I'm talking about the appnotes of TL783, don't you asked you to find it for TL783?
So I missed your point?

AFAIKR: that's a high voltage part - but its still a 3-terminal regulator.

Your post that I was referring to depicted a 317 - if you have the attention span of a goldfish, you might try your hand at something less demanding.

 

AFAIKR: that's a high voltage part - but its still a 3-terminal regulator.

Your post that I was referring to depicted a 317 - if you have the attention span of a goldfish, you might try your hand at something less demanding.

Actually in the beginning, I want to suggest the circuit as this or using two 2N2955 to be the power bjt, they are quite normal, I didn't expect the LM317 will designed like the one I linked on page #5, I just thought that the circuit should be like the AnalogKid suggested, but I will suggest to using two 2N3055 and in series with two small current limiting resistor for each Emitter, so that's my fault, sometime if we didn't look more carefully will cause the mistake like this.

 

To the TS: Please ignore the last 10 posts. Texas Instruments datasheets are not very user-friendly, and often cause such squabbling.

The TL783 operating parameters are in the datasheet if you know where to look. The minimum dropout voltage is shown in figure 10, and the minimum load current required to maintain regulation is shown in figure 12. Note that the dropout voltage is significantly higher that that of the LM317, which increases the minimum thermal load of the regulator. I recommend the circuit in post #12; that guy is smart.

ak

 

To the TS: Please ignore the last 10 posts. Texas Instruments datasheets are not very user-friendly, and often cause such squabbling.

The TL783 operating parameters are in the datasheet if you know where to look. The minimum dropout voltage is shown in figure 10, and the minimum load current required to maintain regulation is shown in figure 12. Note that the dropout voltage is significantly higher that that of the LM317, which increases the minimum thermal load of the regulator. I recommend the circuit in post #12; that guy is smart.

ak

You can't ignore your last posted, otherwise probably you have to blow for his power bjt.

 

You can ignore your last posted, otherwise probably you have to blow for his power bjt.

blow - ???

 

blow - ???

Sorry.
What I want to say is you can't ignore the posted at your "last" posted and that is my posted, I mentioned that the circuit you suggested should use two 2N3055, otherwise you may need to use your mouth to blow out the wind to dissipate heat for 2N3055, because only use one 2N3055 then it will be too hot.

Most of application, 2N3055 only use for 2A or a little more current is the best way, of course, you can challenge it, and I saw someone use 2N3055 for 10A.

 

The 2N3055 is rated for 15A continuous collector current, and this application calls for 5 A, so it is well under the "50% rule". Yes it will get hot, and depending on the input voltage it might get *very* hot. But whether there is one boost transistor, or two, or ten, the total power dissipated in the regulator is the same, and the need for a large heatsink and fan is the same.

ak

 

The 2N3055 is rated for 15A continuous collector current, and this application calls for 5 A, so it is well under the "50% rule". Yes it will get hot, and depending on the input voltage it might get *very* hot. But whether there is one boost transistor, or two, or ten, the total power dissipated in the regulator is the same, and the need for a large heatsink and fan is the same.

ak

Since that is your suggestion, so it's your problem now, not mine, my suggestion is using two or three pcs of 2N3055, at least two pcs.

 

The TS says the equipment needs a voltage of 45 to 55 volts. With the right transformer, about 38 VAC secondary voltage, he should be able to get by with no regulator at all.

The grid waveform nowadays is somewhat flat-topped, so the traditional 1.414 times the Vrms of the secondary gives a result that is higher than the actual peak voltage.. He could use a small variac on the primary and get just what he needs.

 

At 5 A output current, a non-regulated supply would need a large filter capacitor. Otherwise, the ripple voltage would be more than the common mode capabilities of a POTS phone can suppress. In round numbers, 8,500 uF would leave 5 Vpp of 120 Hz ripple in the power out to the phones. Reducing this to 1 Vpp would take 42,000 uF. At 100 V, those are big caps.

A large series inductor between two caps would reduce the ripple even more, and more importantly reduce the harmonic content to make it far less audible, but I think both the cost and the weight would be greater than a solid state regulator.

ak

 

A 50,000 uF computer grade cap would be about the size of a big can of Red Bull; doesn't seem all that big compared to a bunch of vintage telephone exchange equipment. If harmonic buzz is a problem, a simple capacitance multiplier can eliminate it.