Help understanding this circuit

[E]

Hey guys,

I'm trying to understand how this circuit works, and what the SCR is for (see attachment).
It's for an old intercom where the button is pressed when you speak and after you release it you hear the person at the door for a few seconds.

The way it looks to me, the cap is being charged when you press the button and then keeps the SCR on for a bit, stopping the audio output when it's empty. But I don't understand how it would do that.

When the button is pressed, the cap sees some voltage until the button is released, at which point it will lose charge over the 1k resistor and the gate of the SCR.
From experience the sound is audible for about 5-10s after releasing the button, but the cap should be empty after a few tens of ms or less if I'm not mistaken.

Anyways, as you can see I have some trouble wrapping my mind around this one. Hopefully someone here can help me understand it.

Thanks in advance,

E

//edit: corrected circuit
//edit2: circuit still wrong: audio in is 2VDC + modulated audio signal with very small ptp for noise and unknown ptp for signal

[c4757p]

I think the cap is basically just filtering the button presses. The SCR will not switch off when its gate goes low, it switches off when the current through it is below a threshold. What exactly does that connector go to? I imagine it's not actually a speaker - surely the circuit doesn't put a speaker across the 12V rails?

[E]

I think you've completely answered my question already. Now it seems obvious that the cap is filtering the button. The 1k resistor is probably there to discharge it in case the SCR is defective. Thanks

Sorry, the +12V symbol is a bit misleading. it's a +12V constant voltage modulated by the audiosignal from presumably the amp on the ground floor. I only have access to the "client"-side circuitry unfortunately.
The connector goes directly to a speaker.

[c4757p]

So the speaker has 12VDC straight up it? What kind of speaker is this?

[E]

It's a standard speaker with a diameter of about 8cm or so. I might be mistaken about the constant 12V, it seems a bit odd.
I measured this thing about half a year ago and I remember that I had this weird DC offset all the time, but didn't put much more thinking in it since I wasn't really interested in the circuit at that time.

The speaker should have 8 Ohms. There's probably a series resistance in the circuit somewhere downstairs. I've never measure the current going through the speaker. Could be that my records are wrong as well tho.

[c4757p]

There must be something else somewhere - first, the SCR will happily run almost 1.5 A through an 8 Ohm speaker connected to 12V, and second, the circuit as I see it does not explain the few second delay.

One thought - this could work (I think - didn't bother testing it with hardware, and I don't have a readily available SCR SPICE model to do a quick software test either): instead of just a speaker, the speaker could be AC coupled, and paralleled with a series RC circuit. The charge current of the RC circuit would keep the SCR conducting. A quick calculation says the capacitor would have to be upwards of 1000-2000 uF, though, to get an appropriate delay, unless the SCR had an unusually small hold current (most small signal SCRs have a hold current around 5 mA). It must be something different.

[E]

Well, the SCR is a EC103a with 5mA hold current and 200uA trigger current.
The whole circuit is nearly 30 years old and only the recieving end of the intercom device. There's probably a lot more stuff down in the basement.
The loudspeaker is simultaneously used as voice output, microphone, and doorbell.

Anyways, thanks for your help 

[MikeK]

I couldn't get the SCR to switch on in Circuit Simulator:

http://www.falstad.com/circuit/#%24+1+5.0E-6+4.078577355933338+59+5.0+50%0A177+352+224+320+272+0+11.999908147043199+11.999334066063195+1.9999999999999998E-4+0.005+50.0%0Aw+320+272+208+272+0%0Ar+320+272+320+368+0+1000.0%0Ac+208+272+208+368+0+2.2E-5+0.5753938309172196%0Aw+208+368+320+368+0%0Aw+320+368+352+368+0%0Aw+352+368+352+272+0%0Ar+208+272+208+176+0+10000.0%0As+208+176+208+80+0+1+false%0Aw+352+224+352+80+0%0Aw+352+80+208+80+0%0AR+352+80+352+48+0+0+40.0+12.0+0.0+0.0+0.5%0Aw+352+272+448+272+0%0Ar+448+272+448+368+0+8.0%0Ag+448+368+448+400+0%0A

Is the circuit drawn properly?

[c4757p]

MikeK, I didn't look at your circuit (because I don't have Java installed and I'm not downloading it just for that), but if you drew it like the original schematic, it should work. I tracked down a SPICE model for 2N5060 and played with the circuit a bit and it worked just fine. (If my workbench weren't such a mess I'd just drag out a real 2N5060, but the model appears to work accurately...)

[hammil]

Falstad circuit simulator is simply awesome. Anyway.. for some reason the other circuit glitched the connection to the SCR's gate, so here is a working one, with added modulation and decoupling:

http://www.falstad.com/circuit/#%24+1+5.0E-6+22.512744558455275+59+5.0+50%0Aw+320+272+208+272+0%0Ar+320+272+320+368+0+1000.0%0Ac+208+272+208+368+0+2.2E-5+6.098254685937832E-5%0Aw+208+368+320+368+0%0Aw+320+368+352+368+0%0Aw+352+368+352+272+0%0Ar+208+272+208+176+0+10000.0%0As+208+176+208+80+0+1+false%0Aw+352+224+352+80+0%0Aw+352+80+208+80+0%0AR+416+80+416+48+0+0+40.0+12.0+0.0+0.0+0.5%0Ar+448+272+448+368+0+8.0%0Ag+448+368+448+400+0%0A177+352+224+336+272+0+-1.437770148916023+-1.4378311314628824+0.01+0.008199999999999999+50.0%0Av+352+80+416+80+0+1+100.0+5.0+0.0+0.0+0.5%0Ac+352+272+448+272+0+1.0E-6+11.524352988294561%0Ao+11+64+0+35+0.01953125+0.003125+0+-1%0A

From my understanding, this whole thing acts as a relay of sorts. The resistors mean that the current to the gate is far too low for the SCR to trigger. However in this case, the power drawn by the speaker is miniscule - in the order of microwatts.

Regarding the capacitor, it seems that any charge that is built up is drawn away at the anode by the SCR's gate, through to ground. If it was allowed to charge, then the current it would supply would be enough for the SCR to trigger fully.

[MikeK]

But as with mine it seems to function the same if you delete the SCR.

[E]

Ok... I'm back to officially being confused. I took out the oscilloscope and checked a few voltages.

The 12V is wrong - it's actually ~1.5V ptp repetetive noisy signal, which you can also hear on the speaker once you pressed the button. I assume if someone would talk into the microphone downstairs, it would have a somewhat higher amplitude. But it's centered around 0V not 12V!

The next thing I measured is the voltage on the cap, which is about 3V when the button is pressed and then very slowly falls to ~0.8V, at which point it drops to 0V very fast which is also the time when the SCR turns off.

The trigger current through the SCR must flow from the Anode to the Gate? I checked the schematics, datasheet and the pcb again for the 20th time, everything should be right. Also, there's definitely no 1uF cap between the speaker and the SCR, nor after it. The only unknowns are what provides the input voltage (presumably an audio amplifier).

 :-/

[mswhin63]

Without looking at the whole thing it looks like a weird VOX circuit, the 2V P-P signal is filter and switches on the gate, no audio no triggering. But it is an SCR  so it would chop the waveform up a bit.

[hammil]

There has to be something else going on

The SCR will only conduct in forward bias, regardless of whether it is switched on or not. So any 0V-centered audio input would be horribly clipped, as mswhin63 mentioned

It will take several mA at least to trigger the SCR, at which point it will continue to conduct until the anode-cathode voltage drops below the threshold.

You're right - the trigger current does flow from the anode to the gate, but there simply isn't enough being conducted to trigger the SCR. Look on the datasheet for the trigger current, hold current and gate-cathode resistance for some values

[E]

Well, there's more circuitry going on on the board, that isn't connected to this except for GND.

I have to correct myself again :-/

I realized that actually the input has a DC offset, just not 12V but about 2V slowly going down to ~0.6V before turning off. The 1.5V ptp I measured before were noise on the ground line. I measured the difference between two channels of my scope this time.
I assume that the audio signal is modulated onto the 2V DC offset, but is so small that my cheapy scope couldn't display it. I also measured on the speaker directly and not surprisingly found pretty much the same thing (mabye a slight voltage difference due to the SCR).

I'll try measuring it again when someone's at the door so I can have an actual signal and not just noise. But it seems to me that if the signal's too small to see on the scope (1V/div) it should be too small to hear on the speaker? No idea how big a signal ought to be to produce a quite audible noise.

--
- So the way I see it now, the button starts the SCR (trigger current 200uA)
- Once the SCR is open, the cap is charged via the A-G current
- The modulated audio signal is audible via the speaker
- If the signal for some reason falls below the cutoff current (5mA), the cap will discharge as G-C current and keep the SCR open
- The cap downstairs giving the DC offset to the audio signal discharged slowly, at some point falls below the cut-off current
- The cap in the diagram discharges via the SCR and audio output is stopped

--
Remaining questions:
1) what's the 10k for (protecting the gate? as voltage divider with the 1k?)
2) what's the 1k for (discharging the cap in case of SCR fail?)
3) is the cap really necessary?

Thanks for all the help so far, and sorry for the confusion. I guess I'm more noob than I thought I'd be