Isolated video buffer...

[bitrex]

The LT1722 is a $1 bargain:

<https://www.dropbox.com/s/045cu5n8afqqadj/vid_amp.JPG?dl=0>

Version 4
SHEET 1 2108 772
WIRE -448 -592 -528 -592
WIRE -288 -592 -368 -592
WIRE -208 -592 -288 -592
WIRE -16 -592 -128 -592
WIRE -288 -528 -288 -592
WIRE -1248 -512 -1248 -560
WIRE -1088 -512 -1088 -544
WIRE -1248 -416 -1248 -432
WIRE -1088 -416 -1088 -432
WIRE -288 -416 -288 -448
WIRE -288 -320 -288 -352
WIRE -1360 -304 -1408 -304
WIRE -1232 -304 -1280 -304
WIRE -1120 -304 -1152 -304
WIRE -1008 -304 -1056 -304
WIRE -832 -304 -928 -304
WIRE -656 -304 -784 -304
WIRE -288 -240 -288 -256
WIRE -640 -224 -784 -224
WIRE -528 -224 -528 -592
WIRE -528 -224 -560 -224
WIRE -320 -224 -528 -224
WIRE -304 -224 -320 -224
WIRE -1408 -208 -1408 -304
WIRE -16 -208 -16 -592
WIRE -16 -208 -256 -208
WIRE 48 -208 -16 -208
WIRE -320 -192 -368 -192
WIRE -304 -192 -320 -192
WIRE -288 -144 -288 -176
WIRE -832 -128 -832 -224
WIRE -656 -128 -656 -304
WIRE -368 -128 -368 -192
WIRE -1408 -80 -1408 -128
FLAG -1088 -416 0
FLAG -1088 -544 +V
FLAG -1248 -416 0
FLAG -1248 -560 -V
FLAG -368 -128 0
FLAG -288 -144 -V
FLAG -288 -256 +V
FLAG -656 -128 0
FLAG -288 -320 0
FLAG -832 -128 0
FLAG -1408 -80 0
FLAG 48 -208 OUT
IOPIN 48 -208 Out
SYMBOL voltage -1088 -528 R0
SYMATTR InstName V1
SYMATTR Value 5
SYMBOL voltage -1248 -528 R0
SYMATTR InstName V2
SYMATTR Value -5
SYMBOL ind2 -768 -208 R180
WINDOW 0 -4 79 Right 2
WINDOW 3 -2 41 Right 2
SYMATTR InstName L2
SYMATTR Value 2m
SYMATTR Type ind
SYMBOL ind2 -848 -208 M180
WINDOW 0 -4 79 Right 2
WINDOW 3 -2 41 Right 2
SYMATTR InstName L1
SYMATTR Value 2m
SYMATTR Type ind
SYMBOL res -352 -608 R90
WINDOW 0 0 56 VBottom 2
WINDOW 3 32 56 VTop 2
SYMATTR InstName R2
SYMATTR Value 47
SYMBOL res -112 -608 R90
WINDOW 0 0 56 VBottom 2
WINDOW 3 32 56 VTop 2
SYMATTR InstName R3
SYMATTR Value 47
SYMBOL res -304 -544 R0
SYMATTR InstName R4
SYMATTR Value 39
SYMBOL polcap -304 -416 R0
SYMATTR InstName C2
SYMATTR Value 470µ
SYMATTR Description Capacitor
SYMATTR Type cap
SYMATTR SpiceLine V=63 Irms=800m Rser=0.081 Lser=0 mfg=\"Nichicon\"
pn=\"UPR1J471MRH\" type=\"Al electrolytic\"
SYMBOL polcap -1120 -288 R270
WINDOW 0 32 32 VTop 2
WINDOW 3 0 32 VBottom 2
SYMATTR InstName C1
SYMATTR Value 220µ
SYMATTR Description Capacitor
SYMATTR Type cap
SYMATTR SpiceLine V=63 Irms=750m Rser=0.16 Lser=0 mfg=\"Nichicon\"
pn=\"UPG1J221MRH\" type=\"Al electrolytic\"
SYMBOL Opamps\\\\LT1723 -288 -272 R0
SYMATTR InstName U3
SYMBOL ind -544 -240 R90
WINDOW 0 5 56 VBottom 2
WINDOW 3 32 56 VTop 2
SYMATTR InstName L2_LEAK
SYMATTR Value 0.5µ
SYMBOL ind -912 -320 R90
WINDOW 0 5 56 VBottom 2
WINDOW 3 32 56 VTop 2
SYMATTR InstName L1_LEAK
SYMATTR Value 0.5µ
SYMBOL res -1264 -320 R90
WINDOW 0 0 56 VBottom 2
WINDOW 3 32 56 VTop 2
SYMATTR InstName Z_SOURCE
SYMATTR Value 75
SYMBOL res -1136 -320 R90
WINDOW 0 -5 59 VBottom 2
WINDOW 3 32 56 VTop 2
SYMATTR InstName R1
SYMATTR Value 75
SYMBOL voltage -1408 -224 R0
WINDOW 123 24 44 Left 2
WINDOW 39 0 0 Left 0
SYMATTR Value2 AC 1
SYMATTR InstName V3
SYMATTR Value \"\"
TEXT -832 -392 Left 2 !K1 L1 L2 1
TEXT -1032 -472 Left 2 !.ac oct 10 10 100Meg
TEXT -872 -360 Left 2 ;Murata DA102JC

 

[bitrex]

On 12/6/2020 1:16 AM, bitrex wrote:

The LT1722 is a $1 bargain:

https://www.dropbox.com/s/045cu5n8afqqadj/vid_amp.JPG?dl=0

Or, showing the isolation:

<https://www.dropbox.com/s/ya4bl2agot9scyo/vid_amp_2.JPG?dl=0>

 

[Bill Sloman]

On Sunday, December 6, 2020 at 5:30:11 PM UTC+11, bitrex wrote:

On 12/6/2020 1:16 AM, bitrex wrote:
The LT1722 is a $1 bargain:

https://www.dropbox.com/s/045cu5n8afqqadj/vid_amp.JPG?dl=0
Or, showing the isolation:

https://www.dropbox.com/s/ya4bl2agot9scyo/vid_amp_2.JPG?dl=0

This isn\'t the right way to represent a real transformer in LTSpice. Leave out the 0.5uH leakage inductors and just write K L1 L2 0.99975 - which is probably unrealistically optimistic.

--
Bill Sloman, Sydney

 

[Bill Sloman]

On Sunday, December 6, 2020 at 5:30:11 PM UTC+11, bitrex wrote:

On 12/6/2020 1:16 AM, bitrex wrote:
The LT1722 is a $1 bargain:

https://www.dropbox.com/s/045cu5n8afqqadj/vid_amp.JPG?dl=0
Or, showing the isolation:

https://www.dropbox.com/s/ya4bl2agot9scyo/vid_amp_2.JPG?dl=0

This isn\'t the right way to represent a real transformer in LTSpice. Leave out the 0.5uH leakage inductors and just write K L1 L2 0.99975 - which is probably unrealistically optimistic.

--
Bill Sloman, Sydney

 

[bitrex]

On 12/6/2020 2:54 AM, Bill Sloman wrote:

On Sunday, December 6, 2020 at 5:30:11 PM UTC+11, bitrex wrote:
On 12/6/2020 1:16 AM, bitrex wrote:
The LT1722 is a $1 bargain:

https://www.dropbox.com/s/045cu5n8afqqadj/vid_amp.JPG?dl=0
Or, showing the isolation:

https://www.dropbox.com/s/ya4bl2agot9scyo/vid_amp_2.JPG?dl=0

This isn\'t the right way to represent a real transformer in LTSpice. Leave out the 0.5uH leakage inductors and just write K L1 L2 0.99975 - which is probably unrealistically optimistic.

Bode plot looks about the same. Performance of the actual circuit on a
sheet of copper-clad is very similar to the sim, the ~1dB peak at about
8.5 MHz is there and the low-frequency -3dB point is a bit higher, about
50Hz.

 

[bitrex]

On 12/6/2020 2:54 AM, Bill Sloman wrote:

On Sunday, December 6, 2020 at 5:30:11 PM UTC+11, bitrex wrote:
On 12/6/2020 1:16 AM, bitrex wrote:
The LT1722 is a $1 bargain:

https://www.dropbox.com/s/045cu5n8afqqadj/vid_amp.JPG?dl=0
Or, showing the isolation:

https://www.dropbox.com/s/ya4bl2agot9scyo/vid_amp_2.JPG?dl=0

This isn\'t the right way to represent a real transformer in LTSpice. Leave out the 0.5uH leakage inductors and just write K L1 L2 0.99975 - which is probably unrealistically optimistic.

Bode plot looks about the same. Performance of the actual circuit on a
sheet of copper-clad is very similar to the sim, the ~1dB peak at about
8.5 MHz is there and the low-frequency -3dB point is a bit higher, about
50Hz.

 

[bitrex]

On 12/6/2020 3:00 AM, bitrex wrote:

On 12/6/2020 2:54 AM, Bill Sloman wrote:
On Sunday, December 6, 2020 at 5:30:11 PM UTC+11, bitrex wrote:
On 12/6/2020 1:16 AM, bitrex wrote:
The LT1722 is a $1 bargain:

https://www.dropbox.com/s/045cu5n8afqqadj/vid_amp.JPG?dl=0
Or, showing the isolation:

https://www.dropbox.com/s/ya4bl2agot9scyo/vid_amp_2.JPG?dl=0

This isn\'t the right way to represent  a real transformer in LTSpice.
Leave out the 0.5uH leakage inductors and just write K L1 L2 0.99975 -
which is probably unrealistically  optimistic.


Bode plot looks about the same. Performance of the actual circuit on a
sheet of copper-clad is very similar to the sim, the ~1dB peak at about
8.5 MHz is there and the low-frequency -3dB point is a bit higher, about
50Hz.

DC resistance of the transformer is about 0.5 ohm

 

[bitrex]

On 12/6/2020 3:00 AM, bitrex wrote:

On 12/6/2020 2:54 AM, Bill Sloman wrote:
On Sunday, December 6, 2020 at 5:30:11 PM UTC+11, bitrex wrote:
On 12/6/2020 1:16 AM, bitrex wrote:
The LT1722 is a $1 bargain:

https://www.dropbox.com/s/045cu5n8afqqadj/vid_amp.JPG?dl=0
Or, showing the isolation:

https://www.dropbox.com/s/ya4bl2agot9scyo/vid_amp_2.JPG?dl=0

This isn\'t the right way to represent  a real transformer in LTSpice.
Leave out the 0.5uH leakage inductors and just write K L1 L2 0.99975 -
which is probably unrealistically  optimistic.


Bode plot looks about the same. Performance of the actual circuit on a
sheet of copper-clad is very similar to the sim, the ~1dB peak at about
8.5 MHz is there and the low-frequency -3dB point is a bit higher, about
50Hz.

DC resistance of the transformer is about 0.5 ohm

 

[Bill Sloman]

On Sunday, December 6, 2020 at 7:00:45 PM UTC+11, bitrex wrote:

On 12/6/2020 2:54 AM, Bill Sloman wrote:
On Sunday, December 6, 2020 at 5:30:11 PM UTC+11, bitrex wrote:
On 12/6/2020 1:16 AM, bitrex wrote:
The LT1722 is a $1 bargain:

https://www.dropbox.com/s/045cu5n8afqqadj/vid_amp.JPG?dl=0
Or, showing the isolation:

https://www.dropbox.com/s/ya4bl2agot9scyo/vid_amp_2.JPG?dl=0

This isn\'t the right way to represent a real transformer in LTSpice. Leave out the 0.5uH leakage inductors and just write K L1 L2 0.99975 - which is probably unrealistically optimistic.

Bode plot looks about the same. Performance of the actual circuit on a
sheet of copper-clad is very similar to the sim, the ~1dB peak at about
8.5 MHz is there and the low-frequency -3dB point is a bit higher, about
50Hz.

What sort of parallel capacitance have you plugged in for L1 and L2? You might need to include a third inductor with a resistive load to represent the current induced in the transformer core. With a nickel-zinc ferrite core the resistance would be fairly high. TDK has got data sheets

https://product.tdk.com/info/en/products/ferrite/info/en/catalog/datasheets/ferrite_summary_en.pdf

but it won\'t come up for me now - if memory serves its a fairly voluminous file.

--
Bill Sloman, Sydney

 

[Bill Sloman]

On Sunday, December 6, 2020 at 7:00:45 PM UTC+11, bitrex wrote:

On 12/6/2020 2:54 AM, Bill Sloman wrote:
On Sunday, December 6, 2020 at 5:30:11 PM UTC+11, bitrex wrote:
On 12/6/2020 1:16 AM, bitrex wrote:
The LT1722 is a $1 bargain:

https://www.dropbox.com/s/045cu5n8afqqadj/vid_amp.JPG?dl=0
Or, showing the isolation:

https://www.dropbox.com/s/ya4bl2agot9scyo/vid_amp_2.JPG?dl=0

This isn\'t the right way to represent a real transformer in LTSpice. Leave out the 0.5uH leakage inductors and just write K L1 L2 0.99975 - which is probably unrealistically optimistic.

Bode plot looks about the same. Performance of the actual circuit on a
sheet of copper-clad is very similar to the sim, the ~1dB peak at about
8.5 MHz is there and the low-frequency -3dB point is a bit higher, about
50Hz.

What sort of parallel capacitance have you plugged in for L1 and L2? You might need to include a third inductor with a resistive load to represent the current induced in the transformer core. With a nickel-zinc ferrite core the resistance would be fairly high. TDK has got data sheets

https://product.tdk.com/info/en/products/ferrite/info/en/catalog/datasheets/ferrite_summary_en.pdf

but it won\'t come up for me now - if memory serves its a fairly voluminous file.

--
Bill Sloman, Sydney

 

[server]

On Sun, 6 Dec 2020 01:16:22 -0500, bitrex <user@example.net> wrote:

The LT1722 is a $1 bargain:

https://www.dropbox.com/s/045cu5n8afqqadj/vid_amp.JPG?dl=0

Will that work in real life? The DC gain is huge.



--

John Larkin Highland Technology, Inc

The best designs are necessarily accidental.

 

[server]

On Sun, 6 Dec 2020 01:16:22 -0500, bitrex <user@example.net> wrote:

The LT1722 is a $1 bargain:

https://www.dropbox.com/s/045cu5n8afqqadj/vid_amp.JPG?dl=0

Will that work in real life? The DC gain is huge.



--

John Larkin Highland Technology, Inc

The best designs are necessarily accidental.

 

[server]

On Sun, 06 Dec 2020 08:09:06 -0800, jlarkin@highlandsniptechnology.com
wrote:

On Sun, 6 Dec 2020 01:16:22 -0500, bitrex <user@example.net> wrote:

The LT1722 is a $1 bargain:

https://www.dropbox.com/s/045cu5n8afqqadj/vid_amp.JPG?dl=0

Will that work in real life? The DC gain is huge.

I think you are accounting for the leakage inductance twice.



--

John Larkin Highland Technology, Inc

The best designs are necessarily accidental.

 

[server]

On Sun, 06 Dec 2020 08:09:06 -0800, jlarkin@highlandsniptechnology.com
wrote:

On Sun, 6 Dec 2020 01:16:22 -0500, bitrex <user@example.net> wrote:

The LT1722 is a $1 bargain:

https://www.dropbox.com/s/045cu5n8afqqadj/vid_amp.JPG?dl=0

Will that work in real life? The DC gain is huge.

I think you are accounting for the leakage inductance twice.



--

John Larkin Highland Technology, Inc

The best designs are necessarily accidental.

 

[bitrex]

On 12/6/2020 11:09 AM, jlarkin@highlandsniptechnology.com wrote:

On Sun, 6 Dec 2020 01:16:22 -0500, bitrex <user@example.net> wrote:

The LT1722 is a $1 bargain:

https://www.dropbox.com/s/045cu5n8afqqadj/vid_amp.JPG?dl=0

Will that work in real life? The DC gain is huge.

Yeah it looks to work quite nice on the copper-clad, the low-end
bandwidth is phenomenal for such a small transformer, the current-feed
topology helps a lot.

It\'s variation on a circuit from an app note on the topology, can\'t find
it now but I think it may have been Renesas? The idea is to use a
transformer with decent low-frequency response but relatively large
leakage L, leverage the high open-loop gain at low frequency to extend
the bandwidth there, but loading on the op amp output increases the
primary-side leakage L as seen by the op amp input, so there\'s peaking
added to boost the high freq gain back up.

Performance on the board is pretty similar to the sim. The 1722 is a
very well-behaved for a 200MHz/70 V/usec amp, they claim this in the
datasheet and it seems true, on copper clad at least in this arrangement
it doesn\'t seem prone to instability at all.

The peaking might not even be needed it has 50 mA drive current, typical
and a full-power bandwidth at 1V peak of 9MHz

 

[bitrex]

On 12/6/2020 11:09 AM, jlarkin@highlandsniptechnology.com wrote:

On Sun, 6 Dec 2020 01:16:22 -0500, bitrex <user@example.net> wrote:

The LT1722 is a $1 bargain:

https://www.dropbox.com/s/045cu5n8afqqadj/vid_amp.JPG?dl=0

Will that work in real life? The DC gain is huge.

Yeah it looks to work quite nice on the copper-clad, the low-end
bandwidth is phenomenal for such a small transformer, the current-feed
topology helps a lot.

It\'s variation on a circuit from an app note on the topology, can\'t find
it now but I think it may have been Renesas? The idea is to use a
transformer with decent low-frequency response but relatively large
leakage L, leverage the high open-loop gain at low frequency to extend
the bandwidth there, but loading on the op amp output increases the
primary-side leakage L as seen by the op amp input, so there\'s peaking
added to boost the high freq gain back up.

Performance on the board is pretty similar to the sim. The 1722 is a
very well-behaved for a 200MHz/70 V/usec amp, they claim this in the
datasheet and it seems true, on copper clad at least in this arrangement
it doesn\'t seem prone to instability at all.

The peaking might not even be needed it has 50 mA drive current, typical
and a full-power bandwidth at 1V peak of 9MHz

 

[bitrex]

On 12/6/2020 1:53 PM, bitrex wrote:

On 12/6/2020 11:09 AM, jlarkin@highlandsniptechnology.com wrote:
On Sun, 6 Dec 2020 01:16:22 -0500, bitrex <user@example.net> wrote:

The LT1722 is a $1 bargain:

https://www.dropbox.com/s/045cu5n8afqqadj/vid_amp.JPG?dl=0

Will that work in real life? The DC gain is huge.




Yeah it looks to work quite nice on the copper-clad, the low-end
bandwidth is phenomenal for such a small transformer, the current-feed
topology helps a lot.

Hard to find good intel on what an appropriate low-frequency cutoff
point for an NTSC composite signal is, exactly. What component extends
down to 60 Hz, like the envelope of the vertical sync-pulse train?

 

[server]

On Sun, 6 Dec 2020 13:53:13 -0500, bitrex <user@example.net> wrote:

On 12/6/2020 11:09 AM, jlarkin@highlandsniptechnology.com wrote:
On Sun, 6 Dec 2020 01:16:22 -0500, bitrex <user@example.net> wrote:

The LT1722 is a $1 bargain:

https://www.dropbox.com/s/045cu5n8afqqadj/vid_amp.JPG?dl=0

Will that work in real life? The DC gain is huge.




Yeah it looks to work quite nice on the copper-clad, the low-end
bandwidth is phenomenal for such a small transformer, the current-feed
topology helps a lot.

It\'s variation on a circuit from an app note on the topology, can\'t find
it now but I think it may have been Renesas? The idea is to use a
transformer with decent low-frequency response but relatively large
leakage L, leverage the high open-loop gain at low frequency to extend
the bandwidth there, but loading on the op amp output increases the
primary-side leakage L as seen by the op amp input, so there\'s peaking
added to boost the high freq gain back up.

If you effectively short the transformer secondary, LF response goes
way down below data sheet values, to w=Ls/Rs or something.

Performance on the board is pretty similar to the sim. The 1722 is a
very well-behaved for a 200MHz/70 V/usec amp, they claim this in the
datasheet and it seems true, on copper clad at least in this arrangement
it doesn\'t seem prone to instability at all.

The peaking might not even be needed it has 50 mA drive current, typical
and a full-power bandwidth at 1V peak of 9MHz

The DCR of that transformer isn\'t specified. The DC gain of that
circuit could be huge, and make a lot of offset at the output.



--

John Larkin Highland Technology, Inc

The best designs are necessarily accidental.

 

[bitrex]

On 12/6/2020 2:15 PM, jlarkin@highlandsniptechnology.com wrote:

On Sun, 6 Dec 2020 13:53:13 -0500, bitrex <user@example.net> wrote:

On 12/6/2020 11:09 AM, jlarkin@highlandsniptechnology.com wrote:
On Sun, 6 Dec 2020 01:16:22 -0500, bitrex <user@example.net> wrote:

The LT1722 is a $1 bargain:

https://www.dropbox.com/s/045cu5n8afqqadj/vid_amp.JPG?dl=0

Will that work in real life? The DC gain is huge.




Yeah it looks to work quite nice on the copper-clad, the low-end
bandwidth is phenomenal for such a small transformer, the current-feed
topology helps a lot.

It\'s variation on a circuit from an app note on the topology, can\'t find
it now but I think it may have been Renesas? The idea is to use a
transformer with decent low-frequency response but relatively large
leakage L, leverage the high open-loop gain at low frequency to extend
the bandwidth there, but loading on the op amp output increases the
primary-side leakage L as seen by the op amp input, so there\'s peaking
added to boost the high freq gain back up.

If you effectively short the transformer secondary, LF response goes
way down below data sheet values, to w=Ls/Rs or something.


Performance on the board is pretty similar to the sim. The 1722 is a
very well-behaved for a 200MHz/70 V/usec amp, they claim this in the
datasheet and it seems true, on copper clad at least in this arrangement
it doesn\'t seem prone to instability at all.

The peaking might not even be needed it has 50 mA drive current, typical
and a full-power bandwidth at 1V peak of 9MHz

The DCR of that transformer isn\'t specified. The DC gain of that
circuit could be huge, and make a lot of offset at the output.

Here\'s some scope shots from the Rigol, input is 1V P2P sine using 50
ohm source and 50 ohm terminator but all other values the same, DC
coupled, acquisition set at 128x averaging:

5MHz:

<https://www.dropbox.com/s/igd53xtf4ekikt9/5MHz.bmp?dl=0>


500 Hz:

<https://www.dropbox.com/s/t3f69ywe43je42k/500Hz.bmp?dl=0>

90Hz:

<https://www.dropbox.com/s/wzyxas7ji33p9dq/90Hz.bmp?dl=0>

Saturation distortion becoming un-rectifiable below about 100 Hz

 

[bitrex]

On 12/6/2020 5:05 AM, Bill Sloman wrote:

On Sunday, December 6, 2020 at 7:00:45 PM UTC+11, bitrex wrote:
On 12/6/2020 2:54 AM, Bill Sloman wrote:
On Sunday, December 6, 2020 at 5:30:11 PM UTC+11, bitrex wrote:
On 12/6/2020 1:16 AM, bitrex wrote:
The LT1722 is a $1 bargain:

https://www.dropbox.com/s/045cu5n8afqqadj/vid_amp.JPG?dl=0
Or, showing the isolation:

https://www.dropbox.com/s/ya4bl2agot9scyo/vid_amp_2.JPG?dl=0

This isn\'t the right way to represent a real transformer in LTSpice. Leave out the 0.5uH leakage inductors and just write K L1 L2 0.99975 - which is probably unrealistically optimistic.

Bode plot looks about the same. Performance of the actual circuit on a
sheet of copper-clad is very similar to the sim, the ~1dB peak at about
8.5 MHz is there and the low-frequency -3dB point is a bit higher, about
50Hz.

What sort of parallel capacitance have you plugged in for L1 and L2? You might need to include a third inductor with a resistive load to represent the current induced in the transformer core. With a nickel-zinc ferrite core the resistance would be fairly high. TDK has got data sheets

https://product.tdk.com/info/en/products/ferrite/info/en/catalog/datasheets/ferrite_summary_en.pdf

but it won\'t come up for me now - if memory serves its a fairly voluminous file.

For parallel C, nothing at the moment. Too simple a circuit to spend
much time futzing with in simulation, just went straight to the
breadboard and rigged it up to test with the values as indicated. The
performance of the real circuit is pretty good, whatever the particulars
of this transformer are.