ICD-U Cannot Detect Chip on Iosofts ER22 (PIC18F452)

Todd · Joined: 17 Oct 2003 Posts: 4

Hi all,
I recently purchased the Iosoft ER22 board (with PIC18F452) and the Iosoft recommended CCS ICD-U. When attempting to do ICSP, the ready light on the ICD is steady green, but the ICD Control Program cannot detect the chip. I've selected the PIC18F452 in the advanced dialog to no avail.

CCS tech support blames the 4.7K resistor between MCLR and 5V and/or the 0.1uF capacitor between MCLR and ground. They claim the 4.7K resistor should be 47K and the capacitor shouldn't be there at all.

Iosoft tech support claims that they based the ER22's ICSP circuit on the PICDEM and are currently using the CCS ICD-U with the ER22 and suspect that the ICD-U version of software and/or firmware may be to blame (i.e., some versions work with this device ... some don't). Unfortunately Iosoft did not specify which version works ... I'm still waiting for a response to that question.

I'd rather not have to modify this $300 board for this prototype as one of the reasons for selecting this board was to prove commercial off-the-shelf ready availability.

Has anyone out there successfully used the ICD-U with the 4.7K resistor/0.1uF Capacitor configuration? If so, what versions of software and firmware work?

Current Versions:
ICD Control Program: V2.6
ICD Firmware: 1.13 Rev #97 for PIC16 PIC18

Thanks
_________________
Todd

Neutone · Joined: 08 Sep 2003 Posts: 839 Location: Houston

Guest · Posted: Wed Oct 22, 2003 2:47 pm

Todd:

I'm of no immediate help to you now, but I ordered a ER22 yesterday and will try my USB ICD2 on it and let you know.... at least you'll have a comparision there if that works and the CCS ICD-U does not.

I'd also be interested in trading notes with you as I go along if your willing. I'm planning on doing a serial gateway initially with it to see if it works as claims, and then look at a comercial license with my own product.

BTW, how long did it take you to get your ER22 after you ordered it?

Regards,

Jeff King

Todd · Joined: 17 Oct 2003 Posts: 4

Jeff,
Thanks. At this point, I'm leaning toward faulting the ICD-U as I find it hard to imagine that I'm the first one to ever try to use the Iosoft recommended CCS ICD-U with the ER22 board, and the ER22 is working as advertised otherwise.

The ER22 actually came in remarkably fast (within two days) considering that its coming from the U.K ... of course for the approx. $40 shipping charge, it should. I think the default, and only, shipping option was next day air.

I look forward to your results with the ICD2.

Todd
_________________
Todd

Todd · Joined: 17 Oct 2003 Posts: 4

email response from Jeremy Bentham:

I rechecked the Microchip documentation and they are still using 4.7K on the PICDEM.net. Our ER22 is manufactured in volume, so we can't change component values without a very good reason.

IIRC, the whole 4.7K/47K saga started when someone was concerned at the overdriven reset line would charge up the 5 volt rail through the 4.7K resistor. On our design, the 5 volt rail is relatively heavily loaded by the network controller and the wireless card, so the current from the reset line can have no such effect.

If the ICD-U reset driver is really incapable of driving a 4.7K load, then that must be
the fault of the ICD-U, not the many boards that use that resistor value.

Jeremy Bentham
Iosoft Ltd.
_________________
Todd

Douglas Kennedy · Joined: 07 Sep 2003 Posts: 755 Location: Florida

Without a definitive statement from Microchip for the MCLR to Vdd pull up when using ICSP many went for doing as they did. The orginal ICD demo board used a 47K resistor. on MCLR of the target processor.

Jeff King · Posted: Thu Oct 23, 2003 1:19 pm

Still no ER22 yet, but I did check my ER21.

It has a 4.7K to MCLR from +5Volts. I've successfully reprogrammed it using a MicroChip ICD2 (USB). So I suspect I'll have no trouble with the ER22 using the MicroChip ICD2.

So, in theory, you can remain "off the shelf" if you use the MicroChip ICD2. Or you can change the 4.7K resistor to 47K and use the CCS ICD-U. It however is not clear to me what the problem is.

Specifically:

Is it that the ICD-U cannot reset the part? (that seems unlikely even with a 4.7K pullup)

Or is it that the ICD-U cannot supply Vihh (programming voltage) against the 4.7K load to +5volts? If you measure it, it should be between 9.0V and 13.25V.

Also, see my question about what cards work on the ER21/22 on the tcp/ip lean thread.

Neutone · Joined: 08 Sep 2003 Posts: 839 Location: Houston

When programming it takes ~0.2 mA to bring the MCLR pin to 13 volts if the pull up resistor is 47K. If its a 4.7K pull up you need ~2 mA to reach 13 volts. This is only taking into account current flowing through the pull up. If the ICD-U was designed to supply up to 0.2mA @ 12V this all makes good sense.

Jeff King · Posted: Thu Oct 23, 2003 3:20 pm

PCM programmer · Joined: 06 Sep 2003 Posts: 21708

I don't use ICD, so I'm not that familiar with it, but I researched this
a little bit and found something that may help:

On this page, http://www.beyond-designs.com/microchip.htm
in this section: Microchip's new In-Circuit Debugger (ICD) they say:

The ICD module is easily modified to meet the requirement for more VPP current by replacing L1 on the ICD board with a 120 uH inductor (such as Delevan 1025-70K).

I believe they're referring to this schematic on page 65 (page 71 in
the Acrobat reader) of this document:
http://www.microchip.com/download/tools/picmicro/icds/mplabicd/51184d.pdf
It shows a 300 uH coil, L1, in a circuit that produces +21v to drive
a 12v regulator. I'm not a switching power supply guy -- I think
R.J. Hamlett could probably answer that question better -- about
whether changing that inductor to 120 uH will increase the current.

Of course, there's another question -- does CCS use this circuit in
their ICD-U ?

Todd · Joined: 17 Oct 2003 Posts: 4

Problem Solved
A lesson in (CCS) quality control ... three of the little springy contacts in the ICD-U's RJ45 connector were not making contact with the cable end, including the one connected to MCLR. Both the cable and ICD-U are brand new from CCS. A little adjustment with a bent paperclip and all appears to be working.

Thanks for the help, checking the suggested voltages lead to this discovery.

Jeff, I'd be glad to collaborate on ER22 stuff, I'm sure I'll have many questions also. My email is [email protected].

Thanks. Very Happy

_________________
Todd

Jeff King · Posted: Wed Jan 07, 2004 2:31 pm

I had a problem last night with the ICD-U40, out of the box, and ended up going back to the ICD2 and things worked fine. This morning, I recalled this conversation, tried the thing with the paperclip, and bang it works the first time.

The thing that worries me is this might be a low quaility RJ-45.... anyone recomend a good quality Rj-45 so I can replace this?

blak3r · Joined: 11 Jan 2004 Posts: 45

I just received a ICD-U40 and was having the infamous "cannot detect target chip" problem as well.

Before i stumbled upon this thread i hadn't considered the possibility of it being a mechanical problem.

In my particular case it was actually the cable itself. I kinda messed up my pins a little doing the paper clip trick so i replaced the RJ-11 connector.

The connector i used was AMP 555165-1, Digikey part number: A9031-ND

Thanks
_________________
"Everything should be made as simple as possible, but not one bit simpler" -- Albert Einstein

http://www.blakerobertson.com