4 transistor discrete opamp (LTspice problem or circuit prob

[Bill Bowden]

Why doesn't this discrete 4T opamp run in LTspice at higher frequencies? It
looks ok at 1Khz, but I get a 5 meg oscillation at 10K or more. There are no
oscillations with the input disconnected.

Version 4
SHEET 1 880 680
WIRE 128 -112 32 -112
WIRE 192 -112 128 -112
WIRE 400 -112 192 -112
WIRE 512 -112 400 -112
WIRE 400 -96 400 -112
WIRE 32 -64 32 -112
WIRE 128 -16 128 -32
WIRE 128 -16 96 -16
WIRE 192 -16 192 -32
WIRE 304 -16 192 -16
WIRE 400 16 400 -16
WIRE 576 16 400 16
WIRE 640 16 576 16
WIRE 400 32 400 16
WIRE 32 80 32 32
WIRE 304 80 304 64
WIRE 304 80 32 80
WIRE 336 80 304 80
WIRE 128 112 128 -16
WIRE 192 112 192 -16
WIRE 512 112 480 112
WIRE 304 128 304 80
WIRE 480 144 480 112
WIRE -64 160 -96 160
WIRE 48 160 16 160
WIRE 64 160 48 160
WIRE 512 192 512 112
WIRE 160 208 128 208
WIRE 192 208 160 208
WIRE 256 208 256 160
WIRE 160 224 160 208
WIRE 304 304 304 208
WIRE 304 304 160 304
WIRE 400 304 400 128
WIRE 400 304 304 304
WIRE 512 304 512 272
WIRE 512 304 400 304
WIRE 48 352 48 160
WIRE 272 352 48 352
WIRE 576 352 576 16
WIRE 576 352 352 352
FLAG 512 -32 0
FLAG -96 240 0
FLAG 256 208 0
FLAG 480 144 0
FLAG 640 96 0
SYMBOL voltage 512 -128 R0
WINDOW 123 0 0 Left 0
WINDOW 39 0 0 Left 0
WINDOW 0 39 60 Left 0
WINDOW 3 40 92 Left 0
SYMATTR InstName V1
SYMATTR Value 15
SYMBOL voltage -96 144 R0
WINDOW 123 0 0 Left 0
WINDOW 39 0 0 Left 0
WINDOW 0 -77 57 Left 0
WINDOW 3 -110 143 Left 0
SYMATTR InstName V2
SYMATTR Value SINE(0 .5 1k 0 0 0 10)
SYMBOL npn 64 112 R0
WINDOW 3 -1 119 Left 0
SYMATTR Value 2N3904
SYMATTR InstName Q2
SYMBOL npn 256 112 M0
WINDOW 0 40 33 Left 0
SYMATTR InstName Q3
SYMATTR Value 2N3904
SYMBOL res 144 208 R0
SYMATTR InstName R1
SYMATTR Value 1k
SYMBOL res 32 144 R90
WINDOW 0 0 56 VBottom 0
WINDOW 3 32 56 VTop 0
SYMATTR InstName R4
SYMATTR Value 1k
SYMBOL res 368 336 R90
WINDOW 0 60 116 VBottom 0
WINDOW 3 32 56 VTop 0
SYMATTR InstName R5
SYMATTR Value 10k
SYMBOL npn 336 32 R0
WINDOW 3 77 31 Left 0
WINDOW 0 76 59 Left 0
SYMATTR Value 2N3904
SYMATTR InstName Q5
SYMBOL res 112 -128 R0
WINDOW 0 29 88 Left 0
WINDOW 3 -45 47 Left 0
SYMATTR InstName R3
SYMATTR Value 470
SYMBOL pnp 96 32 R180
WINDOW 0 84 13 Left 0
WINDOW 3 35 44 Left 0
SYMATTR InstName Q4
SYMATTR Value 2N3906
SYMBOL res 176 -128 R0
WINDOW 0 43 82 Left 0
WINDOW 3 39 45 Left 0
SYMATTR InstName R2
SYMATTR Value 470
SYMBOL res 288 112 R0
SYMATTR InstName R8
SYMATTR Value 360
SYMBOL res 384 -112 R0
WINDOW 0 -52 65 Left 0
WINDOW 3 -38 29 Left 0
SYMATTR InstName R11
SYMATTR Value 1k
SYMBOL voltage 512 176 R0
WINDOW 123 0 0 Left 0
WINDOW 39 0 0 Left 0
WINDOW 0 -71 36 Left 0
WINDOW 3 -67 71 Left 0
SYMATTR InstName V3
SYMATTR Value 15
SYMBOL res 288 -32 R0
WINDOW 0 -44 50 Left 0
WINDOW 3 38 54 Left 0
SYMATTR InstName R6
SYMATTR Value 12k
SYMBOL res 624 0 R0
SYMATTR InstName R7
SYMATTR Value 1k
TEXT -152 328 Left 0 !.tran .01s



--- news://freenews.netfront.net/ - complaints: news@netfront.net ---

 

Technically you got the noninverting input grounded and are using the inverting. sopmething has to stop it. You don't have a cap to do that.

Look up "Miller capacitor" and hopefuly that (WIKI maybe) should explain it. If not, someone around here should be articulate enough.

 

[Jim Thompson]

On Thu, 5 Jun 2014 19:00:20 -0700, "Bill Bowden"
<bperryb@bowdenshobbycircuits.info> wrote:

Why doesn't this discrete 4T opamp run in LTspice at higher frequencies? It
looks ok at 1Khz, but I get a 5 meg oscillation at 10K or more. There are no
oscillations with the input disconnected.

Version 4
SHEET 1 880 680
WIRE 128 -112 32 -112
WIRE 192 -112 128 -112
[Full listing at Message-ID: <lmr7ai$1u4v$1@adenine.netfront.net>]
SYMBOL res 624 0 R0
SYMATTR InstName R7
SYMATTR Value 1k
TEXT -152 328 Left 0 !.tran .01s

Looks like S.E.B doesn't allow attachments, so see...

<http://www.analog-innovations.com/SED/OpAmp_Bowden.pdf>

Other than GROSS DC imbalance and -81° phase shift when it crosses
0dB, it's just fine ;-)

LTspice, nor most any other Spice for that matter, will not show
oscillations in a static .TRAN analysis unless you push start them.

A .AC analysis of the Loop Gain and Phase shows that it fails the
criteria for stability.

What are you attempting to do? Or is this just a self-educational
exercise?

Suggestion: Go back and do your "disconnected" situation, but ramp up
the supplies from zero... then it'll (most likely) show oscillations.

...Jim Thompson
--
| James E.Thompson | mens |
| Analog Innovations | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| San Tan Valley, AZ 85142 Skype: skypeanalog | |
| Voice480)460-2350 Fax: Available upon request | Brass Rat |
| E-mail Icon at http://www.analog-innovations.com | 1962 |

I love to cook with wine. Sometimes I even put it in the food.

 

>"Reducing R3 to 82 ohms seemed to cure the oscillation problem. "

That reduced the open loop gain of the ciruit.

 

[Bill Bowden]

"Jim Thompson" <To-Email-Use-The-Envelope-Icon@On-My-Web-Site.com> wrote in
message news:64m9p95hspaebm6ivgp9e2lpmda4kej4t0@4ax.com...

On Thu, 5 Jun 2014 19:00:20 -0700, "Bill Bowden"
bperryb@bowdenshobbycircuits.info> wrote:

Why doesn't this discrete 4T opamp run in LTspice at higher frequencies?
It
looks ok at 1Khz, but I get a 5 meg oscillation at 10K or more. There are
no
oscillations with the input disconnected.

Version 4
SHEET 1 880 680
WIRE 128 -112 32 -112
WIRE 192 -112 128 -112
[Full listing at Message-ID: <lmr7ai$1u4v$1@adenine.netfront.net>]
SYMBOL res 624 0 R0
SYMATTR InstName R7
SYMATTR Value 1k
TEXT -152 328 Left 0 !.tran .01s



Looks like S.E.B doesn't allow attachments, so see...

http://www.analog-innovations.com/SED/OpAmp_Bowden.pdf

Other than GROSS DC imbalance and -81° phase shift when it crosses
0dB, it's just fine ;-)

LTspice, nor most any other Spice for that matter, will not show
oscillations in a static .TRAN analysis unless you push start them.

A .AC analysis of the Loop Gain and Phase shows that it fails the
criteria for stability.

What are you attempting to do? Or is this just a self-educational
exercise?

Suggestion: Go back and do your "disconnected" situation, but ramp up
the supplies from zero... then it'll (most likely) show oscillations.

...Jim Thompson

It's just self-educational.. The DC imbalance Q2/Q3 did seem to be a problem
as you say. Reducing R3 to 82 ohms seemed to cure the oscillation problem.
I was attempting to create an opamp with the fewest transistors. I thought
the minimum might be 4, but after looking at it, I didn't see a need for the
PNP transistor and did it again using just 3 NPNs (Q4 and R3 are gone). And
I think there is a way using just 2 transistors, but it wouldn't have both
inverting and non-inverting inputs. I don't have much experience with
LTspice, the only AC analysis I've done is sweeping from frequency A to B.

-Bill



--- news://freenews.netfront.net/ - complaints: news@netfront.net ---

 

>"Try this... "

That's an interesting implementation of Q2. Pretty sure I've not seen it done that way before. what advantage does that have rsather than the standard collector loaded current source ?

 

[Jim Thompson]

On Mon, 09 Jun 2014 18:17:08 -0700, Jim Thompson
<To-Email-Use-The-Envelope-Icon@On-My-Web-Site.com> wrote:

On Mon, 9 Jun 2014 17:43:38 -0700, "Bill Bowden"
bperryb@bowdenshobbycircuits.info> wrote:


"Jim Thompson" <To-Email-Use-The-Envelope-Icon@On-My-Web-Site.com> wrote in
message news:64m9p95hspaebm6ivgp9e2lpmda4kej4t0@4ax.com...
On Thu, 5 Jun 2014 19:00:20 -0700, "Bill Bowden"
bperryb@bowdenshobbycircuits.info> wrote:

Why doesn't this discrete 4T opamp run in LTspice at higher frequencies?
It
looks ok at 1Khz, but I get a 5 meg oscillation at 10K or more. There are
no
oscillations with the input disconnected.

Version 4
SHEET 1 880 680
WIRE 128 -112 32 -112
WIRE 192 -112 128 -112
[Full listing at Message-ID: <lmr7ai$1u4v$1@adenine.netfront.net>]
SYMBOL res 624 0 R0
SYMATTR InstName R7
SYMATTR Value 1k
TEXT -152 328 Left 0 !.tran .01s



Looks like S.E.B doesn't allow attachments, so see...

http://www.analog-innovations.com/SED/OpAmp_Bowden.pdf

Other than GROSS DC imbalance and -81° phase shift when it crosses
0dB, it's just fine ;-)

LTspice, nor most any other Spice for that matter, will not show
oscillations in a static .TRAN analysis unless you push start them.

A .AC analysis of the Loop Gain and Phase shows that it fails the
criteria for stability.

What are you attempting to do? Or is this just a self-educational
exercise?

Suggestion: Go back and do your "disconnected" situation, but ramp up
the supplies from zero... then it'll (most likely) show oscillations.

...Jim Thompson

It's just self-educational.. The DC imbalance Q2/Q3 did seem to be a problem
as you say. Reducing R3 to 82 ohms seemed to cure the oscillation problem.
I was attempting to create an opamp with the fewest transistors. I thought
the minimum might be 4, but after looking at it, I didn't see a need for the
PNP transistor and did it again using just 3 NPNs (Q4 and R3 are gone). And
I think there is a way using just 2 transistors, but it wouldn't have both
inverting and non-inverting inputs. I don't have much experience with
LTspice, the only AC analysis I've done is sweeping from frequency A to B.

-Bill



--- news://freenews.netfront.net/ - complaints: news@netfront.net ---

You need a capacitor in there to keep it stable under all feedback
conditions.

I'll modify your basic structure with a few gimmicks I've learned
after more than 50 years of designing OpAmps... maybe tomorrow if I
have enough time... I have a design review tentatively scheduled.

...Jim Thompson

Try this...

<http://www.analog-innovations.com/SED/OpAmp_Bowden_JT.pdf>

...Jim Thompson
--
| James E.Thompson | mens |
| Analog Innovations | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| San Tan Valley, AZ 85142 Skype: skypeanalog | |
| Voice480)460-2350 Fax: Available upon request | Brass Rat |
| E-mail Icon at http://www.analog-innovations.com | 1962 |

I love to cook with wine. Sometimes I even put it in the food.

 

[Bill Bowden]

"Jim Thompson" <To-Email-Use-The-Envelope-Icon@On-My-Web-Site.com> wrote in
message news:n0hep9tcfsbtfqaatmq0kko83su6p6ufuk@4ax.com...

On Mon, 09 Jun 2014 18:17:08 -0700, Jim Thompson
To-Email-Use-The-Envelope-Icon@On-My-Web-Site.com> wrote:

On Mon, 9 Jun 2014 17:43:38 -0700, "Bill Bowden"
bperryb@bowdenshobbycircuits.info> wrote:


"Jim Thompson" <To-Email-Use-The-Envelope-Icon@On-My-Web-Site.com> wrote
in
message news:64m9p95hspaebm6ivgp9e2lpmda4kej4t0@4ax.com...
On Thu, 5 Jun 2014 19:00:20 -0700, "Bill Bowden"
bperryb@bowdenshobbycircuits.info> wrote:

Why doesn't this discrete 4T opamp run in LTspice at higher
frequencies?
It
looks ok at 1Khz, but I get a 5 meg oscillation at 10K or more. There
are
no
oscillations with the input disconnected.

Version 4
SHEET 1 880 680
WIRE 128 -112 32 -112
WIRE 192 -112 128 -112
[Full listing at Message-ID: <lmr7ai$1u4v$1@adenine.netfront.net>]
SYMBOL res 624 0 R0
SYMATTR InstName R7
SYMATTR Value 1k
TEXT -152 328 Left 0 !.tran .01s



Looks like S.E.B doesn't allow attachments, so see...

http://www.analog-innovations.com/SED/OpAmp_Bowden.pdf

Other than GROSS DC imbalance and -81° phase shift when it crosses
0dB, it's just fine ;-)

LTspice, nor most any other Spice for that matter, will not show
oscillations in a static .TRAN analysis unless you push start them.

A .AC analysis of the Loop Gain and Phase shows that it fails the
criteria for stability.

What are you attempting to do? Or is this just a self-educational
exercise?

Suggestion: Go back and do your "disconnected" situation, but ramp up
the supplies from zero... then it'll (most likely) show oscillations.

...Jim Thompson

It's just self-educational.. The DC imbalance Q2/Q3 did seem to be a
problem
as you say. Reducing R3 to 82 ohms seemed to cure the oscillation
problem.
I was attempting to create an opamp with the fewest transistors. I
thought
the minimum might be 4, but after looking at it, I didn't see a need for
the
PNP transistor and did it again using just 3 NPNs (Q4 and R3 are gone).
And
I think there is a way using just 2 transistors, but it wouldn't have
both
inverting and non-inverting inputs. I don't have much experience with
LTspice, the only AC analysis I've done is sweeping from frequency A to
B.

-Bill



--- news://freenews.netfront.net/ - complaints: news@netfront.net ---

You need a capacitor in there to keep it stable under all feedback
conditions.

I'll modify your basic structure with a few gimmicks I've learned
after more than 50 years of designing OpAmps... maybe tomorrow if I
have enough time... I have a design review tentatively scheduled.

...Jim Thompson

Try this...

http://www.analog-innovations.com/SED/OpAmp_Bowden_JT.pdf

Interesting gimmick using the Vbe drop of Q5 across the 1k resistor to set
the current constant at 0.7/1k = 700 uA. The collector just dumps any extra
current passing through R8. I was wondering how to keep the current constant
with lower suppy voltages. The way I had it would only work at +/- 15 volts.
Nice job. How did you arrive at 10pF for the capacitor?

-Bill




--- news://freenews.netfront.net/ - complaints: news@netfront.net ---

 

[Jim Thompson]

On Wed, 11 Jun 2014 17:33:08 -0700, "Bill Bowden"
<bperryb@bowdenshobbycircuits.info> wrote:

"Jim Thompson" <To-Email-Use-The-Envelope-Icon@On-My-Web-Site.com> wrote in
message news:n0hep9tcfsbtfqaatmq0kko83su6p6ufuk@4ax.com...
On Mon, 09 Jun 2014 18:17:08 -0700, Jim Thompson
To-Email-Use-The-Envelope-Icon@On-My-Web-Site.com> wrote:

On Mon, 9 Jun 2014 17:43:38 -0700, "Bill Bowden"
bperryb@bowdenshobbycircuits.info> wrote:


"Jim Thompson" <To-Email-Use-The-Envelope-Icon@On-My-Web-Site.com> wrote
in
message news:64m9p95hspaebm6ivgp9e2lpmda4kej4t0@4ax.com...
On Thu, 5 Jun 2014 19:00:20 -0700, "Bill Bowden"
bperryb@bowdenshobbycircuits.info> wrote:

Why doesn't this discrete 4T opamp run in LTspice at higher
frequencies?
It
looks ok at 1Khz, but I get a 5 meg oscillation at 10K or more. There
are
no
oscillations with the input disconnected.

Version 4
SHEET 1 880 680
WIRE 128 -112 32 -112
WIRE 192 -112 128 -112
[Full listing at Message-ID: <lmr7ai$1u4v$1@adenine.netfront.net>]
SYMBOL res 624 0 R0
SYMATTR InstName R7
SYMATTR Value 1k
TEXT -152 328 Left 0 !.tran .01s



Looks like S.E.B doesn't allow attachments, so see...

http://www.analog-innovations.com/SED/OpAmp_Bowden.pdf

Other than GROSS DC imbalance and -81° phase shift when it crosses
0dB, it's just fine ;-)

LTspice, nor most any other Spice for that matter, will not show
oscillations in a static .TRAN analysis unless you push start them.

A .AC analysis of the Loop Gain and Phase shows that it fails the
criteria for stability.

What are you attempting to do? Or is this just a self-educational
exercise?

Suggestion: Go back and do your "disconnected" situation, but ramp up
the supplies from zero... then it'll (most likely) show oscillations.

...Jim Thompson

It's just self-educational.. The DC imbalance Q2/Q3 did seem to be a
problem
as you say. Reducing R3 to 82 ohms seemed to cure the oscillation
problem.
I was attempting to create an opamp with the fewest transistors. I
thought
the minimum might be 4, but after looking at it, I didn't see a need for
the
PNP transistor and did it again using just 3 NPNs (Q4 and R3 are gone).
And
I think there is a way using just 2 transistors, but it wouldn't have
both
inverting and non-inverting inputs. I don't have much experience with
LTspice, the only AC analysis I've done is sweeping from frequency A to
B.

-Bill



--- news://freenews.netfront.net/ - complaints: news@netfront.net ---

You need a capacitor in there to keep it stable under all feedback
conditions.

I'll modify your basic structure with a few gimmicks I've learned
after more than 50 years of designing OpAmps... maybe tomorrow if I
have enough time... I have a design review tentatively scheduled.

...Jim Thompson

Try this...

http://www.analog-innovations.com/SED/OpAmp_Bowden_JT.pdf



Interesting gimmick using the Vbe drop of Q5 across the 1k resistor to set
the current constant at 0.7/1k = 700 uA. The collector just dumps any extra
current passing through R8. I was wondering how to keep the current constant
with lower suppy voltages. The way I had it would only work at +/- 15 volts.
Nice job. How did you arrive at 10pF for the capacitor?

-Bill

10pF capacitor determined by running...

<http://www.analog-innovations.com/LoopGain.zip>

And twiddling until I got a happy 45° phase margin.

...Jim Thompson
--
| James E.Thompson | mens |
| Analog Innovations | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| San Tan Valley, AZ 85142 Skype: skypeanalog | |
| Voice480)460-2350 Fax: Available upon request | Brass Rat |
| E-mail Icon at http://www.analog-innovations.com | 1962 |

I love to cook with wine. Sometimes I even put it in the food.