| Author |
Topic  |
|
ll81
Junior Member
 
 124 Posts |
Posted - 26/08/2007 : 15:36:57
|
I ended up with this through a series of unexpected occurrences. After my first linear dump I decided I needed to go switch mode to keep transistor dissipation down. I then used a sensor circuit to feed a Velleman DC-PWM kit. Whilst operating I noticed that the sensor would go from a high-low output due to a capacitor on the input taking time to dischage and charge.
I wonder what happens if I ditch the Velleman kit and run the FET direct off the oscillating sensor I thought. Scoping shows a relativly clean pulsed output that keeps transistor dissipation minimal. A small heatsink is all that is needed on the transistor.
The load must be large enough to put enough current drain on the batteries that there is some voltage drop upon operation for the circuit to oscillate.
This can be built cheaply and easily by absolute begginers with a good soldering iron, the necessary components and a piece of vero board (strips of copper on a piece of board with lines of holes in it). Just adjust the potentiometer to start the load dump at the required voltage.
I have actually got this very circuit as the load dump on my turbine I finished constructing yesterday.

woo-hoo
|
|
|
Victor
Average Member
  

United Kingdom
615 Posts |
Posted - 28/08/2007 : 05:12:20
|
Good stuff! Only thing I can add to the circuit is the type of variable resistor VR1 as I know there is log and linear potentiometers...
One has a steeper curve than the other......and it might be better using one of the types of resistors that does not have such a steep curve where you can gradually adjust it to the desired level.....Id also keep the transistor IRF as close to the Ua741 as possible and all lead lengths small as it might effect the circuit in some way.....
I know with radio circuits some of these components can self osscilate if the lead lengths are too long......
Dont really know much on the electronic side of things it`s been years since I have done anything in that line.... |
 |
|
|
ll81
Junior Member
 

124 Posts |
Posted - 28/08/2007 : 11:11:07
|
Good points victor. Yes a linear pot is desired here over a log pot. The IRF540 will be ok as long as it isn't point-point wired with a long length. It should be easy enough to make sure it is on the main piece of board. FETs can be troublesome devices in linear applications but in this instance it is clamped either on or off.
I have done a price up and this entire circuit can be built for under a tenner even with singular purchase of components (including case, PCB screw type cable connectors and heatsink).
That's a heck of alot cheaper than I can find any comercial units for.
woo-hoo |
 |
|
|
fungus
Moderator
 

308 Posts |
Posted - 28/08/2007 : 17:06:44
|
Looks good, pretty common design with an opamp/comparator.. One note I would make is to keep the power and the sense leads separate as the voltage drop in the wires made when the load is switched on can cause it to oscillate at high freq. (up to 10khz...) as it senses the voltage drop. |
 |
|
|
Victor
Average Member
  

United Kingdom
615 Posts |
Posted - 28/08/2007 : 17:52:28
|
I purchased electronic design studio in 1999. Known as quickroute up until then...If I can dig it out would it be worth making a pcb design for this?
I think there is also a few packages on the net better than my one for it like orcad etc......Seemingly some last for 30 days and are more mordern than mine.
Quickroute 3 would even do it and it`s a 1 meg floppy program..
just an idea!
But pcb boards would add onto the price of a finished product. one sided should not be too bad. |
 |
|
|
ll81
Junior Member
 

124 Posts |
Posted - 28/08/2007 : 18:32:20
|
I use express PCB to make boards with. The press-n-peel blue film method is easier although slightly more expensive than photoresist. It utilises the toner from a laser as the etch resist after ironing it to the plain copper clad board. Could be done with dalo pen and a ruler easy enough.
I just tried with seperate and same wire feeds all seems fine. The time constant of the RC network will only let it oscillate so fast. At low dumping it is slow but as current increases it does increase in frequency considerably until it stays permanently on. Stays pretty cool though.
woo-hoo |
 |
|
|
Victor
Average Member
  

United Kingdom
615 Posts |
Posted - 28/08/2007 : 20:02:33
|
Impressive!
This is good. |
 |
|
|
fungus
Moderator
 

308 Posts |
Posted - 28/08/2007 : 21:10:41
|
You can do the toner transfer method without the expensive press n peel stuff; Print the design from a laser printer onto glossy photo paper Clean the copper board well and iron the design onto the copper with plenty of pressure for about 5 mins soak in soapy water and gently lift the paper off-it leaves the toner on Etch |
 |
|
|
ll81
Junior Member
 

124 Posts |
Posted - 28/08/2007 : 21:26:29
|
I heard about that, never tried it though.
woo-hoo |
 |
|
|
BushWhacker
Junior Member
 

Canada
347 Posts |
Posted - 29/08/2007 : 06:59:23
|
Ummm... you guys are from Greece aren't you? Cause this is all Greek to me. One day I will catch up, a bit.
Some questions for from the guy sitting in the corner of the class wearing the pointy hat.(that would be me) 1) How much amperage will this circuit dump at 13.8 volts? 2) Rather than the 50-70 watt halogen lamp,could you substitute a coil style heater (block heater)? 3)What is, a)IFR 540? b)Q1 c)X1,UA741 d)A 3v9
Keep writing and I'll keep reading. With a hope and a prayer I will start to catch on.
Thanks, BW
Everything should be made as simple as possible, but not one bit simpler. - Albert Einstein |
 |
|
|
ll81
Junior Member
 

124 Posts |
Posted - 29/08/2007 : 10:23:49
|
OK.
The circuit at this form is good for 6 amps. With a biger heatsink on the current transistor (Q1 generic part number IRF540) it could handle 15 amps. To get more current you can simply add more transistors with a little modification.
Yes a heating element could be used, parallel the old heating coils of a 240v convection heater to make a 2 ohm resistance. I like using a bulb as you get a visual and audiable (it clicks) indication as the circuit begins to operate. It will run slightly differently but still work. A bulb will last a long time you just have to keep an eye on it.
An IRF540 (Q1) is a MOSFET transistor. IRF540 is the part number and defines it's ratings. Ua741 (commonly LM741) is an 8 pin operational amplifier. Put simply the output will go to the most positive input. If + input is more positve the output goes positive (high). If - input is more positive the output goes negative (low). It can be used other ways but we need not worry about them for this use.
The thing marked 3v9 is a zener diode. It is a reverse biased (put in the wrong way round) diode that has been specially made to conduct when over 3.9 volts is across it. LEDs (all semiconductors actually) have a forward voltage to conduct, standard green LEDs need around 1.9 volts. Just use two of them in series (not a different colour, 12v devices, high brightness or low current, just standard green LEDs. They will be sold as such) In this simple circuit it forms a constant voltage reference for the LM741 to "compare" the battery voltage to no matter what the battery voltage is.
Hope that helps.
woo-hoo |
 |
|
|
BushWhacker
Junior Member
 

Canada
347 Posts |
Posted - 30/08/2007 : 02:49:38
|
ll81 it's still a bit foggy but better than when I started. I'll get there one day!
Thank you for your patience. :)
Cheers, BW
Everything should be made as simple as possible, but not one bit simpler. - Albert Einstein |
 |
|
|
gotwind
Forum Admin
   

United Kingdom
1374 Posts |
|
|
ll81
Junior Member
 

124 Posts |
Posted - 04/09/2007 : 16:40:07
|
I would like to see his exact circuit, looks interesting.
woo-hoo |
 |
|
|
gotwind
Forum Admin
   

United Kingdom
1374 Posts |
|
|
ll81
Junior Member
 

124 Posts |
Posted - 04/09/2007 : 19:48:23
|
Interesting simple circuit. It's allways nice to see someone elses approach to get the same result. I would suggest 47 ohm resistors between the gate of each power FET and the PCB and not to direct connect two FET gate terminals together as high gate capacitance is likely to lead to parasetic oscillation and preamture destruction.
woo-hoo |
 |
|
Topic  |
|