And here's a plot of that data, using the 4 top units (i.e. 2,9,7,4) and the 4 bottom units (8,10,15,20) to show the spread.
The X axis is log current, and the Y log Resistance
Notice how tight they are at higher currents, and more spread at lower currents. Also how one or two units are "weird" and don't follow the general trend.
The X axis is log current, and the Y log Resistance
Notice how tight they are at higher currents, and more spread at lower currents. Also how one or two units are "weird" and don't follow the general trend.
And here's the attenuation and channel mismatch calculations. The attenuation was calculated by using the first series data point and the last shunt data point, and continuing, i.e. 2nd series data point and 2nd to last shunt point, etc. Formula used - attenuation = shunt R/(series R + shunt R)
The data at bottom swaps the series and shunt columns, to show you how the results would be if you swapped the L and R series LDR units in the volume control. It's valuable to try that.
This data was not in any way the best I could do, I just grabbed a few to LDRs to develop and test the spreadsheet.
One thing is quickly apparent - you really don't need a matched quad of LDRs.
What you need are a series matched pair, and shunt matched pair, that can possibly be selected to compliment each other for best match throughout the range, or at least for where it is most important. A tool like this really helps to do that.
The data at bottom swaps the series and shunt columns, to show you how the results would be if you swapped the L and R series LDR units in the volume control. It's valuable to try that.
This data was not in any way the best I could do, I just grabbed a few to LDRs to develop and test the spreadsheet.
One thing is quickly apparent - you really don't need a matched quad of LDRs.
What you need are a series matched pair, and shunt matched pair, that can possibly be selected to compliment each other for best match throughout the range, or at least for where it is most important. A tool like this really helps to do that.
Last edited:
These are measured using current source or voltage source supply? I assume current source.
It seems that if one point was chosen for a match, possible that channel balance can be maintained in reasonable range.
It seems that if one point was chosen for a match, possible that channel balance can be maintained in reasonable range.
Last edited:
Measurement incremental entire range of Shunt/Series from Off, Channel A is first set, with a second LDR Left and right measured as Channel B, and showing slightly better still L/ R balance
This shows there may be advantage in starting the volume setting slightly later akin to log response, as godfrey raised and I suggested 33k being a suitable value running from anodes of the top set to gnd.
I will run through another group of measurements with resistor in place, to check if variances L and R tighten up even further still.
Cheers / Chris 🙂
This shows there may be advantage in starting the volume setting slightly later akin to log response, as godfrey raised and I suggested 33k being a suitable value running from anodes of the top set to gnd.
I will run through another group of measurements with resistor in place, to check if variances L and R tighten up even further still.
Cheers / Chris 🙂
To everyone who has enjoyed this controversial thread, ... a BIG Thank You.
The addition of a resistor anode to gnd upsets channel balance. So follow schematic at this post and at post 215 It achieves good if not very good result with channel balance, Volume rate on is quite nice to use not too abrupt. Transistors are BC547 or BC847 surface mount. Voltage can be raised to Vout of LM317 U1 to 5.45v by using an additional 220R ie 1.25 x 740/220 + 1 rather than 520/220.
I recommend use with Silonex NSL32SR2 LDR's that are matched pairs for Left- Right Shunt (2) and matched pairs Left- Right Series (2) Op amps are TL072, however NE5532 and LF353 , also appear to function with very early tests. ( Finally a great use for these ancient models 🙂 )
PSU for op amp supplies,I recommend cap of at least 220uf and the usual bypassing necessary for op amps. The circuit works very well with a NPN based cap multipliers for each op amp supply, And a plugpack supply single rail DC positive of 12v @ preferably more than 200ma
Cheers / Chris 🙂
The addition of a resistor anode to gnd upsets channel balance. So follow schematic at this post and at post 215 It achieves good if not very good result with channel balance, Volume rate on is quite nice to use not too abrupt. Transistors are BC547 or BC847 surface mount. Voltage can be raised to Vout of LM317 U1 to 5.45v by using an additional 220R ie 1.25 x 740/220 + 1 rather than 520/220.
I recommend use with Silonex NSL32SR2 LDR's that are matched pairs for Left- Right Shunt (2) and matched pairs Left- Right Series (2) Op amps are TL072, however NE5532 and LF353 , also appear to function with very early tests. ( Finally a great use for these ancient models 🙂 )
PSU for op amp supplies,I recommend cap of at least 220uf and the usual bypassing necessary for op amps. The circuit works very well with a NPN based cap multipliers for each op amp supply, And a plugpack supply single rail DC positive of 12v @ preferably more than 200ma
Cheers / Chris 🙂
Last edited:
Yes, measured with a variable current source, at 10 current levels.
Yes, once the measurements are in excel, you can sort on any current value to arrange the parts close to each other at that current value.
The requirement of sensible unloaded input impedance 10k to 15k is met by shunt op amp receiving 2.7ma and Series op amps ( both TL072 dual ) a darlington diode drop away.
4.4k resistance arriving at cap multiplier base then sets total limit of current approx 2 ma through potentiometer and LDR's, This current limit can be neatly arranged by making Shunt op amp V+ available via connection to base of cap multiplier transistor, and op amp arranged to feed the series LDR anodes by the emitter of that cap multiplier. 2N6427 recommended ( amazing hfe )
In practice measurement between the anodes of series and shunt LDR shows that when the potentiometer is varied a polarity change occurs, hence the cap polarity is negative at Shunt and cap polarity positive at Series LDR
More complex arrangement of cathodes being connected to negative of inverting op amps
requires current shunt transistors as in Model 7b published earlier. This schematic has a nice simplicity 🙂
NE5532 are being tested for suitability.
Cheers / Chris
4.4k resistance arriving at cap multiplier base then sets total limit of current approx 2 ma through potentiometer and LDR's, This current limit can be neatly arranged by making Shunt op amp V+ available via connection to base of cap multiplier transistor, and op amp arranged to feed the series LDR anodes by the emitter of that cap multiplier. 2N6427 recommended ( amazing hfe )
In practice measurement between the anodes of series and shunt LDR shows that when the potentiometer is varied a polarity change occurs, hence the cap polarity is negative at Shunt and cap polarity positive at Series LDR
More complex arrangement of cathodes being connected to negative of inverting op amps
requires current shunt transistors as in Model 7b published earlier. This schematic has a nice simplicity 🙂
NE5532 are being tested for suitability.
Cheers / Chris
Last edited:
All the LDR anodes are always positive with respect to ground. The cap connected to the shunt LDRs should be reversed.
Hi Godfrey
Thanks Yes agreed with respect to ground, I am away at the moment but will test any differences in approx 7 days, particularly with channel balance arising from implementing either way and report.
Cheers / Chris
Chris,
Sorry, but your circuit seems to be missing links or somethimg - can't figure out where the current from the strange C-muliplier (?) powers the circuit, and if the 50k pot distributes the current to the leds, what do the ICs actually do?
Also, could you turn the upside down parts of the circuit up the right way - might make it easier to understand it's operation.
Sorry, but your circuit seems to be missing links or somethimg - can't figure out where the current from the strange C-muliplier (?) powers the circuit, and if the 50k pot distributes the current to the leds, what do the ICs actually do?
Also, could you turn the upside down parts of the circuit up the right way - might make it easier to understand it's operation.
There was a lot of heated discussion about his circuits earlier in the thread. Let's just say that Chris likes to use opamps in [ahem] highly unconventional ways.
This is actually the least bizarre circuit I've seen in this thread so far. If you think that's a wild claim, have a look at the circuit in post 3, for example.
This is actually the least bizarre circuit I've seen in this thread so far. If you think that's a wild claim, have a look at the circuit in post 3, for example.
Last edited:
Yeah Godfrey, seems like an exercise in a SIM program, like the guys used to do with those craziest series regulators that were never worked well in practice.
At least with Uriah's (LighterNote) , Paul's (CCS & remote), Alan's (W/speed), ZenMod's, etc, there is a clear current path to follow and each cct advantage can be seen -
Here, not sure how the ICs function - current buffers?, current sources?
What's the advantage/improvements of the extra components in the power supply? Something new?
At least with Uriah's (LighterNote) , Paul's (CCS & remote), Alan's (W/speed), ZenMod's, etc, there is a clear current path to follow and each cct advantage can be seen -
Here, not sure how the ICs function - current buffers?, current sources?
What's the advantage/improvements of the extra components in the power supply? Something new?
You are making the mistake of assuming that these circuits operate in a conventional way, with a conventional explanation based on circuit theory. Here it is all Alice in Wonderland stuff, such as circuits which take an output from the input pins of opamps (presumably via the protection diodes).
The advantage of this method is to provided approx 2ma from the base of the darlington to the shunt op amp V+ that then controls the total current maximum varied by the potentiometer to the shunt and series LDR anodes.
Thats all the current that is needed in this configuration to then arrive at input impedance of 10- 15k.The small amount of current provided is much kinder on potentiometers.
The quest is simplicity with excellent L/R channel difference.
Cheers / Chris
Thats all the current that is needed in this configuration to then arrive at input impedance of 10- 15k.The small amount of current provided is much kinder on potentiometers.
The quest is simplicity with excellent L/R channel difference.
Cheers / Chris
"Providing current from the base of the Darlington"
On Planet Earth we take current from the collector or emitter of a Darlington, but on Chris's planet you can take output from the base of a Darlington? I guess this fits quite well with taking output from the input pins of opamps. Consistency is a good design principle!
On Planet Earth we take current from the collector or emitter of a Darlington, but on Chris's planet you can take output from the base of a Darlington? I guess this fits quite well with taking output from the input pins of opamps. Consistency is a good design principle!
It works well, delivering current to the shunt op amp, whereas the emitter feeds the series op amp. Current then is no more than 2ma. Try it and see.🙂
Cheers / Chris
If you don't mind I will pass on that one.
I mainly use valve circuits now - should I try to take output from their control grids? (Yes, I have heard of one circuit which does this for very particular reasons).
What a great idea!!! You mean something like this?
Channel balance should be superb with the midpoints of the heaters tied together, and used to bias the Darlington. Current control to the Darlington is regulated by two of the cathodes, while the output of the other two is nicely smoothed by the RC filter.
It should sound wonderful, I can't wait to build it!
Channel balance should be superb with the midpoints of the heaters tied together, and used to bias the Darlington. Current control to the Darlington is regulated by two of the cathodes, while the output of the other two is nicely smoothed by the RC filter.
It should sound wonderful, I can't wait to build it!
Attachments
Last edited:
Actually, I was thinking of a circuit which used an 'inverted 6AU6' (IIRC). Used grid as anode and anode as grid. Something to do with astonishingly high input impedance for an electrometer, as the anode is well away from the cathode space charge. That was a genuine circuit. I can't remember where I saw it.
- Status
- Not open for further replies.