ARM-USB-OCD not reliable ?

Development Boards MicroMod
1

Hi,

we have some annoying problems with our ARM-USB-OCD

JTAG interfaces from OLIMEX.

They work fine when connected to a PC with a short

thick USB cable.

Using our normal somwhat thinner and longer cables

they turn themselves off.

I have the larger voltage drop on the cable under suspicion,

but not quit sure.

In the USB spec I found that every USB device has to work

from (alt least) 4.75V to 5.25V.

As there goes some dynamic signaling on, its not easy to

measure the voltage correctly.

But when the ARM-USB-OCD goes idle I measure 5V on the

longer cabel an 5.06V on the shorter one.

I read here something about connecting the interface to an

USB hub.

I wonder what the Olimex people say to this.

From my point of view this device is not designed

to work on a standeard USB port!

Maybe there is some hardware patch available, especially

when not using the interface as a power source?

I think the FTDI chip works with lower voltage level (4,35V)

when powered from USB.

Regards,

Joerg

2

Joerg,

ARM-USB-OCD have FTDI chip inside, we also have noticed that FTDI chips (not only F2232 but also FT232 which is used in PIC-MCP-USB) will not work on long cable with high resistance or work not reliable, but it’s hard to say which is out of the specs FTDI chips or the cheap cable you try to use :slight_smile:

Best regards

Tsvetan

3

Hello Tsvetan,

thanks for your reply.

It seems that the high inrush current is responsible

for reset of USB connection.

I’ve replaced the big 250uF capacitor with a 100uF

and got stable connections with longer cables.

Output voltage has large ripple of course, but I don’t use it.

I measured a voltage drop of about 1,2V on VCC of the FTDI,

when the FET for the switching regulator closes (PWREN#).

But something around 4V should be enough anyway.

Adding a bypass capacitor to VCC had no effect.

So I still wonder if the FTDI chip resets because of

the larg voltage drop when switching on the regulator,

or if newer motherboard hardware is more sensible to current peaks.

I still think Olimex should have done more testing

on this device. :frowning:

As I need no output voltage I want a patch to switch

off the regulator. If it’s possible, I’ve seen some magic around powering

the 74AC244…

Will there be a new hardware version addressing these problems?

Regards,

Joerg

4

When decoupling a power source wich drives a switched load, use low ESR capacitors. And it is better to place two 47uF in parallel, instead of one 100uF.

Add a ceramic 10nF, 100nF or 1uF capacitor to help filtering HF noise.

PCB design could be also an issue. Don’t forget the basic rules for power supply lines: short and large connections :lol:

Angelo

5

Hi,

despite the poor support from Olimex,

I found a patch that works for me,

and hopefully for everybody who has problems

running ARM-USB-OCD on their PC.

As already said the problem lies in the high current

when the FTDI chip turns on the switching power supply.

It does this by driving FET1 low.

(at the USB connecter right above the wire jumper)

Replacing R35 (1k) with a 10k resistor rises the turn on time

of the FET enough to reduce the current peak significantly.

Alternatively you can increase the capacitance of C19

by soldering an additional capacitor backpack (about 200nF).

These changes helped me running this device at several

PCs with normal USB cables.

Maybe you need to play around with the part values a bit.

Maybe you send your devices back to Olimex, so

they do it for you.

Regards,

Joerg