723 psu mod

[p.larner]

Would swaping q1 bd139 for a tip 122 run 6x 2n3055  (not 4 as n the pic) pass transistors ok?

[tggzzz]

Would swaping q1 bd139 for a tip 122 run 6x 2n3055  (not 4 as n the pic) pass transistors ok?

Pls mk u r sntncs an qs more coherent.

[coromonadalix]

@tggzzz   can you write  "badder" than that 

[p.larner]

why cant you guys answer the question rather than leave childish comments,fwiw the schematic pic was a copy/paste from the net.Q1 is marked bd-139.

[Zero999]

why cant you guys answer the question rather than leave childish comments,fwiw the schematic pic was a copy/paste from the net.Q1 is marked bd-139.

Accept people will take the piss when you clearly make no effort with your spelling, punctuation and grammar.

This is also your third 723 thread this month. You'll get better responses, if you keep it all in one thread.

To answer the question: have you checked the data sheets? TIP122 is a Darlington pair, whilst the BD139 just a single NPN BJT.

It depends on the requirements. A Darlington pair has much more current gain, yet less transconductance, double the base-emitter voltage and a collector-emitter voltage drop of at least 0.7V, with 2V being more typical at higher currents.

[CaptDon]

Your total load current divided by the HFE of a single 2N3055 should give an approximate guess as to the current your driver will have to pass. Scale that up by a safety factor of 2:1. Multiply that current value by the anticipated voltage drop across Q1 to get the power dissipation. If voltage, current, S.O.A. and total power specs are not exceeded then you should be ok. The unknown here is the surge current when your supply comes up from zero volts at turn on although the LM723 should be able to keep surge current and fault current under control. There must be some feeble or fake LM723 / ua723 parts floating around as replacements because in the ham radio Astron 30 amp 13.2vdc those 723's are popping like popcorn!! In my stash of 723's aquired over the years from many different sources it turns out some of mine won't even work at all in the Astron (and knock offs) supplies. I have two here in front of me going to a ham radio buddy and he has already killed two over the years.

[CaptDon]

It's all up for grabs swapping a darlington for a normal BJT. I did fully comprehend your first post in this thread.

[Terry Bites]

Heatsink deisgn!

[p.larner]

i wanted to know if the tip 122 darlington would have enough guts to drive 6 pass as oposed to the bd tranistor circuits 4  "without killing the 723"?.

[Zero999]

i wanted to know if the tip 122 darlington would have enough guts to drive 6 pass as oposed to the bd tranistor circuits 4  "without killing the 723"?.

Have you looked at the collector current ratings on the data sheets?

No, it won't kill the 723 because a Darlington pair has a higher current gain, thus a lower base drive requirement. The extra voltage drop could be more of an issue, depending on the input voltage.

[p.larner]

Am i looking at this correctly,the tip 122 can conduct 5a max so does that mean it will drive 6 transistors at 0.8a each,assuming an beta of 20 for the 3055's the psu will output about 16amps max?

[p.larner]

can folks tell me if my thinking is correct on my last post?,or am i way off?.

[nali]

Not enough info:-
What power do you expect it to dissipate
What heatsinking are you going to use
What is the expected junction temperature
Look at the TIP122 datasheet (ideally BEFORE posting...). Especially SOA and temperature derating
Then you have your answer.

[Ian.M]

Also look at the 2N3055 datasheet SOA and derating graphs.   For a 0-30V PSU, your input voltage needs to be over 34.4V in the ripple troughs (35.1V with a Darlington driver), and assuming a max. heatsink temperature of 75°C, that means the most a 2N3055 is good for is 2.3A, with no safety margin.   If the input voltage is higher that drops considerably, e.g 2A @ 40V in, so your six transistor PSU will at best be good for about 14A continuous into a shorted output, and possibly only 12A.  Therefore, unless you implement fold-back current limiting, so at higher output voltages it can take advantage of the more favourable lower Vce side of the SOA, expecting 16A out is unreasonable.

I doubt you need a Darlington driver.  If the original circuit was good with a single BD139, simply use two, each driving three of the 2N3055, and drive both their bases from the '723 VOUT pin.  Worst case the total load on VOUT will be under 10mA if you select for the crappiest low gain 2N3055 and BD139 transistors, and if you use decent average gain transistors, it will only be a couple of mA.

To minimise voltage drop and thus unwanted dissipation, I'd use the voltage drop across the emitter degeneration (current sharing) resistors for the current sense.  Simply chose them for the desired limit current PER TRANSISTOR and combine all the emitter voltages via individual 100 ohm resistors to the 723's CURRENT LIMIT pin.   The arrangement with the 500 ohm pot between CURRENT LIMIT and output, wiper to CURRENT SENSE is no good.  Turn the wiper too far towards the CURRENT LIMIT end and it disables current limiting!  Also it works by having too much sense resistance, and 'potting down' the voltage seen by the 723's internal limiting transistor which results in excessive voltage drop across the sense resistance at high currents.  A better approach is to provide a stabilised voltage that can sink current  1V below the output voltage, and connect one end of the current limit pot to the output and the other via a trim resistor to set min. output current, to the 1V lower rail.  The voltage from the pot wiper then adds to the voltage across the sense resistor(s), resulting in limiting at a lower current.

[p.larner]

Do you fancy modding the schematic for that ?.

[Zero999]

Do you fancy modding the schematic for that ?.

The last time I posted a schematic for you. You didn't acknowledge it. No response, just ghosted me.

How about posting your own schematic this time and we'll tell you whether you're on the right track or not?

[p.larner]

Just testing my 10 turn pots,my 500 ohm one is open,can i use a 1k instead?.

[Ian.M]

There's no point in using a multiturn pot to set a '723's current limit using the internal limiting transistor.  The limiter has too soft a knee,  and is too inaccurate and temperature sensitive.  Use a regular pot as a multiturn will just be annoying!

N.B. to avoid further degrading the current limit accuracy the impedance seen by the CURRENT SENSE pin (emitter of the limiting transistor) should be fairly low.  500 ohms from a 1K pot mid travel probably isn't enough to significantly decrease the slope to the right of the knee but will probably make the knee itself softer.

[p.larner]

will a normal 1k pot work then?,its all i have to hand.

[Ian.M]

will a normal 1k pot work then?,its all i have to hand.

That depend on your current circuit, either 'suck it and see' and report results here or take the advice you have been given:

How about posting your own schematic this time and we'll tell you whether you're on the right track or not?

if you want our further input!

[MathWizard]

What about junction capacitance's, when would that start to matter for being driven from the LM723 ? I know good emitter follower's have little output capacitance, but I've never really worked around the loop of anything besides the simplest op-amp +driver circuit.

[p.larner]

i am thinking of trying a 1k pot with a resistor in series with one leg so it is not 0 ohm at one extreem of the travel so the foldback still works,would that be better?.

[Ian.M]

For some small value of better limited to 'does not immediately go bang!' . . . .

Foldback limiting requires a different circuit to ordinary CC limiting, too complex to discus without a good schematic.
 
For CC limiting, the end with the fixed resistor would be the pot's max current setting and there would be 0.65V (approx) across the fixed resistor at that setting.  Make it 1K and the min. current at the other end of the pot would only be half the max.  Make it 250 ohms and the min.  current would be 1/5 the max.  Its impossible to make a current limit control that way that can go down to zero, or even close to zero.  Another major problem is the voltage drop across the current sense resistor.  To get 0.65V between CURRENT LIMIT and CURRENT SENSE to trigger the limiting at max. current you need (Rpot/Rfix +1)*0.65V across the current sense resistor.  Therefore with a 250 ohm fixed resistor in series with the pot there would be 3.25V drop across it at max. current, and for your proposed 16A PSU, over 50W dissipation in the sense resistor!  Its hard enough to manage the sense resistor dissipation for a high current '723 PSU with only 0.65V across the sense resistor, so 'potting down' the current sense voltage isn't practical.  Amplifying the current sense voltage with variable gain to permit a far lower burden voltage for the sense resistor is a theoretical solution but there are many ways to get it wrong causing instability and possibly high power oscillation!
 
I have already suggested a way to avoid these pitfalls . . .

[Zero999]

Looking at the circuit again, it appears as though adjusting R11 too far will completely disable the over-current protection. Another resistor should be connected in series to prevent this from happening.