BBD Questions - Vgg and Shielding

[MarshallPlexi]

Hey Guys!

I'm working on a CE-2 chorus and had a couple of questions I have not found the answer to here after searching.

1) Regarding Vgg for a MN3007, if the supply is +15V so that PIN 1 is 15V and PIN 5 is GND, should Vgg be tied to 14/15 of PIN 5 (1V) or just taken to ground? I know the Zombie Chorus just takes it to GND. I've tested this with a trimmer pot but I can't hear any difference between GND and 1V. Is there something that I could be missing here?

2) I see people shielding their clocks (MN3102's, CD4046 etc) but do the BBD's themselves need to be shielded? It would seem that as insidious as clock noises are, it might be a better layout to have the BBD's shielded near the clock drivers and then run wires from the audio section over to this the isolated clock/BBD section. Perhaps even use shielded cables for this run.

Thoughts?

Thanks again for all of the information over the years. I've spend 100's of hours reading this forum and always come away wanting to learn more!

[Kevin Mitchell]

In the respect of positive voltage, VGG would be ~0.6v for the MN3007 on a +9 volt supply while others (3207) would be 14/15 from the positive side (~8.4v).  VGG does have effect on the biasing of the device and is usually complimentary provided with the relevant clock device. Since .6v isn't too far from 0v, you probably won't hear much of a difference.

As for shielding, keep the clock traces/wiring short and away from the audio path. Isolate with a ground trace/plane. This is in effort to keep the square clock pulses from leaking/crosstalking into the audio path, through it usually isn't too bad with the limitations of a MN3101 or 3102 driver's frequency/clock speed range.

Typically in custom layouts it's approached as "I'll cross that bridge when I get there".
So don't sweat shielding too much 
Yet.

[MarshallPlexi]

Roger that about Vgg.

So keep the BBD's away from the audio as well given that they have the clock signal on them?

[fowl]

I "cheated" and built one on the Madbean Pork Barrel PCB.  MN3007.  I didn't shield anything, other than with the enclosure itself.  No noise issues.

[ElectricDruid]

1) Regarding Vgg for a MN3007, if the supply is +15V so that PIN 1 is 15V and PIN 5 is GND, should Vgg be tied to 14/15 of PIN 5 (1V) or just taken to ground?

Re-reading the datasheet, it doesn't actually give the 14/15ths figure, just implies it from a -15V supply and a -14V Vgg.
What it is clear about is that Vgg should be Vdd+1volt, so just tying it to ground is probably not the best idea.

2) I see people shielding their clocks (MN3102's, CD4046 etc) but do the BBD's themselves need to be shielded? It would seem that as insidious as clock noises are, it might be a better layout to have the BBD's shielded near the clock drivers and then run wires from the audio section over to this the isolated clock/BBD section. Perhaps even use shielded cables for this run.

Don't go crazy, but keep the high-speed (for us pedalbuilders, anyway!) clock traces nice and short and neat. Keep them away from any traces carrying audio as far as possible. Like that, it'll be fine. If a BBD circuit is noisy, it's not usually because of bad layout!

[MarshallPlexi]

You got it. 1V it is.

[Vivek]

The spec sheet for MN3207 says Vgg = 14/15 Vdd

MN3007 spec sheet says Vgg = Vdd +1

It is fun designing a PCB that can accept either a MN3007 or MN3207 depending on 3 jumper switches.

[Kevin Mitchell]

VGG is +1 volt on a 15 volt supply - not a 9 volt supply. The datasheets are assuming one is using the dedicated clock and running on the typical supply voltage.
Here's a snippet from the MN3101 clock datasheet;


The importance of the 400mv difference running on 9 volts is on you guys 
Obviously not detrimental. So long as it's not too far off.

Again, VGG isn't "+1 volt", it's 14/15 the supply voltage. If your supply voltage is 15 volts, it's going to be 1 volt. If it's 9 volts, it's going to be 0.6 volts.

[bean]

VGG is +1 volt on a 15 volt supply - not a 9 volt supply. The datasheets are assuming one is using the dedicated clock and running on the typical supply voltage.
Here's a snippet from the MN3101 clock datasheet;


The importance of the 400mv difference running on 9 volts is on you guys 
Obviously not detrimental. So long as it's not too far off.

Again, VGG isn't "+1 volt", it's 14/15 the supply voltage. If your supply voltage is 15 volts, it's going to be 1 volt. If it's 9 volts, it's going to be 0.6 volts.

The datasheet doesn't show the complete calculation, I think. Maybe because it was designed primarily for positive ground?

In an MN3101 negative ground circuit, Vgg works out to be
Vgg = Vdd-Vdd(14/15)
= Vdd(1-14/15)
= Vdd*0.067.
So, for 9v we get 9*0.067 = 0.6v, 15v we get 1v.

For MN3102 (which can only be done negative ground)
Vgg = Vdd(14/15)
So for 9v we get 8.4v.

I've round up in all these since Vdd is usually only a few hundred mV lower than Vcc, depending on how it's created in a circuit. Here are some real world measurements of the MN3101 I've taken in different circuits with different Vdd.

[Kevin Mitchell]

Nice!

Here's some simple math based off of bean's measurements. We'll do VGG = 1/15 of the supply voltage in respect of negative ground - I know these can get confusing, must use a bit of imagination and assume VDD and GND pins are swapped on the MN300X devices for these measurements.

First row;
VDD = 8.76v
VGG is calculated to be 8.76/15 = 0.584
VGG = 0.59v


Second row;
VDD = 14.24v
VGG is calculated to be 14.24/15 = ~0.949
VGG = 0.96v

Third row;
VDD = 11.11v
VGG is calculated to be 11.11/15 = ~0.740
VGG = 0.755v

We can blame the slight variations on what I assume is loose tolerances of an internal resistor divider - being the easiest way to produce these voltages

[thomasha]

Does anybody know why it has to be 14/15 or 1/15 ?
Has it something to do with limiting something to less than the supply voltage?
Compensate for a diode somewhere, that might result in a slight asymmetry?

IMHO once one knows the reason behind the setting, it is easier to decide if it is actually important or can be neglected depending the type of circuit.

I built circuits where it is just tied to ground or others where there is a voltage divider. What I do not know, and would really be interested in, is if it makes any difference and I could obtain a clearer sound out of it.

[ElectricDruid]

All this talk about 9V supplies with the MN3007 is a bit weird. For one thing, the datasheet recommends a 14V minimum supply, so 9V is way out of spec. And secondly, what's the point? If you want to run the thing at 9V and lose the noise and headroom advantages of the 15V supply, you use a MN3207. At least that chip was *designed* to work at that voltage. No?

[bean]

Does anybody know why it has to be 14/15 or 1/15 ?
Has it something to do with limiting something to less than the supply voltage?
Compensate for a diode somewhere, that might result in a slight asymmetry?

IMHO once one knows the reason behind the setting, it is easier to decide if it is actually important or can be neglected depending the type of circuit.

I built circuits where it is just tied to ground or others where there is a voltage divider. What I do not know, and would really be interested in, is if it makes any difference and I could obtain a clearer sound out of it.

Both versions of the clock should generate the correct Vgg voltage regardless so there's no need to set it externally. Just use a small decoupling electro for stability. You really only need to mess with it when using something other than the MN3101 or MN3102 for a clock.

Real world example, here's how the DMM creates Vgg using a CD4047 as a clock and the actual measurements taken off my own build. It more or less follows what you expect, Vgg=Vdd*14/15 (being a positive ground circuit):

[Kevin Mitchell]

All this talk about 9V supplies with the MN3007 is a bit weird. For one thing, the datasheet recommends a 14V minimum supply, so 9V is way out of spec. And secondly, what's the point? If you want to run the thing at 9V and lose the noise and headroom advantages of the 15V supply, you use a MN3207. At least that chip was *designed* to work at that voltage. No?

It's not uncommon for some MN3007 based effects to run off 9 volts. In this thread we are indeed talking about the old CE-2 circuit - which is typically powered by a 9 volt battery.

You are correct that the datasheet voltages suggest the best signal-noise ratio, which goes down with a lower VDD. But that didn't stop Boss/Roland. And obviously it didn't turn out too bad.

Looking back at the OP, they did specify a 15 volt supply. So my details become irrelevant though true!

Does anybody know why it has to be 14/15 or 1/15 ?
Has it something to do with limiting something to less than the supply voltage?
Compensate for a diode somewhere, that might result in a slight asymmetry?

As far as my rather limited understanding goes, it's to bias the fets. I'm only reciprocating some of my limited research on this question as I have asked myself the same thing in the passed. One could trim VGG to verify.

[ElectricDruid]

It's not uncommon for some MN3007 based effects to run off 9 volts. In this thread we are indeed talking about the old CE-2 circuit - which is typically powered by a 9 volt battery.

You are correct that the datasheet voltages suggest the best signal-noise ratio, which goes down with a lower VDD. But that didn't stop Boss/Roland. And obviously it didn't turn out too bad.

True enough! Interestingly, the CE-3 moved to the MN3207/3201 combination but is otherwise pretty much identical. I wonder why they made the change? Of course, there's lots of possible non-technical reasons (cost, availability, etc).