Hi CDRIVE,

Thanks a lot for your circuit.

I like open source forums and this forum too.

Best regards.

jm

 

I can't seem to resist the challenge of discrete circuits, so here's yet another more complex design that guarantees saturation. The theory of operation is described in the attachment. This circuit enables use of much smaller values of C1. Increasing C1 to 500uF will produce a 5 second delay.

Longer delays:
Because of the feedback nature of this circuit there is still quite a bit of headroom left for increasing the values of R1 and R2. I spiced this circuit to a to over 15 seconds with C1 = 500uF R1= 200k and R2 = 100k, but this assumes a perfect Cap with no leakage, which is unrealistic for an Electrolytic. It also assumes Q1 beta >=215 & Q2 beta >=277. For certain, the circuit is operating at extreme limits with R1 & R2 values this high.

Most transistor models do not include Vbe breakdown. Q3 is vulnerable, and will also screw up your timing, as C1 will charge up to around 6V almost immediately.

 

Most transistor models do not include Vbe breakdown. Q3 is vulnerable, and will also screw up your timing, as C1 will charge up to around 6V almost immediately.

Ouch! Ron, you're absolutely correct. What the bleep was I thinking? Much more to say about this (including the Spice aspects) when I get back from the pub. To maintain thread integrity I will leave that schematic as is with an error message to read your post. This is quite embarrassing, to say the least!

 

Ouch! Ron, you're absolutely correct. What the bleep was I thinking? Much more to say about this (including the Spice aspects) when I get back from the pub. To maintain thread integrity I will leave that schematic as is with an error message to read your post. This is quite embarrassing, to say the least!

Hey, it happens to the best of us. Me, too.

 

Good info in this thread.

I'm thinking about a POR circuit, (power-on reset), but that wouldn't source/sink enough current for the relay.

 

Most transistor models do not include Vbe breakdown. Q3 is vulnerable, and will also screw up your timing, as C1 will charge up to around 6V almost immediately.

I spiced a modified version using two NPN's which is attached, but then I realized that all that was needed was a change to the value of R7 to a value that would keep Q3 off while not exceeding the max Vbe rating. At worst case Vbe doesn't exceed +1.92V with R7 = 6.8K. Thanks again to Ron for pointing out my error.

Attached, are two circuits that perform a function that a 555 can do much better and a Picaxe 08M can do even better.

 

Most transistor models do not include Vbe breakdown.

I'm using Tina Classic Design Suit which, as you say, doesn't provide for those parameters. Tina calls it Stress Analysis and if I was willing to fork out an additional $800.00, or so, I could have it in their Pro version. Too rich for my meager budget though.

 

I'm using Tina Classic Design Suit which, as you say, doesn't provide for those parameters. Tina calls it Stress Analysis and if I was willing to fork out an additional $800.00, or so, I could have it in their Pro version. Too rich for my meager budget though.

Yeah, for 800 bux, I'd remain an amateur! I think I have kluged in an ideal zener in parallel with the B-E occasionally in LTspice. Something like this:

.model My6VZener D(Ron=5 Roff=1e12 Vfwd=6)

I didn't simulate this, but it should work. It doesn't add capacitance, or any conduction in the other direction, as would happen if you used a real zener mode. The only problem is, if you use a diode symbol, the schematic looks all wrong.

 

.model My6VZener D(Ron=5 Roff=1e12 Vfwd=6)

Ron, this is clever, but admittedly, I need a primer on PCode. I can examine existing code, dissect it and edit some of it, but I've not written any from scratch. Though I understand your code explicitly, I wouldn't know how to properly integrate it into an existing transistor model. Do you, or anyone following this thread, have a link that explains the basic PCode syntax structure for better understanding and creating my own models?

BTW, Tina Classic does allow me to create a macro using discrete components or PCode, or both, and then assign it to an existing model package, or create my own. Fortunately, creating a macro is very easy and it vastly increases the power of Tina Classic.

Thanks,
Chris

 

I don't know how to integrate it into a transistor model.
I could make a subcircuit for a particular transistor, which would include the models for the transistor and the "zener". It probably would be inferior to a proper transistor model that had parameters included for B-E breakdown.

 

Come to think of it, it's not a Macro option that my Tina provides me, (it does have Macro models though) it's a 'Subcircuit' option too.

BTW, since both Zeners and Schottkys represent a Diode junction when rev biased, I believe an additional Diode (rev biased) would also be required (in series) to totally devorce the Zener or Schottky when the base voltage is in a forward biased state..... I just spiced this and yes, the Diode is a must. I'm not sure if 5V schottkys exist, as I don't have any that low in my library, so I just changed a 20V model to 5V.

 

Come to think of it, it's not a Macro option that my Tina provides me, (it does have Macro models though) it's a 'Subcircuit' option too.

BTW, since both Zeners and Schottkys represent a Diode junction when rev biased, I believe an additional Diode (rev biased) would also be required (in series) to totally devorce the Zener or Schottky when the base voltage is in a forward biased state..... I just spiced this and yes, the Diode is a must. I'm not sure if 5V schottkys exist, as I don't have any that low in my library, so I just changed a 20V model to 5V.

Yeah, but the model I proposed is a forward biased diode, with a fwd voltage of 6V (instead of the usual 0.6V) in this case, and virtually zero reverse conductivity (10^12Ω). So no series diode is required.
If you use a zener model, then you will require a series diode, as you mentioned. However, real zener models come with some undesirable baggage (e.g., capacitance), which you don't want to add in parallel with the Vbe junction.

 

I made a 2N3904 subckt with Veb=6V. Below is the results of a simulation of an astable multivibrator, with and without BE breakdown modelling.
Here is the subcircuit, for use in LTspice.

*
.subckt 2N3904 1 2 3
Q1 1 2 3 0 2N3904
D1 3 2 IdealZener6V
.model D D
.lib C:\Program Files\LTC\SwCADIII\lib\cmp\standard.dio
.model NPN NPN
.model PNP PNP
.lib C:\Program Files\LTC\SwCADIII\lib\cmp\standard.bjt
.model IdealZener6V D (Ron=10 Roff=1e12 Vfwd=6)
.ends

Copy and paste in Notepad, then save in SUB folder as 2N3904.sub.
To use it, you have to place an NPN, select the 2N3904 model, then ctrl-right click on the transistor and change the prefix to X. Add a .lib 2N3904.sub directive to the schematic.

If you run the sim, note that the reverse base current spikes to over 80mA on each cycle. I believe that this can eventually reduce the beta of the transistor.

 

Uh, yes... I'm brain dead. I told you yesterday that I explicitly understood your PCode but this old brain must have forgotten... literally OVERNIGHT...duh!

Yes, I can do that without writing any PCode. Tina lets the user change model parameters for the models that are in my circuit, from a properties panel. I can change Cjo and Vj without affecting the values stored in the model libraries.

 

I made a 2N3904 subckt with Veb=6V. Below is the results of a simulation of an astable multivibrator, with and without BE breakdown modelling........I believe that this can eventually reduce the beta of the transistor.

yes.
I was ordered to never allow the emitter-base junction of a transistor to have avalanche breakdown at my first job with Philips. They wanted the beta to stay high. During avalanche breakdown, hot spots form on the junction and slowly destroy it. The same as with a zener diode but it is built to dissipate the heat and doesn't have beta.

 

yes.
I was ordered to never allow the emitter-base junction of a transistor to have avalanche breakdown at my first job with Philips. They wanted the beta to stay high. During avalanche breakdown, hot spots form on the junction and slowly destroy it. The same as with a zener diode but it is built to dissipate the heat and doesn't have beta.

It's been many years since I used one but high end curve tracers (like Tektronix) tested diodes and transistors for breakdown. They limited the current to uAmps, so breakdown didn't destroy the junction.