quote:

---

Hi Will,
If the inverter works properly (I don't know which circuit that you are talking about), the 24V winding's center tap is connected to +12V, so when a transistor grounds one side then the other side will swing to +24V. Select a 24V center tapped transformer that has a high voltage winding that is rated for whatever output voltage that you need, not double what you need.
To get 120VAC output, choose a 120V to 24V ceter-tapped transformer, which is also written as 120V to 12-0-12.

---

quote:

---

Hi Will,
Didn't you read my analysis of this website's 12V to 120V Inverter project that I posted here in May?
Are you talking about people getting an output of only 65VAC into a 75W load?
This simple circuit cannot give much output because the transistors do not have enough current gain and are being forced into avalanch breakdown of their reverse-biased base-emitter junctions.

To provide 75 Watts, the 13.5V battery must supply a current of 5.6A. Each transistor in the inverter must switch this 5.6A through its side of the transformer winding. At 5.6A of collector current, a 2N3055 transistor has a typical current gain of 25, so its base current will be 224mA. But the 180 ohm base drive resistors can supply only 68mA!
With 68mA of base current, the collector current of a typical 2N3055 transistor is only about 3A. So the typical output power is only 40.5W, much of which is used to heat the capacitors and to cause avalanch breakdown of the transistors.
This circuit doesn't work!

---

quote:

---

Hi Will,
That is a good idea. If you use 47 ohm base resistors then this inverter that uses typical 2N3055 transistors should have an output power of a whopping 75 Watts!
You can expand your theory further if you use only 6 ohms for the resistors so that the base current is 2A. Now there will be enough base current so that even 2N3055 transistors that have their minimum guaranteed gain of only 5 will have a collector current of 10A, and the inverter's output will be 120W.

>> I don't think a 2N3055 will do 10 amperes at 12 Vdc. Without looking at the spec sheet, this min. current rating starts to drop off at 5 Vdc. The most I've ever seen them used at was 5 amps IC for 12 Vdc circuits.

Now that we are talking about 10A collector current, we must not forget that the transistors are part of the oscillator, and have capacitors that must charge and discharge into the transistors' bases. Using 6 ohm resistors, those capacitors will be huge.

>> Yes, your correct. The 10 amp rating for one 2N3055 I don't think will work. That Will in fact go up in smoke. Personally, I've never seen one used in any power circuits with a collector current of over 5 amperes or so. That's why I was holding it back. If I were to want 10 amperes IC then I'd use two in parallel.

When the 1st transistor conducts, it turns off the 2nd transistor by its collector cap. When the 2nd transistor turns off, its collector voltage rises to about 27V by center-tapped transformer action.

>> It will actually be higher than this at turn on because of the spike created from the sudden shift in the windings. Thus, the zener diode clamps across the collectors and emitters will throw this to ground and hold it at a reasonable amount. They would have to be at least 24 V + 1 or 2 volts.

The cap that is connected from the 2nd transistor's collector to the 1st transistor's base must charge without any current-limiting and its current will exceed the 7A maximum base current rating of the 1st transistor for a moment. You could add 3.9 ohm resistors in series with the caps to limit the momentary base current to about 6.7A.

>> Exactly. If the base current is not limited, it will run away and destroy the transistor. I'd prefer to use two 2N3055's at each side for this and limit each to 1/2 the current. Here would be a good place to implement that kick start thing I was mentioning. Use a 33 ohm on one base and say a 39 ohm on the other. The slight difference will cause the circuit to always oscillate and not latch up.

Do you understand about avalanch breakdown of a silicon transistor's reverse-biased base-emitter junction?

>>I reckon I do, if not, I flushed two years of school and 23 years experience down the toilet.

All silicon transistors have a maximum reverse base-emitter voltage rating of about 7V, whtch if exceeded causes the junction to avalanch like a zener diode and conduct massive current.

>>Yes it does, this has to be limited in the circuit.

This current creates hot-spots on that junction that is not cooled well by the transistor's case since the other junction's collector is bonded to the case.
That is bad news for the transistor since a massive current flowing through an uncooled base-emitter junction that has about 7V across it creates a very high temperature at that junction. Poof!

What about the ripple-current rating of the capacitors? That is what caused them to blow-up (even with the correct polarity) in the original circuit. Using 6 ohm resistors, the charge and discharge currents of the caps will be huge. Poof!

>>Use a large enough, and correct type of capacitor. When the voltage rating of an electrolytic is increased, so is it's current capacity. This is done by increasing the plates size and the coating thickness on the one plate (anode). Let's say for practical purposes, use a 220 uF @ 450 Vdc cap instead of a 220 uF @ 35 Vdc. The difference in physical size is huge. The larger cap will dissipate heat way better and would probably hold up I would think. I forget what the current carrying capacity is for that cap but it's a good amount. This type I'm speaking of is the radial screw top, computer types.

So most of the power from this inverter will be used to destroy its transistors and capacitors, and very little power will remain for its load.
It might work a lot better and with improved reliability if the transistors are replaced by Mosfets.

>> The Mosfets is a good idea because they're voltage dependant, not current dependant. The capacitance polarity and spikes still have to be corrected though as above.

---

quote:

---

Hi again Will,
I forgot to explain how avalanch breakdown occurs in this circuit.
When the 1st transistor conducts, the capacitor at its collector has been charged to 27V, and the cap's end that is connected to the 2nd transistor's base will attempt to drive the base to about -26V. The cap is very quickly discharged with a very high current because it has a high-current collector of the 1st transistor driving its other end, and the avalanching base-emitter junction of the 2nd transistor is limiting its negative end. So lots of current through the cap and lots through the junction. Poof! Poof!
The other cap and junction have the same problem. Poof! Poof!
This simple inverter is just a poof-maker and is good only for self-destruction!

---

quote:

---

Hi Will,
Yeah, I have built unbalanced multivibrators too, to guarantee start-up.
But when used with a supply voltage more than 5V, I always add diodes in series with the emitters to prevent avalanching.
I don't think that limiting the current during avalanching is good enough, after all, the manufacturers just say, "Max reverse base-emitter voltage is 7V. Don't exceed it". In my book that means "Dont do dat!"

>> Actually, the diodes off the emitters is a good idea and I forgot about that myself. Anyhting to keep the transistor from breaking down.

The users won't learn much when authors post projects like this one, then say to use more powerful trasistors and transformer to get more output. Only the guy who tried 300V/30A transistors and a rewound transformer from a microwave oven learned that it didn't make any difference.

>> The only way to get more output voltage wise is to increase the turns in the winding, i.e., a higher voltage secondary. However, the current will drop accordingly. Wattage going in has to equal wattage going out minus about 5% due to losses in the transformer itself. So if say 300 watts were created by the primary, It takes 315 watts from the secondary. The old saying goes here, you cant get something from nothing.

>>To me, it's best to have more than enough switching transistors/fet's to do the job and keep them running cool instead of pushing their limits. I think the FET idea is the best yet using healthy capacitors. However the current draw wont be there on the capacitors like on the bipolars.

>> Really, the only advantage to this circuit is to set an oscillating frequency. However, in a lot of invertors used to control a DC circuit on the other end, the frequency is raised. The higher the frequency, the more efficient the transformer is. Actually, the amount of iron needed drops as the freq. rises. For portable TV's, etc which use DC circuits anyways, it will work on higher frequencies. But, motors, etc. won't.

Will

---

quote:

---

The 500W inverter project that I fixed was reported to deliver 720W.
I would never push transistors so hard, therefore rated the circuit at only 500W.
The link to that schematic is posted again here:
http://www.electronics-lab.com/forum/attachments/500Watts_Inverter.gif

---