Timmy asymmetrical vref voltage divider

[Clint Eastwood]

This must have been asked before, although I couldn't find a relevant thread:
What exactly is the point of the Timmy overdrive using an asymmetrical (8.2k/10k) vref voltage divider?

[antonis]

IMHO, to offset upwards Vref to counterbalance voltage drop across 510k bias resistor due to input bias current..

[Rob Strand]

This must have been asked before, although I couldn't find a relevant thread:
What exactly is the point of the Timmy overdrive using an asymmetrical (8.2k/10k) vref voltage divider

A good number of opamps clip more symmetrically when Vref is a little above Vcc/2.

If you look at the Timmy the second stage has gain.   With the diodes limiting the input to the second stage the gain of the second stage is very close the clipping the second stage.

Things like the OCD push the second stage a bit harder.

I don't think people have experimented enough with the second stage gain.   Of course, some opamps will sound better than others when pushed.

[m4268588]

It was measured under such conditions. displayed in vdiv=2V.
(VCC=9.92V. NOT 9V!)

I haven't tested 1458 and NJM4559, use 741, 4558 instead.

μa741(TI)	NJM4558

IMHO, to offset upwards Vref to counterbalance voltage drop across 510k bias resistor due to input bias current..

The differential input device of NJM4559 is PNP. In that case, will be offset upwards.

If NFB resistance of the second-stage is smaller, offset will be more than expected.

[Rob Strand]

The differential input device of NJM4559 is PNP. In that case, will be offset upwards.

If NFB resistance of the second-stage is smaller, offset will be more than expected.

The offset on the first opamp due to the 510k should be less than 0.1V.    Since the feedback pot is 1M, if the pot is set to the center then on average the offsets from the 500k and 510k should cancel.


Interestingly, the Timmy can be powered from battery.  So when the battery goes flat the chances of the second opamp clipping goes up and that should change the tone.   I don't think many people come across that because most people use power supplies these days.

[Clint Eastwood]

Thanks for the replies folks.

It was measured under such conditions. displayed in vdiv=2V.
(VCC=9.92V. NOT 9V!)

I haven't tested 1458 and NJM4559, use 741, 4558 instead.

μa741(TI)	NJM4558

IMHO, to offset upwards Vref to counterbalance voltage drop across 510k bias resistor due to input bias current..

The differential input device of NJM4559 is PNP. In that case, will be offset upwards.

If NFB resistance of the second-stage is smaller, offset will be more than expected.

Ok, but one could also in stead of connecting the feedback network of the second stage to Vref, connect it to ground via a capacitor. That would also give very little offset.
Is this just a way to use one less capacitor, or would there be a sonic difference with the feedback to Vref?

[m4268588]

I don't think good.

Code Select Expand

Version 4
SHEET 1 800 600
WIRE 432 -336 384 -336
WIRE 384 -320 384 -336
WIRE 48 -304 32 -304
WIRE 128 -304 112 -304
WIRE 208 -304 128 -304
WIRE 336 -304 208 -304
WIRE 128 -288 128 -304
WIRE 208 -288 208 -304
WIRE 432 -288 432 -336
WIRE 448 -288 432 -288
WIRE 320 -256 272 -256
WIRE 336 -256 320 -256
WIRE 576 -256 528 -256
WIRE 592 -256 576 -256
WIRE 272 -240 272 -256
WIRE 528 -240 528 -256
WIRE -16 -224 -96 -224
WIRE 432 -224 432 -288
WIRE 480 -224 432 -224
WIRE 576 -224 576 -256
WIRE 592 -224 576 -224
WIRE -96 -208 -96 -224
WIRE 128 -192 128 -208
WIRE 128 -192 112 -192
WIRE 320 -192 320 -256
WIRE 336 -192 320 -192
WIRE 432 -192 432 -224
WIRE 432 -192 416 -192
WIRE 320 -176 320 -192
WIRE 480 -176 464 -176
WIRE -16 -160 -16 -224
WIRE -96 -112 -96 -128
WIRE -64 -112 -64 -144
WIRE -64 -112 -96 -112
WIRE 464 -112 464 -176
WIRE 480 -112 464 -112
WIRE 576 -112 576 -224
WIRE 576 -112 560 -112
WIRE -96 -96 -96 -112
WIRE 32 -96 32 -304
WIRE 464 -96 464 -112
WIRE -64 -80 -64 -112
WIRE 320 -80 320 -96
WIRE -96 0 -96 -16
WIRE -64 0 -64 -16
WIRE -64 0 -96 0
WIRE -16 0 -16 -80
WIRE -16 0 -64 0
WIRE 32 0 32 -16
WIRE 32 0 -16 0
WIRE 208 0 208 -208
WIRE 208 0 32 0
WIRE 272 0 272 -160
WIRE 272 0 208 0
WIRE 320 0 320 -16
WIRE 320 0 272 0
WIRE 464 0 464 -16
WIRE 464 0 448 0
WIRE 32 16 32 0
FLAG 32 16 0
FLAG 32 -304 Sig
FLAG -16 -224 VCC
FLAG -64 -144 Vref
IOPIN -64 -144 Out
FLAG 384 -240 0
FLAG 528 -160 0
FLAG 112 -192 Vref
IOPIN 112 -192 In
FLAG 592 -256 Out
IOPIN 592 -256 Out
FLAG 448 0 Vref
IOPIN 448 0 In
DATAFLAG -96 -112 ""
DATAFLAG 448 -288 ""
DATAFLAG 592 -224 ""
SYMBOL signal 32 -112 R0
WINDOW 123 0 64 Invisible 2
WINDOW 39 2 104 Left 2
SYMATTR SpiceLine Rser=0
SYMATTR InstName V1
SYMATTR Value SINE({Ofs} {Level} {Freq})
SYMBOL voltage -16 -176 R0
SYMATTR InstName V2
SYMATTR Value {VCC}
SYMBOL res -112 -224 R0
WINDOW 0 6 40 Right 2
WINDOW 3 6 56 Right 2
SYMATTR InstName R100
SYMATTR Value 8k2
SYMBOL res -112 -112 R0
WINDOW 0 6 40 Right 2
WINDOW 3 6 56 Right 2
SYMATTR InstName R101
SYMATTR Value 10k
SYMBOL cap -80 -80 R0
SYMATTR InstName C100
SYMATTR Value 47u
SYMBOL cap 48 -288 R270
WINDOW 0 32 46 VLeft 2
WINDOW 3 0 18 VRight 2
SYMATTR InstName C1
SYMATTR Value 47n
SYMBOL res 112 -304 R0
SYMATTR InstName R1
SYMATTR Value 510k
SYMBOL e 384 -336 R0
SYMATTR InstName E1
SYMATTR Value 100k
SYMBOL current 208 -208 R180
WINDOW 0 -6 48 Right 2
WINDOW 3 -2 32 Right 2
SYMATTR InstName I1
SYMATTR Value {Ib}
SYMBOL current 272 -160 R180
WINDOW 0 -6 48 Right 2
WINDOW 3 -2 32 Right 2
SYMATTR InstName I2
SYMATTR Value {Ib}
SYMBOL res 432 -208 R90
WINDOW 0 8 56 VBottom 2
WINDOW 3 24 56 VTop 2
SYMATTR InstName R2
SYMATTR Value {Gain}
SYMBOL res 304 -192 R0
SYMATTR InstName R3
SYMATTR Value 3k3
SYMBOL cap 304 -80 R0
SYMATTR InstName C2
SYMATTR Value 47n
SYMBOL e 528 -256 R0
SYMATTR InstName E2
SYMATTR Value 100k
SYMBOL res 576 -128 R90
WINDOW 0 8 56 VBottom 2
WINDOW 3 24 56 VTop 2
SYMATTR InstName R4
SYMATTR Value {Rf}
SYMBOL res 448 -112 R0
SYMATTR InstName R5
SYMATTR Value {Rf}
TEXT 0 56 Left 2 !.OP\n*.TRAN 0 50m 48m ;25u\n \n.Param VCC=9V OFS=0V Level=100mV Freq=1kHz\n+ Rf=3k3 Ib=25n Gain=1/(1/500k+1/3k3)+500k\n.STEP Param Gain 1m 1Meg\n+ Param Ib List 25n 500n

[Clint Eastwood]

don't think good.

I'm afraid I don't understand you last post..

[amptramp]

If you take a look at the specs for the popular TL07X series, the common-mode input voltage range when operating from ±15 VDC is worst case ±11 VDC but the typical case is +15 to -12 VDC so moving the Vref voltage above Vcc/2 is reasonable.  Other op amps may have different centres of common-mode input range, so it has to be tailored to the device used.

Note also that for the TL07X, this implies that the worst case common-mode input range when operating from a 9 VDC supply would be +4 to +5 VDC.  It's a good thing we don't run into worst case conditions all that often.

[Rob Strand]

I'm afraid I don't understand you last post..

It's the main parts of the schematic redrawn with two Ib current sources added.

The idea is to show the effect of the opamp's input bias current on the DC bias voltage at output.

The current sources Ib are added to each input of the first opamp.  The current is stepped over the range of bias currents in the datasheet then the effect is seen on the DC voltage at the output.  Actually  the two plots show a different values of Ib and the gain pot value is stepped.

The DC output will change depending on the gain pot setting.  The AC gain changes with the pot but the DC gain does doesn't, it stays at unity.  What changes with the pot setting is how much Ib cancels, it cancels out when the gain pot equal 510k.

This comes under the topic of opamp input bias currents.