Magna-Chordion Vibrato

[ljudsystem]

Hi guys

I don't know if you know about this. I found the schematic for the Magnatone 240 Magna-Chordion - a solid state accordion amp from the 60's that had a vibe/vibrato effect. Similar to the tube vibratos but solid state, done with transistors and LDRs.

Here is the schematic (drawn by me):



As you can see it's pretty similar to the Magnavibe. I've got it on the breadboard right now and it sounds good! Better than the Magnavibe, don't know if that is just because of the extra ldr stage.

[duck_arse]

nice work. 22V - that'll get people excited. what transistor type numbers? and what does the pot furthest right do, offset?

[ljudsystem]

Thanks 

The transistors are 2N3721 "orange" (?). I found a datasheet that says 60 - 600 hfe so I used 2N2222's in mine.

I'm not sure about the trimpot, I think it biases the transister which works as a LED driver, right? I've breadboarded the circuit with the magnavibe lfo instead since I couldn't get the lfo on the schematic to work. Don't know if it's because I ran it with only 9v or because I made a mistake somewhere.

[Kipper4]

Sweet. Got any samples of the accordion please ?
I'd love to hear it.
What value R is the Rc and Re please?

If you have it on the breadboard you could try some different value caps on the collector i.e. Smaller than 100nf ergo 47n 20n see what you think.
Rich

[ljudsystem]

Sure, I'll try to record a video when I get home.

The Re and Rc values are 3.3 Kohms though I suppose those should be changed when running it at 9v. I measured the voltage at the collector and it was above 6V. That voltage should be about half of the supply V, right?

I'll try some different cap values and see what it does. I tinkered a bit with the circut last night and found that the 100k paralell to the ldrs makes a big difference - it really smoothes out the vibrato and makes it less choppy.

[duck_arse]

ahh, it's a trimpot? so it would set offset for the led drive. might make a panel-mount mod, with stoppers each end.

on the original not working oscillator copy - if'n you ditch the two base resistors and replace w/ a 1M5 [?] from collector to base, you might get it going. it needs a big electro to bypass that emitter resistor, tho.

[ljudsystem]

Ok, thanks. I'll try that it would be interesting to be able to offset the led.

I made a recording straight into the computer where I fiddle with the speed and depth, Enjoy! 

https://drive.google.com/file/d/1YIrT4qgyCMyxoWUco2JtlO74n_1YxKY5/view?usp=sharing

[Kipper4]

Interesting. Thanks for that.
It's much more subtle than I thought it might be.
I guess I expected a p90 sound from a p45.
Very nice, great recording.

[j_flanders]

Sounds subtle but very nice and smooth.
If you mix in the dry you have a p45-ish phaser effect. You can try it without soldering if you have a mixer or a something like a Boss LS2 pedal.
What LDR's (light/dark resistance?) and LED(green?, yellow?) did you use?

[ljudsystem]

Yeah, it's subtle but nice.

The Ldrs are something like 20k to 2M (I tried to match them as close as possible) the led is actually a rather weak red led. I've tried it with a superbright green led but I didn't like it. The effect was much more extreme but very choppy, square wave-ish and it becomes more of a bad phaser than a vibrato. Also, as mentioned, the 100k resistors across the ldrs dampen the effect some but also smoothes out the lfo.

One idea I have is to split the signal as on a Phase 45 and mix it back but highpass the vibrato and lowpas the dry signal. I'm thinking that might get me into Leslie territory, what do you guys think?

[Kipper4]

I either roll my own 3mm red and 500k dark or if the 500k dark isnt cutting it use Gl5539.
I usually experiment on the breadboard with the vactrols.

Sounds really good.

[duck_arse]

sounds good to me, nice double swirly.



don't call me swirly.

[ljudsystem]

So, I'm not going to run this circuit at 20V... should I bias the transistors differently for a 9v power supply? I'm searching the net but all the math makes my brain hurt.

Is there a more hands on approach to biasing transistors? like, using a breadboard and multimeter?

[Kipper4]

The manga vibe biases the concertina splitter with Rc and Re 7k5 if I remember right on a +9V supply.
2x 10k will work just fine too
Not forgetting the 1M+ resistor C-E this will bias the base. I don't know how many times I neglected to includes this and the countless hours at the breadboard scratching my head.

There was some discussion about a slightly smaller Re a while back.

Do a forum search for the manga vibe. Lots of good info there mate.
Ergo

https://www.diystompboxes.com/smfforum/index.php?topic=113411.0

[duck_arse]

So, I'm not going to run this circuit at 20V... should I bias the transistors differently for a 9v power supply? I'm searching the net but all the math makes my brain hurt.

Is there a more hands on approach to biasing transistors? like, using a breadboard and multimeter?

breadboard, multimeter. tell us your measures. seeing as the transistor will bias itself, what would you change? if you used a two resistor voltage-divider bias, it would be different, but still easy to work out. 30% of V+, plus 0V6 should set the emitter to ~30% and collector to ~60% of supply.

but don't " me

[PRR]

> I measured the voltage at the collector and it was above 6V. That voltage should be about half of the supply V, right?

No! For one output (the usual amplifier), lightly loaded, half is a good start. So with -two- outputs, 1/4 3/4 is a similar good start.

With load, the transistor will pull "in" real well but the resistors pull out not so well. 30%/70% even 33%/66% may be better.

> should I bias the transistors differently for a 9v power supply? ..makes my brain hurt.

How does this bias *work*? Ratio of base resistor, divided by hFE, to load resistors. This much will work the same at ANY supply voltage. There is also a Vbe in there, but if supply is >>Vbe then we can ignore it. 9V is not SO much greater than 0.6V, but we can still turn a blind eye to it. The variation in hFE from one part to another is greater.

And if you would use the intended 20V supply, it would not matter, because signal level is much-much less than supply. (In this case, I know that, because the LDRs will distort for >1V signal.) You could bias at 3V and 17V, or 7V and 13V, and signal would pass clean. Yes, at 9V you may have to take care with bias.

> Is there a more hands on approach to biasing transistors? like, using a breadboard and multimeter?

Crank the biggest signal you expect into it. 'Scope the works. Is it clipping? Which side? Adjust Rb for less clipping. If that can't fix it, reduce "the biggest signal you expect" until it works, or drowns in hiss (unlikely).

[ljudsystem]

ok, I measured the voltage across the transistors and got this:

T1 and T2:

E - 2.5V
B - 3V
C - 6V

T3:

E - 3.7V
B - 4.3V
C - 8.5V