PIC24EP512GP806 - Use port B w/programmer connected [CLOSED]

benoitstjean · Posted: Thu Nov 24, 2022 6:58 am

Device: PIC24EP512GP806
Compiler: 5.026

Hi all,

This may sound like a trivial question but I've never really looked into it quite honestly.

I need to use a 16-bit port in input and the only two 16-bit ports on this PIC are port B and C.

I can't use port C because RC12/15 are used for the external oscillator (a 29.4912MHz oscillator).

As for port B, well B3/6/7 are used for the programmer. I want to leave the programmer connected while still be able to use the circuit after its been programmed. The device connected to port B will send it 16-bit data and I don't want this data to go to the programmer just like I don't want the programmer to send signals to the connected device.

What is the simplest way to achieve this with minimal parts? I was thinking of using low drop-out schottky diodes but this will lower the voltage a bit. I don't think it will matter since most likely the voltage will still be in the acceptable range but perhaps someone has a better solution.

Thanks,

Ben

temtronic · Posted: Thu Nov 24, 2022 7:11 am

Microchip has ( had ??) an AB on using ICSP and 'stuff' connected to those pins....yeah, decades ago but I bet it's still there..

In the early years (think 16C84..) I recall using a 4PDT toggle switch. 100% isolation,no voltage drop or other 'weird' stuff....
4th pole fed 2 LED ..red for pgming, grn for go (running)

course this was back in the wirewrap dayze with PLENTY of room for parts...

alan · Joined: 12 Nov 2012 Posts: 357 Location: South Africa

Programming are not on C12/C15

Ttelmah · Joined: 11 Mar 2010 Posts: 19545

Realistically, as jay says, a switch is the only reliable and simple way to do
this.
The actual data pin for programming is bi-directional, so not easy to do
with diodes. You have three pins involved, PGEC, PGED & MCLR.
MCLR really needs to be disconnected from the normal reset circuit. The
programmers normally do not like it if this is pulled up by anything more
aggressive than perhaps 20KR, and certainly do not like a capacitor on
this pin. PGC is only an input to the PIC, so could probably be handled by
some sort of diode setup, depending on what you are using the pin for.
PGD is the data pin and is used for both input and output, so a physical
switch is by far the easiest way to do this. Many of the old prototype
boards just had jumpers that you could move to change the pin use.
Physically disconnecting with a switch also removes the possibility of
damaging the programmer.
Now because the chip supports moving the pins to different locations, they
are called PGECx and PGEDx. Why not just use the alternate pins?. You
are already avoiding port C, so use C13 & C14. You can use any pair of
the PCEC and PGED pins for programming. If you want to use them for
ICD, you have to specify which pair you are using with the ICSP=x
setting in the code.
You have the choice to use B6/B7, C13/C14, or B0/B1.

On MCLR, look at section 2.4 of the data sheet. Note how it shows a
jumper being used to disconnect the capacitor.

benoitstjean · Posted: Fri Nov 25, 2022 7:31 am

Hi guys, ok thanks. Dip switches is the route I will take if I realize that what I am trying to accomplish is not feasible with my current setup.

Thanks!

Ben

temtronic · Posted: Fri Nov 25, 2022 9:31 am

One possible problem if you don't get the isolation correct, is what happens to whatever is connected to the I/O pins, say during programming.

PGMMING Clock and data could get transferred to relays or valves causing no end of 'fun'..... Smile

newguy · Joined: 24 Jun 2004 Posts: 1909

PCM programmer · Joined: 06 Sep 2003 Posts: 21708

The Microchip documents for programmers show these pullup sizes:

ICD4 - 10K
ICD3 - 4.7K to 10K typical

PicKit4 - 10K to 50K
PicKit3 - 50K typical

newguy · Joined: 24 Jun 2004 Posts: 1909

Actually, we also use an ICD4 with a 4.7k PU on /MCLR with no ill effects.

Ttelmah · Joined: 11 Mar 2010 Posts: 19545

Theree used to be a problem years ago, with one of the ealy programmers
not being able to handle <20K. Microchip at that time recommended 22K.
I've seen some problems even more recently when you go much lower.
So I prefer to keep the presented resistance high.
Remember on really old chips, the Vpp on MCLR, went up to over 20v.
If you used 4.7KR on these the programmer had to deliver about 5mA.
Also on small boards this could result in the actual supply rail being drawn
up.
So for modern chips using low voltages to program, 4.7K is OK. But on
old chips, you need to use much higher resistance values.