Programming at 3.3V or below

[crystal_lattice]

Hi.

I have a Microchip ICD1 running CCS frimware, I followed all the mods listed in the ETN21 from microchip.

The ICD does not want to connect to the target at 3.3V but only starts to connect and work at about 3.5V I want to use the ICD for a project using a SD card running at 3.3V so 3.5V is way too high.

The voltage regulator I'm using is a LM317 connected to my PSU and adjusted to 3.3V the input has been adjusted from 5V-15V but no change.

The ICD utility from CCS reads the voltage as about 3.65V-3.72V once it has connected (voltage adjusted to about 3.5V)

Any suggestions??

Thanx in advance

[Ttelmah]

What processor?.
What connections to the ICD? (the resistor values used, have to change on some low voltage chips).

Best Wishes

[mskala]

I'm probably missing something about your setup, but in general I have used programming at 5V on a 3.3V system by either 1) using a diode on the power to the microchip device or 2) using a dedicated power isolating chip, so you can program without powering the rest of the system.

[ljbeng]

I have 24 to 5 to 3.3v on my board. I run 5v to the ICD and 3.3v to the chip (18F2620) and SD card. I connect the icd to the chip like normal, MCLR,B6 and B7. This works great for me.

[Ttelmah]

Yes.
The key question is 'what chip'. If you look at things like the 16F883 (as an example), though the chip runs down to 2v, it requires 4.5v minimum for programming. Some other chips are happy to program to their minimum operating voltage. Generally the solution is to diode connect as has been outlined above.

Best Wishes

[crystal_lattice]

Hey thanks for all the replies!!

The chip in question is 18LF4620, I'm using a development system from my university so i can't remember the resistor config on the idev board, if you could be more specific Ttelmah then I can check tonight (Am at work at the moment) would such resistor info be in the mcu datasheet?

The diode config sounds easy but may I be so blind as to ask how??

The dual voltage also sounds good but i don't want to mod my dev board. on the final product board it might be an option but i'm planning on using a bootloader with the SD card (Anybody got working code/examples??) so there should not be any need for ICD support.

I have been thinking that the LM317 voltage reg add-on might be problematic as well since i didn't add all the smoothing caps as required since it is fed from my PSU (which doesn't give any problems on its own) I bought some tantulum caps...

PS. Even with no chip in my dev board the ICD does not connect, is this still related to resistors on dev board?? resistors as stipulated in ETN21 have been replaced.

Any help welcome!!

Thanks

[crystal_lattice]

Hi guys.

I have plastered my LM317 regulator add-on with caps (Tant, elec, poly and ceramic) but still no connection from ICD. I checked the resistors on the board but don't think that will prevent ICD from connecting if no target is in dev board. Resistors are as follows :

5V----- >|-- /\/\/\/---|--/\/\/\/------o o------- 0V

>| = diode 1N4148
/\/\/\/ = First res =10K
/\/\/\/ = Second res =100R
--o o-- = reset switch
--|--- = junction that tees off to MCLR/Vpp through a 1K series res.

sorry for the primative ASCII scetch... a picture says a thousand words...

[crystal_lattice]

Hi all. I got my Microchip ICD1 working with CCS on a target running on 3.3V. The solution?? Upgrade the 232 level converter from a MAX232 to a MAX3232 pin compatible low voltage version!!! ...and follow instructions in the ETN21 note...

After reading the ETN note numerous times i finally clicked that the MAX232 can't operate below about 4V. I ordered my samples from maxim by end of November and only got them early January but is was worth the wait. I tested ICD1 on a 18LF and works like a charm without any cap changes.

Hope this helps somebody else trying their hand at low voltage ICD1 programming.

Kind Regards