current boss pedal power filtering examination

[m_charles]

hello. a while back someone online traced their "wazacraft" boss DS-1. The circuit is unremarkable, but what had me intrigued was the power supply filtering. This seems like a lot of components to achieve circuit protection and a stable supply. I do understand that the setup is meant to accomplish those tasks with as little voltage drop as possible (probably why no series resistor is used). However, I assume the engineers at boss know much more than I do about this so wondering if you folks had any thoughts? I'm always looking for new ways to filter DC and protect a circuit.
Obv r18, r19, and c16 need no explanation. they are the voltage divider.

Two caveats.

1) This is a trace someone did so no guarantee that it's 100% correct although I don't see anything that looks glaringly wrong.
2) no values given for c1 or c9. My guess is that c9 is a ceramic 100n as that would make sense. Unsure of what c1 would be. maybe another 100n?

I know boss does tend to err on the side of using more parts than needed sometimes but anyone with more knowledge than me understand why this is all happening just to get clean vcc?
Thx!

[Rob Strand]

Boss started using the circuit a while back in various pedals.   You can see a few variants here (and there's a few more),
https://www.diystompboxes.com/smfforum/index.php?topic=132064.msg1284964#msg1284964

[antonis]

This is a trace someone did so no guarantee that it's 100% correct although I don't see anything that looks glaringly wrong.

IMHO, D6 being 9V1 Zener is redundant..
It can't stabilize anything 'cause voltage on its Cathode can't reach its nominal voltage rating due to D3 and R14..
It could work for either lower nominal voltage or D3 of higher one..

P.S.
Without it, Q6, R14 & C10 form a healthy capacitance multiplier..
(actually, they form a LPF of corner frequency f= 0.159 / (R14*C10*h_FE)

[m4268588]

It can't stabilize anything 'cause voltage on its Cathode can't reach its nominal voltage rating due to D3 and R14..

???

[Rob Strand]

MHO, D6 being 9V1 Zener is redundant..
It can't stabilize anything 'cause voltage on its Cathode can't reach its nominal voltage rating due to D3 and R14..
It could work for either lower nominal voltage or D3 of higher one.

It acts as over an voltage voltage protection.  If you plug in a 9V unregulated supply (12V unloaded) the zener kicks in and the pedal gets about 9V.

[antonis]

It acts as over an voltage voltage protection.  If you plug in a 9V unregulated supply (12V unloaded) the zener kicks in and the pedal gets about 9V.

Of course that stands for D3..
I was talking about D6..

[antonis]

It can't stabilize anything 'cause voltage on its Cathode can't reach its nominal voltage rating due to D3 and R14..

Can't get your query..
Is there a voltage drop across R14, or not..??

[m_charles]

My main question though is why the setup? In my rudimentary builds I've never used a BJT transistor in the filter section. Can you explain why they did it this way? Trying to understand the main benefit. Thx!

[m_charles]

Boss started using the circuit a while back in various pedals.   You can see a few variants here (and there's a few more),
https://www.diystompboxes.com/smfforum/index.php?topic=132064.msg1284964#msg1284964

Thanks Rob, reading those posts now. Can you give any insight as to why this setup removes the need for a series resistor?

[Rob Strand]

Of course that stands for D3..
I was talking about D6..

I was talking about D6 being for protection.

As for D3 I've got doubts it's a zener at all.

The early pedals with Boss ACA supply (unregulated 9V, 12V unloaded) used an 11V zener in the D3 position.  That protected against overvoltage.  They relied on a series diode and resistor in the negative rail of the pedal to limit the zener current.

A few Boss ACA pedals replaced the zener with a 1N4004 type diode and lost the over-voltage protection.

The later pedals with Boss PSA supply (regulated 9V) tended to only used rectifier diodes in the D3 position.  So that's why I suspect D3 isn't a zener.  Also 9.1V for D3 is pushing things too fine for a 9V supply.  For the ACA pedals many don't have the zener for over-voltage protection but some do.

There's also cases where the transistor + zener is used for lower than 9V rails, like ~5V rails on the Boss pedals with the discrete opamps and for low voltage BBD etc.

There might a trace of this pedal on fsb.  There's a guy there who has been tracing a lot of pedals and Waza versions and his traces are very accurate.

[Rob Strand]

Thanks Rob, reading those posts now. Can you give any insight as to why this setup removes the need for a series resistor?

For filtering the active circuits (those power supply circuits with the transistor with or without a zener), effectively move the series resistor to R14.

However, without D3 a series resistor before D3 protects D3.  If you have D3 for reverse protection without a series resistor then often the diode and PCB fries when you apply reverse voltage.   Unfortunately low valued series resistors often fry anyway under reverse voltage because they don't have a high enough power rating.  The addition of D2 stops all that.

With D2 in place, it only makes sense for D3 to be a zener.   However there's no series resistance so that's risky as well.  Also generally Boss use a rectifier diode for the D3 position on PSA pedals, and it has no series resistor.   So perhaps D2 is an addition on the Waza units and D3 is a left over.

Check out the circuit in this thread,

Boss - SD-1w Super OverDrive (Waza Craft)
https://www.freestompboxes.org/viewtopic.php?t=29267

It's got what looks like a rectifier diode for D3; 1SR154-400.  So perhaps D3 is a left over from adding D2 on the Waza pedals,

1SR154-400 datasheet
https://docs.rs-online.com/1ab3/0900766b80dcd82a.pdf

[PRR]

why this setup removes the need for a series resistor?

Everything *should* have a series resistor. Infinite current is bad. The smoke tells you where the problem was.

The dynamic impedance ("plate resistance") of a BJT is about infinity. This into a cap or Zener makes much better filtering than your 470r resistor.

Alternatively: 'Capacitance multiplier'. Your 47uFd cap and a hFe=100 transistor acts a little like a 4,700uFd cap.

I see the double-Zener topology as a 12 cent approximation to blow-up proof. Not utterly unkillable; just harder to kill at a price musicians can pay.

Also note that Zeners are not brick-wall, their Vz varies with current. The first may be holding 9.3V semi-steady which the second running lower current Vz=8.6V is a second smoothing action.

Go ahead and lose it all, plug to a 9.36V battery or a mystery ??V wall wart. It may be fine 9 out of 10 times. Certainly enough pedals were sold on the assumption nothing bad would ever happen.

[m_charles]

Thanks Rob, reading those posts now. Can you give any insight as to why this setup removes the need for a series resistor?

For filtering the active circuits (those power supply circuits with the transistor with or without a zener), effectively move the series resistor to R14.

However, without D3 a series resistor before D3 protects D3.  If you have D3 for reverse protection without a series resistor then often the diode and PCB fries when you apply reverse voltage.   Unfortunately low valued series resistors often fry anyway under reverse voltage because they don't have a high enough power rating.  The addition of D2 stops all that.

With D2 in place, it only makes sense for D3 to be a zener.   However there's no series resistance so that's risky as well.  Also generally Boss use a rectifier diode for the D3 position on PSA pedals, and it has no series resistor.   So perhaps D2 is an addition on the Waza units and D3 is a left over.

Check out the circuit in this thread,

Boss - SD-1w Super OverDrive (Waza Craft)
https://www.freestompboxes.org/viewtopic.php?t=29267

It's got what looks like a rectifier diode for D3; 1SR154-400.  So perhaps D3 is a left over from adding D2 on the Waza pedals,

1SR154-400 datasheet
https://docs.rs-online.com/1ab3/0900766b80dcd82a.pdf

thanks for the reply. I wasn't asking about the protection aspect, I was referring to why there is no series resistor needed to filter out wall hum with this setup. I'm assuming that if the setup failed to properly do this, they wouldn't use this. I've found I pretty  much always need a 47 or 100 ohm in series with the +9vDC or I get hum. Of course with some circuits this can cost you a lot of voltage drop.

[Rob Strand]

thanks for the reply. I wasn't asking about the protection aspect, I was referring to why there is no series resistor needed to filter out wall hum with this setup. I'm assuming that if the setup failed to properly do this, they wouldn't use this. I've found I pretty  much always need a 47 or 100 ohm in series with the +9vDC or I get hum. Of course with some circuits this can cost you a lot of voltage drop.

There's no need because resistor R14 and capacitor C10 do the same job.    The transistor allows high currents to flow with large values of R14.  If the zener is conducting then it's a little different but in this case the zener stops the ripple getting through.

In the thread I linked earlier I compared the RC filter with the active filter and an LC filter.
https://www.diystompboxes.com/smfforum/index.php?topic=132064.msg1284995#msg1284995

The main thing to see is the active filter does filter without a resistor in series with the supply.

[m_charles]

why this setup removes the need for a series resistor?

Everything *should* have a series resistor. Infinite current is bad. The smoke tells you where the problem was.

The dynamic impedance ("plate resistance") of a BJT is about infinity. This into a cap or Zener makes much better filtering than your 470r resistor.

Alternatively: 'Capacitance multiplier'. Your 47uFd cap and a hFe=100 transistor acts a little like a 4,700uFd cap.

I see the double-Zener topology as a 12 cent approximation to blow-up proof. Not utterly unkillable; just harder to kill at a price musicians can pay.

Also note that Zeners are not brick-wall, their Vz varies with current. The first may be holding 9.3V semi-steady which the second running lower current Vz=8.6V is a second smoothing action.

Go ahead and lose it all, plug to a 9.36V battery or a mystery ??V wall wart. It may be fine 9 out of 10 times. Certainly enough pedals were sold on the assumption nothing bad would ever happen.

Taking out the protection aspect I'm still not understanding why they don't need it. I understand WHY they didn't want to use it. Because it doesn't cost them the voltage drop. I'm assuming it's not because none of them realized that a series resistor would do the trick. The newer boss pedals I own are usually dead quiet regardless of what power supply I use be it my isolated voodoo labs brick, or a $7 wallwart from amazon. I'm guessing a large part is due to this particular arrangement (along with good layout practices, etc). I realize I can just copy it and not worry about it, but I'm trying to understand the why of it... 

[m_charles]

thanks for the reply. I wasn't asking about the protection aspect, I was referring to why there is no series resistor needed to filter out wall hum with this setup. I'm assuming that if the setup failed to properly do this, they wouldn't use this. I've found I pretty  much always need a 47 or 100 ohm in series with the +9vDC or I get hum. Of course with some circuits this can cost you a lot of voltage drop.

There's no need because resistor R14 and capacitor C10 do the same job.    The transistor allow high currents to flow with large values of R14.  If the zener is conducting then it's a little different but in this case the zener stops the ripple getting through.

In the thread I linked earlier I compared the RC filter with the active filter and an LC filter.
https://www.diystompboxes.com/smfforum/index.php?topic=132064.msg1284995#msg1284995

The main thing to see is the active filter does filter without a resistor in series with the supply.

Ahh, thats the answer I was looking for! thanks Rob. What do you reckon C1 is doing and what value is it most likely? not used to seeing a cap to ground pre-zener diode

[Rob Strand]

Ahh, thats the answer I was looking for! thanks Rob. What do you reckon C1 is doing and what value is it most likely? not used to seeing a cap to ground pre-zener diode

In 99.99% of cases it's doing very little.  Caps in that position are more precautionary.  They help very high frequency ripple, RF and they can help protect against ESD.

If you look at the BD-2, even Boss aren't consistent.  Only the SMD version has it.
https://www.freestompboxes.org/viewtopic.php?p=292877#p292877

On the earlier versions you will see the C1 position is a 100uF electrolytic.   That's more about fending off hum and noise.   Again it's only helping in some marginal cases and that's probably why Boss have ditched the large cap in that position on a lot of later stuff.  One less large electro on the layout as well.

It generally doesn't hurt to add caps here and there but if you have to justify them it can come down to some very fine and obscure arguments.    If you were forced to show a case where you needed certain caps it would require some very specific conditions, conditions which are themselves difficult to justify!

[m_charles]

Ahh, thats the answer I was looking for! thanks Rob. What do you reckon C1 is doing and what value is it most likely? not used to seeing a cap to ground pre-zener diode

In 99.99% of cases it's doing very little.  Caps in that position are more precautionary.  They help very high frequency ripple, RF and they can help protect against ESD.

If you look at the BD-2, even Boss aren't consistent.  Only the SMD version has it.
https://www.freestompboxes.org/viewtopic.php?p=292877#p292877

On the earlier versions you will see the C1 position is a 100uF electrolytic.   That's more about fending off hum and noise.   Again it's only helping in some marginal cases and that's probably why Boss have ditched the large cap in that position on a lot of later stuff.  One less large electro on the layout as well.

It generally doesn't hurt to add caps here and there but if you have to justify them it can come down to some very fine and obscure arguments.    If you were forced to show a case where you needed certain caps it would require some very specific conditions, conditions which are themselves difficult to justify!

Last question. What's your personal preferred way to do it if you need minimal voltage drop?

[Rob Strand]

Last question. What's your personal preferred way to do it if you need minimal voltage drop?

It's not a clear decision until you pin things down a bit.  In many cases there's not *that* much in it you can go either way.  Small points tip you one way or the other.  It depends on how much filtering you need, the maximum drop you can handle, the how big do you want to go on the filter cap, how much 50Hz/100Hz PSU ripple you expect in practice.

If you expect significant ripple that means the active filter won't work unless you add another resistor from the base to ground to increase the voltage loss.   If 0.7V to 1V loss is too much then that also throws out the active filter.    At high currents you need a small resistor for the RC filter, then you ask yourself will the filtering be enough with a 100uF cap or 220uF cap on the supply.    That's the type of things that go through my head.

If you want very high rejection then the active filter can do that with reasonable cap sizes, there's a point where the caps become too big on an RC filter.   You should also consider two filters in cascade, or a voltage regulator.

For SMD designs it favors the active filter as it avoids large caps and electrolytics altogether.   

So there's very fuzzy lines of decision.

You might want minimal drop but how much filtering is acceptable.  At one extreme use a wire, there's no drop but no filtering.   So back that up a bit and say we use a 10 ohm to 22 ohm resistor for minimal drop.   Is the filtering enough with a 100uF cap?   it depends on the circuit.   In many case it's fine.    For many pedals 100R + 100uF works fine and has better filter than 10R.

[m_charles]

For SMD designs it favors the active filter as it avoids large caps and electrolytics altogether.   

but the active filter still needs those electrolytics unless I'm mis understanding? Im assuming we're considering 100uf "big".