DC/AC Inverter project, need some help with design

[Eddie805]

Hey well im definatley going to try the compressed air idea. BTW the engine was running. Also i added a 5 AMP mini c/b on the DC side of this circuit and will trip every time about 20-30 seconds of me turning the circuit on. Also i tested the transistors and they are fine. Im so blown away by this circuit. And Thanks alot for all the help and information.

 

[buloi123]

I recommend: try with darlingtons, and give a small assymetry for base-bias.

If I need 60hz, 220VAC output, what would I need to change? What values, in particular.

 

[Exo]

just the transformer i would guess.

 

[Sebi]

This circuit more stable, and always work....

 

[vik1501]

Sebi,

Could u plz mention the wattages of the resisters used and resister value at the base of transister bd680 :?:

 

[Sebi]

The 2N3055 base resistor 1W, the others 0.5W.

 

[vamboger]

I recommend: try with darlingtons, and give a small assymetry for base-bias.

boys... I have realized this outline resistor 10ohm is 20W not disponible 5W

but I do not succeed to pull much Watt I do not succeed to ignite not even a tv color from 42W like never? I have modified the resistances to 1.200ohm and 1.00ohm and have put 2 resistances to 180ohm the rendering is greater but he goes in block and he becomes all unstable... He arrives himself to times to of Ultrasonic the absurd frequencies... nearly burst the 2n3055 you help me I would want to realize a inverter that he succeeds to distribute 500W thousands Thanks Vamboger

 

[TheOne]

The original design is a very crude design, but it should work (getting the phasing on the windings right is critical for any operation). This can be tested separate by connecting the primary to 120VAC and you should read around 24VAC over the outer connections. Load regulation will be poor though.

The 68uF caps should be replaced by non-polarized caps or at least using 2 electrolytic caps in series (twice the value) connected back to back. The ripple current rating (depends on the ESR of the cap) can be improved by using several caps in a series-parallel configuration to make up the required value. I often use this method when making up cross-overs for high power speaker systems. As is, the 2 caps would be less reverse-biased over a cycle with the +ve's connected to the collectors.

The 2 diodes are used to clamp voltage spikes caused by leakage inductance at turn-off to 12V. As the peak collector voltage will be at about 24V for normal operation from auto transformer action, it will be better to use a zener in series with the diode to only return energy from spikes over 24V (that may be dangerous to the transistor) otherwise you burn a lot of unnecesary energy in the resistor. It is normal to set the clamp at about 2.5 times the input voltage (~30V) or to a safe level below max Vce of the transistor.

Zero references marked by the yellow lines

 

[TheOne]

Apart from the very crude design, the design according to me is flawed in the following ways.

1/ The diode clamps burn a lot of energy and places a lot of unnecessary extra load on the circuit. (Note in the 1st diagram how it affects the oscillation and distort the waveforms).
2/ The b-e junctions of the 2N3055 are subjected to reverse voltages which is quite high and could damage them. (look at the waveforms in my previous post)
3/ The capacitors are reverse biased during the cycle and that is not good for them.

I would attempt the following mods to improve the already bad circuit.

1/ Replace the clamp with a diode zener configuration to only clamp the leakage energy and not to interfere with normal transformer action. The collector winding must be allowed to swing to twice the input voltage, basic theory of a push-pull circuit!
2/ Insert schottky diodes in the bases to prevent b-e reverse breakdown. The caps across the diodes will still aid in switching off the device quickly.
3/ Replace the single polarized cap with a series-parallel combination for non-polarized operation and to increase ripple current handling. The 180 Ohms resistors could be slightly adjusted to get the cycle time correct.

From the waveforms it can be seen that switching is clean with very low reverse voltages over the b-e junctions. I have made the voltmeter's impedance low enough (100 Ohms) to be a load in the output.

(The 10k resistors on the primary are there cause the simulator wants to see some path to GND to converge)

 

[vamboger]

Apart from the very crude design, the design according to me is flawed in the following ways.

1/ The diode clamps burn a lot of energy and places a lot of unnecessary extra load on the circuit. (Note in the 1st diagram how it affects the oscillation and distort the waveforms).
2/ The b-e junctions of the 2N3055 are subjected to reverse voltages which is quite high and could damage them. (look at the waveforms in my previous post)
3/ The capacitors are reverse biased during the cycle and that is not good for them.

I would attempt the following mods to improve the already bad circuit.

1/ Replace the clamp with a diode zener configuration to only clamp the leakage energy and not to interfere with normal transformer action. The collector winding must be allowed to swing to twice the input voltage, basic theory of a push-pull circuit!
2/ Insert schottky diodes in the bases to prevent b-e reverse breakdown. The caps across the diodes will still aid in switching off the device quickly.
3/ Replace the single polarized cap with a series-parallel combination for non-polarized operation and to increase ripple current handling. The 180 Ohms resistors could be slightly adjusted to get the cycle time correct.

From the waveforms it can be seen that switching is clean with very low reverse voltages over the b-e junctions. I have made the voltmeter's impedance low enough (100 Ohms) to be a load in the output.

(The 10k resistors on the primary are there cause the simulator wants to see some path to GND to converge)

this schematic increase Wattage???
speak Italien???

 

[vamboger]

you have realised schematic with 4 2n3055 visit precedent post......

ok??

 

[TheOne]

Might be "easier" to realize with Power FETS, although there's a bit more components. It may have several advantages.

1/ The current through the caps won't be a issue.
2/ The gate resistors R1/2 can be 1/4W and be made tunable with a pot in series to set the correct operating frequency.
3/ FETS could be operated in parallel (for more power) by using (~2W) balancing resistors in the source connections.
4/ Design seem to be much more stable (frequency variation) with load change, compared to the transistor design.

Precautions:
1/ Zener diodes must be added to the gate's for protection.
2/ Because of the higher voltage drop it may be better to use a 9 + 9 volt transformer.

Diagram shows the design running with 110VAC output drawing about 1A current. First graph shows current and voltage waveforms of 1 winding. Last graph shows current drawn from the supply and switching current through C1/2

 

[vamboger]

im have disponible 2n3055.....

disponible schematic Pwm Inverter???


this software analisis is???

im use electronic workbench (problem in use oscilloscope) ((

 

[Roff]

im have disponible 2n3055.....

disponible schematic Pwm Inverter???


this software analisis is???

im use electronic workbench (problem in use oscilloscope) ((

Translation for all you non-Spanish speakers (including me):

disponible=available

 

[vamboger]

speak Francaise e italienne...

Ops..... :lol:

hihihih


cherche inverter puor alimentation di Pc un Auto (car) Tres puissance...
500W \ 1000W

hihihi

schematic 500W witch 2n3055 \ schematic PWM ????

 

[TheOne]

Here I have hooked up a previously posted circuit of mine https://www.electro-tech-online.com/threads/proteus-6-models.13098/ to add some form of output voltage setting by adjusting the duty cycle with RV1. There's very little frequency change while varying the duty cycle.

It is possible to use a 9 9 transformer and adjust the duty cycle for the correct voltage output. Simulating with a 9V transformer gave me about 140V at max duty cycle with a 12V input.. This could also be implemented automatically with a standard pwm chip, but then feedback stability etc. come into play and must be analyzed. Not so easy for beginners.

You could however use the standard pmw chip in a similar open loop control by applying a variable voltage to the compensation pin. So if you don't feel like throwing a couple of CMOS gates together, use a 3525 or something. If you are a beginner, building the discrete control circuit, you can learn a lot by probing with a scope! The max and min on times is set by the fixed resistors on either side of the pot. This dead time at the max on-time will prevent both devices conducting at the same time.

 

[audioguru]

Hi Guys,
Two other websites have hundreds of complaints about this inverter. I see that it has problems:
1) The capacitors got hot and exploded because the turning-off transistor's reverse-biased base-emitter junction was avalanching at about 7V (all silicon transistors) so the 23V charge of the cap was quickly discharged to about 8V with a very high current flow. TheOne's diodes at the bases prevents this.

2) The original circuit's capacitors are backwards because the transistors' collectors swing up to +24V or +28V with a charging battery. When the caps are discharged they get only about -0.6V across them, a little more with the base diodes added.

3) Single power transistors don't have enough current gain to use 180 ohm base resistors. To get only 100W from this inverter the transistors must conduct about 10A, so their base current should be about 1A. Therefore you will need 180V across the 180 ohm resistors to provide 1A. Of course, only 23V to 27V is available.
Since the transistors have higher gain at lower currents, if this project has base diodes added to allow it to work then it might have a max power output of only 20W.

4) The low voltage secondary winding of most transformers doesn't have enough coupling to the core to be used as an efficient primary winding, and therefore is nearly a dead short to the transistors.

 

[TheOne]

Audioguru, you are right that in practice it would be hard to get 100W out of this with the 2 x 3055 transistors. There's a much better chance with FETS. Higher power transistor self oscillating designs are of the Royer variety where you have extra windings for feedback. In general BJT's is not good for a push-pull type structure and often lead to failure because of other surprising reasons.

My main intension was to show that it is theoretically a working concept (even if at lower power) once a few changes were made. The real power handling capability would be difficult to establish on the simulator with all the unknowns in the equation.

 

[vamboger]

no aviable Pcb???
part list?


Vamboger

 

[vamboger]

insert in document zip schematic for proteus?
to apply my mod et My test???

sending schematic Best regard Vamboger