Maker Pro
Maker Pro

Dumb question re. simple inverter

Dougster

Nov 8, 2014
21
Joined
Nov 8, 2014
Messages
21
I have built a simple inverter (12vdc to 120 ac) using a cd4047 to drive (2) n channel mosfets. The output seems very weak, and I noticed I only have about 2.5v of gate voltage. Could anyone recommend an amplifier to put between the IC and the gates? Would adding mosfets in parallel help? I'm dealing with pulsating DC, if that matters. As you can tell, I'm just a beginner. Thanks for any advice you can give me.

Notes.
 

KrisBlueNZ

Sadly passed away in 2015
Nov 28, 2011
8,393
Joined
Nov 28, 2011
Messages
8,393
When you've removed R1, the 4047 will be able to operate properly. But it's not designed to drive MOSFET gates directly; its maximum output current is only around 10 mA. If you find your MOSFETs are getting hot, it could be because they aren't being switched cleanly, because of the low drive current. In that case you should use a dual "low side" MOSFET gate driver IC.

Here's a selection filter for Digi-Key. Any of these devices should be suitable. Some of them have one inverting driver and one non-inverting driver; for those, the inputs should be connected together and driven from one of the 4047 outputs.

http://www.digikey.com/product-sear...nSort=1000011&stock=1&quantity=1&pageSize=250
 

Dougster

Nov 8, 2014
21
Joined
Nov 8, 2014
Messages
21
When you've removed R1, the 4047 will be able to operate properly. But it's not designed to drive MOSFET gates directly; its maximum output current is only around 10 mA. If you find your MOSFETs are getting hot, it could be because they aren't being switched cleanly, because of the low drive current. In that case you should use a dual "low side" MOSFET gate driver IC.

Here's a selection filter for Digi-Key. Any of these devices should be suitable. Some of them have one inverting driver and one non-inverting driver; for those, the inputs should be connected together and driven from one of the 4047 outputs.

http://www.digikey.com/product-search/en?FV=fff40027,fff8020b,401993,1140050,15c0003,33c0031,33c018f,4500031,450009e,45000f1,45001e3,4500208,450023d,450024d,450024f,4500410,4500411,45006e0,4500d70&ColumnSort=1000011&stock=1&quantity=1&pageSize=250
Thanks so much. That's exactly the kind of info I was looking for. I will definitely try your suggestion and let you know. By the way, I copied the basic schematic off of the internet. Since everyone seems to agree R1 should be removed, why do you think it is included in the first place?
 

KrisBlueNZ

Sadly passed away in 2015
Nov 28, 2011
8,393
Joined
Nov 28, 2011
Messages
8,393
A lot of the designs I see on the net have problems like that. The people who post them just don't have enough understanding or experience to produce designs that work properly.

My pet peeve is instructables.com; most of the electronic projects I've seen have significant problems, and I sometimes post lists of corrections. I emailed the instructables people and offered to check their electronics projects for them, but they weren't interested. I think that's a pretty irresponsible attitude.

I see you got that design from circuitstoday.com. That is another site that generally has poor quality designs.
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
One of the issues with that circuit is that the signals to turn the mosfets on and off are simple complements of each other.

That means one mosfet is turning off as the other is turning on. This allows current to flow through both mosfets. Because this effectively cancels out the inductance of the primary winding, only resistance limits the current.

In this case the mosfets turn on and off quite slowly. Whilst this means they tend to run warmer, it also means that it is unlikely that both mosfets will be substantially turned on at any time.

To improve the efficiency of the inverter, you need you improve the switching speed of the mosfets. This is done by using a mosfet driver that can provide significantly more gate current to the mosfets. As you do this, the chance of both being turned on simultaneously increases. This is cured by introducing dead time to the mosfet switching. Essentially this means that the turning off of the mosfet happens as soon as possible, but the turning on is delayed (perhaps by a few μs).
 

Dougster

Nov 8, 2014
21
Joined
Nov 8, 2014
Messages
21
Thanks for your help guys. I removed R1 and added a driver for the gate circuits, but I still have a very weak circuit. It won't even drive a 5 watt night light I use for testing w/o the voltage dropping to about 90V.The problem Steve mentions seems to fit the symptoms I'm seeing, as the mosfets are often hot and don't seem to be working properly, but I have no idea how to introduce "dead time", or delayed switching. Is it a simple procedure? Bear in mind you're conversing with a tech from the stone ages. Thanks again. I LOVE this forum!
 

Arouse1973

Adam
Dec 18, 2013
5,178
Joined
Dec 18, 2013
Messages
5,178
The MOSFETs getting hot is probably due to them not being turned on fast enough as already mentioned by Steve. Do you still have the large value gate resistors? Try reducing them by a factor of 10. To turn on the MOSFET fully in 1us will require approx. 72mA. If it needs to be quicker then this will require more current from the driver so just make sure the driver you are using can supply enough current for the gate cappacitance.
Adam
 

Dougster

Nov 8, 2014
21
Joined
Nov 8, 2014
Messages
21
I dropped the gate resistors way down, but the output only got weaker. I got to thinking about Steve's theory that the Mosfets are both conducting together and defeating the coil inductance in the transformer. I didn't know how to sharpen their turn-on/turn-off time, so I played with the frequency since that decides the switching time, and VOILA!! I found a .47uf to be close to ideal. That's a hell of a lot more than the .01uf shown in the schematic. I don't know what's inside the CD4047, but the shown RC network by itself equates to about 250khz, no? Maybe at a high frequency, a slow mosfet sees the gate as always on? Anyway, it's working great now. Thank you for all your help and suggestions. The next time I copy a schematic from the web, I'll ask you guys for advice first!
 

KrisBlueNZ

Sadly passed away in 2015
Nov 28, 2011
8,393
Joined
Nov 28, 2011
Messages
8,393
If you're using a mains transformer in reverse, you should run it at around 50~60 Hz.
 

KrisBlueNZ

Sadly passed away in 2015
Nov 28, 2011
8,393
Joined
Nov 28, 2011
Messages
8,393
Are you still using the basic design from post #1 or are you using the BJT-based design that you sent me in your PM? If the latter, can you repost the schematic and questions from your PM on this thread?
 

Dougster

Nov 8, 2014
21
Joined
Nov 8, 2014
Messages
21
It's the basic design from post one except I added. the gate driver you suggested, removed R1, and installed a bigger rc network clocking the IC. I'm happy with it now (finally)! I might add another gate driver and maybe 4 more mosfets in parallel. I'm already using a 500 VA transformer. Is it OK to parallel more mosfets? I read somewhere not to daisy chain them as each one should have its own gate resistor. Any tips?
Thanks again!
 

duke37

Jan 9, 2011
5,364
Joined
Jan 9, 2011
Messages
5,364
The frequency is 1/(4.4*R*C) someone has borrowed my slip stick so you will need to work out the frequencies yourself.

390k C= 1E-8
390k C= 0.47E-8

The frequency must be within the capability of the transformer. You can make a transformer from a TV line output transformer core with 1 turn/volt. Run over 20kHz to prevent whistle.
 

duke37

Jan 9, 2011
5,364
Joined
Jan 9, 2011
Messages
5,364
The frequency is 1/(4.4*R*C) so 390k and 0.01μF is 58Hz and 390k with 0.47μF is 1.2Hz which is far too low. Is the resistor actually 390k?
 

duke37

Jan 9, 2011
5,364
Joined
Jan 9, 2011
Messages
5,364
Thanks for the welcome. I have been locked up, mending my broken spine. I managed to walk (or stumble) without a stick yesterday. My electronic hobby has been put on hold but I am off the pain killers and some parts of my brain are working again.
 

Rleo6965

Jan 22, 2012
585
Joined
Jan 22, 2012
Messages
585
Dougshter

It seems I recognized your the simple inverter diagram. It was from the website that had intentionally change value of components in its schematic diagram. Probably to test the circuit analysis capability of the person that was interested on the circuit. They usually add "K" or add "0" to components value. For example R3 and R4 should be 22 ohms.

Also keep in mind that the circuit was simple inverter or household inverter. This means it ran on frequency between of 50-60hz as KrisBlueNz mentioned. Knowing the ac output frequency will help you compute to change value of RC timing components.
 
Top