Radiation Detector

 Radiation Detector With Geiger Muller Tube

For ages I have wanted to build a radiation detector using a Geiger Muller tube.  No real reason, just wanted to build one.  I purchased a tube from ebay and set about designing the detector.  The only design goals where it had to run from a single cell (1.5V) and be built from what I had.

The left hand side of the diagram is all power supply.  It generates about 380V from the 1.5V cell.  It draws around 2mA from the cell.

TR1 has three windings.  About 1500 turns form its secondary, 12 turns each for primary and feedback.  The output is trippled and regulated by feedback through the neon bulbs.  The 'neon bulbs' I have aren't neon because they glow green due to a phosphor coating, probably emitting UV.  Ideally these 'neons' would be zener diodes, but I don't have any with a high voltage knee.

The pulse duration from the tube is extended by Q4 & Q3 to provide a good 'click' from the speaker.  The speaker I used has a coil resistance of 45 Ohms.

All transistors are BC550C and BC560C, these have quite a high gain and substitutions with a lower gain device may need some tweaking.

Here it is built...

 

It doesn't have a CPM meter yet, that will come later.

High Input Impedance & Voltage Buffer

 High Voltage & Impedance Buffer

I am finally designing a Geiger counter, which requires some high voltage of about 300V.  This voltage has a fairly high source impedance and measuring it with a normal digital volt meter is enough to drag the voltage down.

I need a high input impedance buffer that can handle the few hundred volts I want to measure, so built this...

It has above 1Gigaohm input impedance and outputs 1V per 1kV input.  Runs off a 9V battery.  One half of the TL082 op-amp works as an oscillator and produces its own negative supply.  R6 can be adjusted to better the output accuracy.

Crystal Radio

 Crystal Radio

When I was a lad, I used to build radios.  So I thought I would try and get the kids doing something else for 15 minutes that didn't involve those Vidiots on youtube.

You will often find diagrams on the internet using a single coil, paralleled with the tuning capacitor, directly connected to the detection diode and then earpiece.  Although this will work, there will be very little selectivity in the tuning, since nothing will be impedance matched.  The diagram of the radio I built is this...

And here is a quick description of each stage...

L1, L2 & L3 are wound on separate cardboard formers which can be moved up and down independently on a ferrite rod.

L1 is 20 turns, L2 is 40 turns and L3 is also 40 turns.

C1 is a 500pF air cored variable capacitor.  In the picture it is ganged with other various capacitors, but it was the first one out the junk box that looked any good.  You will notice one end of this is also earthed.  This is because the capacitor has a metal frame and tuning shaft.  It is this frame which is earthed to prevent my hand changing the resonant frequency of this part of the circuit while tuning.

D1 is an OA91 germanium diode, again out of the junk box.  This diode must be a germanium diode, silicone diodes wont work.  I have heard tale of Schottky diodes working here, but have yet to try it.

R1 is 100k.  This is included here because the earpiece (SPK) I have is quite capacitive and so needs somewhere for the charge within it to dissipate.  The earpiece is a high impedance ceramic type.  The high impedance is important since there is very little energy available to drive any sort of passive load.

Some diagrams will include a capacitor across R1 to filter out the RF, but like I said, my earpiece is capacitive enough to do this job.

The aerial (ANT) was about 15 Meters of cable running from an upstairs window in the house to a tree opposite.  The earth connection was to a copper radiator pipe.  Many modern central heating systems are piped with plastic tubing, so this wouldn't work with those.

I initially had a 2 Meter metal stake hammered into the grass outside to provide the earth, forgetting that underneath the top foot of soil was just rubble and hardcore, not a good earth and not a good reception.

The tuning range of the radio was about 500kHz to 1.2MHz and me and the lad could tune in a couple of stations.  There are another couple just above 1.2MHz so I'll have to do some adjustments to get them.

Selectivity of tuning was very good.  The positions of L1, L2 & L3 on the ferrite rod didn't matter too much, as long as L1 was close to the other two windings.  I will experiment a bit more with turns ratio but the next step will be tuning in those missing stations.

Obviously my son wanted it louder so we will probably be building an amplifier as well.  This will need a battery.  It is at this point I explained our radio didn't currently use any form of local power supply and the energy came out of the sky, that boggled his mind.

Factory Reset A Citaq H10

 So, I had one of these Citaq H10 EPOS Machines...

Unfortunately it was locked and stuck with the 'Just Eat' POS software.

This machine runs android and on taking the back off there is a button marked 'recover'.  Pressing this and resetting the unit gets the usual android recovery menu, but the unit has no volume up / down buttons to navigate the menu.

The processor inside the unit is a Rockchip RK3368.  On their website is an example circuit of how to use this chip.  It is very similar to the EPOS unit!

This bit...

...shows the recover button is read by an A/D converter and checking the linux device tree file (from the github repository for the RK3368 evaluation board) we find ADC values corresponding to the different buttons which can be on this pin....

Volume Up - 1

Volume Down - 170

Menu Key - 355

Home key - 746

Back key - 560

Camera key - 450

So I tacked wire across the recover switch which led to the outside of the unit...

The voltage across these wires measures 1.8V and is pulled up via a 10K resistor, like the schematic shows.

The ADC inside this chip is 10 bits so to get a value of 170 (Volume down) I need a resistor to pull the ADC pin down to 0.3V ( ((1.8V / 1023) * 170) = 0.299V )

A 2k resistor will do this, so booting into the recovery menu and using the 2k resistor to move the menu down to 'factory reset' and then selecting, does a factory reset :)

After rebooting, the unit boots into android with no Just Eat !

After installing YouTube from the Play Store...

I don't know if there was a software way of doing this, but now I can add volume buttons to the EPOS unit.  I can't get the printer to work yet, since it is a serial printer and doesn't expose its self to the android OS as a printer.

Replacement bandsaw switch

 So, I switches on my bandsaw and BANG! This thing...

explodes.  Inspecting the switch reveals it had shorted out.  Googling around provided plenty of replacements at cheap prices.  The replacement switches were all of a similar quality, so if I replaced like for like, is it going to go BANG! again in a few months?

So I'm not replacing like for like.  I designed a simple push button operated relay which requires a long, one second push to switch on the saw and any push, while the saw is on, will switch it off.

I initially disregarded a microcontroller for this simple task and came up with...

Powered directly off mains supply using a capacitive dropper.  It was going back in the saw so the circuit was going to be encased safely.

Works fine if your supply voltage comes up quickly, doesn't work so well if the supply is slow.  So it would require various reset circuitry adding etc, etc... Sometimes its just easier to go with a £0.50 microcontroller from the start.  It simplified down to this...

Works great.  Now that ATtiny44 is in there I think I might make it periodically 'beep' while the mains is on, to remind me to switch it off at the plug.  I used an ATtiny44 because that's what I had in the drawer.

Here it is working...

Box full of lights and buttons (or simon says)

A box full of lights and buttons...

The little lad wanted a box with lights and buttons, so I got some from fleabay for arcade machines.  They were cheap and I was surprised how good the quality of them is.  Each button has a microswitch on the back and an integrated LED with current limit resistor.  Looks like the LED's are set up for 12V operation, they work fine at 9V.

I was going to charlieplex the lot but the processor I was going to use had tonnes of IO so I just wired each switch to a pin and wired each LED to a transistor to a pin.

I used darlington drivers to switch the LED's 'cos I had them and it makes building the board a bit easier.

That's it really, simples!

IR Remote Control Pop Tin Car

IR Remote Control Pop Tin Car

When I was at school I remember building something like this...

Two cans suspended on a frame driven by a motor and a rubber band.  The kids like making stuff and

drilling holes in wood, so we made this one weekend.  It didn't survive long because once its trundled underneath the sofa and distractions happen, the batterys go flat quickly driving into a stalled motor.  So, more for my enjoyment, I added remote control using an old IR remote.

Now it can go in TWO whole directions, forwards and back under control of the IR Remote.

It also shuts down into low power mode after some time of inactivity.

Schematic of IR Receiver and Motor driver.

The motor is driven by an H Bridge with the load in the collector of the transistors to minimise volt drop.  Each of the H bridge transistors is controlled by the ATtiny44 microcontroller.  The micro and the bridge must be powered from the same supply as there is no level shifting done.

This was built out of junk I had in the box so the power supply to the IR receivier isn't ideal.  The receiver I had shut down below about 4.5V, so to make the most of the battery, the supply is doubled and regulated back down to 5V via that LDO regulator.

Here it is in full action....

Inductance Meter

Quick Inductance Meter

I needed an inductance meter and had already spent my months spending money so I couldn't simply buy one.  I built one instead.

Luckily I had all the bits I would need to build the meter in the junk box.

The circuit is very simple.  Just an LC oscillator with a known tank capacitor.  The micro counts the oscillator frequency and works out the connected inductance.

The formulae for the resonant frequency of the parallel LC tank is..

f = 1 / (2 * Pi * sqrt( L * C )) 

Rearranged for L is...

L = 1 / (4 * Pi * Pi) * C * ( f * f )

The circuit was cobbled together on a bit of board...

And a window, for the display, CNC'd into a plastic box.

Whoops, n00b, lol.
Clamp, clamp.
Then Clamp some more...

Fixed, nice label :)

I was stressing about powering the thing off a 1.5V battery, but I had a 9V battery and the case takes a 9V battery so I used an XC6202 3V linear regulator to power it.

I added a power button which briefly powers up the device, does a reading and then switches its self off after 10 seconds.

The circuit as built looks like this...

Source code and hex file are available here
The meter measures from 1uH to 10mH.

And a quick video showing the operation.

Making that video put me over the budget for this quick project, I had to buy a holder for my camera.

Pressure activated speed controll

Controlling drill speed with pressure

I have a small hand held drill similar to a Dremel.

I need a way of controlling the speed of the drill while I am using it.

I thought some sort of pressure sensor mounted on the drill would allow me to vary the drill speed without needing an external speed controller.

To achieve this I used a piece of silicone pipe mounted on the drill which leads back to a pressure sensor, connected to a PWM power controller.

The pressure reading is taken by the PIC10F320 and converted to a PWM signal, which is connected to a MOSFET, which drives the drill.

Component choice is down to what I had to hand at the time.

Wired Home Automation

After mucking around with the idea of wireless nodes around the house controlling various things, I dumped the idea and instead ran an RS485 network around the house.

So far I have two temperature and humidity nodes...

The hardware and software are really simple compared to any wireless solution.  These two nodes are polled every now and then and the average temperature in the house is calculated.  This will eventually switch the central heating on and off.

The 'controller' is simply a PC connected to the RS485 network via a bit of hardware.  The hardware does the automatic switching of the RS485 transceiver between Rx & Tx.  There is also a current limited regulator to provide 8V at 200mA.  All this is piped round the house through CAT5 cable.

The PC software is written in JavaScript and executed with node.js.  The user interface is via a browser on any capable device on the IP network.

The two temperature / humidity nodes were installed quickly and I soon required more nodes.  Hand building prototypes is fine but when more units were needed I did a PCB using EasyEDA and ordered them from JLCPCB.  Two weeks later, PCB's arrived.

Nice.  The PCB's were good quality and its not really worth etching boards at home anymore unless you need a PCB within the next few hours!

The PCB contains the RS485 comms stuff and a 5V regulator.  'Functionality' such as the temperature sensors can be added in the 'prototype' area.

The boiler and front house lights require a relay switch to operate them.  Rather than having a relay eating into the power budget of the system I used relays with a coil rated at 12V and ran them at 8V.  The circuit below was used to boost the voltage to the coil at switch on, reducing the coil voltage once it had switched.

Voltage boost relay driver

This circuit provides a voltage pulse of just under twice its supply voltage to the coil.  It also provides a turn off delay, determined by the charging of C1.

Stepper motor speed controller (Tesla coil winder)

I have been wanting to build a tesla coil forever, winding the coil has always put me off.  I think its about time I built one.

I was going to build the coil winder using a drill powered by a 300W, 0 - 40V variable voltage supply.  I built the supply using a variac and large torroid transformer.  The supply worked and is nearly indestructable, then I remembered I had some huge stepper motors...

...so I used those instead.

At least I now have another power supply :)

The stepper motor is driven by a microcontroller (ATtiny2313) with coils driven by mosfet's (BUK9840).  The speed and direction of rotation is varied using a rotary encoder.  The micro spits out information on speed / direction / revolutions on its UART.  The UART connects to an LCD.
The motor coils are rated for 3 volts.  To trade off some torque for speed the motor coils are driven from a 12V supply via some resistors.  These get quite warm in use, the only ones I had with a large enough power dissipation were 25W aluminium resistors, so no trouble there.

I was going to use a shift register to drive the LCD from a couple of pins, but didn't have any.  I have lots of ATtiny2313's so I stuck one on a board with the LCD and used the UART to shift data to the display.  The LCD microcontroller has a 32 character buffer into which any characters received on the UART are stored.  Characters below an ASCII space reset the buffer index to zero.

Very simple, now I've written the code for that it will be useful for other stuff.

To turn the piece of pipe which will be the secondary's former, I made two 'centers' out of the only stuff I had, concrete.  I used two plastic funnels for a mold, which also held the aluminium shaft central while the concrete went off.

Here, the bottom funnel is full of concrete and its stuffed into a bucket of sand while it sets.

The finished winder with first coil secondary.  The two blue concrete 'centres' are holding the piece of plastic pipe.

This only took about 15 minutes to wind, I was expecting to spend at least an hour on it.

USB STK500 AVR Programmer

So, I had one of these AVRISPMk2's for programming AVR's

It was very good.. I was using Atmel Studio 6.  For some reason I upgraded to Studio 7.  I then wasted an entire day messing with various drivers trying to get it working again.  I eventually got everything going again and connected the project I was working on to a 24V power supply, blowing the programmer completely.

I have a working PICkit3 and some PIC16F1455's, so I built a replacement programmer using these parts.  It pops up in windows as a com port and emulates an STK500.  No driver issues there.
The code is based upon a project found here AVR Programmer which I ported over to the PIC and added the necessary code to drive the USB.  As you can see from the below schematic, there isn't much in there.

I have done a PCB and ordered a few in EasyEDA.  The sourcecode can be downloaded here...
https://drive.google.com/open?id=0B0Dxv1kRlKo8MVJwU2tuczJkYmM

I can now program my AVR's again :) and if I blow the programmer up, its cheap enough to replace.

OWON PDS5022S fixin'

Found one of these.....

thrown out.  It was dead and is a basic scope, but it has a battery power option and would be good to have as a "disposable" 'scope.

Opening it up and checking the supply voltages around the PCB revealed nothing surprising.  Everything looked OK.  The processor on the main PCB is a BGA packaged part.  Taking a well educated guess I repowered the 'scope while pressing down on the processor, it worked :) so after applying foil around the chip and heating it with a hot air gun I had a working 'scope.

When I found the thing it was wet through which hadn't done the LiPo battery any good and it looked a bit 'puffy'.  I have plenty of NiMh cells lying around so I thought I would make up a battery pack from these to replace the LiPo battery.  First I checked the PSU to see what kind of provision had been made for charging the batteries....

The output stage of the PSU looks like this...

The battery is just connected across the output via the main switch.

The 'scope draws...
360mA, 8.4V, LCD Backlight.
380mA, 8.4V, Main board.
80mA, -8.4V, Main board.

After trying out a NiMh pack and realising it was stupid, I decided to dump the entire power supply and have the thing powered externally by a ready made 19 volt or so ( old laptop psu ) brick.  Clearing out the power supply left lots of space.  I can get lead acid batteries for free and there is room for a small one plus its charging psu plus all the other power stuff I am going to need.

While I'm modding this thing I might as well swap out that LCD backlight for an LED one, that will save 200mA !!!

This was easy to do :)

The replacement power supply is fed with 19V and provides 13.8V for charging the lead acid battery, this stage is followed by a further regulator which outputs +/-8.4V.
The backlight is powered from the 8.4V via two 47R paralleled resistors which gives it 20 to 30mA, which is plenty bright enough.

Metal Forge / Induction heater

YouTube suggested I watch this video.. Grant Thompspn - "The King of Random"

Ooo, thats good I thought. I'll make one.  So I purchased some of the stuff to make this small charcoal fired forge, then remembered I was an electronics engineer and thought why am I not making an electric one?  I pondered this for a while and then thought why am I not building an induction heated forge.  So thats what I have started doing....

Induction heating involves inducing eddy currents in the work piece.  These currents are big, so the work piece heats up.

Wikipedia has the explanation here https://en.wikipedia.org/wiki/Induction_heating

The output stage of this heater will probably connect to rectified mains so I would like all the control electronics isolated from this.  I started with what I had 'in the drawer' to construct an oscillator and gate drive transformer driver.  The IR2153 is a self oscillating half bridge driver, so sounded ideal for what I was trying out.

IR2153 Half Bridge

The board is supplied with 24V for the half bridge, regulated down to 12V for the IR2153.  This provides a +/-12V square wave output to drive a gate drive transformer.

Out of curiosity I connected this...

...to the output of the driver via a 1uF Cap.  Using a 'scope I could tune the IR2153 to get the coil resonating.  This was able to heat my screwdriver up to above 100°C, I know this because my finger sizzled when I touched it.

From initial measurements on the above coil of wire I could tell the final coil would be conducting 100's of amps, so water cooling would be a good idea.  I got some 8mm copper tubing and found a paint tin to wrap it around.  I couldn't just wrap the tube around the tin, it would buckle at the first bend, so I perused the WWW and plenty of people suggested filling the tube with salt.  This I did and the tube bent very nicely...

...Getting that salt out was difficult and took ages, tap tap tap, rotate, tap tap tap, rotate, Zzzzz
I tried buzzing it out with a sander and all sorts.

I needed a smaller coil.  I wasn't going to use the salt method again.  Water doesn't compress, so I capped one end of tube with a compression fit stop end, filled the tube with water and capped the remaining end.  The tube requires a slight tweak to take up any air that got in, but after that the tube wrapped up into a coil nicely.  Getting the water out was easy, just release the compression fittings and empty the water out.  Using the compression fittings 'wastes' a couple of olives and some tube but saves a lot of time.

The coil gets hot without cooling, so I purchased a water CPU cooler from ebay.  It was 'untested', but after cleaning it out, removing pipework and re-piping it to the coil, it pumped water round no problem.

The pump and cooling fan required an extra 12V supply, the pump at 200mA and the fan at 1.5A.  I decided to under run the pump and fan at 10V and constructed a 2A buck regulator from a TL494 and associated components.

The heater as it was built was getting hard to work on so I put the lot into a metal enclosure..

I am planning on adding a 'wall' down the middle of the box to separate the electronics from the half that has water pumping round it, just in case :)

To test a theory about using Litz wire for the work coil I obtained some of these...

1uF, 1500V Polypropylene capacitors.  Manufactured by ICW Ltd. I bolted three together to make 3uF...

Very nice.

Full of stars

I thought it would be good to flash some LED's.

I wanted either red or white LED's flashing from a 2 or 3 volt source.  White LED's would require some boosting while red LED's would run quite happily down below 2V.

I put this together at lunch time to prove the idea and adjust flash rates...

The 74HC14 is spec'd down to 2V, I ran the thing from about 2.5V.  The LED I used was a random white LED.  With the components shown the flash rate was a few tens of ms on and a few seconds off.  C2 is used to boost the voltage across the LED.

That works ok, now I need to switch it on/off depending on ambient light....

Swapping the 74HC14 for 74HC132 makes this a bit easier.

Adding a simple 'dark' detector (Q1, D3 & R5) switches the flashing off during the day.  While 'off' current draw is below 10uA.