Searching \ for '[EE] Hot plugging daughter board' in subject line. ()
Make payments with PayPal - it's fast, free and secure! Help us get a faster server
FAQ page: techref.massmind.org/techref/index.htm?key=hot+plugging+daughter
Search entire site for: 'Hot plugging daughter board'.

Exact match. Not showing close matches.
PICList Thread
'[EE] Hot plugging daughter board'
2011\09\13@201650 by Brent Brown

picon face
Having a little problem, thought I would share for interest and maybe someone has been here before.

Two boards, both of my deisgn. The base board has Silicon Labs C8051F120 processor running on 3.3V (not a PIC, hence EE tag). Recent revision added an 8 way header to allow a daughter board (factory option) to be plugged in to add extra featrues. Header provides +12V (motors etc), +5V, +3.3V and GND connections to supply power, and RXD and TXD for comms (3.3V level).

Newly designed daughter board sits on top of previously described board and connects by way of 0.1" header pins. Also has Silicon Labs processor (C8051F121) runs off 3.3V supply, communicates over RXD and TXD to main processor (also running at 3.3V), no need for any regulators etc on daughter board.

Problem: Hot plugging daughter board onto the main board fries the microcontroller on the main board. Happened twice now. First time it was me: forgot that 12V supply to main board is battery backed, so installed daughter board while I thought power was off. Ok, make a note: don't do this again.

Second time customer shipped the product, end user noted shipping had casued the daughter board to come loose, re-seated it without removing 12V battery first. Same result, dead main processor (won't respond to JTAG programmer, replace chip all is fine).

What I strongly suspect is happening is that plugging into header pins the GND pins are not contacting first (there is nothing to guarantee the order of connections) and when +12V pin contacts is couples power through discharged elelctrloytic caps on daughter board to GND rail, then "backfeeds" to +3.3V rail through internal body diodes of chips on that rail, pushing up the +3.3V rail and popping the processor on the main board. Similar could be happening on RXD, TXD and +5V rail... but nature of 2 failures so far suggest +3.3V rail more likely.

Solutions:
1. Don't hot plug the board
2. Fix the daughter board in place more solidly
3. Raise the height of the GND pins so they are more likely to contact first

Other:
- Have opportunity now to modify daughter board design now before going to full production, but seems would have to add an unreasonable amount of circuitry to isolate during power up +3.3V, +5V, RXD and TXD.
- Prefer not to re-design main board, even so I suspect it would not be trivial to add comprehensive protection.

For I'm going with 1 and 2 above, looking at the feasibility of 3. Keen to hear any other ideas~!

-- Brent Brown, Electronic Design Solutions
16 English Street, St Andrews,
Hamilton 3200, New Zealand
Ph: +64 7 849 0069
Fax: +64 7 849 0071
Cell: +64 27 433 4069
eMail:  spam_OUTbrent.brownTakeThisOuTspamclear.net.nz

2011\09\13@205715 by Josh Koffman

face picon face
On Tue, Sep 13, 2011 at 8:16 PM, Brent Brown <.....brent.brownKILLspamspam@spam@clear.net.nz> wrote:
> What I strongly suspect is happening is that plugging into header pins the GND pins
> are not contacting first (there is nothing to guarantee the order of connections) and
> when +12V pin contacts is couples power through discharged elelctrloytic caps on
> daughter board to GND rail, then "backfeeds" to +3.3V rail through internal body
> diodes of chips on that rail, pushing up the +3.3V rail and popping the processor on
> the main board. Similar could be happening on RXD, TXD and +5V rail... but nature
> of 2 failures so far suggest +3.3V rail more likely.

Ok, I'm definitely not an expert in this area, but would a diode on
the 3.3V line in between the connector and the rest of the circuit on
the daughterboard help?

Josh
-- A common mistake that people make when trying to design something
completely foolproof is to underestimate the ingenuity of complete
fools.
        -Douglas Adams

2011\09\13@211728 by Isaac Marino Bavaresco

flavicon
face
a) Ensure the ground connection connects before any other signals,
perhaps making it longer or using a special connector;
b) Put a series resistor (between 33R and 100R for fast signals, more
for slow signals) in each logic signal that crosses from one board to
another;
c) If possible, put a diode in series with the power signals;
d) Put a TVS from each signal to ground.


Best regards,

Isaac



Em 13/9/2011 21:16, Brent Brown escreveu:
{Quote hidden}

>

2011\09\13@222213 by Mark Hanchey

flavicon
face
On 9/13/2011 8:16 PM, Brent Brown wrote:
> Other:
> - Have opportunity now to modify daughter board design now before going to full
> production, but seems would have to add an unreasonable amount of circuitry to
> isolate during power up +3.3V, +5V, RXD and TXD.
> - Prefer not to re-design main board, even so I suspect it would not be trivial to add
> comprehensive protection.
>
> For I'm going with 1 and 2 above, looking at the feasibility of 3. Keen to hear any
> other ideas~!
  Add a zener diode on each of the 3.3V, +5V, RXD and TXD lines so it clamps the voltage at whatever is safe for those pins. Put a 10K + resistor across +12V power rails so that when the +12V battery power is removed those capacitors self discharge quickly enough to prevent someone from plugging in another board before they discharge normally.  Also consider adding an LED to the board near the connector that shows when the board is powered to remind people when plugging in another board.

Good Luck
Mark

2011\09\13@235320 by Brent Brown

picon face
Thanks for the repsonses, see comments below.

On 13 Sep 2011 at 22:22, Mark Hanchey wrote:

>    Add a zener diode on each of the 3.3V, +5V, RXD and TXD lines so it
> clamps the voltage at whatever is safe for those pins

Similar to d) below

>  Put a 10K + resistor across +12V power rails so that when the +12V battery
> power is  removed those capacitors self discharge quickly enough to prevent
> someone from plugging in another board before they discharge normally.

Yes, not a bad idea, but battery powered means I'm reluctant to add waste power unnecessarily. I think capacitor charging current is more likely the cause of the problem.

> Also consider adding an LED to the board near the connector that shows
> when the board is powered to remind people when plugging in another board..

Clever, but in as it's battery backed (12V 2AHr) even an LED @ 1mA or so would waste too much power.

On 13 Sep 2011 at 20:56, Josh Koffman wrote:

> Ok, I'm definitely not an expert in this area, but would a diode on
> the 3.3V line in between the connector and the rest of the circuit on
> the daughterboard help?

Yes, it would help, but affect other things, see c) below.

On 13 Sep 2011 at 22:17, Isaac Marino Bavaresco wrote:

> a) Ensure the ground connection connects before any other signals,
> perhaps making it longer or using a special connector;

Yes, will try to achieve that with standard 0.1" head pins by leaving the GND pins slightly higher.
> b) Put a series resistor (between 33R and 100R for fast signals, more
> for slow signals) in each logic signal that crosses from one board to
> another;

Good idea. The only signals are TXD and RXD, running at about 900kbaud, traces maybe 100mm long, could try 100R.

> c) If possible, put a diode in series with the power signals;

I'll consider that a bit more. I have +3.3V and +5V analog supplies derived from the power signals - will have to think through exactly what this will affect and by how much.

> d) Put a TVS from each signal to ground.

Would have to be on the main board which is less desireable to modify. Clamping voltage likely not be close enough to be effective.  Thanks again.

-- Brent Brown, Electronic Design Solutions
16 English Street, St Andrews,
Hamilton 3200, New Zealand
Ph: +64 7 849 0069
Fax: +64 7 849 0071
Cell: +64 27 433 4069
eMail:  brent.brownspamKILLspamclear.net.nz

2011\09\14@093210 by Isaac Marino Bavaresco

flavicon
face
Em 14/9/2011 00:53, Brent Brown escreveu:
>> a) Ensure the ground connection connects before any other signals,
>> perhaps making it longer or using a special connector;
> Yes, will try to achieve that with standard 0.1" head pins by leaving the GND pins
> slightly higher.

Simply displacing some pins up may not work, they may not be soldered
correctly to the board or they may limit the insertion of the ones that
weren't displaced and those may not make good contact. Besides, the
amount of displacement would be very limited, perhaps not ensuring that
the ground pins always make contact first.

Right now I'm designing with a bottom-entry 0.1" receptacle and I think
it may help you.

The bottom entry receptacle doesn't limit the length of the pins. If a
pin is too long it will appear over the top of the receptacle, while the
other pins still are correctly connected and at their normal height.

All you need is to remove some (ground) pins from the header and replace
them with pins of a header with longer pins. This way, the ground pins
will be taller than the others and will make contact first. As you push
the daughter board down, the ground contacts will slide along the ground
pins, ensuring good grounding while the remaining pins begin to make
contact.


Best regards,

Isaac

2011\09\14@180857 by Dwayne Reid

flavicon
face
At 06:16 PM 9/13/2011, Brent Brown wrote:

>What I strongly suspect is happening is that plugging into header
>pins the GND pins
>are not contacting first (there is nothing to guarantee the order of
>connections) and
>when +12V pin contacts is couples power through discharged
>elelctrloytic caps on
>daughter board to GND rail, then "backfeeds" to +3.3V rail through
>internal body
>diodes of chips on that rail, pushing up the +3.3V rail and popping
>the processor on
>the main board. Similar could be happening on RXD, TXD and +5V
>rail... but nature
>of 2 failures so far suggest +3.3V rail more likely.

Just a thought - can you add a power switch (MOSFET or whatever) to the 12V line as it enters the daughter board?  The 12V rail would be dead until something tells it that its OK to power-up.

Note that this switch is on the daughter board, since you mentioned that you would like to avoid making design changes to the main board.

For example, you could make a simple one-transistor controller that monitors for about 3V between the daughter-board ground and 3.3V rail.  If either ground or 3v3 was missing, don't allow the MOSFET to turn on.

dwayne

-- Dwayne Reid   <.....dwaynerKILLspamspam.....planet.eon.net>
Trinity Electronics Systems Ltd    Edmonton, AB, CANADA
(780) 489-3199 voice          (780) 487-6397 fax
http://www.trinity-electronics.com
Custom Electronics Design and Manufacturing

2011\09\14@202109 by Oli Glaser

flavicon
face
On 14/09/2011 23:08, Dwayne Reid wrote:
{Quote hidden}

I was thinking along those lines too - you could also add an RC filter on the base to bring the rail up slowly.

2011\09\15@020642 by Brent Brown

picon face
{Quote hidden}

Thanks Isaac, good idea, might go with that nex time. Can't change receptacle easily now as it would require a board mod.

I did have a quick go at offsetting the pin heights on the standard 0.1" header pins I'm using. I found I can extend the GND pins out 0.75mm before they hit the stop in the receptacle. Will still solder OK as board is D/S PTH. The rest of the pins could be pulled back say 1.25mm (there is at least 2.5-3mm of "contact sliding" distance) making my GND pins stick out a usefull 2mm further than the others. Could feasibly make a simple jig to press the pins to the correct lengths. Undecided about wether I'll do this or not for production, depends on how favourable the assembly company is to the idea.

-- Brent Brown, Electronic Design Solutions
16 English Street, St Andrews,
Hamilton 3200, New Zealand
Ph: +64 7 849 0069
Fax: +64 7 849 0071
Cell: +64 27 433 4069
eMail:  EraseMEbrent.brownspam_OUTspamTakeThisOuTclear.net.nz

2011\09\15@022145 by Brent Brown

picon face
{Quote hidden}

Thanks Dwayne (and Oli). I had somehow overlooked that option, initially because the 12V rail handles a fair bit of current (up to 4 x DC motors and 2 x linear actuators) and I didn't want any voltage drops. Designing it into the board now using a IRF7424 P-channel MOSFET (30V, 10A, 20mOhm, SO8 package, low enough price), and using the single remaining spare I/O line on the micro control it. Will put power up delay in code before turning 12V rail on, nice to be able to turn it off also to reduce quiescent current. Need a couple of BJT's and a few resistors for level shifting and to make sure it doesn't turn on during reset (8051 based chip has weak intenal pullups enabled by default). Have included RC network of 5ms time constant, to limit inrush current of some electrolytic capacitors.

Technically speaking the 5V rail could also couple to the 3.3V rail if GND pins not connected, but these rails are through regulators so current is limited, and take away a diode drop and the voltage difference between them is much less likely to cause the same kind of damage the 12V rail does.

-- Brent Brown, Electronic Design Solutions
16 English Street, St Andrews,
Hamilton 3200, New Zealand
Ph: +64 7 849 0069
Fax: +64 7 849 0071
Cell: +64 27 433 4069
eMail:  brent.brownspamspam_OUTclear.net.nz

More... (looser matching)
- Last day of these posts
- In 2011 , 2012 only
- Today
- New search...