Voltage drop on fume extractor, unknown cause

Thread Starter

Marcus2012

Joined Feb 22, 2015
403
Hey guys



I’m having a problem with a fume extractor I bought in that it experiences a voltage drop at full power. It’s a 5V USB fume extractor with lights and a control unit for motor on/off, lights (shades and brightness). However, when the lights are on full and the fan is on the fan slows down, NOT the lights only the fan, lights seem unaffected either way). Lights still cause a voltage drop of 0.5V when operating alone but it appears unnoticeable. I’ve tried opening the control unit and using a direct 5VDC 20A ATX supply and the drop still happens. If I lower the brightness the fan speed with increase again, so something is a bottleneck. I had assumed the USB cable was too high a resistance but apparently not as I’ve gone right to the board to supply it now.



Looking at the control board I think they all share a common positive and have switched grounds. But I am a little confused about it. I am thinking either

  • The board IC limiting current (not sure why only the motor though)
  • Transistors are limiting (but why not all the time?)
  • The common supply wire is too thin and there is too high a voltage drop at 2A.
  • Something I know not of.


The LEDs draw approximately 1.6A and the fan 0.4A. The cable for the common supply looks to be 24AWG.



Anyone got any ideas on this? I got a half refund from the seller so I could hack at it but I’m not entirely sure what’s at fault at the moment. To make matter worse I cannot make sense of the IC they used and how it controls brightness and all 3 FET gates. It’s Vcc pin looks like it’s a potential divider.



As always, any and all help is appreciated.



Thanks in advance :)





TLDR Voltage drops when lights used in fume extractor/light unit. Doesn’t appear to be wiring as a jumped the common with a thicker wire, no change.
 

Attachments

Marc Sugrue

Joined Jan 19, 2018
213
What USB charger are you using? Is it man enough for the Job? It needs a 2A+ Capability which is the Top end of most standard chargers. Try using a >11W USB charger which is fast charger teritory i believe. It sound like the USB voltage regulation is starting to current limit and the voltage is rolling off.
 
Last edited:

jpanhalt

Joined Jan 18, 2008
11,088
Now, concerning your detective work:
I’ve tried opening the control unit and using a direct 5VDC 20A ATX supply and the drop still happens. If I lower the brightness the fan speed with increase again, so something is a bottleneck. I had assumed the USB cable was too high a resistance but apparently not as I’ve gone right to the board to supply it now.
I got a half refund from the seller so I could hack at it but I’m not entirely sure what’s at fault at the moment.
1) I am not surprised there is current limiting. What if there wasn't and someone plugged it into the USB port of a laptop or PC? Then insted of geting half your purchase price back, the seller might be held responsible for a lot more. Nothing is fool proof.

2) You got it at half price. It does what it is designed to do. Maybe that is not what you want it to do. Next time buy one that has a dedicated supply and doesn't present a risk to those PC's with low-current UBS ports. Caveat emptor.
 

jpanhalt

Joined Jan 18, 2008
11,088
I realize that. Read the second and third lines of my response. However, remember that the thing you purchased was not designed specifically for your use. The designer had to consider what would happen it someone else plugged it int a PC's USB port.

Stop complaining. You got it it at half price, and it works as advertised. Of course, you are free to destroy it trying to make it better. If you want to bypass those safety features, you need to provide more information, like a reverse-engineered schematic for it.
 

Thread Starter

Marcus2012

Joined Feb 22, 2015
403
Stop complaining. You got it it at half price, and it works as advertised.

Ok so what the <****> This forum has gone downhill massively if I get told not to complain when I have not ONCE complained. Fume extractor that doesn’t extract fumes. Sound like that works as intended you condescending nasty person?

Expletive Deleted: Moderation
 
Last edited by a moderator:

ericgibbs

Joined Jan 29, 2010
12,884
hi Marcus,
What rating and length of cabling are you using between the power source and the unit and at what point are you measuring voltage drop of 0.5V?
E
 

DickCappels

Joined Aug 21, 2008
7,421
Please remember this quote from our User Agreement. There are links to the User Agreement and the Terms of Services at the bottom of every page, including this one.

Appropriate conduct. Debates should be a civil activity and can be both enlightening and entertaining, but always keep discussion to the facts and the opinions. Ad hominem tactics and directed abuse are always "out-of-bounds".

Honestly, we are trying to improve the forum and the cooperation of everybody is necessary.
 

Thread Starter

Marcus2012

Joined Feb 22, 2015
403
Can you show more pictures with more identifying information (motor #, characteristics, front of circuit-board, etc).

Hi thx, sorry I forgot to upload the PCB photo, i've done that now. Nothing to identify the fan unfortunately it's label has been removed. All I know is it's 5V and approximately 400mA.

hi Marcus,
What rating and length of cabling are you using between the power source and the unit and at what point are you measuring voltage drop of 0.5V?
E
Hi thx, I’m measuring the voltage across the PSU input/output to the control board. Wires to the board are now 18awg and I think the wire from the board to the LEDs motors is 24awg, at most 22awg. Voltage doesn't drop for independent operation of the fan. But when I operate the LEDs independently of the motor the voltage will drop to 4.6V (however this is not noticeable in operation, they appear bright). I thought maybe the motor and the fan share a PWM gate control but this would not explain how fan speed is indirectly proportional to LED brightness.
 

Attachments

BobaMosfet

Joined Jul 1, 2009
1,767
God I regret asking, didn't realise it had got so leet around here
@Marcus2012 - it hasn't. Please ask. There are people here that want to help. Even jpanhalt is a good one, but perhaps something has him off kilter today (happens to all of us).

I took a look at your circuit board, and it looks like (possibly, just my SWAG) a PIC12F629 MCU controlling 3 MosFETS doing switching. The current limiting aspect may be part of the firmware on the MCU. Most USB ports only provide 500mA to 1A at most, and a fan uses current, as do the LEDs, plus there is an inductive aspect to the fan.

Normally, if voltage drops, too much current is being drawn (either more than a power-source can provide, or more than it is being allowed to provide). I suspect your problem is that their design inherently limits current through the PCB. IMHO.

If you want, you can snip the fan & LED leads, make them so you can reconnect them, and put them through a DMM in current mode, and see exactly what current level is being drawn based on what you do, which in and of itself may be very revealing.
 
Last edited:

Thread Starter

Marcus2012

Joined Feb 22, 2015
403
I took a look at your circuit board, and it looks like (possibly, just my SWAG) a PIC12F629 MCU controlling 3 MosFETS doing switching. The current limiting aspect may be part of the firmware on the MCU. Most USB ports only provide 500mA to 1A at most, and a fan uses current, as do the LEDs, plus there is an inductive aspect to the fan.

Normally, if voltage drops, too much current is being drawn (either more than a power-source can provide, or more than it is being allowed to provide). I suspect your problem is that their design inherently limits current through the PCB. IMHO.

Thank you!! :) I didn't think it would be an MCU but that makes complete sense when I think about the pin count and functionality. I had assumed they used a USB plug to save money on a PCU for the unit and using a higher current source than USB would be ok. But sadly it would appear, as you say, the PCB is configured for USB power only and limits current accordingly.

So, I have been thinking of just removing the control unit, running some 18awg to the fan unit. install switches for the fan and lights seperately (if I have room) with the light switch being a SP centre off to change colour. Abandon the dimming and simplify it as much as possible. Only two conductors running the length and a mains 5V 2-3A supply. The only worry here would be overvoltage for the LEDs but they appear to have current limiting resistors in their annular PCB around the fan. Do you think this may be a better way of getting the functionality I'd like?
 
The LEDs draw approximately 1.6A and the fan 0.4A. The cable for the common supply looks to be 24AWG.
I'm surprised it works at all. it soundslike this junk: https://www.amazon.com/s?k=usb+coffee+cup+warmer USB Coffee Cup Warmer. 5W from USB.

USB gives you 0.1A unless you do handshaking, If you so handshaking 1A. Motors draw current initially based on the winding resistance for a short time because the current in an inductor cannot change instantaneously.

USB3 would be a different story,

2Ais bigger than 1A (the supply) and the supply doesn't allow for startup of a motor.

There is a USB "Y" cable which is sometimes included with disk drives and CD-ROMS. https://www.amazon.com/Onvian-Splitter-Adapter-Charging-Syncing/dp/B01KX4TKH6/ref=asc_df_B01KX4TKH6/

I have no idea if it would work.

More than likely, it's an ersatz fan.
 

jpanhalt

Joined Jan 18, 2008
11,088
The transistor number (A09T) cross references to a A03400 mosfet in an SOT233L package (datasheet attached).

1612007136320.png

If you apply 5V directly to the gate with the negative connected to common it will turn on hard. If you apply voltage to the gate, use a resistor in series to limit current that might damage the MCU. A 200Ω resistor in series would limit the current into the MCU to 25mA. A 1kΩ would be safer. Mosfets are voltage controlled. A larger resistor will still cause it to turn on fully, but a little slower. Since you aren't PWM'ing it, that won't matter. EDIT: On second thought, with the MCU or whatever in the circuit, it is probable that it could still pull the gate low. Shorting or using a resistor from drain to source is more certain.

Alternatively, shorting the drain to source will put whatever it is controlling fully on. The latter would probably be safer from the standpoint of the MCU controlling it. You could add a resistor of suitable wattage between drain and source will give some control. Of course, the short can be directly to common or battery negative rather than the small source pin.
 

Attachments

Last edited:

BobaMosfet

Joined Jul 1, 2009
1,767
Thank you!! :) I didn't think it would be an MCU but that makes complete sense when I think about the pin count and functionality. I had assumed they used a USB plug to save money on a PCU for the unit and using a higher current source than USB would be ok. But sadly it would appear, as you say, the PCB is configured for USB power only and limits current accordingly.

So, I have been thinking of just removing the control unit, running some 18awg to the fan unit. install switches for the fan and lights seperately (if I have room) with the light switch being a SP centre off to change colour. Abandon the dimming and simplify it as much as possible. Only two conductors running the length and a mains 5V 2-3A supply. The only worry here would be overvoltage for the LEDs but they appear to have current limiting resistors in their annular PCB around the fan. Do you think this may be a better way of getting the functionality I'd like?
I actually think you're head in the right direction. You could easily do your own PCB to control those items so long as you know a few things about the fan and LEDs in terms of their loading (voltage & current draw). The fan will be harder than the LEDs, since they are pretty straightforward. You could power it using a common brick or wallwart.

I'm sorry if I missed it in the thread- did you provide where you got this device so we can go look at it ourselves and see what we can learn about it?
 

Thread Starter

Marcus2012

Joined Feb 22, 2015
403
I'm sorry if I missed it in the thread- did you provide where you got this device so we can go look at it ourselves and see what we can learn about it?
I didn't sorry it was AliExpress so I didn't expect much really Link (big one)

It's also sold on amazon and a few other places.

It's a great concept but the build quality is awful. With the exception of the frames, they're great metalwork lol. It really should have been a 12V device imo. So..that's what I'm making it into. I've disassembled the fan/light assembly and, as expected, it used a standard 80mm case fan (25mm depth) but 5V, no label. So that fan is getting replaced with a high static pressure Sunon 12VDC 3 pin I have left over from my PC case modding. There also appears to be room for me to install a small DC-DC converter, something like this. It's made for cars but should be ok and I'm only using half its rating. This would make it all 12V in essence, lower current needed from a wall wart and actually has the switches on the fan assembly.

Think that sounds doable? No dimming, but do I need it?

it's an ersatz fan.
I'd never heard that term before hehe but I like it. Good point too, though I'm changing the fan with a sunon monster. :)
 

Thread Starter

Marcus2012

Joined Feb 22, 2015
403
Finally got this done. Gutted it entirely and ran a length of 20AWG pair from a DC connector to the unit so it's all now 12V straight from mains wall wart. Changed the fan for a leftover high static pressure Sunon model that was way too overpowered for my PC and added a 12V-5V DC-DC converter for the LEDs. No dimming just a simple STDT (CO) for the colours. Works sooo much better now and I am quite surprised at the efficiency of the converter.

White LEDS = 1.38A @ 5VDC = 6.9W
0.66A @ 12VDC = 7.92W
Efficiency 87.12%

Not bad at all

Thanks to all for the help with this. I would still be trying to get the PCB controller working had you not clued me into the fact it actually had an MCU. :)
 
Top