Negative 1V5 at 100mA is so expensive?

C

cupoftea

Advanced Member level 6
Joined
Jun 13, 2021
Messages
3,041
Helped
62
Reputation
124
Reaction score
139
Trophy points
63
Activity points
15,919
Hi,
Just looking for the cheapest, fairly well regulated NEG1V5 supply from a 10V source. Load from 0 to 100mA.
The attached is the cheapest i can get so far.
Any cheaper ideas?
Charge pumps are not regulated (not the cheapos) and so cant really do it even though cheap.

Its just a NEG1V5 rail which acts as a turn off voltage for NPN's....due to the sheer expanse of the circuit, and wiring etc, the grounds may not be the same everywhere, so we want a slight negativised turn-off voltage to assure we deffo turn off when we want to.

LTspice and PNG attached.

BTW we would use the cheaper UCC28C43 rather than the pin4pin LT1243.
 

Attachments

  • CUK NEG 1V5.png
    79.1 KB · Views: 33
  • cuk.zip
    2.2 KB · Views: 15
Last edited:

Sort by date Sort by votes
Beware overthinking such s crude pull-to supply. You have about 5-6V to play in (BVebo less margin). As this is only a "peaking" supply with a fixed-ish charge slug per cycle an aux winding charge pump with a two-diode-series shunt "regulator" could be the ticket.
 
Upvote 0 Downvote
Thanks , right I see, the 555 can do it...

I suppose you just do it in on/off control to get the rough reg neg 1v5?

Our customer has TC1044S as a stock part, so we would also like to think about using that, but we are studying the datasheet to see if we can on/off control
it to give a reg'd 1V5.
The datasheet only shows open loop things, and the output Z is so high and has such wide tol that you cant be sure of any voltage output
particularly.

TC1044S
 

Upvote 0 Downvote
If you regulate the output with a transistor to cap instead of a diode on the 555, the gain of the error amp can reduce the 300 Ohms Zout plotted by TI by the d.c. fraction of the GBW. That will give you decent load regulation. But it may start as underpowered.
 
Upvote 0 Downvote
A level shift ckt that eventually pulls down on pin 5 will regulate a -1v5.

using a xtor Vbe as the ref.

for 100mA out you can buffer the 555 o/p with an emitter follower.
 
Upvote 0 Downvote
Easy peasy said:
A level shift ckt that eventually pulls down on pin 5 will regulate a -1v5.

using a xtor Vbe as the ref.

for 100mA out you can buffer the 555 o/p with an emitter follower.
Click to expand...
An emitter follower can only source + current. The load draws -ve current from cap @ V-.
 
Upvote 0 Downvote
Hi,
Thanks, ill go into that 555 with pull down on pin5 next....at the mo you gave me ideas and the reg'd cuk is a little lesser in cost...
Not a nobel prize winner attached , but a little cheaper.
 

Attachments

  • CUK NEG 1V5 _vbe reg.png
    50.7 KB · Views: 21
  • cuk vbe reg.zip
    1.6 KB · Views: 10

Upvote 0 Downvote
Hi,

I´m not sure what accuracy and noise you need..

But a half bridge driven with 15% duty cycle will make an average of 1.5V from 10V.
LC filtered may give smooth DC.
And this may be way more efficient than the linear solution (which dissipates 85% of the input power, or about 6 times the output power)
Maybe in your circuit there somewhere is a microcontroller with a spare PWM output... for free

Klaus
 
Upvote 0 Downvote
Thanks, yes , and there are some ti.com things that do something like that with haf bridge, so ill dig deeper there.
Mean time, had to drop the NPN reg as it was taking COMP more neg than -0.3v

So attached brings back the opto

Accuracy is like Tony Stewart kindly allures, its just literally a neg rail to assure turning off NPNs.......ie, to solve the wandering ground potentials between boards wired to each other
 

Attachments

  • cuk vbe reg OPTO1.zip
    1.5 KB · Views: 16
  • CUK WITH VBE OPTO REG.png
    48 KB · Views: 13

Upvote 0 Downvote
Easy peasy said:
respectfully - we are buffering the 555 with the emitter follower - charge pump to -ve volts - I'm not sure you are following the thread . . .
Click to expand...
respectfuly - I'm not sure if you understood what I said... please explain your circuit. This is what I thought you said.
 
Upvote 0 Downvote
Seems to me that any requirement for accuracy
is unstated. Maybe unexplored. Crude pull-to,
I think doesn't care much. Why not -2V? Is -1V
too short?
 
Upvote 0 Downvote
Aah - OK - you need the pnp underneath it, bases connected, emitters connected, npn coll to Vcc, pnp coll to 0v

this gives proper buffering - whatever the load connection the emitter follows it ( minus one Vbe ).

 
Last edited:
Upvote 0 Downvote
That buffer loses 2Vbe of pump amplitude, maybe that's the idea, to drop voltage along the rectifier chain? Still linear-lossy.

What's your inductor like? Open bobbin, how many spiral turns of PCB trace inductor would it take to grab a whiff and FWB? Solution values might include not-needing clean supplies or other resources to get it on.
 
Upvote 0 Downvote
...thanks, so now to concentrate on the reg bit where the 555 control pin gets pulled down so as to reg the output to neg 1v5 (or therabouts). Presumably this is adjusting the duty of the 555....ill work on it soon...
 

Upvote 0 Downvote
If a charge pump can push-pull +/-150mA with < 1V drop it may be enough to sink 100 mA into 12+ Ohms at -1.2V (1.44W).
Then it just needs to be regulated.

If I were doing this in 70/80's I'd be using cheap discrete and small-scale CMOS I might have used a CMOS Schmitt trigger CD40106 for 95% duty and AC coupled totem pole discretes for low pulse impedance.
Easy peasy said:
respectfully - we are buffering the 555 with the emitter follower - charge pump to -ve volts - I'm not sure you are following the thread . . .
Click to expand...
The LM555 has a Totem pole output rated for 200 mA but plotted to 1V/100 mA = 10 Ohm driver.

An emitter follower needs to source as much current as the 100 mA load with Re=12 ohms about the same as the load = 1/2V/12 Ohms thus <= 50% efficient.


But I agree with @FvM that a buck-boost reg is better.

Here with < 4% d.f. unregulated https://tinyurl.com/26e38odg
 
Last edited:
Upvote 0 Downvote
You must log in or register to reply here.

Similar threads

Menu
Log in
Register
Cookies are required to use this site. You must accept them to continue using the site. Learn more…