LDRs as pots?

[tbryanh]

Binary weighting lets you have a 16 tap "pot" with only 4 relays, but this doesn't work for non-linear tapers. If you're not using this trick then I expect board sizes in the square feet too.

Remember, we are only building a preamp here as part of a rack-mount system, so even if the board is in the square feet, there is plenty of room for it.

We are building from scratch. My rack mount units will be 19" wide, 12" deep, and up to 5" high. I have plenty of room for relays.

I think the cost is more of an issue that anything else. A decent relay is probably $10.00 (I'm just guessing). That can add up pretty quick.

As I mentioned before, it might be the only reason Mesa made the LDR Pots is because it was cheap.

[tbryanh]

Cheap transistor crap for $2000.00

[Don Moose]

My rack mount units will be 19" wide, 12" deep, and up to 5" high. I have plenty of room for relays.

Um, 17" between the rails. Only the front panel is 19".

... don't want that to be a surprise for you late into layout.

[loudthud]

Here is a little something I quickly drew up. As a volume pot, 6 relays gives a first step of -36dB. Each relay adds 6dB to that so 8 relays would be -48dB. Not as good as I would like. Maybe there is a way to get better resolution on the low end like -60dB with 1 or 2dB steps.

[tbryanh]

Wow, the binary ladder looks awesome. I have never seen one before. So you push one resistor in below the “wiper” at the same time you pull the same value resistor out above the “wiper.” The total resistance between the ends of the “pot” stayed the same, but the “wiper” moved up. Killer.

This is much different than the Mesa Pot. (See Attached.) The Mesa Pot uses a string of resistors in series, which does not really appear good for converting to use with relays. (Too many relays are needed.) In the Mesa Pot, all of the outputs in device 10 and 11 (X0-X7) are in the high impedance state except for one. That output is at 5v, which selects the adjustment of the “wiper.” There are only 16 possible settings of the “wiper” in the Mesa Pot. Not much really. Of course this could be expanded if needed.

Is it possible to use one set of resistors for linear taper and a different set of resistors for log taper in the binary ladder? Or is it impossible to come up with a set of resistors for log taper in the binary ladder?

If it is impossible to come up with a set of resistors for log taper, if there is a high enough resolution in the ladder, maybe one set of resistors could simulate both linear or log taper. It might just be a matter of skipping some of the possible selections if log taper is desired. Increasing the resolution appears easy. Adding one relay doubles the resolution.

[tbryanh]

Is it possible to use one set of resistors for linear taper and a different set of resistors for log taper in the binary ladder? Or is it impossible to come up with a set of resistors for log taper in the binary ladder?

Well, to answer my own question, the values that are in the ladder now, 15K, 30K, 62K, 130K, 270K, and 560K cause an exponentioal taper, base 2 raised to some power. I don't know if this is the same as audio taper, but it sure is not linear taper.

It seems that linear taper could be made simply by using the same value of resistor in each position of the ladder.

Probably any taper that is desired could be made using the resistor ladder.

[Steve Conner]

No, those values cause a linear taper with steps of 15k. You're not moving a tapping point along a series string of them: you're switching binary-weighted combinations of them in and out of circuit.

I don't know whether the R-2R ladder is usable as a variable resistor. I've only ever used it as a voltage divider.

I'm also happy designing hybrid circuits where the tubes clip before the ICs do, not to mention transistor circuits with soft clipping that "fizz" even less than overdriven preamp tubes. I don't see why transistors, digital pots, DACs, whatever, should be banned from a modern amp design just because they're not tubes.

I am reminded of that guy who built an AC30 clone entirely out of op-amps and Zener diodes, and got bought out by Korg.

[Enzo]

And I would say in defense of JFETs, when you say they are in the signal path, that is correct, but in use as a switch, they act more like a resistor - maybe 200 ohms. They would not be used as an amplifier stage. Only thing on their gates would be a control voltage - on or off. How well they would work or not here would depend on a lot of things.

[Tiago]

And I would say in defense of JFETs, when you say they are in the signal path, that is correct, but in use as a switch, they act more like a resistor - maybe 200 ohms. They would not be used as an amplifier stage. Only thing on their gates would be a control voltage - on or off. How well they would work or not here would depend on a lot of things.

If going on switched resistor arrays, I would put my bet on that one.
Jfets or Multiplexers are worth to try them before going on the mechanical stressing relays.

But again, thats what digital potentiometers are, switched resistor arrays...

[tbryanh]

No, those values cause a linear taper with steps of 15k.

Given the resistors are 15K, 30K, 62K, 130K, 270K, and 560K, I think if you progess through the 64 combinations one at a time in sequence (from 15K to 30K, then from 30K to 45K, then from 45K to 62K, etc.), then the effect is linear taper as you say, but if you switch from 15K to 30K, then from 30K to 62K, then from 62K to 130K, then from 130K to 270K, and finally from 270K to 560K, the effect is exponential. Its all in how you progress through the ladder.

[loudthud]

I updated the schematic attached to my previous post for a little more generic view.

You can't taper the resistors because you will disturb the binary sequence which is what allows the reduction in the number of relays. One approach that would work is basically a relay that would switch in a 20dB attenuator such that the circuit still sees a pot with the same end to end resistance. The penality would be an increase in source resistance with the attenuator switched in.

Consider the attenuation of the standard Linear and 20dB Log pots. These are theoretical numbers, not measurements on actual parts. Log pots have some smoothing so the jump between 5 and 6 is not so extreme.


..............Linear Pot.............20dB Log Pot

10...........0dB.....................0dB
9...........-0.9dB.................-1.7dB
8...........-1.9dB.................-3.9dB
7...........-3.1dB.................-6.7dB
6...........-4.4dB.................-11.1dB
5...........-6.0dB.................-20dB
6...........-8.0dB.................-21.9dB
3...........-10.5dB...............-24.4dB
2...........-14dB..................-28.0dB
1...........-20dB..................-34.0dB
0...........infinite.................infinite


Notice how the the jumps get bigger at the bottom around the 1 and 2 positions. Even the log pot has a 6dB jump. Now here's some numbers from the 6 relay Binary Ladder pot.


63...........0dB
56..........-1.0dB
50..........-2.0dB
44..........-3.1dB
39..........-4.2dB
35..........-5.1dB
31..........-6.2dB
22..........-9.1dB
16..........-11.9dB
11..........-15.2dB
8...........-17.9dB
7...........-19.1dB
6...........-20.4dB
5........... -22.0dB
4...........-23.9dB
3...........-26.4dB
2...........-30.0dB
1...........-36.0dB
0...........infinite

So, when replacing a log pot, the binary ladder pot should work fairly well on settings where the analog pot setting was higher than 2 or so.

[tbryanh]

And I would say in defense of JFETs, when you say they are in the signal path, that is correct, but in use as a switch, they act more like a resistor - maybe 200 ohms. They would not be used as an amplifier stage. Only thing on their gates would be a control voltage - on or off. How well they would work or not here would depend on a lot of things.

This might be true because the 200 ohms on-resistance of the FET does not cause much voltage to drop across the FET. You don't have the luxuray of being fully on or fully off in the Mesa Pot. The Mesa Pot operates fully on and fully off and everything else inbetween. It is when the Mesa Pot is midrange, say 10K or 100K that it drops a large voltage accross itself. This large voltage is what causes the LDR to distort. Look at the Perkins Elmer datasheets.

[tbryanh]

If going on switched resistor arrays, I would put my bet on that one. Jfets or Multiplexers are worth to try them before going on the mechanical stressing relays.

But again, thats what digital potentiometers are, switched resistor arrays...

I would like to use FETs too since they probably do not distor when fully on or fully off, but it does not appear FETs can be use in the resistor ladder. The ladder requires the FETs to be installed in series with each other. If you hve 6 FETs in series, you might want all FETs on except for one in the middle. It does not appear possible for this condition to exist. Likewise you might want the one in the center to be on and the others to be off, another condition that does not appear possible.

Look at the Mesa Pot, it does not use a resistor ladder, it uses a sring of resistors connected in series. We are talking about using a resistor ladder, not a resistor string.

It might be that FETS could be used with a resistor string, but it does not appear FETs can be used with a resistor ladder.

I am not sure what you mean by the term "resistor array."

[tbryanh]

I am reminded of that guy who built an AC30 clone entirely out of op-amps and Zener diodes, and got bought out by Korg.

Sure, and lets remember how much money Korg made compared to what they would have made if they used expensive tubes and OTs, etc. in the circuit instead.

The public is seeing through the deception. The return to so-called vintage amps is a return to the fundamental principal that tubes are superior to transistors when it comes to guitar amplifiers.

[tbryanh]

I don't see why transistors, digital pots, DACs, whatever, should be banned from a modern amp design just because they're not tubes.

You can make a tube sound as bad as a transistor, but you cannot make a transistor sound as good as a tube.