PullDown resistor values?

[Pommie]

I have a board and part of it contains this,

The problem is that this is a board for teaching programming so the I/O pin starts as input and if you touch the I/O pin with your finger the LEDs light. I thought the 10K to ground would fix the problem but it doesn't. I'm thinking of trying 4K7s next to see if that prevents it.

Am I missing something obvious here?
Anyone any suggestions?

Thanks,

Mike.
Edit, Micro running at 5V.
Edit2, A little more info, if I connect the board to my PC it stops happening. I suspect this is because my PC earths the ground line and the 12V supply I'm using doesn't. Is there anyway to tell if a particular supply is earthed or not?

 

[augustinetez]

Personally, I use 470 from micro to transistor (or 220 if more drive current is required) and 4k7 at the base of the transistor to ground.

 

[Pommie]

I've just changed the pulldowns to 2K2 with the same results. I can't change the 2k2 base resistors because they are 0402 SMDs and I don't have superman's vision. It seems to be the fact that the whole board is floating, if you earth it in any way the problem goes away. I wonder if I can find some non isolated 12V 1A supplies. It's not really a problem but an annoyance, the kids use laptops so even the USB cable won't provide an earth. I may have to build this from discrete components to try and solve the problem before getting new boards made.

Mike.

 

[augustinetez]

I just had a look at the current Altronics and Jaycar power pack offerings and none of those are earthed in the 12v 1A range

The ones that use an IEC mains connector may be earthed but they are overkill for your use.

Something weird is happening - haven't got a ground track from the micro missing by any chance?

 

[danadak]

TH BC847 can have quite a lot of G, aggravating the problem, which group part do you have ?

Regards, Dana.

 

[AnalogKid]

I wonder if I can find some non isolated 12V 1A supplies.

No no.

I know what you mean, but that's not what you said. Always stay with fully isolated power supplies. Then, connect the output return to the input AC GND. Not the neutral, and obviously not the line. On many supplies there is a pc board mounting screw location on the secondary side that is within or has a trace going to the output side return. Inserting the screw makes the connection to the power supply chassis. Another method is a third output terminal so you can optionally install a short jumper between the output return and earth ground. This is very common on lab/bench supplies. And (geezer reference), some open-standard industrial computer bus systems of the 80's and 90's.

Separate from that, as above the 847 has about twice the minimum gain of a 2222 or 4401. Also, it comes in three gain ranges, and the highest has a typical gain over 500.

Finally, re-check the connection between the transistor circuit GND and the microcontroller GND.

ak

 

[ronsimpson]

I usually make the circuit like this because it increases the turn_on_voltage.

if you touch the I/O pin with your finger

Then don't touch the I/O pins

 

[Pommie]

Sorry, non isolated was the wrong description. I did, of course, mean with the DC negative terminal grounded.
As for checking connections, I have and the board works fine when programmed. If I set the pins to output and low the problem disappears.
The transistor is a BC847A so max gain 220 typ. 180. with a 2k2 base resistor and 5V I/O pin it gets ~2mA base current.
Guess I'll knock it together on a breadboard and see if I can find a way to fix it.
Weird thing is, some other sections are using 2N7000s and the pullup/down fixes this completely.

Mike.
BTW, sorry for the late reply. Been at my daughters this weekend (lambing season) which is 2 hours West of Brisbane = (almost) no internet.

 

[Pommie]

Just received some new 12V 1A wall warts and they appear to be earthed. Problem solved.

Mike.
BTW, I went down to 1K pulldowns with no success.