48v slow regulator question...

[TomWaterman]

So sorry chaps didn't mean to cut and run after starting this topic.
Can't believe its been almost 4weeks, ugh time flies...

Matt, sorry for not being more specific - this is for a mic pre/eq that I'm kind of designing from scratch based upon an old british design. I was going to do a powersupply for it with a separate +52VDC rail for phantom. +52VDC so each pre board with local slow regs would still provide a nice clean +48V after min. reg drop of 3V.

I was going to use a slow-reg based upon the LM317 datasheets, but subbing in the TL783C as you rightly suggested. Farnell seem to have them in stock again now BTW.

This PDF shows my original slow reg plus a simple passive RC one nabbed from JLM...

After some thought I've decided to scrap the PSU as I don't want to waste time on it when JLM make a nice one available for DIY'ers at a good price. So my design criteria has changed a bit. There will be a max of two channels in one box, a single JLM Powerstation providing individual +-V for each audio channel and a global +48V feed for both channels phantom (the 48V will also supply a separate TL783 for relays/leds and maybe two JLM DIs.

After reading yours, brads and mediatechs thoughts, the problem that any active slow regs on the preamp will drop a couple of volts means that I may end up with +45V. However I undertand that the DIN phantom spec is +-4V @ 48V so 44-52V is acceptable(?).

For sake of overall discussion here is Brad's TL783 linear ramp circuit that you all may like to see.

bcarso's 48V Reg PDF

I have enough board space left for a TO-220 reg or a darlington/JFET one...but am unsure of what to do.

Maybe the standard passive RC ramp shown as 'B' in the first PDF is adequate. I understand that the circuit decays and charges to 37/63% from 48v in a time of t=RC in sec, and approx 5RC for 1%, at least that what H+H says.

I also read somewhere about adding a 'bleeder' so that it will discharge quicker...any thoughts on that?

Really do appreciate all of your collective genius..

Thanks Tom

 

[Matt Syson]

Hi
As you will have a clean supply from the TL regulator you only need a 'slow switch' so you can use a darlington or fet as indicated prevously. You would need to experiment a bit to ensure you don't loose excessive volts although using a PNP type and drive it from ground would get it to saturate easily losing only 0.5 volt or less.
Yes spec for phantom is quite wide but there are a handful of old mics that are fussy and need about 50 volts.
PM me if you want specifics.
Matt s

 

[sodderboy]

I have not seen it mentioned, but is there a transformer at the front end? I would not build a PSU without one. A diode can short and wall neutral can get directly to a mic, or an innocent gofer, and there is no fuse? :shock:
There are some simple circuits searchable with a 50 vac transformer, rectification, filtering, and a zener. And a fuse! I can supply if interested. KISS for me.
Or I get two regulated 24VDC wallwarts from MCM or All Electronics and connect them in series. Even at 250mA, enough for 8 of the hoggiest mics, TLM 170. They draw about 30mA, where most others draw less than 10.
Safety, Mon!

 

[Gus]

The real answer is to place the circuit in the microphone.

 

[TomWaterman]

He, good idea Gus ? have you tried it ? are there any commercially available mics that do that?

Sodderboy ? the circuits will be fed a pre-regulated DC source from a powersupply so they do not require a transformer?

So Matt, I'm trying to get a handle on this....miserably I might add!

SonsofThunder?s schematic:

Is it essentially an emitter follower switch using a NPN darlington with an RC network of R1 (R2) and C1 to create a ramp time with a time constant (t=RC) of 1sec. Approx 5RC or 5-ish secs to ramp/drain correct? R2 acts as a drain for C1 in the off position and also turns of Q1?

If I understand correctly then D1 acts as safety incase of problems with the load and emitter resistor R3 is selected so that almost all of the +48v feed passes to the load?(so I take that to mean that this is where the slow switch drops a few volts to the phantom line, hence the need for a PNP circuit to minimize voltage drop?).

In the 'on' position R1 is setting a small base current of (48v-1.4)/100000 = 0.46mA and this base current creates a larger current of Ic through the collector and emitter. With the NPN darlingtons we get a large current gain, maybe Hfe of 10000? What current do we get out and what is the effect of R3?

To make the circuit work as PNP can I substitute a PNP emitter follower for the NPN one as below?

[Edit] just realised I was being a silly boy and my schem didnt make any sense at all![/edit]

If someone could explain how to calculate the circuit/component values I would be really stoked.
Also in terms of PNP darlingtons, I imagine a TO-220 package would be ideal with a heatsink?..BD680?

Thanks for the help
Cheers Tom

 

[TomWaterman]

Hmmm,

Is this more like it? A common emitter switch with a PNP transistor...I dunno I could use some help...spent ages trying to read up on this, trouble is google returns a bunch of contradictory pages and reading it gets all the more confusing.

If this is right, where should I put the cap to control ramp up and down, after R1 before the switch? Also PNP darlingtons like the BD680 seem to have a high Vce sat, like 2.5v, does that mean I could end up with 45.5v as my phantom supply.

I understand that the darlington makes the switch slower but can't I just use a normal PNP BJT like BD140?

Cheers Tom

 

[Matt Syson]

Hi
Tom's circuit above should be pretty close. The ramp cap goes onto the base of transistor to ground. Max current through a 6K8 shorted would be slightly over 7mA so you do not need a heatsink. I kind of had a compound configuration rather than a darlington with PNP as the 'main' switch, this would reduce drops. As you only want 15mA with the thing shorted saturation would only be very small voltage (0.3V at worst).As shown Tom's circuit would be OK. Using a big darlington would make it virtually bomb proof although it would be happy with smaller varieties.
The 2.5 volts quoted for saturation is with it passing 5 AMPS or so!
Matt S

 

[TomWaterman]

Thanks Matt,

Just to check the ramp cap goes off the base of the PNP before R1 or after R1, before the switch?

I realised after I posted that the current is small enough to not require a heatsink..cool. Thanks for the info on Vcesat...I think I will try with a BD680 darlington to see what drop I get, but if thats going to be bomb proof thats just great!

When you say compound, you mean a smaller transistor switching a larger PNP that switches on the 48v?...do you think there are other advantages with this arrangement as opposed to using the PNP darlington (apart from perhaps less drop)?

Also I have one final question - in the circuit I posted how would one modify it to enable differing ramp up/down times as in Sonsofthunders original NPN scheme?

Cheers Tom
:guinness:

 

[Matt Syson]

Hi
The ramp cap goes to the base of the transistor (+) and the - goes to ground.
As the darlington will have a gain of at least 1000 (probably a lot more) and you only want 15mA max in theory the base current should be 13mA divided by 1000. Since we want saturation perhaps double or even 10 times the base current, this is still pretty small.
Increase the 100 K resistor on the base to change time or increase the capacitance.
You can change the on / off timing by adding a diode in parallel with the 100K but add a 1K or so in series with the diode for current limiting. Depending which way you fit the diode it will speed up the switch on, or the switch off.
I must say that I don't think the project is really wothwhile (except for a learning excercise) as mic's are used to just being plugged in and the maximum current draw is only 7mA per leg anyway.
Try looking for circuits of the 'internals' of the mics that use phantom and see what goes on in the mics.
Have fun
Matt Syson

 

[TomWaterman]

Thanks again Matt,

I'm clear that the cap goes to ground I just meant base side of R1 or after R1 to ground...

I doubt I'll bother with the diodes but I will recalculate the values for a BD680 at avg. Hfe...

Cheers Tom

 

[Matt Syson]

Hi
It should join the 100K and the base of the transistor.
Matt Syson

 

[TomWaterman]

Cheers Matt - thats what I thought.

Tom

 

[Matt Syson]

Hi
When you have built it, let me know how you got on as I am curious. Unfortunately I have too much on to play with this myself.
Matt Syson