Question about usage of crowbar circuits

[bushidov]

Hi all,

So, I was playing on my breadboard this little circuit that I whipped up:


It works perfectly fine in regards to it does its job. If the voltage goes much higher than 9V, the Zener goes over, trips the SCR into shunting and the PTC resettable fuse opens and does it relatively fast, protecting the 1044 IC.

I was curious to know why I don't see this circuit in guitar pedals that need that kind of over-voltage protection? I hear people with 1044 charge pumps putting in Zener diodes, but they just blow the zener completely, requiring it to be replaced.

I'm sure there is probably some reason. BOM count, or PCB real estate maybe? Your thoughts?

[antonis]

If the voltage goes much higher than 9V, the Zener goes over, trips the SCR into shunting and the PTC resettable fuse opens and does it relatively fast, protecting the 1044 IC.

TC1044 has 12V max input..
Any PS of 9V nominal output capable to go over 12V is a REALLY BAD PS..

[GGBB]

Isn't there some kind of engineering rule/principle that suggests you shouldn't use a device within a certain range of its limits? If a pedal is intended for 9V use - or in the real world of guitar pedals more like ~9.5V - you probably shouldn't be using a device without safeguards who's max input voltage is 10V. Your idea is a great safeguard, but isn't it just a band-aid for poor design - using a MAX1044 in a 9V pedal? The primary purpose of the MAX1044 is +5V conversion to -5V in logic circuits. Use something like an LT1054 instead if you want to operate at 9-10V. Maybe we don't see this kind of thing being done in commercial pedals is because it's a poor design.

[antonis]

Gord, I think 12V is maximun input voltage range ..

http://ww1.microchip.com/downloads/en/devicedoc/21348a.pdf

[bushidov]

The MAX1044 was just to throw something in there as an example. Just an IC that some people use for charge pumps for overdrives and boosts. I've heard people put in charge pumps to boost op-amp rails, but then the user of the pedal takes an 18V power supply and applies it to the pedal, smoking the IC.

The main question was the  actual act of using a crow-bar. Like does using a crow-bar add noise to the circuit in some odd-ball scenario I am not aware of, or catches fire if you do something weird? Stuff like that.

I am guessing the reason I don't see it when a specific voltage range is demanded is that it adds to the BOM and/or PCB real estate.

[GGBB]

Gord, I think 12V is maximun input voltage range ..

http://ww1.microchip.com/downloads/en/devicedoc/21348a.pdf

I assumed MAX1044 - didn't see the part number on the schematic - my mistake.

'MAX1044' is 10V.

https://datasheets.maximintegrated.com/en/ds/ICL7660-MAX1044.pdf

`LTC1044` even worse - 9V:

https://www.analog.com/media/en/technical-documentation/data-sheets/LTC1044_7660.pdf

The LT1054 is the one we want to use in 9V pedals - 15V:

https://www.ti.com/lit/ds/symlink/lt1054.pdf

[GGBB]

The MAX1044 was just to throw something in there as an example. Just an IC that some people use for charge pumps for overdrives and boosts. I've heard people put in charge pumps to boost op-amp rails, but then the user of the pedal takes an 18V power supply and applies it to the pedal, smoking the IC.

The main question was the  actual act of using a crow-bar. Like does using a crow-bar add noise to the circuit in some odd-ball scenario I am not aware of, or catches fire if you do something weird? Stuff like that.

I am guessing the reason I don't see it when a specific voltage range is demanded is that it adds to the BOM and/or PCB real estate.

I think the question "Why don't we see better power protection in commercial effects pedals?" is a good one. I don't have an answer to suggest other than the obvious - cost.

I *think* manufacturers tend to add disclaimers to their manuals about external power supply usage - that way they don't have to worry about what happens when somebody messes up, and then don't have to make the pedal more expensive to build.

[bushidov]

I think the question "Why don't we see better power protection in commercial effects pedals?" is a good one. I don't have an answer to suggest other than the obvious - cost.

I'm guessing you are correct, too. I just didn't know if there was some other reason I wasn't considering, like "No, in guitar pedals, we do this..." I know from an industrial bench-top power supply, we do, do things like this (albeit with glass or other kind of fuses, not typically resettable, unless it's a low current channel/rail), but as I've been playing with analog audio circuits, I've been learning that some of those rules of thumb don't cross over. That's why I ask you experts, as I am still new to this world.

[anotherjim]

Confidence is another issue. To the non-technical, the pedal might stop working for no obvious reason and mysteriously start working again later. Will the owner realize it isn't the pedal's fault?
You might need a positive indication of an overvoltage trip, which will add extra cost.

[Rob Strand]

While it looks like it works on paper with nominal values  you can run into problems when you consider tolerances.

In order not to false trip the *minimum* expected zener voltage (plus the SCR gate drop) needs to be higher than the *maximum* expected supply (in fact the peak of the ripple on an unregulated supply).   Not false tripping is an important requirement.

After choosing a suitable zener you then need to consider the maximum zener voltage.  The maximum trip point needs to be below the limits of the device you are trying to protect.   In the case of some of those converter devices the maximum is sailing close to the wind and will find you can't meet the upper limit on those old school converters (10V or 10.5V max).   Basically the safety margin between the maximum supply voltage and the maximum device voltage is too close for any practical tolerances to work.   The only way out is to add series diodes to drop the voltage to the converter.  Then of course your converter voltage rails will be a little lower.

Instead of shorting out the rail you can design similar zener-based protection schemes which open a MOSFET switch when the supply is too high.   It doesn't get around the tolerance issue though.

A different angle is to use a regulator but then you have to make sure the output voltage is set low enough not to drop out, or the regulator has to be able to dropout without creating other problems.

Another way to buy more margin is to use switchers which handle higher voltages.  Some are 12V or 13V and the LT1054 is 16V (please check datasheets).  That does solve the tolerance issue.

[PRR]

A mature, well-engineered approach:
https://www.analog.com/en/products/ltc4381.html
https://www.analog.com/media/en/technical-documentation/data-sheets/ltc4381.pdf
As-is, it allows +32V on its load, for up to +72/-60V input. So not quite pedal-scale, and costs $4 in small quantity which some builders think is offensive. (Analog Devices is not about den/garage projects.)

[niektb]

I see you already have a PTC, why would you need a crowbar? You just need a Zener that is big enough to handle the short power surge (until the PTC trips which is often a few seconds) (and omit R1 ofc). And in the same solution you can probably remove D2 as the Zener also handles reverse polarity.

[antonis]

you can design similar zener-based protection schemes which open a MOSFET switch when the supply is too high.

It will also work quite well for PS inrush current limiting with the addition of a cap..

[bushidov]

Confidence is another issue. To the non-technical, the pedal might stop working for no obvious reason and mysteriously start working again later. Will the owner realize it isn't the pedal's fault?

That seems like a pretty good explanation.

I have a friend who is doing boutique pedals and he makes an overdrive that uses a TC1044, like what they have in Klon clones, and he was tired of his customers putting 18V power supplies on his circuit, which ultimately blew the Zener diodes on it, which meant the customer had to send in their pedals back to him to go fix. A crowbar with a resettable fuse means once the customer realizes, "hey this isn't working" and contacts my friend, they don't have to send it back for repair.

As of the tolerances, I probably shouldn't have used the MAX1044 for the example and used the TC1044 instead to illustrate. It can take up to a 12V supply. As of the noise, that's why I keep C1 as close to the anode of the SCR as possible to prevent false triggering (as added to the notes on the schematic). The same is true about keeping C2 as close to the gate as possible. Because the 1N5239 is a 9.1V Zener, but note that according to the datasheet that it can max out at around 9.555V (assuming I am using the ON Semiconductor version, per its datasheet), and the 800mV gate trigger of the 2N5064 SCR, that 10.35V maximum voltage will be enough to protect the TC1044 from over voltage with some room to spare.

As of current levels, I chose 100mA, as that's a good PTC fuse for stuff in guitar pedal range. Unless I am doing something more complicated or a digital pedal that requires a bit of oomph, I usually don't see pedals going much higher than 20 - 50 mA. The fuse is rated for less current than the SCR, so that should be fine for current tolerances.

A mature, well-engineered approach:

That's a pretty nifty switcher. I have used Analog Devices stuff for industrial electronics (mostly comparators and voltage references for PID controllers), but haven't used any of their switchers. In boost or overdrive, I would have to see how easy it would be to clean up the switching noise/harmonics.

[Rob Strand]

It will also work quite well for PS inrush current limiting with the addition of a cap..

Actually I put one of those MOSFET protection circuits in a medical product and I added caps to limit inrush/soft-start.  No stress on any parts and no MOSFET failures in 10 years.