Coming soon to runoffgroove.com: Thunderbird

[~arph]

It's powered by 9V. there is a charge pump involved. (it even says 9V on the PCB  )

[deadastronaut]

ahhh i didn't see the pcb either...gotcha!.

i'm in then...

[midwayfair]

24v....i'm out then... 

i only have a 9v daisy psu..

Voltage tripler = turns your 9V into 24V*.

*I'm sure this involves some sort of new math I'm not good at.**

**Okay, it's because a charge pump drops a volt or two.

[~arph]

Three diode drops to start with.. and then perhaps a regulator. ( Q1 78L24 ? )

[stm]

Voltage tripler = turns your 9V into 24V*.

*I'm sure this involves some sort of new math I'm not good at.**

**Okay, it's because a charge pump drops a volt or two.

If you really want to know how we can make 9*3 = 24, read on:

Let's start from something simpler until we get to the voltage tripler:


1) Voltage inverter formula:  Vo = -(Vin - Vloss)

where Vloss is the voltage loss internal to the charge pump IC at the output current; if you look at the LT1054 datasheet, Vloss is aprox. 0.5V at 15mA output current (this is estimated from the graph, as it actually shows the values for 10 and 50 mA only).

Using a 9V battery you get:

Vo = -(9 - 0.5) = -8.5 V


2) Voltage doubler formula: Vo = 2*Vin - Vloss - 2*Vd

where Vd is the voltage drop of the additional diodes; assuming we are using 1N5819 power schottkys for increased efficiency, we can assume Vd = 0.25V

Using a 9V battery you get:

Vo = 2*9 - 0.5 - 2*0.25 = 18 - 0.5 - 0.5 = 17 V


3) Voltage tripler formula: Vo = 3*Vin - 2*Vloss - 4*Vd

In this case the internal voltage loss affects twice because there are two capacitors that will be charged to a reduced voltage.  Also, there are four diodes instead of two, so this also reduces the output voltage.

Using a 9V battery you get:

Vo = 3*9 - 2*0.5 - 4*0.25 = 27 - 1 - 1 = 25 V


4) Voltage tripler formula with series supply polarity protection diode: Vo = 3*(Vin - Vd) - 2*Vloss - 4*Vd

Yes, the polarity protection diode will triple-tax the circuit perfomance and has to be taken into account in the formula.

Using a 9V battery you get:

Vo = 3*(9 - 0.25) - 2*0.5 - 4*0.25 = 26.25 - 1 - 1 = 24.25 V


5) Other factors

We are dealing with real components, so the capacitor ESR (Equivalent Series Resistance) will add to the losses too, so in practice you end up with a voltage that can be 24V or slightly less.

----

The previous analysis has been done using some key components for better performance and to minimize losses, namely:

a. Recommended charge pump IC is the LT1054N device, which is intended for lower losses at higher currents.  Using MAX1044 charge pump will have much higher losses and is not recommended for this particular application.

b. Supply polarity protection diode and charge pump diodes are 1 Ampere schottkys to minimize losses.  Recommended device P/N ar 1N5819 (40V reverse breakdown voltage).  1N5818s can be used as well (30V reverse breakdown voltage).  1N5817 are not suitable, as their reverse breakdown voltage is just 20V.
EDIT 07-JAN-2015:  1N5817 diodes are indeed suitable for the task as pointed below by ggedamed.  In this particular circuit the reverse voltage of the charge pump diodes is approximately 9V, however I had assumed (without verification) that it would be close to 24V as is the case for the capacitors.

c. Some capacitors involved in the voltage tripler are specified to be solid tantalum for low ESR (as explained above).  These capacitors must be rated for 35V minimum to avoid trouble.

[ggedamed]

I don't think that 1N5817 will have any trouble in the multiplier circuit, since the voltage across any of the diodes will not go over Vin (9V). Did you really had any trouble with 1N5817 or you're being cautious?
I fed 1N5817 doubler circuits with 18V - only for a minute or so, though - and they were OK. I'll make some measurements to convince myself.

[stm]

I don't think that 1N5817 will have any trouble in the multiplier circuit, since the voltage across any of the diodes will not go over Vin (9V). Did you really had any trouble with 1N5817 or you're being cautious?
I fed 1N5817 doubler circuits with 18V - only for a minute or so, though - and they were OK. I'll make some measurements to convince myself.

Thanks for pointing this out!  I added the pertinent ammendment to my post.

[jubal81]

I was experimenting with a tripler for 24V recently and did it with a 317 regulator on the input to cover over voltage issues and keep the level stable. Works great, but the drawback to me is the real estate all that circuitry takes up. I've been thinking the split supply switching bricks that run off 5V make a lot more sense.

[midwayfair]

I don't think that 1N5817 will have any trouble in the multiplier circuit, since the voltage across any of the diodes will not go over Vin (9V). Did you really had any trouble with 1N5817 or you're being cautious?
I fed 1N5817 doubler circuits with 18V - only for a minute or so, though - and they were OK. I'll make some measurements to convince myself.

Thanks for pointing this out!  I added the pertinent ammendment to my post.

Thanks for the clarification ... I've been using 1N5817s in all my pedals with charge pumps for a while and I got worried! 5819 often have very slightly lower Fv, though, so you could squeeze out a smidgeon more voltage!

I noticed we're talking about a "tripler" here to achieve 24V. Was any consideration given to just using a doubler and the -9V output of the LT1054? Both methods require the addition of a Vref voltage divider for the op amp bias, but the -9V method would have saved you a few diodes and given you a higher voltage (closer to 26V total)? Or even just a doubler for both rails (same number of components as the tripler, but as much as ~36V!) and using ground as Vref, which actually would probably have been slightly fewer components in the long run.

[stm]

I noticed we're talking about a "tripler" here to achieve 24V. Was any consideration given to just using a doubler and the -9V output of the LT1054? Both methods require the addition of a Vref voltage divider for the op amp bias, but the -9V method would have saved you a few diodes and given you a higher voltage (closer to 26V total)? Or even just a doubler for both rails (same number of components as the tripler, but as much as ~36V!) and using ground as Vref, which actually would probably have been slightly fewer components in the long run.

Yes, -9V and +18V (or so) are another way of getting a 24-ish total supply voltage.  I agree it is more efficient in terms of voltage loss, however your reference voltage (typically Vcc/2) now becomes around 4.5V (because (-9 + 18)/2 = 4.5), and now you have to deal with four different voltage rails in your circuit (-9, 0, +4.5 and +18) instead of three (0, +12 and +24), which obscures things a bit.

To be honest, I don't have a stronger prefence towards any of these alternatives, both will work adequately and both have their pros and cons.

[Bucksears]

So........are they waiting for the NAMM show to release the T'bird?


Also, is it fabbed-PCB only, or will there be an etchable layout?

[GGBB]

5819 often have very slightly lower Fv, though, so you could squeeze out a smidgeon more voltage!

Jon - is that based on measurements you've made at low pedal-type currents? My interpretation of the data sheets led me to believe the opposite - that 5817s have slightly lower Vf at low currents. Now I'm worried! 

[midwayfair]

5819 often have very slightly lower Fv, though, so you could squeeze out a smidgeon more voltage!

Jon - is that based on measurements you've made at low pedal-type currents? My interpretation of the data sheets led me to believe the opposite - that 5817s have slightly lower Vf at low currents. Now I'm worried! 

Just based on the ones I have around the house, using my multimeter. I have no idea what happens at higher currents! Although ... my 5818s at least are a very different package from the 5917s I have (which are from Tayda) and the 5819s I have (which are from Mouser). So this could simply be a manufacturer thing. Please don't take me as an expert on these things!

[B Tremblay]

So........are they waiting for the NAMM show to release the T'bird?


Also, is it fabbed-PCB only, or will there be an etchable layout?

Release is imminent! We want everything to be just right.

Yes, there is an etch layout, artfully created by Josh at 1776.

[GGBB]

So this could simply be a manufacturer thing.

Out of curiosity I just took a quick look at a handful of manufacturers' 1N5817-5819 datasheets. When provided, graphs of instantaneous Vf versus instantaneous If consistently show 1N5819 to be higher than 1N5817. At 0.1A the difference is usually around 0.2V. 1N5818 is between the two. Extrapolation of the graph data would suggest less difference at lower currents (a couple of the graphs go as low as 0.01A and still show differences), and some manufacturers provide only one plot for all three diodes.

However, there is a big range across manufacturers. For 1N5817, the range was from ~01.V (MultiComp) to ~0.45V (Diotec). Fairchild was ~0.3V, Vishay was ~0.17V (for all three diodes). Vishay provided a second plot for higher temperature (125C vs. 25C) which lowered Vf by about 0.9V.

So apparently the manufacturer can make a big difference - enough that one's 5819 could be lower than another's 5817. But this is instantaneous only - there was no data specific to continuous - not sure what difference if any that would make other than raising the temperature which according to Vishay would be helpful for our purposes.

[K Zustang]

FINALLY!!!

Thunderbird is out of the cage.

http://www.runoffgroove.com/thunderbird.html

PCB available for $13 at 1776 and I AM GETTING ONE FOR SHURE!!!

[J0K3RX]

Nice! I was curious about this one... I didn't see the LM833's comin, caught me by surprise!    I give 100 points for originality and another 100 points for achieving your goal, and then surpassing it! Damn fine work fellers!!

[akc1973]

Awesome!!!

[newperson]

I think the top second side art has been left off the build document PDF?  Unless they made 2 different layouts.  Commercial and home etch single side?

Looks like a nice project!

[~arph]

Cool, I like the dynamic level shifting and following clipping!
for me three treble controls is a bit too much though  

EDIT: Btw, yes I think there are two layouts.. the 1776 PCB is double sided.