Saturday, October 25, 2014

Pure Sine Wave Inverter with IC 4047

              Pure Sine Wave Inverter with IC 4047

Here's a screen capture of the pure sine wave output using 4 mosfets (two mosfets in parallel) from the pure sine wave inverter with the IC 4047 circuit, designed by Mr. Swagatam Majumdar:

Pure Sine Wave Inverter with the IC 4047 designed by Mr. Swagatam Majumdar

I've been working on the design for a while now, and this is my first attempt at adding mosfets in parallel. As you can see, I have been trying to figure out on my own (unsuccessfully) why there would be spikes on the output. I have tried a flying diode configuration on pins 10 and 11 of the 4047 (which seems to help diminish the spikes), but the spikes are still with me. Currently using a blue toroid with 5.46┬ÁH inductance factor and a permeability of 4300 with 32 turns of 22 guage wire. It looks like I may have to ask for an answer to that one (from Mr. Majumdar).

Currently I am using a 10K potentiometer at P1, the C1 capacitor at 0.1uF, and about 200K at R1 (as opposed to 180K off of pins 2, 6, and 7 of the middle 555 IC1).
From my initial testing, these are the values that I have found useful for tuning the inverter (prior to the addition of additional mosfets):

 For R1 (off of pin2 of the 4047) and C1 (between pins 1 and 3 of the 4047), with only two mosfets from the original design, these values were useful in frequency adjustments:

R1 ~=  100 with C1 = 1.0 uF
R1 ~=  470 with C1 = 0.22 uF
R1 ~=  15K with C1 = 0.1 uF
R1 ~=  37K with C1 = 0.047 uF
R1 ~= 202K with C1 = 0.01uF

This was for the 100 watt, two mosfet configuration using a 10k potentiometer at pins 1 and 3 of the 4047. That is, I only tested this one up to 100 watts.

However, after adding mosfets in parallel, the frequency does not appear to be as stable under light loads. I have not tested heavier loads enough yet, but the frequency appears to be more stable under heavier loads from what I have seen thus far.

My test configuration for mosfets in parallel:

Pure Sine Wave Inverter with the IC 4047 (proposed mosfets in parallel)
Update: 10/27/2014 - 

After having melted some wire insulation and breadboard fittings, it looks like I will need to figure out how to get this off the breadboard for further testing. The point being that I am seeing the need to bump up the amperage level, and in turn, bump up the resistor an potentiometer (P1 and R1 on pins 1 and 3 of the 4047). I ended up the day today with R1 set at 100K, P1 at 100K, and testing the various capacitors for best response to frequency adjustments. 

Adding the mosfets in parallel is changing what I need to get the targeted frequency (which I believe, for the U.S. should be 120 Hz). I can get to 60 Hz, but with 4 mosfets in place, I am having difficulty getting to 120 Hz. I am thinking Mr. Swagatam Majumdar intended the frequency to be for phased power, but keep forgetting to ask him about that.

Yet another update:   10/29/2014 -

I just wanted to post this image of the sine wave this morning.

Pure Sine Wave Inverter with the IC 4047 - 60 Hz no load

This is with two IRF 3205 mosfets in parallel (4 fets total) and 0.047uf for capacitor C1, although the image is similar for C1=0.22uF, etc.

Main thing there is that the spikes are gone. I believe the two main items that I changed there were:

a) The resistance at R1 and P1 on pins 1 and 3 of the 4047. 
     I bumped both up to 100k Ohms in an effort to control the frequency more easily.
     Of course, bumping up these means I need to work with more battery amps.

b) The 1F capacitor linked to the 557 transistor between IC2 and IC2 (555's).
    I am now using a Metalized film capacitor 250 WVDC max that I picked up from Radio Shack.

c) I am now testing with batteries with more amperage. I don't have a variable power supply.

Not sure which one made the difference in the presence of spikes, but if I were to guess, I would say it was the 1uF capacitor or the increase in amperage. Or, it could simply be that I have no appreciable load for that screen capture. I don't know, this is new territory for me.

Without bumping up the resistance, the frequency began running away from me. The increased resistance helped me to keep the frequency around 50-60 Hz. I still have not figured out the exact resistor values needed to get that to around 120 Hz

Increasing the amperage power of the battery is still giving heartburn with melting wire insulation and breadboard fittings, but I hope to figure something out so that I can determine how much power in watts is being delivered by mosfets in parallel (4 and 6 count).

Which means, I may simply build the prototype, just to get it off the breadboard, and enable me to test the mosfets. That would mean I would be crossing my fingers that I have the resistance in the required range. 

Currently, the following settings appear to provide 60 Hz frequency (with the 4 mosfets and the 0.1uF capacitor at pins 1 and 3 of the 4047):  

R1 ~=  44K with C1 = 0.1 uF using 4 mosfets and a 100k Ohm 1/4 watt potentiometer at pins 2 & 3 of the 4047

However, my 4 amp transformer is getting hot easily, which I presume tells me I need a transformer with more amps with the 4 mosfet configuration. 

Finally got a wiring arrangement that permitted me to run from the breadboard with the wires being slow to heat up. Here, I am running a 250 watt flood light at 120 Hz using 4 mosfets, C1 = 0.047uF, R1 = 15K, with P1 = 100K 1/4 watt from a 4 Amp 12-0-12 transformer:

Pure Sine Wave Inverter Using the IC 4047 at 120 Hz, 250 Watts using 4 mosfets
Problems not yet addressed:

In the current configuration, I have the mosfets isolated on a separate breadboard, so they are yet to be hard wired separately. I still have a wire heating problem, although I am now using heavier wire on the high current areas. And, the 4 Amp 12-0-12 transformer still gets fairly hot very quickly. 

I have ordered a few specialty light bulbs so that I can do a reality check on the measurement of the watts. I believe that I will need to get a transformer with more Amps, and use still better insulated wire for the high current areas. Alternatively, I could try increasing the number of mosfets in order to see if the transformer holds up under more power. I suspect not, but I do not know that yet.

As you can see, I have not yet gotten rid of the spikes under a heavier load. My current thoughts on addressing that is to use yet heavier wire on the toroids that I have added off of pins 10 and 11 of the 4047.

Update: 10/27/2014 - 

Well, in my review of Mr. Swagatam Majumdar's notes about the 4047, it turns out that he has probably already suggested a remedy for the spike problem that I am having.

In his comments regarding the Mr. Daniel Adusie after connecting a 0.22uF/400V capacitor and a suitable load, Mr. Majumdar suggests an inductor on pins 1 and 3 (of the 4047), along with the capacitor.

I've been running a flyback diode on pin 11, with a balancing toroid on pin 10. I had done that in response to the power problem with my substitute mosfet, IRFP150. Just too much power there for me, which was one of the reasons that I had decided to switch to the IRF3205 style mosfet. I was getting a good sine waveform from the 3205 and the toroids, and the frequency was responding to the R1 and P1 settings.

Therefore, my understanding would suggest a trial of an inductor off of pins 1 and 3 of the 4047.

Now, I don't have a collection of inductors, but I do have a few toroids. That should enable me to test the number of windings, the wire thickness, whether or not I need the toroids off pins 10 and 11, and whatever variance in resistance that might involve.

Finally, I only have one 22 Amp hour battery that enables me to run at 250 watts in my test mode. One day of testing should run down the charge on that battery. Therefore, my game plan this weekend is to test the inductor configuration on pins 1 and 3, checking for number of windings, thickness of coil, and whether or not the toroids on pins 10 and 11 are really needed for the waveform. Since I have limited battery power, I will test from the 2 mosfet, 100 watt configuration on with batteries that are not as powerful.  That should help me get an approximation of what inductor (toroid) structure is needed for the 4 mosfet configuration.

Of course, not to forget that another option for me is to see how the configuration works with the IRFP150 (as opposed to the IRF150 in Mr. Majumdar's schematic).

Additionally, I will be looking at heavier insulated wires, and perhaps shopping for a new transformer (with more amps) if the transformer heat problem does not improve. I am also looking at overload protection, adding a fan, and adding a voltage regulator. It would be nice to add some sort of battery monitor as well. The design has an ammeter area, and I am currently using a BeesClover amperage and voltage meter at that location (between pins 2 and 3 of IC1 and IC2). That little BeesClover ammeter has been helpful here.

Very nice to see this design has a number of options, and that it holds so well up to my abusive testing. The design has been fairly durable thus far.

Lots of options there for me to explore this weekend.

update 11/03 - Corrected number of windings on toroid to about 32 turns. Now checking to see if a different number of turns will correct the spikes.

Update  11/12/2014:

This one appears to be responding to what I believe is called "passive" filtering, with respect to the 'spike' problem. I have removed the toroids inductor(s), and removed the flyback diode, and have placed 10 Ohm resistors and 0.1uF capacitors (630 volt poly metal film) between the mosfets and the transformer. I have a schematic of what that looks like:

Pure Sine Wave Inverter Circuit Using IC 4047 With Passive_Filter

I plan to use four levels of this resistor/capacitor filter, but just so that you can see the difference, here is a view of a level 3 passive filter example from a 4 watt light bulb, using just the two mosfets, as in the original 100 watt version of the inverter:

Pure Sine Wave Inverter Circuit Using IC 4047_W_0.1uF_75Hz_2FETS_55K_P1_Level3_filter

Needless to say, the spikes are really no longer a problem at level 3 filtering, but I do hope to eliminate the spikes altogether by taking the filter to a fourth capacitor/resistor configuration. In the image above, I am using IRFP150 mosfets.

P1 = 10K, R1 ~=  55K with C1 = 0.1 uF, no toroids, using passive filtering

 I am having problems locking in on the frequency, as I have not experimented with this configuration much yet. Currently using C1 = 0.1uF, R1 = 55k at pins 1, 2, and 3 of the 4047. For the filter configuration and voltage less than 5, the frequency wants to hover at about 1.5 KHz, which is something I would like to improve. The frequency for higher voltages looks O.K. for the moment, but I don't want to show that until I have tested these low voltages sufficiently.

As for the filter, I have already looked at using 1 uF for the capacitor, but that appears to decrease the efficiency, requiring more Amps from the battery. I have also looked at using a 0.01 uF filter capacitor, but that is making it a bit difficult to get a handle on the frequency. 

This also appears to have relieved some stress on the transformer. At the moment, I am still using a 4 Amp, 12-0-12 center tapped transformer.

Perhaps with some combination of C1 on pins 1 & 3 of the 4047, and the right selection of the filter capacitor, it might be possible to come up with a trade off between a stable frequency and battery efficiency.

Pure Sine Wave Inverter with IC 4047 Schematic 

Pure Sine Wave Inverter with IC 4047 on a Breadboard


In the video, I should note that I added a voltage regulator (LM7808) and a silicone controlled rectifier (SCR, 2N6507).
The multimeter was not connected to the power outlet, as demonstrated when I shorted out the connection (and the light went out at about 9 minutes into the video). So, there was a little wireless power going from the outlet to the multimeter probes.


  1. Hello
    Impressive work on swagatam circuit but I think there are little suggestions.
    1) don't drive the MOSFETs directly with pin 10 and 11 of 4047
    Once connected to diode use a MOSFET driver like ir2110 or tc427 or dedicated push pull MOSFET driver so that frequency and response will be stable.
    Simply connect the pin 10 and 11 of 4047 into the mosfet driver.
    If you drive MOSFET directly from pin 10 and 11 it will be unstable because diode cannot drive mosfet faster under heavy load.its too slow.
    Use 1k resistor at 10 and 11 then connect diodes and then connect directly into mosfet driver then drive MOSFET through driver.
    If you use ir2110 then use as "dual low side driver" Google "using ir2110 as dual low side driver tahmid" to see how.
    Also you cannot run appliance at 120hz at the output. Output AC should be 60hz or 50hz depending on country.
    So pin 10 and 11 should be 60 hZor 50hz depending on country while pin 13 will be 100hz when pin 10 and 11 is 50hz and 129hz when pin10 and 11 is 60hz.
    If output AC is 120hz it will burn appliance soon.
    Your transformer is too small and so inefficient at high current.

    1. Yes, I would agree that the circuit is very nice work by the various authors, and I appreciate that Mr. Swagatm Majumdar has posted it to his blog.
      O.K., I will try the IR2110 or TC427 as a mosfet driver.
      Also, I will try dropping down the frequency to 60 Hz (here in the U.S.).
      I was thinking that Swagatam was suggesting a smaller transformer, but I do have larger transformer that I can try.

      Thank you, I appreciate the comments.

  2. One more thing. Don't put filter at DC side of inverter. Filter should be at output of Inverter and should be LC filter.
    Don't use resistor as filter other wise it willbe inefficient due to high current under heavy load.
    You can first try ordinary AC capacitor as filter at inverter output. Then later you use LC filter.

  3. Right, I can tell that using a resistor as a filter is fairly inefficient.
    I will try the AC capacitor and see how it goes.
    Thank you for the comments.

  4. Hello
    Concerning the transformer, you should know that to effectively get a good sine, if you are using a 12v battery supply then your transformer should be around 7-0-7/230or 8-0-8/230.or for US is 120V I think so it should be 7-0-7/130 or 8-0-8/130
    This is because the peak voltage of a sine wave is 1.4141 x R.M.S voltage so for operiating at 12vdc the primary should be less than 12/1.4141 = 8.5v so less than 8.5 should be 8v.
    So you should use 8-0-8 or 7-0-7.
    I recomment you use 7-0-7/170 or 8-0-8/170
    But 7-0-7 would be better.
    The reason is this we need to induce at least twice the required AC voltage we need so for U.S 120v sine wave you need to get at least 240v square wave but more is recommended at about 280V square wave so that when you apply diode at gate of MOSFET it will chop or carve it down very well into well space PWM sine waves and also it will come in handy when we want to regulate voltage.
    So basically aim to get more than twice your required sine wave voltage when diodes are not present so that when diodes are put the voltage will reduce and then we can control it well and it will be perfectly carved.
    Remember that when we apply diodes gate voltage reduces.
    Hope you get the whole concept I explain.

  5. Try mr swagatam circuit with sg 3525 its better than 4047...i have tried it its better... And you can adjust output easily.

    1. Hello hisham,

      I have my review of the 3525 Pure Sine Wave here:

      However, I did not include pictures of it on a breadboard.

      Let me see, I included the waveforms, but I did not tune it for the waveform or frequency. Oh, since I did not finish tuning it, I never got around to boxing it up in an enclosure.

      I preferred to work on the 4047 Pure Sine Wave because it has no low pass filter on the end. I just love to play around with this one.

      Fun stuff.

  6. very nice project sir have u ever think of trying it with a h bridge design. I have tried it but running into some problems

    1. Thank you, I admire the design work of Mr. Majumdar.

      I have not tried an inverter with an H-Bridge design, but I have done a solar tracker circuit on a breadboard with an H-Bridge design.

      Here, to increase the wattage capability, I have used more powerful mosfets and have had modest success at damping the spikes with the use of inductors or flyback diodes.

      I have not tried an H-Bridge with this one.

  7. thank u for ur reply I appreciate this keep up the great work and I continue to follow ur blog

  8. Great Post, I read it well. very useful info Thank you so much for sharing this one really well defined all peaceful information...
    Manufacturer of batteries | Solar battery manufacturers

  9. If you want your ex-girlfriend or ex-boyfriend to come crawling back to you on their knees (no matter why you broke up) you need to watch this video
    right away...

    (VIDEO) Win your ex back with TEXT messages?