FJN3303F

[TimeWaster]

hi,

i have already build a Geiger counter out of an Arduino Uno, a Geiger counter shield and many other parts (GPS, SDCard an so on), but i have problems with the voltage of the counter shield, so i want to build this part from scratch.

The layout of http://mightyohm.com/blog/products/geiger-counter/ is by far the simplest and best i found and i would like to recreate the high voltage and detection parts.

i have 2 questions:

1.: Unfortunately the FJN3303F is impossible to buy in Germany, and even the suggested replacement type (http://mightyohm.com/forum/viewtopic.php?f=15&t=501) i cannot find. Does anyone know another replacement type?

2.: The Arduino Uno runs on 5v, so i don't have another voltage at hand, is it possible to run this circuit on 5v?

thanks in advance for your help, if i am able to adapt this circuit i will definitely credit MightyOhm on my Blog with a Link.

[mightyohm]

I'm not aware of other substitutions, but you are looking for a high breakdown device with VCBO at least 600V, and VCEO as high as possible (400V for the FJN3303F). The current capability doesn't need to be high, 1.5A is plenty. (Actual current in the circuit is <50mA) The on-time and off-time probably matter a bit. You can compare the values in the FJN3303F datasheet (http://www.fairchildsemi.com/ds/FJ/FJN3303F.pdf) with other devices.

The circuit will work at 5V, but I would increase R6 (labeled R5 in the schematic) to 1K and R11 to 330 ohms. You might need to play with R8 to get a decent voltage range as you adjust VR1.

[TimeWaster]

thanx for your answer!

in the schematic (http://mightyohm.com/files/geiger/geiger_sch.png) R5 is already at 1K, but R6 is only at 330, did you mean R6?

[mightyohm]

Hi,

The labels for R5 and R6 are swapped in the layout vs. schematic. I was referring to the labels in the layout. The schematic is wrong, R5 is 330, and R6 is 1K in the current design. (And coincidentally you could probably just swap those two resistors to get the circuit working at 5V.)

[TimeWaster]

don't you think correcting the schematic would be good for users who want to recheck their pcb?

[mightyohm]

Yes.

[TimeWaster]

hm...

i ordered a kit from your site, build it on a breadboard, fed it 5 volts and must say i am confused.
i rechecked the circuit three times, i am positive that it is correct.

no matter if i set R5 to 1K and R6 to 330 or R5 to 330 and R6 to 1K, i never get more than 311 volts. (variable resistor on maximum).
in one configuration i get 250 volts and in the other 311 volts.
and yes, i measured it with a 10M resistor in series with the plus sensor of the multimeter which has a 10M internal resistance, so it is a voltage divider with a 2:1 ratio, i rechecked that with a 9 volt battery, i am measuring exactly half the voltage.

can you tell me
1. which resistor (ohm value, not the label, the label switching is extremely confusing) should go to the diode and which to the transistor in the circuit board (if i want to feed the circuit 5v)?
2. why do i only get 311 volts max? (no matter if the tube is connected or not)

[mightyohm]

10M is not enough series resistance. In your tests, you are assuming that your source resistance is low (like a 9V battery), but it is not - the equivalent source resistance of the HV supply is many megaohms. You need a larger resistor, at least 100M, preferably 1G, in series with your meter.

If the schematic is confusing, look at the photos of the kit on my site and compare them with the PCB layout. That will show you which resistor is installed where in the circuit. As I said, R5/R6 are reversed in the schematic. There shouldn't be any other differences between the schematic and layout.

Also, using a breadboard (the kind where you push in the leads) may affect how the HV supply behaves, so the voltage may not come out exactly right. Breadboards add undesirable parasitics due to the way they are constructed (not to mention the possibility of loose/bad connections). I recommend building the circuit on a small piece of perfboard instead, if you haven't done that already.

[TimeWaster]

i now rebuild the circuit on a perfboard.

result:
- the voltage is very hard to adjust (i'm adjusting after CPM since my 1Gohm resistor will take a couple of days to ship). there is a range where the CPM is alright, but it is only 1 degree of angle on the variable resistor, if i only twist slightly too far the CPM is going through the roof (a couple of thousand CPM). i think that there is something producing feedback in the circuit.
- if i hold my hand over the circuit the CPM changes. that is not acceptable, it would produce wrong measurements if i am holding the geiger counter.

i have not the slightest notion about HF and HV circuits, so i have no idea what i did wrong. i again rechecked the circuit it seems to be right.
could you give me some clues where to look at?

p.s.: 1Gohm resistors are VERY hard to find, and i have never in my life before paid over 20 euros for 2 resistors. unbelievable.

[mightyohm]

Hi,

I can't provide much help as I don't know what components you used or the design of your perfboard.

It might be useful for you to build up one of my kits as I designed it, to get a feel for how it works and how component substitutions affect the operation of the high voltage power supply.

Can you share a schematic of what you have built, with actual component values? I can't promise that I will know the answer, but I can try. One thing to check is that your variable resistor is 10 ohms, not 100 ohms.

It does sound like you have some unwanted feedback in your circuit, or noise coupling into the microcontroller. I went to great lengths to separate the microcontroller from the HV supply in my design, putting them as far apart as possible, to reduce the possibility of this coupling. And still, if you put your finger on some sensitive nodes, you can cause false CPM readings in my design. So this is a challenging problem, but one that can be solved with proper attention to coupling and good bypassing and grounding of the digital section.