Multivibrator mystery.

[Skywave]

I cannot recall that I have ever had the need for the traditional and basic multivibrator circuit which uses a pair of transistors cross-coupled with a pair of capacitors - I'm sure you know the circuit. However, a requirement has arisen and despite all my efforts at changing the values of various components - it doesn't work: one transistor stays hard on; the other hard off. The transistors I am using are BC109C. Are these unsuitable for this purpose? And if they are, (which I believe), what's the magic trick to get such a simple circuit to work?
(N.B. I have deliberately avoided the 555 i.c. since in all the circuits I have ever tried, the 555 always self-destructs! )

Al.

[Duke_Nukem]

Pretty much any transistor will work in such a circuit, the BC109 included. Can you post a diagram with circuit values? Small cap values could be a problem if you are trying to get higher frequencies.

TTFN,
Jon

[ukcol]

Hi Al

This article shows an inproved circuit using steering diodes.

[ukcol]

Hi Al,

Going back to the basic circuit, it has belatedly occurred to me that it is strange that one transistor is completely turned OFF. If the MV is not running the DC conditions for that transistor will be determined by component values and hfe. If one transistor is not drawing any collector current at all, that seems to be the area to look for your problem.

[G6Tanuki]

I recall a similar sort of 'lock-up' sometimes happening with the old transistor power-inverters used to produce the HT for things like Pye Cambridge two-way radios - the design was essentially a transformer-coupled push-pull multivibrator and it depended on a gain-discrepancy between the 2 transistors to make sure one drew a bit more current at switch-on than the other so the oscillation started.
Sometimes there was insufficient, or too great, a gain-discrepancy and you could end up with one transistor hard-'on' and the other not conducting at all, or both of them passing lots of current but no oscillation taking place.
I usually fixed this with a 0.01uF capacitor between the base and collector of one of the transistors...

[TrevorG3VLF]

Early multivibrators used pnp junction transistors which can stand an appreciable base/emitter voltage. Modern transistors such as the BC109 will leak if the base/emitter junction is reverse biassed more than about 6V. Thus the power supply should not exceed 6V unless a diode is included in the base or emitter.

If the 555 fails presumably you are loading it too much.

The simplest oscillator is a NAND Schmitt trigger (4093 or 40106) with a resistor fron output to input and a resistor input to ground. F = 2.7/RC. This can drive only a few mA.

[Cobaltblue]

Just thinking sideways a bit it seems your after an Astable Multivibrator you didn't built a Bistable Multivibrator or Monostable Multivibrator by accident did you

I built a load of Bistables when I was 12 but both transistors seemed to be on

It was several days later I realised that connecting the meter was making them toggle

Such is the learning process ( that and only having one meter )

Cheers

Mike T

[mhennessy]

Yes, I'd be suspicious about reverse-VBE breakdown. What is the supply rail, and does it work on 5V? I wonder if the transistors are being damaged by during the first few cycles?

That is a fascinating PDF Colin - I'll make time to read it properly later...

It might be worth saying that a simpler way to make one of these work from higher rails is to put diodes in series with the base of the transistors. The calibrator in my Telequipment DM63 does just that... It also uses different values of capacitor and base resistors, which presumably is to help prevent a lock-up situation.

[n_r_muir]

Hi,

Unless you need a symmetrical output, make the two base resistors slightly different in value e.g. R2 = 43K and R3 = 47K so that one transistor will have slightly more drive than the other and get thing going.

[mhennessy]

Yes, as I said. Although I'd say that the resistor values in the example I cited are much more different that you suggest - they are 68k and 150k. This is because the effect is at the mercy of the Hfe of the transistors, so the effect of making them just slightly different will be swamped by the effect of the variations in Hfe.

There is a classic truism at play here: "Amplifiers oscillate, oscillators don't".

They seem so simple at a first glance

[kalee20]

The classic Eccles-Jordan multivibrator does have a lock-up mode, both transistors hard on. When that happens, there is less than unity loop gain around the circuit, so it is stable like that.

I've only found it a problem if the capacitors are very small - such as for above audio frequency applications. Basically, on switch-on, the power rail needs to come up much faster than the two capacitors charge.

As others have said, watch for reverse base-emitter breakdown. But it shouldn't stop it completely (unless you have zapped your 3-legged fuses).

[n_r_muir]

In the 1960's I used discete transistors to produce thousands of multivibrator, monostable and other pulse circuits I do not recall ever having any reports of multi vibratos "latching Up" i.e. both transistors on at at the same time.

[Skywave]

Thanks for your responses. I'll now reply to the earlier ones; the later ones require me to follow them up first.

The circuit I am using is as per post #4.
Initially, I had R1 = R4 = 1k2; R2 = R3 = 100k; C1 = C2 = 0.1 µF. Result: no oscillation.
Then I tried various combinations: R2 = R3 = 1M; C1 = C2 = 1 µF; R1 = R4 = 4k7. Any permutation of those did not result in oscillation.
Vcc was initially set at +12v. I tried Vcc at +5v: again, no oscillation.
I tried several other BC109C in each position (changing one at a time): again, no oscillation.
The transistors are BC109C: their B-E, B-C and C-E junctions have been checked: they're O.K.

Under all the above variations of 'no oscillation', one of the transistors is 'hard on'; the other is 'hard off'. At no time are they both 'hard on' or 'hard off'

The next thing I will try is choosing R and C values that result in a asymmetrical design - as has been suggested. But I do hear a small voice in the back of my head saying "inadequate Hfe."

Aside.
Is there a law in electronics that states that the degree of difficulty will be found with any given active circuit is inversely proportional to that circuit's simplicity? 'Cause it feels like that sometimes!

Al. / May 8th. //

[n_r_muir]

Bc109's will work ok. You are connecting the cross coupling capacitors correctly? the supply voltage is OK and of the correct polarity? no wiring errors, no solder bridges shorting out etc.

Really there is nothing much to go wrong and hfe's really isn't a problem provided it is greater than about 30, which is a fairly rubbish transistor anyway.

[Skywave]

Problem solved: a 'mechanical fault'. Had to be something really fundamental, didn't it?

I'll explain. For bread-boarding purposes, I use two parallel rows of tag strip, the individual tags of which are mounted horizontally on paxolin strips. These strips are mounted on a slab of copper-sided laminate using insulated stand-offs, 0.25" high. This arrangement is convenient for 'picking up' a ground / 0v. and R.F. applications. But underneath a pair of adjacent tags - therefore hidden to the eye - was a miniscule splash of solder, bridging those tags, one of those tags being at 0v. / ground.

What led me down that path was that no matter what changes I made to the component values, it was always the transistor in 'position 2' that stayed 'hard on' - even when the two transistors were swapped over. And, of course, I had every faith that the cct. should work and all the components were good.

Needless to say, things are now working as they should. And I now have a lot less hair left! I think my 'signature' below says it all.

So thanks guys for all your responses. There are a lot of useful remarks in those which will be useful to me as I develop this circuit for its intended application. But if anyone now considers that I've wasted their time, I can only apologise.

Al.

[Alistair D]

For one transistor to be off it's base is not getting enough current to turn it on or there is leakage of some kind from the on transistor's collector holding the base low. Try the circuit with the caps removed to prove if both transistors turn on.

Al

Originally Posted by Skywave

If anyone now considers that I've wasted their time

Not that I replied, no time wasted at all, we all need our feet back on the ground every now and again. Reminds me of a failed microwave oven, I was there with tools and measuring instruments (christmas day too after a few sherrys) SHMBO said 'what about the (plug) fuse, guess what it was.

[FrankB]

Use .01 mfd for the capacitors, and 10K and 100K for the resistor pairs.
Its basically the same circuit I "stole" and modified from a 70's vintage ARRL handbook (Keyer schematic, IIRC)to make a signal injector from.
I used 2N404 germanium transistors, and 1.5V flashlight AA battery to power it.
IIRC those were PNP's, to use NPN's reverse the battery polarity.
I must have built 15 of those signal injectors for friends over the decades.
I did make one with silicone xistors also, but it never worked as good as the one with germanium xistors.( I also put an .01 mfd 1KV blocking capacitor on the output of the injector).