Dropping a few tenths of a volt problem

The 3.9V is a problem. Maybe add a transistor buffer in front? Schottky or transistor at A2? ...Jim Thompson

Show your full schematic, including the "zener tripping" you mentioned.

I'm off to my every few years cardiology exam, so I'm won't reply until around noon time. ...Jim Thompson

I couldn't open your netlist schematic, but the schmitt thing is probably fixable. Can you post an image?

Will you get back to us with more requests for ideas when you discover that the cells can pump out useful current down to a voltage of 0.9V or so?

On a sunny day (Fri, 21 Mar 2014 14:42:18 GMT) it happened Jan Panteltje wrote in :

Actually more seriously you system has severe limitations. Do the math: if you 'beeper' works at 4 kHz, it will not likely be able to make less that one period. that already limits your output frequency, and you won't be able to detect or notice a nearby nuclear blast. While just toggling output (of some transducer, dynamic, piezo, whatever), imposes no real limit on the counts per minute, 1000 counts gives 1000 clicks, and 10,000 counts from the nuke next door gives 10,000 Hz, Trust me (tm) I have tried both methods.

Even Radio Shack carries the TLC555 (at $2 !!!) but it has really wide operating range of 2-15 Volts VDD,so that and a transistor gets you a compact pulse stretcher/extender at really low currents depending on your load. impedance.

There's really nothing more to it than this zener which just limits the amplitude of the higher input pulses to the first gate . Again, for schematic purposes only:

Version 4 SHEET 1 1168 680 WIRE 480 -16 -416 -16 WIRE 480 48 480 -16 WIRE 48 128 -192 128 WIRE 176 128 48 128 WIRE 176 144 176 128 WIRE 192 144 176 144 WIRE -416 160 -416 -16 WIRE 352 160 256 160 WIRE 480 160 480 128 WIRE 480 160 416 160 WIRE 640 160 480 160 WIRE 48 176 48 128 WIRE 800 176 704 176 WIRE 800 192 800 176 WIRE -192 208 -192 128 WIRE 480 208 480 160 WIRE -416 320 -416 240 WIRE -192 320 -192 288 WIRE -192 320 -416 320 WIRE 48 320 48 240 WIRE 48 320 -192 320 WIRE 160 320 48 320 WIRE 192 320 192 176 WIRE 192 320 160 320 WIRE 480 320 480 272 WIRE 480 320 192 320 WIRE 640 320 640 192 WIRE 640 320 480 320 WIRE 800 320 800 272 WIRE 800 320 640 320 WIRE 160 336 160 320 FLAG 160 336 0 SYMBOL Digital\\schmitt 192 80 R0 SYMATTR InstName A1 SYMBOL voltage -192 192 R0 WINDOW 123 0 0 Left 2 WINDOW 39 24 124 Left 2 SYMATTR SpiceLine Rser=600 SYMATTR InstName V1 SYMATTR Value PULSE(0 3.9 0 10n 10n 5m 300m 500) SYMBOL diode 416 144 R90 WINDOW 0 0 32 VBottom 2 WINDOW 3 32 32 VTop 2 SYMATTR InstName D1 SYMATTR Value 1N4148 SYMBOL res 464 32 R0 SYMATTR InstName R1 SYMATTR Value 180k SYMBOL cap 464 208 R0 SYMATTR InstName C1 SYMATTR Value 80n SYMBOL Digital\\schmitt 640 96 R0 SYMATTR InstName A2 SYMBOL res 784 176 R0 SYMATTR InstName R_Load SYMATTR Value 400k SYMBOL voltage -416 144 R0 WINDOW 123 0 0 Left 2 WINDOW 39 24 124 Left 2 SYMATTR SpiceLine Rser=500 SYMATTR InstName V2 SYMATTR Value 4.5 SYMBOL zener 64 240 R180 WINDOW 0 24 64 Left 2 WINDOW 3 24 0 Left 2 SYMATTR InstName D2 SYMATTR Value 3.9v Zener SYMATTR Description Diode SYMATTR Type diode TEXT 240 56 Left 2 !.tran 0 1.4s 0 .1m

I'll see what I can do, but don't hold your breath!

In the mean time, fig. 12 towards the end of this app note is the actual circuit I'm following:

On a sunny day (Fri, 21 Mar 2014 16:32:58 +0100) it happened Cursitor Doom wrote in :

but have you ever looked at his:

Normally you would use and extra 220pF and 5k6 to ground to reduce possible LF pickup. In that case you can slightly bias the NPN base to get more sensitivity, in this setup the sensitivty is Ferry Goot.

• logic | [ ] |--------------- c GM toob ----- b NPN e | ///

• logic | [ ] |--------------- 220p c GM toob ------||------------ b NPN | | e [ ] 100k [ ] 5k6 | | | /// /// ///

Trust me (tm) I have tried both ways.

Because I can't see how it could be any better or simpler than fig. 12 here, which is what I'm using:

Well, we all knew it would happen one day. ;-)

Your schematic that you posted as a text was compatible with LTspice = came through VERY well.

That's small amount of text you copied starting with 'version 4' at the top.

just take schematic.asc open it with any text editor and scoop it up and transfer it to the bottom of your posting. Works great and it's right here, and peole cn modify approiately and post it back for you.

Just curious, doesn't seem like a heavy load for the pen cell. Do you have room in there for a SMPS? that way, you could use one battery, not more than that and use all the space to put in a 'regulated' supply, so your electronics always has an ideal voltage. You could drag that battery down below 0.9 Vdc and still be operating.

On a sunny day (Fri, 21 Mar 2014 15:44:24 GMT) it happened Jan Panteltje wrote in :

| e

Forgot the 100 kilos

• logic

On a sunny day (Fri, 21 Mar 2014 16:45:25 +0100) it happened Cursitor Doom wrote in :

GM toob pulse stretcher:

• logic | [ ] 3k3 R2 |--------------- c | GM toob --------- b NPN === | e | Cx [ ] | /// | 100k /// ///

When GM toob pulses tyen transistor Ic = beta x GM toob current discharghes Cx. (resisister also com-sumes some current) Cx will then slowly charge again via R3, time constant R2 Cx do the math... Trigger counter on negative edge.

Hello?

Now let me see.... You say I'm being overly complicated, but then you suggest I fix it in software??? There's a contradiction here somewhere!

Actually it sounds *really* good in practice, so it's just as well I didn't do the math. :-)

My GC uses a Philips toob.

John said he had problems with it. I've no idea why.

The only load here is a sub-min peizo buzzer so naturally it won't be heavy. BTW, your suggestion takes up way more space than my original 'solution' of changing the 3 AAA cells to a pp3 plus variable voltage reg. IC. Space is at a real premium here!