Resistance required to effectively block signal?

[johnnyreece]

I'm slowly making my way through the planning stages of my SLO-inspired project. One thing I'm having trouble with is what vactrols I can use for channel switching. The majority of the ones in the circuit are in-line with the signal and rely on a high resistance to keep signal from one channel to bleeding into the other (one is a shunt to ground). My question is, how high does the resistance need to be to effectively keep the signal out? I'm aware of the Xvive vactrols being preferred replacements for many circuits, but $10/each is a little salty when I want to use five in one build. I'm not opposed to rolling my own, but again, finding the right LDR/LED combo is a bit overwhelming. I appreciate any insight out there!

[award70]

The Vactrol VTL5C1 has on off resistance of 50M.

[johnnyreece]

Yes, I was just curious how low the resistance could practically be while still keeping the signal inaudible. I would imagine that 50M is a bit overkill. Or, I'm completely wrong...

[nuke]

It sort of depends on the circuit.

Maybe we set a target of 60dB between the "on" and "off" signal in terms of RMS voltage. That's pretty much "off" for audio in a guitar amp switching scenario.

60dB is a difference of 1000x in voltage.

If you have 1v rms AC signal on the "on" state, you want to attenuate by a factor 1000, so you arrive at 1mv as the highest allowable floor to arrive at the desired attenuation. More attenuation would be more betterest of course.

If you're using common-cathode triode pre-amp stages, calculate the input impedance of the stage being fed. In general, the grid input impedance is roughly equal to the value of the grid resistor. The Miller capacitance comes into play as frequency rises, but will be kind of small at the frequencies of concern, and we're trying to go from on-to-off so sledgehammer calculus is good enough.

Now we look at the source impedance of the stage that feeds it. Assuming it is also a common-cathode triode, the source impedance is essentially the parallel of the plate resistor and the tube's plate resistance at its operating point. For typical 12AX7 Rp is around 68k and typically, 100k plate load resistor, so more or less, about 47k of source impedance, which appears as a resistor in series with a voltage source. (again, cowboy calculus).


So it is (stage-1) plate AC voltage X -> 47k-in-series -> 1M-to-ground -> (stage-2 grid).


Now we have a simple model to understand what we need from a vactrol to switch the effective signal on/off. We can put the vactrol to ground right in between 47k and 1M, and just do a simple DC resistance calculation. Eazy-peezy. We can add to the 47k, or lower the 1M to get the attenuation we want.

Generally, we want to put the vactrol to ground to turn signal off, and in series if we want to isolate the signal away from a shunt-to-ground so we can steer it somewhere else without cutting it too much.

Got a snippet of the proposed circuit?

[pdf64]

Why not use relays?

[Snicksound]

Why not use relays?

Indeed, Soldano no longer uses LDRs. Most likely he got that idea from Mesa, who got that idea from Fender. All of them use relays now.

That said, some will argue that the very very very low clean signal that mixes in with the overdrive signal is part of the sound... So if you want to pay 50$ more for that it's up to you. Mike Soldano himself doesn't think so.

Note that there are other changes in modern SLOs (so if you find that the new ones don't sound the same, it's probably not because of that)

SLO Clone has two versions of their layout: LDR and relays. So you don't even have to figure out yourself how make the change.

If you must have the bleed, just put a 50M resistor bypassing the relay. I bet you won't notice.

[johnnyreece]

Thanks for the info to chew on. To be honest, I've never done an amp with channel switching, so I was lazily trying to just keep with the schematic I was borrowing from. The version I was looking at switches a high resistance at both the preamp and the MV for each channel, plus, the OD channel shunts to ground. Seems like a lot of redundancy, but again, I've not done anything with switching like this before, so I was trying to not deviate too much. I do have some random relays hanging around. I believe I have 12v and 5v. I'll see about joining the SLO Clone board and getting that info. I appreciate the help!

[R.G.]

It might be more useful to use one of Toshiba's MOSFET-output photo couplers. An example is the TLP170 series from Toshiba. Toshiba calls these "photo-relays". If the LED is off, they are very high resistance. The datasheet shows a minimum of 60M ohm, and I suspect the typical is var larger. The on-resistance is about 20 ohms.

They're about US$1.50 at Mouser.
See
https://www.mouser.com/datasheet/3/105/ ... 069C77.pdf

[Snicksound]

It might be more useful to use one of Toshiba's MOSFET-output photo couplers. An example is the TLP170 series from Toshiba. Toshiba calls these "photo-relays". If the LED is off, they are very high resistance. The datasheet shows a minimum of 60M ohm, and I suspect the typical is var larger. The on-resistance is about 20 ohms.

They're about US$1.50 at Mouser.
See
https://www.mouser.com/datasheet/3/105/ ... 069C77.pdf

That's interesting! Ironically, wondering if the 1ms switching time is "too fast" for this use (LDRs were favored because of their slow switching that reduced popping, then they figured out how to switch without popping and they're back to relays...).

As an alternative to relays though, color me curious. Would save some board space, flexible voltage and very low current.

[johnnyreece]

It might be more useful to use one of Toshiba's MOSFET-output photo couplers. An example is the TLP170 series from Toshiba. Toshiba calls these "photo-relays". If the LED is off, they are very high resistance. The datasheet shows a minimum of 60M ohm, and I suspect the typical is var larger. The on-resistance is about 20 ohms.

They're about US$1.50 at Mouser.
See
https://www.mouser.com/datasheet/3/105/ ... 069C77.pdf

Not gonna lie, the price is enticing and the package looks sturdy. Unfortunately, I'm also an idiot with anything MOSFET (among other things, but that's another story). For the life of me, I can't see where it shows the off resistance on the datasheet. I see there's an off state output voltage of 60V, but I have no idea what that means in this case, either. No matter, they're so cheap, I've ordered ten of them for just over $18 after tax/shipping. Not bad! I plan to play with them and see if I can build this circuit in a temporary form. If it works, I'll use them in my build. I will have five spares, so if I fry five of them, I will still have five left for the build.

[R.G.]

That's interesting! Ironically, wondering if the 1ms switching time is "too fast" for this use (LDRs were favored because of their slow switching that reduced popping, then they figured out how to switch without popping and they're back to relays...).

Good question. I actually think that you could slow down the current ramp into and out of the LED and fade these. The internal structure is LEDs driving what amounts to solar cells to make a gate source voltage for the MOSFETs. Lower currents do seem to affect the conductance.

On the other hand, these have a really good isolation between the LED drive circuit and the MOSFET resistive elements. There's not going to be control voltage feedthrough to click. And there is no internal DC offset to click. This is very much not like CMOS analog switches.

Fast switching itself does offer click problems. The absolute worst for this is ... relays. They go from hugely, massively open to hard-metal contact instantly. The MOSFET switches in photocouplers are a kinder, gentler thing than relays.

[R.G.]

For the life of me, I can't see where it shows the off resistance on the datasheet. I see there's an off state output voltage of 60V, but I have no idea what that means in this case, either.

Yeah. I didn't find it either. I inferred-computed that based on the leakage at 60V being under 1uA. R = V/I says that it would be a 60 Mohm resistor. That is a specified minimum resistance in the sense that it's the most current at a high voltage. I think the real effective resistance will be much higher in real units.

[nuke]

Maybe do what we do with relays and physical switches that are sensitive to popping, and put a 10 megohm resistor across the switching terminals?

I think the opto-isolator is designed to switch as rapidly as possible to avoid operating the MOSFETs in the resistive range, and keep them either full-on or full-off to minimize dissipation through the MOSFET. It will be an "off the sheet" usage scenario to try to slowly ramp up the LED current in an attempt to slow the rise/fall time.

[johnnyreece]

Well, I got them in the mail and was able to do some testing. When off, the resistance was too high for my meter to measure. When on, it was around 28 ohms (70mA current approximately). It was hard to hook up. I didn't look at the specs very closely before ordering. These are tiny! The body is about 3mm x 5mm. The legs are very fragile. I broke one off the first piece I tried to test. It would seem that these *should* work if I can get them mounted properly.

[nuke]

Here ya go:

SO6- SDPI6 adapter.

https://www.ebay.com/itm/152377253239

ChipQuick makes a whole lot of nice adapters for SMD to pinned/breadboard work. Also low-temperature solder to make taking SMD's and plate-through-hole eazy-peezy to get done.