Support
To: CPB22
Sent: Thursday, January 31, 2008 9:59 AM
Subject: Re: Faulty ICD2


Hi Chris,
I'm sorry if my e-mail offended you.
How on earth I would know if you put blank PIC or not?
We have *any* kind of customers and some of them are experienced some are not.
Some of our customers find blink LED code on the internet web pages and blame our programmers when they try to burn the HTML pages with the blink LED examples in PIC and it doesn't blink!
Please read again what you wrote below and tell me if you were on my place what you would suggest?
On first place you shouldn't replace the on board PIC. I have no idea what is blueroomelectronics, but our ICD2 have custom bootloader which works with "A" version of PICs as original Microchip firmware support only PIC16F877 without "A" suffix,
If selftest pass the problem is not in the ICD2, check your voltage levels: Target VCC, VPP, do you supply target voltage from ICD2? in this case switch over to external power supply of your target.

Best regards
Tsvetan

----- Original Message -----
From: CPB22
To: OLIMEX Support
Sent: Thursday, January 31, 2008 11:48 AM
Subject: Re: Faulty ICD2


Tsvetan,

I am not an idiot. You must think me very stupid if you think that I would expect a blank chip to work.

Im fairly experienced with using pics now, I have owned many programmers.
Far from having no 'glue' what to do, I was able to flash a replacement chip.
The reason i blame the ICD2 for this, is because the ICD2 is at fault.

I didnt do anything. It just stopped working. I havent used it for about 25 days, and when i got it out of its box, the problems started.

PIC

I replaced the pic with a 16F877a (the same chip), I flashed the new chip with firmware that i obtained from blueroomelectronics.
When i put the replacement chip in, it functions as well as the old chip. but the problem is still there.

After i put the new chip in, i was able to use MPLAB to download the latest operating system onto it. Strangley there was no problems encountered during this transfer.

This has allowed me to succesfully rule out the pic being broken.
I have now put the original undamaged pic back in.

CYPRESS CHIP
There is also NO problem with the cypress chip. Usb functionality works fine.

CABLES

I have tried multiple power cables, 2 (proven) usb cables, 2 (proven) ICSP cables all of these have no effect. (as before, only about 3/10 of the programming attempts work).


TESTING

I have tried numerous different boards that I have made over the last few months. All these boards previously worked with the ICD2. Now id say that only about 3/10 attempts to program any board, are successful.
When i try and attempt to debug, it allways says "cannot enter debugging mode." 0/10 attempts to debug succeed.

SELFTEST
Self test passes for some reason.


Thanks,
Chris.

----- Original Message -----
From: OLIMEX Support
To: CPB22
Sent: Thursday, January 31, 2008 6:20 AM
Subject: Re: Faulty ICD2


Dear Chris,
The information you provide is not enough to make any speculations what happened with your ICD2.
It worked but now doesn't work may means anything: broken cable, power supply problem, burned ICD2, how could we help you based on this information?
Tell us what you did when the ICD2 stopped working, does it connect to USB when you plug it, what selftest show?
Verify failed may also mean just weared PIC on the target i.e. broken PIC which can't be programmed anymore, why blame the ICD2 for this?
Changing the onboard PIC with the firmware is the worst thing you could do if you have no glue what to do. What did you put as replacement: blank PIC? How do you expect ICD2 will work with the new PIC?
Best regards
Tsvetan / Olimex
----- Original Message -----
From: CPB22
To: info@olimex.com
Sent: Wednesday, January 30, 2008 7:49 PM
Subject: Faulty ICD2


Hi,
I bought an olimex iCD2 from sparkfun less than a year ago ago.

It now refuses to program anything, allways saying "Failed to verify". It will not enter debugging mode either. I have tried changing the 16F877A, but the same problem is encountered.

Please could you provide support.

Many thanks.
Chris

 

[3v0]

Given that the voltage levels look right and it works some of the time it may not be broken.

You need to go back and check the basic things. Please do not be insulted.

Have you tried it with a shorter/different ICSP cable. Could just be an intermittent cable.

Have you tried it with more then one target.

When you did the tests I suggested you seen the correct voltages but did it also go back to gnd. MCLR functioned correctly based on the three pins that control it.

 

[HerbertMunch]

Given that the voltage levels look right and it works some of the time it may not be broken.

You need to go back and check the basic things. Please do not be insulted.

Have you tried it with a shorter/different ICSP cable. Could just be an intermittent cable.

Have you tried it with more then one target.

When you did the tests I suggested you seen the correct voltages but did it also go back to gnd. MCLR functioned correctly based on the three pins that control it.

No insult taken!

As mentioned in my emails to olimex:

CABLES

I have tried multiple power cables, 2 (proven) usb cables, 2 (proven) ICSP cables all of these have no effect. (as before, only about 3/10 of the programming attempts work).


TESTING

I have tried numerous different boards that I have made over the last few months. All these boards previously worked with the ICD2. Now id say that only about 3/10 attempts to program any board, are successful.
When i try and attempt to debug, it allways says "cannot enter debugging mode." 0/10 attempts to debug succeed.

Ill test it again to see if returns to gnd.

thanks.

 

[HerbertMunch]

it does.

here is the latest from olimex:

Hi Chris,
If your selftest pass I can blame your target (PIC).
Another issue is if you burned somehow partially your buffers on the ICD2 so they keep good voltage levels in static mode but fail on timing.
Our buffers are on sockets so they can easily be exchanged.
Best regards
Tsvetan / Olimex

Is this likely?

 

[HerbertMunch]

Hiya Herb,
I had a problem with my ICD2 clone from olminex, it turned out I had pullups on the pgc, pgd and when I tried to program I blew the 3 transistors right below the yellow 20K pot in the pic Bill put up. The one right next to the pot and the 2 underneath it. I did a post on the sparkfun forum and both Sparkie and olimex chimed in to offer help. fortunatly by simply replacing those transistors got the clone working again and its been perfect ever since

now the fun part what did you do right before it stopped working ?

hope this shed's some light on the problem

Cheers Bryan

Sorry mate, i must have skipped past your post when i posted, probably because it landed on the second page.

Im not so lucky when speaking to olimex! he thinks that my ICD2 is not to blame, even though it definately is.

I think that im probably just going to replace the buffers as he suggested, and if that doesnt work, i will replace everything!

I would like to buy an inchworm+ as everyone speaks highly of it, not to mention that its creator is a top bloke and a regular here
The problem is that i am but a lowly student. A student who is waiting for his next 'hit'/infusion of student loan money

 

[3v0]

We were all lowly students at one time. I remember digging it the sofa to find a few cents to buy resistors at Radio Shack.

You may be happier with a Junebug.

If you are at a U can can possibly get you hands on a digital scope. Take a look at what the various signals look like in real time.

 

[blueroomelectronics]

The schematic posted here is for your programmer, lets see if we can make it function.
Like 3V0 mentioned a scope would be a great tool but if you don't have one do you have a multimeter or logic probe?
What target board / PIC are you using? Can you post a photo?
It would appear to be the output stage of your ICD2.
First thing I would do is swap the 74HCT126 ICs and try it out, is the error the same or different.
Next remove all socketed ICs, check for pin or socket damage (a pencil eraser is handy here) and reseat the ICs.
Also carefully check for a dry solder joint, connectors can get them from constant unplugging / plugging so reheating the solder joint won't hurt it.

 

[donniedj]

Has anyone else had any problems with the olimex clone, if so did you manage to fix them?

Yes, and solved them with a hook shot to the trash. Went PICkit2 and wished I had a time machine to correct the error of my ICD2 clone ways.

 

[HerbertMunch]

Thanks everyone for your replies.

We were all lowly students at one time. I remember digging it the sofa to find a few cents to buy resistors at Radio Shack.

You may be happier with a Junebug.

If you are at a U can can possibly get you hands on a digital scope. Take a look at what the various signals look like in real time.

The schematic posted here is for your programmer, lets see if we can make it function.
Like 3V0 mentioned a scope would be a great tool but if you don't have one do you have a multimeter or logic probe?
What target board / PIC are you using? Can you post a photo?
It would appear to be the output stage of your ICD2.
First thing I would do is swap the 74HCT126 ICs and try it out, is the error the same or different.
Next remove all socketed ICs, check for pin or socket damage (a pencil eraser is handy here) and reseat the ICs.
Also carefully check for a dry solder joint, connectors can get them from constant unplugging / plugging so reheating the solder joint won't hurt it.

Yes, and solved them with a hook shot to the trash. Went PICkit2 and wished I had a time machine to correct the error of my ICD2 clone ways.

I have access to only an analogue scope, but unfortunately Im doing the wrong course at uni: Computer science. Havent even had a whiff of electronics yet.

Whats better about the junebug? Can you still debug targets?

I have ruled out the boards involment in the problem. I have checked many differenent boards, they all will no longer go into debug mode, and they all program only about 30% of the time.

I have included my most recent board anyway.

Once again,
thanks very much.
Chris.

 

[3v0]

Yes the Junebug will debug targets. It is a clone of the PICkit2. See Microchip's site for the list of targets the PICkit2 can debug.

The Junebug includes an onboard target with a few do-dads you can play with. See

The schematics are included in the build instructions.

 

[blueroomelectronics]

Your old scope should work fine, try watching the PGC & PGD lines on the programming connector. Your target schematic looks good. Everything seems to be there and in the right place.
Does MPLAB think your VPP is ok? Is VPP switching properly *use MPLAB to test reset/run
Also did you swap the 74HCT126 ICs? Try it with the scope.
Also look at the Inchworm+ schematic on my site, it's simple design will show you just the basics and the 16F877 is the same so use your scope to look at any pins that go to PGC & PGD.
Does your school have access to a logic pulsar?
If not I would remove the 16F877 and with a 1K resistor and a short length of wire build a simple logic pulsar, toggle those programming pins at the 16F877 socket and watch the results on the programming connector.
And yes the Junebug debugs most 16F & 18F PICs and is just as fast as the ICD2 (not the 24F 30F or 33F yet...)
Edit:
The ICD2s 16F877 uses PORTC for controlling the rest of the programmer so it's those pins you'll want to toggle.

 

[mvs sarma]

Though,Part of the schematic doesn't tally, the 9Vac input point, it is very handy and helps debugging fast. -- olimex has bridge and perhaps the resistor is 4E7 / 2watt in series with input to regulator. U11 (DG411 could not be located on top of board- perhaps it may be on the solder side as SMD--? Even many component numbers and others would change from the board layout.
regarding debugging, Herbert Munch may get the detachable 74HCT devices checked outside on a bread board and he could find faulty gates if any. few components like transistors may perhaps be checked in-situ.

 

[blueroomelectronics]

Good eye mvs_sarma, the DG411 is probably not on the version in the photo. Seems that one may use transistors for VPP switching.

Edit: just noticed Chris's board is similar to my Unicorn board.
**broken link removed**

 

[HerbertMunch]

thanks for the input everyone.

I havent changed the dips yet. I dont have replacement ones. Im just in the business of checking the connections. I have seen a few dull joints, that need a bit of heating.

That unicorn LCD looks exactly like the ones I got from china.

cheers.

 

[blueroomelectronics]

Don't change the dips yet, the pair of 74HCT126 chips are identical, simpy swap them and see if anything changes.

 

[mvs sarma]

Good eye mvs_sarma, the DG411 is probably not on the version in the photo. Seems that one may use transistors for VPP switching.

Edit: just noticed Chris's board is similar to my Unicorn board.
**broken link removed**

Thanks Bill
the JUNEBUG is Nice and healthy sir, and a friend of mine at Bangaluru ( formerly Bangalore, in south India) wanted to test it, so it is there and he writes PIC programs.
Nice Rev2 of assembly manual. I would post a photo of the assembled version that I took before sending the board. I have perhaps to re-install the Kodak Create and share software, after a mobo swap to Intel945GCNL.

Ofcourse, I have to now check the USB voltage which i had a problem with my earlier board.

 

[blueroomelectronics]

I look forward to a photo, Krumlink posted his and it looks great. Like yours & Futz he built his up from part he supplied.