| Author |
 Topic  |
|
wasssup1990
Nobel Prize Winner
A Land Down Under
2261 Posts |
 Posted - Jun 01 2011 : 11:02:22 AM
|
quote:
Originally posted by codingplanet
Thanks a lot for the help Johnny. I'll go for the one with higher VR.
You're welcome! 
Yeah, just go for the high VR one. By the way, are you sure your fan doesn't already have one of these diodes in it? Have you been able to detect the back EMF on a scope? |
When one person suffers from a delusion it is called insanity.
When many people suffer from a delusion it is called religion. |
Edited by - wasssup1990 on Jun 01 2011 11:03:02 AM |
 |
|
|
codingplanet
Mad Scientist
United Kingdom
195 Posts |
Posted - Jun 01 2011 : 11:52:18 AM
|
| I don't know. I haven't particularly looked for it. If I connect the fan to a power supply, unplug the PSU with the scope still connected to the fan, should I see a spike (if there is no diode)? |
|
|
|
wasssup1990
Nobel Prize Winner
A Land Down Under
2261 Posts |
Posted - Jun 01 2011 : 8:35:56 PM
|
quote:
Originally posted by codingplanet
I don't know. I haven't particularly looked for it. If I connect the fan to a power supply, unplug the PSU with the scope still connected to the fan, should I see a spike (if there is no diode)?
Well if there is any back EMF at all it'll show up as a voltage spike across the fan terminals when the falling edge on the Gate occurs. I'll assume you know how to use your own scope and let you decide how you should measure it with your particular scope. You can do it your way as well but make sure you switch off fast, just like the MOSFET would do. Use a good switch or push button and don't just tap wires together. |
When one person suffers from a delusion it is called insanity.
When many people suffer from a delusion it is called religion. |
Edited by - wasssup1990 on Jun 01 2011 8:50:09 PM |
|
|
|
codingplanet
Mad Scientist
United Kingdom
195 Posts |
|
|
wasssup1990
Nobel Prize Winner
A Land Down Under
2261 Posts |
Posted - Jun 02 2011 : 06:40:26 AM
|
quote:
Originally posted by codingplanet
Here is the voltage across the MOSFET: http://www.benchtec.co.uk/forums/attachment.php?attachmentid=147&d=1304523155&thumb=1
Is the small peak just before the falling edge what you are talking about?
Oh, you have a CRO. Your time per division looks too large. You'll need to decrease the time per division on the scope to see any spikes and maybe even turn up the beam intensity to see it. Your MOSFET has a relatively high voltage Zener diode between source and drain - if you see the spike, it'll be clamped by this Zener.
I'm assuming Vmax of that waveform is 12V and Vmin is 0V. That would mean your GND scope connection is at the source and the probe is on the drain. In this case the falling gate signal would correspond to a rising VD signal. You're trace seems to show a very rounded curve so, to me, it doesn't even look like you have a spike. But do what I recommended anyway just to get a closer look. |
When one person suffers from a delusion it is called insanity.
When many people suffer from a delusion it is called religion. |
Edited by - wasssup1990 on Jun 02 2011 06:55:34 AM |
|
|
|
codingplanet
Mad Scientist
United Kingdom
195 Posts |
Posted - Jun 02 2011 : 06:57:43 AM
|
| Ok, I'll scope it again when I get time. Thanks. |
|
|
|
codingplanet
Mad Scientist
United Kingdom
195 Posts |
|
|
wasssup1990
Nobel Prize Winner
A Land Down Under
2261 Posts |
Posted - Jun 08 2011 : 12:58:34 AM
|
quote:
Originally posted by codingplanet
http://i547.photobucket.com/albums/hh471/UnknownReactionz/pwm.png?t=1306412447
Voltage measured between +12v and MOSFET source pin: http://i547.photobucket.com/albums/hh471/UnknownReactionz/100_1448.jpg
The main line shows 12v. 2v per division. That is with 8% duty cycle on the PWM.
Okay, so it looks like you have got some noticeable ripple on your 12V supply line. You said a V/div of 2V, this means you have a 2.5V drop in your 12V power supply going by your graph. The sharp trace is caused by the inductance in your wires - it's called "ringing" or "resonance" and its oscillatory frequency is combination of the inductance of your wires and what little capacitance there is. You should be able to minimize the ripple by making your wires as short as possible and putting a capacitor (experiment with that value. Try >= 470uF) in parallel with your 12V to ground as close to your fan circuit as you can. Be aware that cheap-ass capacitors might have high ESR so choose capacitors that have low ESR. This will reduce your ripple even further. I don't imagine this kind of ripple is too healthy for your computer.
Good luck! |
When one person suffers from a delusion it is called insanity.
When many people suffer from a delusion it is called religion. |
|
|
|
codingplanet
Mad Scientist
United Kingdom
195 Posts |
Posted - Jun 08 2011 : 05:32:27 AM
|
I must mention that the circuit is currently constructed on a breadboard. The PSU I am using is an old ATX PSU that has nothing else but my circuit connected. Surely that ~22v spike isn't good?
Sorry, I forgot to mention the time base. I'll have to check another time as I don't have the equipment with me right now. |
|
|
|
wasssup1990
Nobel Prize Winner
A Land Down Under
2261 Posts |
Posted - Jun 08 2011 : 06:54:09 AM
|
quote:
Originally posted by codingplanet
I must mention that the circuit is currently constructed on a breadboard. The PSU I am using is an old ATX PSU that has nothing else but my circuit connected. Surely that ~22v spike isn't good?
Sorry, I forgot to mention the time base. I'll have to check another time as I don't have the equipment with me right now.
Having your fan circuit run on another electrically isolated power supply is good but you can have it on the same power supply as your computer if you professionally design your circuit and put it in a shielded container - assuming your power supply can handle your fan's loading. Just be aware that you're transmitting a RF signal with your beefy fan driver and if you don't minimise it at the source (your fan circuit) you could induce noise into radios and nearby computer circuits. Trust me, it is definitely possible to cause a computer to crash/restart even if your circuit is a few meters away from your computer.
Some people will say that you don't need a low pass filter to drive your fan. I suppose that's up to you - the engineer. You have said your fan is a 12V, 3A type. That's 36W - not your typical computer fan.  |
When one person suffers from a delusion it is called insanity.
When many people suffer from a delusion it is called religion. |
Edited by - wasssup1990 on Jun 08 2011 07:21:49 AM |
|
|
|
codingplanet
Mad Scientist
United Kingdom
195 Posts |
Posted - Jun 08 2011 : 07:51:41 AM
|
I was going to run it off the main PC PSU when it was finished. I am just using a separate one for now while I test.
What is the source of the RF? And what are the disadvantages of not using a low pass filter?
And what about the 22v spike? Is that the back emf? |
|
|
|
wasssup1990
Nobel Prize Winner
A Land Down Under
2261 Posts |
Posted - Jun 08 2011 : 09:00:22 AM
|
quote:
Originally posted by codingplanet
What is the source of the RF? And what are the disadvantages of not using a low pass filter?
You are switching at 30kHz with a square waveform that has a variable duty cycle right? There's your source of RF. It won't just radiate RFI and EMI at 30kHz, it'll do it at all the harmonics too! If you don't use a LPF, the wires going from your circuit to your fan and the fan itself will radiate RFI and EMI which could cause radio interference for people around you and possible computer instability if the motherboard picks up those signals. The LPF should be placed as close to the fan circuit as possible so that you end up with a noise free output from it which will appear as a relatively clean DC signal to your fan. I would put the circuit in a cast-iron/aluminium or similar enclosure - one that will seal in RFI and EMI.
quote:
Originally posted by codingplanet
And what about the 22v spike? Is that the back emf?
Actually that could be more due to the inductance in the power supply leads as mentioned. If you want to measure fly-back from a fan you should put your probe across the fan terminals. Like I said before, I'm assuming you know how to use your own CRO so that you don't damage you probe and CRO when you place that ground lead... If you know what I mean.
This is bad:
12V
o
|
|
Probe(Tip)
|
|
+
FAN
-
|
|
Probe(GND)
|
|
__|__D
G__| |
|_ _|
| S
|
_|_
///
Why? Because the GND (Ground) of the power supply is connected to the earth wire on your power outlet and so is the GND on your probe. NEVER put the GND of your probe on a positive or negative supply rail. Just so you know. 
|
When one person suffers from a delusion it is called insanity.
When many people suffer from a delusion it is called religion. |
Edited by - wasssup1990 on Jun 08 2011 09:02:35 AM |
|
|
|
codingplanet
Mad Scientist
United Kingdom
195 Posts |
Posted - Jun 08 2011 : 09:08:10 AM
|
Are you saying that the 12v rail should be almost flat once I connect a bypass capacitor?
"If you want to measure fly-back from a fan you should put your probe across the fan terminals."
But then you go on to say I can't do this. So how am I supposed to measure the flyback from the fan? |
|
|
|
wasssup1990
Nobel Prize Winner
A Land Down Under
2261 Posts |
Posted - Jun 08 2011 : 10:06:20 AM
|
quote:
Originally posted by codingplanet
Are you saying that the 12v rail should be almost flat once I connect a bypass capacitor?
Just to be clear, don't confuse "LPF" with "bypass capacitor". The LPF would go at the output of your circuit and the bypass capacitor would go at the input. Yes, a bypass capacitor will smoothen your 12V supply rail.
http://www.seattlerobotics.org/encoder/jun97/basics.html
quote:
Originally posted by codingplanet
"If you want to measure fly-back from a fan you should put your probe across the fan terminals."
But then you go on to say I can't do this. So how am I supposed to measure the flyback from the fan?
That is why I was assuming you knew your own scope and you yourself could tell me why you haven't done this measurement already. I don't know what features your scope has. Does it have two probe sockets? Does it have a feature for subtracting one channel from the other? If it has these two things then you are set. If I didn't have this feature on my scope, I have ways to overcome this problem. Let me know how you go. There is the "poor-man's" approach which doesn't require a scope. |
When one person suffers from a delusion it is called insanity.
When many people suffer from a delusion it is called religion. |
|
|
|
codingplanet
Mad Scientist
United Kingdom
195 Posts |
Posted - Jun 08 2011 : 10:16:39 AM
|
I understand bypass caps and lowpass filters (to a lesser degree) and the difference between them. I was not confusing them.
Yes my scope does have those features. And now you have mentioned that, I understand what you were getting at. I can take that measurement sometime this week. |
|
|
|
Topic |
|
USB contraption
