Good point, a visual graph, plotting balance and gain position curve would be nice to see.
Do you suggest I achieve this with straight DMM measurement, which may take some time
for all resistance points to assemble a plot for each channel, or perhaps I could generate a sine wave tone, play it through CD player and get each channel showing on scope, that would be asked to show relative amplitude....for attenuator position, requiring a dual scope trace, and recording same as a short movie as attenuator transgresses range. I could make said movie available at my blog site, as diy audio file size would I think be insufficient. I would love one day a audio precision one... other commitments first though.
Cheers / Chris
Last edited:
A 2x pot version is quite possible, but from what myself and a friend are hearing with this present version - ie the earlier model ( ... reminder must update schematic here )
would take very fine movement on each channel level though to get it as good as the single pot version. six of one, half a dozen of the other, ...but I am able to accommodate what people might need.🙂
Cheers / Chris
would take very fine movement on each channel level though to get it as good as the single pot version. six of one, half a dozen of the other, ...but I am able to accommodate what people might need.🙂
Cheers / Chris
Last edited:
11 points equally spaced give a pretty good idea. There are pots that already have stepping mechanism built in, and should be quite adequate. PC based scopes are adequate for visual comparison, and you can generally monitor at least two channels at the same time.
Currently I think you base lots of claims on listening only, this is very unreliable because our ears are not even perfectly balanced, and may even change from time to time.
Chris,
Quickest and easiest and probably most revealing would be just resistance numbers, from straight DMM measurement, for all four CdS resistances, in circuit but with nothing connected to input or output.
First, just do it with the volume knob approximately centered and report the four resistance numbers, here, please.
Even if your circuit somehow gets the LED currents matched, you really can't get around the need for matching the LED/LDR combinations between channels.
One set of four resistances might not show the problem. But if it doesn't, one or two more sets from about a quarter-turn away on the knob will show it, unless your "not-extensive matching" is better than you thought.
Cheers,
Tom
Quickest and easiest and probably most revealing would be just resistance numbers, from straight DMM measurement, for all four CdS resistances, in circuit but with nothing connected to input or output.
First, just do it with the volume knob approximately centered and report the four resistance numbers, here, please.
Even if your circuit somehow gets the LED currents matched, you really can't get around the need for matching the LED/LDR combinations between channels.
One set of four resistances might not show the problem. But if it doesn't, one or two more sets from about a quarter-turn away on the knob will show it, unless your "not-extensive matching" is better than you thought.
Cheers,
Tom
Hi Tom
Yes this is quite revealing measuring at approx half volume, unloaded.
Series and shunt behave in opposite fashion to each channel as I mentioned in an earlier post through the attenuation range, correcting themselves to behave as an aggregate, as I understand as they go.
. so first off here is the aggregate for LEFT 33.8k
RIGHT 30.6k
made up of LEFT 19.9k series 13.86 shunt
RIGHT 22.16 series 8.50 shunt
So indicates overall channel imbalance of 3200 ohms, at half volume setting unloaded
how does this fare to other LDR circuits you have measured unloaded at half volume or close to being set at 21k series ? , or other attenuators. ?
Cheers / Chris
Last edited:
The series/shunt attenuators are simple resistive voltage dividers. The whole input voltage is across both resistors. but the output voltage is taken from across only the shunt resistor. So the output voltage is a portion of the input voltge, with the portion being = Rshunt / (Rseries + Rshunt).
For your left channel, the output voltage would be about 41% of the input voltage.
For your right channel, the output voltage would be about 28% of the input voltage.
So the left channel's output voltage is 3.3 dB higher than the right channel's output voltage.
I assume that you did not move the knob, until after all of the measurements were taken.
In reality, the input impedances of your amp channels would be in parallel with the shunt resistances, which would change the attenuation amounts, somewhat. But the large mis-match points out the need for matched LED/LDR devices.
Edit: With a 25 kOhm amp input impedance in parallel with each shunt resistor, the left and right attenuation factors would be 31% and 22%. i.e. Left would be 2.98 dB more than right.
Last edited:
I wonder how much improvement one could get with the existing unmatched devices by simply matching the LED currents between channels. At the moment the currents are not matched.
I believe matched devices are matched by having same current. I may be wrong, but even though the matching process uses voltage source, they are let to stabilize before the value is logged. There are many issues involved with sound quality. I remember just replacing a linear pot with a log pot of same value, and the sound was really lifeless.
From a balancing point of view, the LDR balance shift can occur when you turn the volume, that means unless you don't mind the balance, which has a great effect on sound imaging, you probably may find it enjoyable depending on how much you think it's worth in $.
From a balancing point of view, the LDR balance shift can occur when you turn the volume, that means unless you don't mind the balance, which has a great effect on sound imaging, you probably may find it enjoyable depending on how much you think it's worth in $.
Last edited:
Thanks for everyones contributions, so far 🙂
Tom has pointed to some valid but difficult ways to achieve sensing which is needed for yet better L/R ability, other than lots of matching of LDR devices.
I think a potentiometer for each channel is the logical way to satisfy any disagreement about channel balance... despite the other single potentiometer being subjectively very good,
My measurements show voltage remains active at 1.4v even at maximum attenuation - which would indicate and current also measured at .054ma, that if voltage is present at the anode then current is also active to keep the LED element possibly zoned in to its task, and capable of responding more quickly .
So current is enabled far more readily than with traditional LDR offerings as a result of voltage being on. and ready to go... so despite channel balance issues that can be solve with separate attenuators for each channel ,there is merit in this way of doing LDR control.
How does it achieve 1.4v, because this is the voltage given up by single rail op amp use, that remains present
So onward to a L attenuator, R attenuator version... is this what people need ?
Cheers / Chris
Last edited:
If current matching is your claim to better sound, then how does your schematic accomplish that better than my simple series wired LED's? Current must be equal in series wired LED's despite voltage. I am not convinced that your claim of lower %THD via constant current is accurate, and I think you have forgotten Ohms law.
Current is available to the LDR anode cathode in my design as a result of voltage always being present at approx 1.4v with .054ma and upward to approx 1.7v to max of 21ma as attenuator level is decreased . For comparison can you provide voltage in your design at maximum attenuation.. lowest volume , what voltage occurs at that point ? and what current occurs ?
Ohms law would tell us that V/R =I therefore if as i suspect there is no voltage, then there can be no current either.
My design accomplishes that the LDR is enabled to function more readily and to almost instantly start processing current, rather than having to slowly turn on as a result of voltage depletion
Regarding THD this is measured by Silonex relating to Signal voltage, rather than voltage to anode and cathode, sorry for any confusion this may have caused.
Cheers / Chris
Last edited:
Some people like for the channels to be matched within 1db gain or smaller so that they can have near perfect balance all the time, some people that have problems with channel balance elsewhere in the system probably would like the capability for some flexible means to adjust balance (I like that when I had a system that was dying). One of the preamps (Audible Illusions) I have has two volume controls, I did not enjoy having to adjust the balance every time I change a volume setting. LDRs are even more difficult because the curve variation can make the balance different at a different volume setting.
I think this statement could present confusion. The need to slowly turning the volume is not because the current lag, but rather it is the current to resistance response. LEDs respond pretty fast, that is why they can be used in display systems. The response characteristics of the light sensitive resistance material is where the slow response is. I do not believe that faster current onset will have effect. But if you are interested in elaborating more how your design overcomes this limitation, I am all ears; of course measured data will also help clarify the issue.
In the series wired LED arrangement, if voltage were to drop to 0V then the optocoupler would be turned off and resistance would shoot to over 1M. This is not the case. Typical input resistance is around 50K even at the lowest volume setting.
You appear to be saying here that the LED can respond more quickly to a need to produce light because it is never switched completely off. Who cares how quickly the LED responds? It is merely shining some light onto an LDR. The LED is not handling the audio signal, the LDR is. The speed of response of the LED is completely irrelevant, and in any case will be very fast. You can signal with an LED at RF frequencies, so even if it was relevant it would be fast enough without any nonsense about responding quickly. Pure moonshine.Chris Daly said:
Your channel balance has now been exposed (by your own measurements) to be worse than a cheap pot.
So let me get this clear: you have a complex circuit requiring some opamps (and a PSU) which drives an LDR attenuator which like all such devices has poor channel matching and introduces some audio distortion, so is worse than a cheap volume pot, yet we are supposed to regard this as an advance? Even worse, I now realise you are trying to flog this stuff to other people?
Do you believe the nonsense you write, or do you just expect your customers to believe it?
Hi Soongsc
Good points, there are some good ideas that Tom raised in post 12 about sensing voltage and current across the signal side to then adjust anode and cathode, this sort of crosses that audio barrier of leaving the two separate, but Tom also discussed having a known table of resistance that could be reflected on with anode and cathode to provide a ruler of sensing,
If we can get that magic 1db channel balance or less.. would be great.. and individual matching is certainly a valid method.
I agree with you about two volume controls they have some user issues for some people.
I consider it better to have voltage available at all attenuation points and my circuit manages variance of just .340v on each half with slight difference to each, A LDR that goes from totally off to partially on must be less than ideal as it is called with its anode and cathode, to be acting similar to a switch.
Cheers / Chris
The opamps aren't required, they just get in the way. Have a close look at the circuit. It's actually quite hilarious. The way the opamps are wired up, they don't have a snowball's chance of doing anything useful.
DF96 you need to calm down a bit. voltage is voltage current is current, cheap volume pots are cheap volume pots we are simply discussing if progress can be made..., maybe it cannot, but we are trying. 🙂
Cheers / Chris
Grab yourself a TLO72 connect it up to a 12v+ single supply and measure voltage at the output pin 1 and pin 7 and voltage at negative in pin 2 and 6, when positive inputs and V-are grounded. Now see how much current can be drawn across the inverting input to ground ?
when a feedback path is connected, and provide that measurement back here. also measure impedance inverting input to output.
And no external voltages other than that arriving at Pin 8 and Gnd Pin 4., I think you may find it is a useful circuit after all. 🙂
Cheers / Chris
Last edited:
- Status
- Not open for further replies.