Hi, I am working on a project that requires some hardware design but I am very much an amateur when it comes to circuits (software eng by trade), and I'm getting some results that I can't wrap my head around. Here is the circuit I will be referencing in my question:





This circuit worked for me when I prototyped it on a breadboard, but when I got it printed to a PCB it's only partially working. The idea is that the reed switch (SW1) opening or closing creates either a VCC or GND output on terminal 3 of Q1, which gets inverted by Q2 - the signals create a pulse respectively across C2 and C3 that, when the pulse is negative, turn on Q3(Q4) and charge an RC circuit, which will after a short time discharge, momentarily turning on Q5(6) to give a momentary pulse to HCS301 inputs 1(2). The circuit that feeds into HCS301 input 1 works fine (when SW1 closes), but input 2 isn't getting a signal when SW1 opens (which follows the lower circuit path).

After some investigation, I found that the collector of Q1 (terminal 3) was 3V when turned "on", but was 2.5 V when "off", as opposed to Q2, whose collector was 3V when "on", but 1.4 when "off". It seems to me that the voltage drop of Q2 is enough to create a negative pulse across C3 to do what I want it to do, but the change of voltage from Q1 doesn't create a large enough pulse across C2. I'm not seeing why these two transistors are acting differently, they both have enough voltage drop across the emitter/base to switch 'all the way' from what I can see, and I'm confused as to why I'm seeing such a high voltage on the collector of Q1 when it is turned "off" (eg when I have 3V on its base and 3V on the emitter).

I should also note that I am currently thinking my high values of resistances might be part of the problem, but I need this circuit to run for a long time on a CR2032 battery, and I don't know how else to get a very low quiescent current. Any thoughts would be very appreciated, I am very new to circuit design and have no formal training on the matter. Thank you!

 

Hi, I am working on a project that requires some hardware design but I am very much an amateur when it comes to circuits (software eng by trade), and I'm getting some results that I can't wrap my head around. Here is the circuit I will be referencing in my question:


View attachment 287505


This circuit worked for me when I prototyped it on a breadboard, but when I got it printed to a PCB it's only partially working. The idea is that the reed switch (SW1) opening or closing creates either a VCC or GND output on terminal 3 of Q1, which gets inverted by Q2 - the signals create a pulse respectively across C2 and C3 that, when the pulse is negative, turn on Q3(Q4) and charge an RC circuit, which will after a short time discharge, momentarily turning on Q5(6) to give a momentary pulse to HCS301 inputs 1(2). The circuit that feeds into HCS301 input 1 works fine (when SW1 closes), but input 2 isn't getting a signal when SW1 opens (which follows the lower circuit path).

After some investigation, I found that the collector of Q1 (terminal 3) was 3V when turned "on", but was 2.5 V when "off", as opposed to Q2, whose collector was 3V when "on", but 1.4 when "off". It seems to me that the voltage drop of Q2 is enough to create a negative pulse across C3 to do what I want it to do, but the change of voltage from Q1 doesn't create a large enough pulse across C2. I'm not seeing why these two transistors are acting differently, they both have enough voltage drop across the emitter/base to switch 'all the way' from what I can see, and I'm confused as to why I'm seeing such a high voltage on the collector of Q1 when it is turned "off" (eg when I have 3V on its base and 3V on the emitter).

I should also note that I am currently thinking my high values of resistances might be part of the problem, but I need this circuit to run for a long time on a CR2032 battery, and I don't know how else to get a very low quiescent current. Any thoughts would be very appreciated, I am very new to circuit design and have no formal training on the matter. Thank you!

If R4 is 10M Ohms then you have a voltage divider between that, the 10k R5 and the base/emitter junction of Q2, so you're never going to get a low voltage on it's collector! Q2's base could do with a pull up to V+ as well!

 

What is "a long time"?

What is your current budget?

How much static current are all of these transistor inverters currently consuming?

Have you considered using an ultra-low power hex inverter such as the AUP family from Nexperia?

 

Let’s start with the inverter you are having trouble with. In addition to it being way more complicated than it needs to be, and the resistor values being all wrong, it is totally redundant. If you want an inverted signal from the reed switch, just swap the switch and the resistor.

 

Looks like he wants a logic pulse on both open and close of the switch.

 

What is "a long time"?

What is your current budget?

How much static current are all of these transistor inverters currently consuming?

Have you considered using an ultra-low power hex inverter such as the AUP family from Nexperia?

Hey, first of all, thanks for the reply. To respond to your questions/comments:

1. What is "a long time": as long as possible; worst acceptable case I would like to be > 6 months.
2. Current budget: Right now I can make the PCB for about a dollar, and about $10 if it's pre-assembled (which for me is a necessity as I don't have the experience / equipment to solder 0402s). I probably wouldn't want to go over $30 per unit (pre-assembled).
3. I don't know if I can give you an awesome answer on this, it's less current than my radioshack handheld multimeter can measure, so less than 0.1mA -- it lasted a few months on a CR2032 but I did a lot of RF transmitting while testing so I don't think its worth doing the math to see the current that relates to, because the testing environment wouldn't emulate what its like in production.
4. Thanks for this tip, after a cursory look it looks promising -- I steered away from digital ICs when my V1 of this solution used an NE555 and the battery lasted only a few hours, but this looks helpful, and could circumvent having to rely on transistors as inverters here. Thank you!

 

If R4 is 10M Ohms then you have a voltage divider between that, the 10k R5 and the base/emitter junction of Q2, so you're never going to get a low voltage on it's collector! Q2's base could do with a pull up to V+ as well!

Ah thank you for this, it's good to know what was causing it, and thanks for the tip for the pull-up!

 

Let’s start with the inverter you are having trouble with. In addition to it being way more complicated than it needs to be, and the resistor values being all wrong, it is totally redundant. If you want an inverted signal from the reed switch, just swap the switch and the resistor.

Thanks for the reply! WBahn is correct about what I'm going for, so unfortunately just getting an inverted signal won't do it. If you don't mind expanding a little bit on the points about it being over-complicated or having wrong resistor values, I would be appreciative

 

Here is a quick sketch of a really simple and clean way to do what I think you want.

This circuit creates a single clean pulse when (1) the switch opens, and (2) when the switch closes.
The Schmitt trigger inputs makes for bulletproof debounce. and pulse forming.

It's really only 8 parts and should cost under $0.25

C2 and C3 set the length of the output pulse.

 

points about it being over-complicated or having wrong resistor values, I would be appreciative

My version using a CD40106

 

I see a lot of 10MΩ circuits driving much lower resistance loads. That is not going to work. You have it backwards.

You want the driver circuit output resistance to be 10 times lower than the input resistance of the receiving circuit.

 

Here is a quick sketch of a really simple and clean way to do what I think you want.

This circuit creates a single clean pulse when (1) the switch opens, and (2) when the switch closes.
The Schmitt trigger inputs makes for bulletproof debounce. and pulse forming.

It's really only 8 parts and should cost under $0.25

C2 and C3 set the length of the output pulse.

This is awesome, thanks for drawing that out for me! Never heard of a Schmitt trigger before, but it looks like the perfect thing to abstract some of the more complicated things I'm trying to do! Thank you!!

 

My version using a CD40106
View attachment 287532

Want to echo the sentiment of my reply to Sensacell, thank you for drawing out a better option, this is super helpful!

 

I see a lot of 10MΩ circuits driving much lower resistance loads. That is not going to work. You have it backwards.

You want the driver circuit output resistance to be 10 times lower than the input resistance of the receiving circuit.

Thanks for the insight, much appreciated. To make sure I'm pulling out the correct info from your post - this would be referencing relatively small resistances going into the base of transistors with respect to upstream resistors of much higher value are connected to vcc and gnd, which mean that the voltages downstream of these low-valued resistors will influence things upstream in ways I'm not accounting for? Admittedly I was creating a series of components and evaluating them as a black box abstraction as I put them together to make the full design. The way that this type of evaluation would be (closer to being) correct is if the input resistance to the "next" downstream sub-circuit is much higher than the resistances connected to vcc and gnd in the sub-circuit I'm evaluating... if that makes sense?

 

Yes.
You cannot have a weak motor pulling a heavy load. It is better to have a strong motor pulling a weaker load.
The rule of thumb is you want the motor to be 10 times bigger than the load so that there is little strain on the motor (circuit-wise speaking).

 

Suppose you have an input voltage connected through a 10M resistor. Now connect the other side of the 10M resistor to ground through a 10K resistor. What voltage must you put at the input to get 1V across the 10K resistor?

 

Suppose you have an input voltage connected through a 10M resistor. Now connect the other side of the 10M resistor to ground through a 10K resistor. What voltage must you put at the input to get 1V across the 10K resistor?

hey Bob thanks for the reply! I get voltage dividers, and that your scenario requires a 1001V input -- I think what was tripping me up was that transistors have resistance across the PN junctions in the same order of magnitude of the resistors I was using in my circuit

 

Want to echo the sentiment of my reply to Sensacell, thank you for drawing out a better option, this is super helpful!

Thank you.

This we call M2M logic, short for Mickey Mouse Logic.

With a 4093 and a few R's and C's - you can solve a LOT of problems, cheap and simple. - enjoy.

 

hey Bob thanks for the reply! I get voltage dividers, and that your scenario requires a 1001V input -- I think what was tripping me up was that transistors have resistance across the PN junctions in the same order of magnitude of the resistors I was using in my circuit

PN junctions act like diodes, not resistors.

 

PN junctions act like diodes, not resistors.

right, i think i was trying to conceptualize the nonlinear relationship between the voltage delta on the base-emitter junction with respect to the current that would flow through that junction, and view that as a resistance in relation to my upstream resistors, especially since it didn't seem like I was able to get Q2 in the saturation region due to my bad upstream resistor values. it sounds like that isn't the right way to think about it, I probably just need to revisit transistor theory