Bipolar supply

[idlefaction]

hi all.

i am having issues with a ring modulator.

the issue i have is with the LFO.  it dumps heaps of signal to ground which results in a whistle that's always there - even when the whole thing is 'true bypassed'!  also, the LFO signal is waaaay to high - the MC1496 chip app note (which is the chip doing the modulating) recommends a modulation signal of 60mV and this thing has a full scale output of 24V p-p.

has anyone got any ideas on how to reduce the voltage output swing of the LFO, and also reduce the power consumption?  i think that dropping the power supply voltage to the LFO circuit will do this, but i'm not sure how to go about it since it's bipolar.

thanks  

[gez]

What type of LFO is it?  If you can't post a schematic, give a description.

If it's a 2 op-amp triangle job then it's pretty easy to reduce the amplitude but you'll have to compensate by tweaking other resistor values (will speed the LFO up).

[gez]

PS  You can just run the LFO on half the supply (12V) and couple it's output through to the modulator.  If you want to reduce current spikes use a regulator to run it even lower.

[idlefaction]

schem is here:
http://www.ele4music.com/rmp1b/rmp%201b.pdf

it's a two-opamp triangle/square job, yep.  i don't like the square one so i'm going to just use the triangle option.

if you could point me at somewhere that explains the maths behind the circuit value choices, that would be super cool    i'd love to get a 1V signal out of this thing instead of 25V.

[gez]

When the Schmidt’s +ve input is just positive or just negative of 0V, it triggers.  The Schmidt’s output will be either +V or -V, so half the supply will be across the resistor providing positive feedback (22k in the schematic you linked to) at the point when it is triggered.  

The resistor between the integrator’s output and the Schmidt’s input forms a divider with the 22k and sets the amplitude of the triangle.  From a +12V and -12V supply (just using this as an example), the output of the integrator has to fall/rise 5.4V (with a 10k resistor as shown) for the input of the Schmidt to be at 0V and trigger - 10k X (12/22k) - so the amplitude of the triangle will be twice this, i.e. 10.9V.

To reduce the amplitude you can reduce the value of the 10k resistor OR increase the value of the 22k.  Both will speed the LFO up so you have to compensate by increasing the value of the resistor from the Schmidt’s output to the integrators -ve input (in this case 470R) OR by using a larger value cap - you’ll have to jiggle values.  

I’ve just posted the following schematic for convenience to demonstrate the formula.  I haven’t calculated the values in the schematic you linked to (I’ll leave that honour to you) but the same formula applies.  Just imagine the inverters are a couple of op-amps and the Vref inverter is ground.  The 10k pot is a common trick used to get a more linear response from the rate pot, it controls the proportion of output voltage applied to the integrator’s input rather than varying the current to control frequency (this would involve making the value of the 470R variable, but chances are everything would bunch up at one end)

[gez]

PS.  Personally, I'd use larger resistor values than the ones shown in that schematic you linked to.  This allows for smaller cap values (tolerance is better for one thing).

From such a large supply you might have problems getting the amplitude down as small as you want it.  You might have to divide down a larger amplitude than you desired.

Edit:  The resistor from pot's wiper to integrator's input has a loading effect on the pot and is generally made 10X larger (at least), unless you're going for a different 'feel' for the rate pot.

[idlefaction]

thanks heaps gez.  you sure have a neat circuit collection  

that explanation was pretty spot on, i'll be using it very soon.  i'm going to start with a +/- 5V supply from some regulators, then use your explanations to get the triangle amplitude down as much as possible while keeping the LFO in a useable range.

will post my results.  this circuit was not very well thought out, both the modulator and signal inputs are referenced to ground so there's no way to get the whole totem pole biased properly while still nulling out both signals.  i've read before that the circuits from this source aren't too reliable.  

thanks again!

[idlefaction]

looking at this circuit board a bit more, i'm going to have to do so much to it to get it how i like it that it's pretty much not worth staying with.  it wastes a hell of a lot of board real estate too, which is always a no-no in my projects  

so back to the drawing board!  i'll do some maths and hopefully come up with a nice ring mod others can build.  

[petemoore]

I for one would really enjoy a nice Ring Mod.
 That circuit looked like quite alot of work!
 It prompted an astute analysis by Gez !!!