Follow along with the video below to see how to install our site as a web app on your home screen.
Note: This feature may not be available in some browsers.
jla...@highlandsniptechnology.com wrote:
====================================
I want to measure the noise of a voltage regulator, probably an LM2941
making 5 volts out. So I need a low noise high gain preamp for a
scope.
Looks like this would work:
https://www.alphalabinc.com/product/lna10/
Any other suggestions?
I could make something, but it\'s marginal on time/cost and this looks
pretty cool.
Maybe I can bypass the upper feedback resistor to reduce noise, and
massively bypass the output to ground, if all that is atable.
** An L/C or even RC filter after the reg will do the job.
Trim the output to make up for any voltage drop.
.... Phil
lørdag den 10. september 2022 kl. 23.34.00 UTC+2 skrev jla...@highlandsniptechnology.com:
I want to measure the noise of a voltage regulator, probably an LM2941
making 5 volts out. So I need a low noise high gain preamp for a
scope.
Looks like this would work:
https://www.alphalabinc.com/product/lna10/
Any other suggestions?
I could make something, but it\'s marginal on time/cost and this looks
pretty cool.
Maybe I can bypass the upper feedback resistor to reduce noise, and
massively bypass the output to ground, if all that is atable.
I want 5 volts at a few hundred mA and wideband noise well below 100
uV RMS. I could make my own regulator if I had to, but LM2941 is easy.
This is going to power a bunch of cmos logic, about 20 ns worth total,
and Vcc noise becomes jitter.
I my math adds up the LM2941 is 150uV typical at 5V, the ancient LM7805 is 40uV
On Sat, 10 Sep 2022 14:49:02 -0700 (PDT), Lasse Langwadt Christensen
langwadt@fonz.dk> wrote:
lørdag den 10. september 2022 kl. 23.34.00 UTC+2 skrev jla...@highlandsniptechnology.com:
I want to measure the noise of a voltage regulator, probably an LM2941
making 5 volts out. So I need a low noise high gain preamp for a
scope.
Looks like this would work:
https://www.alphalabinc.com/product/lna10/
Any other suggestions?
I could make something, but it\'s marginal on time/cost and this looks
pretty cool.
Maybe I can bypass the upper feedback resistor to reduce noise, and
massively bypass the output to ground, if all that is atable.
I want 5 volts at a few hundred mA and wideband noise well below 100
uV RMS. I could make my own regulator if I had to, but LM2941 is easy.
This is going to power a bunch of cmos logic, about 20 ns worth total,
and Vcc noise becomes jitter.
I my math adds up the LM2941 is 150uV typical at 5V, the ancient LM7805 is 40uV
I do want low dropout to minimize dissipation. We\'ll probably switch
down to +6.25 maybe and then LDO to make a very quiet +5.
This is important so maybe I\'ll make my own regulator, which would
likely be lower noise than an IC reg.
On Sat, 10 Sep 2022 15:55:47 -0700, jlarkin@highlandsniptechnology.com
wrote:
On Sat, 10 Sep 2022 14:49:02 -0700 (PDT), Lasse Langwadt Christensen
langwadt@fonz.dk> wrote:
lørdag den 10. september 2022 kl. 23.34.00 UTC+2 skrev jla...@highlandsniptechnology.com:
I want to measure the noise of a voltage regulator, probably an LM2941
making 5 volts out. So I need a low noise high gain preamp for a
scope.
Looks like this would work:
https://www.alphalabinc.com/product/lna10/
Any other suggestions?
I could make something, but it\'s marginal on time/cost and this looks
pretty cool.
Maybe I can bypass the upper feedback resistor to reduce noise, and
massively bypass the output to ground, if all that is atable.
I want 5 volts at a few hundred mA and wideband noise well below 100
uV RMS. I could make my own regulator if I had to, but LM2941 is easy.
This is going to power a bunch of cmos logic, about 20 ns worth total,
and Vcc noise becomes jitter.
I my math adds up the LM2941 is 150uV typical at 5V, the ancient LM7805 is 40uV
I do want low dropout to minimize dissipation. We\'ll probably switch
down to +6.25 maybe and then LDO to make a very quiet +5.
This is important so maybe I\'ll make my own regulator, which would
likely be lower noise than an IC reg.
The classic dodge is a switcher driving a linear regulator, followed
by Phil H\'s favorite capacitance multiplier, with RC filters all
along.
Joe Gwinn
On Sat, 10 Sep 2022 14:49:02 -0700 (PDT), Lasse Langwadt Christensen
langwadt@fonz.dk> wrote:
lørdag den 10. september 2022 kl. 23.34.00 UTC+2 skrev jla...@highlandsniptechnology.com:
I want to measure the noise of a voltage regulator, probably an LM2941
making 5 volts out. So I need a low noise high gain preamp for a
scope.
Looks like this would work:
https://www.alphalabinc.com/product/lna10/
Any other suggestions?
I could make something, but it\'s marginal on time/cost and this looks
pretty cool.
Maybe I can bypass the upper feedback resistor to reduce noise, and
massively bypass the output to ground, if all that is atable.
I want 5 volts at a few hundred mA and wideband noise well below 100
uV RMS. I could make my own regulator if I had to, but LM2941 is easy.
This is going to power a bunch of cmos logic, about 20 ns worth total,
and Vcc noise becomes jitter.
I my math adds up the LM2941 is 150uV typical at 5V, the ancient LM7805 is 40uV
I do want low dropout to minimize dissipation. We\'ll probably switch
down to +6.25 maybe and then LDO to make a very quiet +5.
This is important so maybe I\'ll make my own regulator, which would
likely be lower noise than an IC reg.
0 dB is 1nV/rtHz.
Am 11.09.22 um 00:55 schrieb jlarkin@highlandsniptechnology.com:
On Sat, 10 Sep 2022 14:49:02 -0700 (PDT), Lasse Langwadt Christensen
langwadt@fonz.dk> wrote:
lørdag den 10. september 2022 kl. 23.34.00 UTC+2 skrev jla...@highlandsniptechnology.com:
I want to measure the noise of a voltage regulator, probably an LM2941
making 5 volts out. So I need a low noise high gain preamp for a
scope.
Looks like this would work:
https://www.alphalabinc.com/product/lna10/
Any other suggestions?
I could make something, but it\'s marginal on time/cost and this looks
pretty cool.
Maybe I can bypass the upper feedback resistor to reduce noise, and
massively bypass the output to ground, if all that is atable.
I want 5 volts at a few hundred mA and wideband noise well below 100
uV RMS. I could make my own regulator if I had to, but LM2941 is easy.
This is going to power a bunch of cmos logic, about 20 ns worth total,
and Vcc noise becomes jitter.
I my math adds up the LM2941 is 150uV typical at 5V, the ancient LM7805 is 40uV
I do want low dropout to minimize dissipation. We\'ll probably switch
down to +6.25 maybe and then LDO to make a very quiet +5.
This is important so maybe I\'ll make my own regulator, which would
likely be lower noise than an IC reg.
I did not measure the LM2941, but the Lm2940 is about the worst
choice possible.
It would require real work to be better than a LT3042. Its main
drawback is the small load current because of the tiny package.
https://www.flickr.com/photos/137684711@N07/24070698809/in/album-72157662535945536/
0 dB is 1nV/rtHz.
In the LT3042 data sheet is a a circuit with an external NPN
to take the heat. It measures about the same as the 3042 alone.
The noise peak climbs to 100 KHz but does not grow in size.
Adding more output C increases peaking. Cset may be increased
at the cost of larger startup time. Cset=100u can make sense and
lowers the 1/f corner to a few Hz. It is not really 1/f but
1/f**2 or 3. More than 100u seems not to help.
https://www.flickr.com/photos/137684711@N07/24070698809/in/album-72157662535945536/
Some empty boardlets are left over. 1.5A@5V from a 6.6V raw
source is ok. The transistor costs a diode drop vs. bare 3042.
Looks like we\'ll need very quiet supplies: +20, +5, +4, -5, all from a
wall wart that might be 12 or 24. And ugly supplies +7, -7, 3.3,
+3ref, 1.8, 1.35, 1.0. Eleven supply rails in a tiny box jammed with
stuff. It\'s a giant puzzle, either fun or a damned nuisance.
On Sat, 10 Sep 2022 15:55:47 -0700, jlarkin@highlandsniptechnology.com
wrote:
On Sat, 10 Sep 2022 14:49:02 -0700 (PDT), Lasse Langwadt Christensen
langwadt@fonz.dk> wrote:
lørdag den 10. september 2022 kl. 23.34.00 UTC+2 skrev jla...@highlandsniptechnology.com:
I want to measure the noise of a voltage regulator, probably an LM2941
making 5 volts out. So I need a low noise high gain preamp for a
scope.
Looks like this would work:
https://www.alphalabinc.com/product/lna10/
Any other suggestions?
I could make something, but it\'s marginal on time/cost and this looks
pretty cool.
Maybe I can bypass the upper feedback resistor to reduce noise, and
massively bypass the output to ground, if all that is atable.
I want 5 volts at a few hundred mA and wideband noise well below 100
uV RMS. I could make my own regulator if I had to, but LM2941 is easy.
This is going to power a bunch of cmos logic, about 20 ns worth total,
and Vcc noise becomes jitter.
I my math adds up the LM2941 is 150uV typical at 5V, the ancient LM7805 is 40uV
I do want low dropout to minimize dissipation. We\'ll probably switch
down to +6.25 maybe and then LDO to make a very quiet +5.
This is important so maybe I\'ll make my own regulator, which would
likely be lower noise than an IC reg.
The classic dodge is a switcher driving a linear regulator, followed
by Phil H\'s favorite capacitance multiplier, with RC filters all
along.
Joe Gwinn
On Tue, 13 Sep 2022 08:04:00 -0700, John Larkin
jlarkin@highlandSNIPMEtechnology.com> wrote:
Looks like we\'ll need very quiet supplies: +20, +5, +4, -5, all from a
wall wart that might be 12 or 24. And ugly supplies +7, -7, 3.3,
+3ref, 1.8, 1.35, 1.0. Eleven supply rails in a tiny box jammed with
stuff. It\'s a giant puzzle, either fun or a damned nuisance.
Does every supply really need to be quiet ?
For instance if the +5 V supply is used for CMOS logic, why does it
have to be so clean ?. If some low power circuitry is also used from
+5 V, why not make a separate feed with better filtering and possibly
linear regulation.
Must the voltages be exactly regulated (with feedback regulation) or
can there be some tolerances (open loop control) ?
How about a transformer coupled power supply with a center taped
primary fed by a push-pull switcher and multiple secondary windings
for various output voltages ?
For those voltages that allow some variation, just rectify one of the
secondary voltages and apply some filtering. For better regulation add
some regulator possibly followed by a capacitance multiplier to reduce
noise.
Having multiple secondary windings makes it possible to keep the
individual grounds separated as possible and only connect the grounds
together at a single point at the load, thus reducing the risk for
various ground noises.
On Sat, 10 Sep 2022 19:19:24 -0400, Joe Gwinn <joegwinn@comcast.net
wrote:
On Sat, 10 Sep 2022 15:55:47 -0700, jlarkin@highlandsniptechnology.com
wrote:
On Sat, 10 Sep 2022 14:49:02 -0700 (PDT), Lasse Langwadt Christensen
langwadt@fonz.dk> wrote:
lørdag den 10. september 2022 kl. 23.34.00 UTC+2 skrev jla...@highlandsniptechnology.com:
I want to measure the noise of a voltage regulator, probably an LM2941
making 5 volts out. So I need a low noise high gain preamp for a
scope.
Looks like this would work:
https://www.alphalabinc.com/product/lna10/
Any other suggestions?
I could make something, but it\'s marginal on time/cost and this looks
pretty cool.
Maybe I can bypass the upper feedback resistor to reduce noise, and
massively bypass the output to ground, if all that is atable.
I want 5 volts at a few hundred mA and wideband noise well below 100
uV RMS. I could make my own regulator if I had to, but LM2941 is easy.
This is going to power a bunch of cmos logic, about 20 ns worth total,
and Vcc noise becomes jitter.
I my math adds up the LM2941 is 150uV typical at 5V, the ancient LM7805 is 40uV
I do want low dropout to minimize dissipation. We\'ll probably switch
down to +6.25 maybe and then LDO to make a very quiet +5.
This is important so maybe I\'ll make my own regulator, which would
likely be lower noise than an IC reg.
The classic dodge is a switcher driving a linear regulator, followed
by Phil H\'s favorite capacitance multiplier, with RC filters all
along.
Joe Gwinn
That adds a lot of dissipation.
This looks pretty good. It\'s actually not bad without R5+C3, kinda
surprising. Using a quiet 3 volt reference helps a lot.
Version 4
SHEET 1 1524 680
WIRE 144 16 -112 16
WIRE 336 16 240 16
WIRE 400 16 336 16
WIRE 544 16 400 16
WIRE 656 16 544 16
WIRE 736 16 656 16
WIRE 832 16 736 16
WIRE 896 16 832 16
WIRE 1008 16 976 16
WIRE 1056 16 1008 16
WIRE 656 48 656 16
WIRE -112 64 -112 16
WIRE 400 64 400 16
WIRE 544 64 544 16
WIRE 832 64 832 16
WIRE 224 96 224 64
WIRE 400 160 400 128
WIRE 544 160 544 128
WIRE 656 160 656 128
WIRE -112 176 -112 144
WIRE 832 176 832 144
WIRE 224 224 224 176
WIRE 336 224 224 224
WIRE 464 224 416 224
WIRE 592 224 528 224
WIRE -80 256 -112 256
WIRE -48 256 -80 256
WIRE 48 256 16 256
WIRE 80 256 48 256
WIRE -112 288 -112 256
WIRE 16 288 16 256
WIRE 368 320 336 320
WIRE 400 320 368 320
WIRE 400 336 400 320
WIRE 592 352 592 224
WIRE 592 352 432 352
WIRE 624 352 592 352
WIRE 736 352 736 16
WIRE 736 352 704 352
WIRE 224 368 224 224
WIRE 368 368 224 368
WIRE 464 384 432 384
WIRE 592 384 592 352
WIRE -112 400 -112 368
WIRE 16 400 16 368
WIRE 400 432 400 400
WIRE 592 496 592 464
FLAG -112 176 0
FLAG -112 400 0
FLAG 400 160 0
FLAG 544 160 0
FLAG 16 400 0
FLAG -80 256 +20
FLAG 48 256 +3
FLAG 336 16 +5
FLAG 656 160 0
FLAG 464 384 +3
FLAG 592 496 0
FLAG 400 432 0
FLAG 368 320 +20
FLAG 832 176 0
FLAG 1008 16 N
SYMBOL Opamps\\\\AD823 400 304 M0
WINDOW 0 70 91 Left 2
WINDOW 3 48 120 Left 2
SYMATTR InstName U1
SYMBOL nmos 144 64 R270
WINDOW 0 22 -62 VRight 2
WINDOW 3 -6 -84 VRight 2
SYMATTR InstName M1
SYMATTR Value SP8K2
SYMBOL res 208 80 R0
WINDOW 0 48 40 Left 2
WINDOW 3 49 69 Left 2
SYMATTR InstName R1
SYMATTR Value 50
SYMBOL voltage -112 48 R0
WINDOW 0 54 49 Left 2
WINDOW 3 38 87 Left 2
WINDOW 123 45 120 Left 2
WINDOW 39 0 0 Left 0
SYMATTR InstName V1
SYMATTR Value SINE(7 100m 1K)
SYMATTR Value2 AC 1
SYMBOL voltage -112 272 R0
WINDOW 0 34 82 Left 2
WINDOW 3 35 108 Left 2
SYMATTR InstName V2
SYMATTR Value 20
SYMBOL cap 384 64 R0
WINDOW 0 46 27 Left 2
WINDOW 3 30 57 Left 2
SYMATTR InstName C1
SYMATTR Value 100µ
SYMATTR SpiceLine Rser=5m
SYMBOL cap 528 64 R0
WINDOW 0 40 9 Left 2
WINDOW 3 40 51 Left 2
SYMATTR InstName C2
SYMATTR Value 10µ
SYMBOL voltage 16 272 R0
WINDOW 0 34 82 Left 2
WINDOW 3 35 108 Left 2
SYMATTR InstName V3
SYMATTR Value 3
SYMBOL res 640 32 R0
WINDOW 0 40 78 Left 2
WINDOW 3 32 107 Left 2
SYMATTR InstName R2
SYMATTR Value 100
SYMBOL res 720 336 R90
WINDOW 0 71 56 VBottom 2
WINDOW 3 78 56 VTop 2
SYMATTR InstName R3
SYMATTR Value 2K
SYMBOL res 608 480 R180
WINDOW 0 59 53 Left 2
WINDOW 3 58 22 Left 2
SYMATTR InstName R4
SYMATTR Value 3K
SYMBOL res 432 208 R90
WINDOW 0 68 81 VBottom 2
WINDOW 3 43 26 VTop 2
SYMATTR InstName R5
SYMATTR Value 1k
SYMBOL cap 528 208 R90
WINDOW 0 71 52 VBottom 2
WINDOW 3 47 9 VTop 2
SYMATTR InstName C3
SYMATTR Value 10n
SYMBOL current 832 64 R0
WINDOW 0 22 89 Left 2
WINDOW 3 -76 158 Left 2
WINDOW 123 0 0 Left 0
WINDOW 39 0 0 Left 0
SYMATTR InstName I1
SYMATTR Value PULSE(0 250m 25m 0 0 10m)
SYMBOL voltage 880 16 R270
WINDOW 0 -68 55 VTop 2
WINDOW 3 -76 54 VBottom 2
SYMATTR InstName V4
SYMATTR Value 5
TEXT 280 456 Left 2 ;OPA197
TEXT 816 408 Left 2 !.tran 50m startup
TEXT 808 296 Left 2 ;T660 5 VOLT LDO
TEXT 824 352 Left 2 ;JL Sep 10 2022
TEXT 560 144 Left 2 ;+5mr
I want to measure the noise of a voltage regulator, probably an LM2941
making 5 volts out. So I need a low noise high gain preamp for a
scope.
Looks like this would work:
https://www.alphalabinc.com/product/lna10/
Any other suggestions?
I could make something, but it\'s marginal on time/cost and this looks
pretty cool.
Maybe I can bypass the upper feedback resistor to reduce noise, and
massively bypass the output to ground, if all that is atable.
I want 5 volts at a few hundred mA and wideband noise well below 100
uV RMS. I could make my own regulator if I had to, but LM2941 is easy.
This is going to power a bunch of cmos logic, about 20 ns worth total,
and Vcc noise becomes jitter.
I have an old TM500 thing in the dungeon, but I doubt it still works.
I guess I\'ll dust it off and see.
I want to measure the noise of a voltage regulator, probably an LM2941
making 5 volts out. So I need a low noise high gain preamp for a
scope.
Looks like this would work:
https://www.alphalabinc.com/product/lna10/
Any other suggestions?
I could make something, but it\'s marginal on time/cost and this looks
pretty cool.
Maybe I can bypass the upper feedback resistor to reduce noise, and
massively bypass the output to ground, if all that is atable.
I want 5 volts at a few hundred mA and wideband noise well below 100
uV RMS. I could make my own regulator if I had to, but LM2941 is easy.
This is going to power a bunch of cmos logic, about 20 ns worth total,
and Vcc noise becomes jitter.
I have an old TM500 thing in the dungeon, but I doubt it still works.
I guess I\'ll dust it off and see.
On Wed, 14 Sep 2022 14:41:16 +0300, upsid...@downunder.com wrote:
On Tue, 13 Sep 2022 08:04:00 -0700, John Larkin
jla...@highlandSNIPMEtechnology.com> wrote:
Looks like we\'ll need very quiet supplies: +20, +5, +4, -5, all from a
wall wart that might be 12 or 24. And ugly supplies +7, -7, 3.3,
+3ref, 1.8, 1.35, 1.0. Eleven supply rails in a tiny box jammed with
stuff. It\'s a giant puzzle, either fun or a damned nuisance.
Does every supply really need to be quiet ?
The four I named. We might want a quiet 3.3 too, as the logic supply
for some PECL comparators, but we can lowpass that from the switched
3.3 with a small LC. ECL prop delay is pretty immune to supply voltage
and the output is differential. CMOS prop delay (+4 and +5 supplies)
is very sensitive to supply voltage, so supply noise becones jitter.
The main reference. +3ref, needs to be good too. I\'ll test some
bandgaps and invest in an ADR363 if we have to.
Low jiter is kind of a bragging point; only a minority of users care
about picoseconds of jitter, but some do.
For instance if the +5 V supply is used for CMOS logic, why does it
have to be so clean ?. If some low power circuitry is also used from
+5 V, why not make a separate feed with better filtering and possibly
linear regulation.
+5 will drive the ultimate customer output pulses. And wobble on those
chips directly modulates output amplitide, which looks like jitter.
And the chip logic thresholds will change too. More jitter. The +5 and
+4 will be super home-made LDOs, the ones I posted Spice models of.
Must the voltages be exactly regulated (with feedback regulation) or
can there be some tolerances (open loop control) ?
We\'ll go for super stable amd low noise; exact voltage can be a few
per cent accurate.
How about a transformer coupled power supply with a center taped
primary fed by a push-pull switcher and multiple secondary windings
for various output voltages ?
Custom magnetics are a nuisance. Buck switchers followed by some
linear regs look reasonable.
For those voltages that allow some variation, just rectify one of the
secondary voltages and apply some filtering. For better regulation add
some regulator possibly followed by a capacitance multiplier to reduce
noise.
Having multiple secondary windings makes it possible to keep the
individual grounds separated as possible and only connect the grounds
together at a single point at the load, thus reducing the risk for
various ground noises.
We\'ll probably use a 6-layer board with two solid ground planes.
Single-point grounding is impractical in complex high-speed products.
Here\'s the gadget.
http://www.highlandtechnology.com/DSS/T560DS.shtml
It\'s time to redesign and improve. It has a 68332 cpu programmed in
assembly, and an FPGA that\'s long EOL.
What\'s bizarre is that we did the fast signal path in a Spartan 3
FPGA, and no modern chips are that fast pin-to-pin. FPGA cores are
getting faster and general i/o is getting slower. So the fast signal
path will be all discrete logic chips.
What\'s also bizarre is that you can still buy 68332, which has
outlasted at least 5 generations of ARM chips.
onsdag den 14. september 2022 kl. 16.59.27 UTC+2 skrev John Larkin:
On Wed, 14 Sep 2022 14:41:16 +0300, upsid...@downunder.com wrote:
On Tue, 13 Sep 2022 08:04:00 -0700, John Larkin
jla...@highlandSNIPMEtechnology.com> wrote:
Looks like we\'ll need very quiet supplies: +20, +5, +4, -5, all from a
wall wart that might be 12 or 24. And ugly supplies +7, -7, 3.3,
+3ref, 1.8, 1.35, 1.0. Eleven supply rails in a tiny box jammed with
stuff. It\'s a giant puzzle, either fun or a damned nuisance.
Does every supply really need to be quiet ?
The four I named. We might want a quiet 3.3 too, as the logic supply
for some PECL comparators, but we can lowpass that from the switched
3.3 with a small LC. ECL prop delay is pretty immune to supply voltage
and the output is differential. CMOS prop delay (+4 and +5 supplies)
is very sensitive to supply voltage, so supply noise becones jitter.
The main reference. +3ref, needs to be good too. I\'ll test some
bandgaps and invest in an ADR363 if we have to.
Low jiter is kind of a bragging point; only a minority of users care
about picoseconds of jitter, but some do.
For instance if the +5 V supply is used for CMOS logic, why does it
have to be so clean ?. If some low power circuitry is also used from
+5 V, why not make a separate feed with better filtering and possibly
linear regulation.
+5 will drive the ultimate customer output pulses. And wobble on those
chips directly modulates output amplitide, which looks like jitter.
And the chip logic thresholds will change too. More jitter. The +5 and
+4 will be super home-made LDOs, the ones I posted Spice models of.
Must the voltages be exactly regulated (with feedback regulation) or
can there be some tolerances (open loop control) ?
We\'ll go for super stable amd low noise; exact voltage can be a few
per cent accurate.
How about a transformer coupled power supply with a center taped
primary fed by a push-pull switcher and multiple secondary windings
for various output voltages ?
Custom magnetics are a nuisance. Buck switchers followed by some
linear regs look reasonable.
For those voltages that allow some variation, just rectify one of the
secondary voltages and apply some filtering. For better regulation add
some regulator possibly followed by a capacitance multiplier to reduce
noise.
Having multiple secondary windings makes it possible to keep the
individual grounds separated as possible and only connect the grounds
together at a single point at the load, thus reducing the risk for
various ground noises.
We\'ll probably use a 6-layer board with two solid ground planes.
Single-point grounding is impractical in complex high-speed products.
Here\'s the gadget.
http://www.highlandtechnology.com/DSS/T560DS.shtml
It\'s time to redesign and improve. It has a 68332 cpu programmed in
assembly, and an FPGA that\'s long EOL.
What\'s bizarre is that we did the fast signal path in a Spartan 3
FPGA, and no modern chips are that fast pin-to-pin. FPGA cores are
getting faster and general i/o is getting slower. So the fast signal
path will be all discrete logic chips.
maybe one of these would work, of the old 68332 dinosaur was enough a cortex M3 should be plenty
https://www.microsemi.com/product-directory/fpga-soc/1639-soc-fpgas
https://wiki.sipeed.com/hardware/en/tang/Tang-Nano-4K/Nano-4K.html
What\'s also bizarre is that you can still buy 68332, which has
outlasted at least 5 generations of ARM chips.
it is probably used in cars
On Wed, 14 Sep 2022 08:40:09 -0700 (PDT), Lasse Langwadt Christensen
langwadt@fonz.dk> wrote:
onsdag den 14. september 2022 kl. 16.59.27 UTC+2 skrev John Larkin:
On Wed, 14 Sep 2022 14:41:16 +0300, upsid...@downunder.com wrote:
On Tue, 13 Sep 2022 08:04:00 -0700, John Larkin
jla...@highlandSNIPMEtechnology.com> wrote:
Looks like we\'ll need very quiet supplies: +20, +5, +4, -5, all from a
wall wart that might be 12 or 24. And ugly supplies +7, -7, 3.3,
+3ref, 1.8, 1.35, 1.0. Eleven supply rails in a tiny box jammed with
stuff. It\'s a giant puzzle, either fun or a damned nuisance.
Does every supply really need to be quiet ?
The four I named. We might want a quiet 3.3 too, as the logic supply
for some PECL comparators, but we can lowpass that from the switched
3.3 with a small LC. ECL prop delay is pretty immune to supply voltage
and the output is differential. CMOS prop delay (+4 and +5 supplies)
is very sensitive to supply voltage, so supply noise becones jitter.
The main reference. +3ref, needs to be good too. I\'ll test some
bandgaps and invest in an ADR363 if we have to.
Low jiter is kind of a bragging point; only a minority of users care
about picoseconds of jitter, but some do.
For instance if the +5 V supply is used for CMOS logic, why does it
have to be so clean ?. If some low power circuitry is also used from
+5 V, why not make a separate feed with better filtering and possibly
linear regulation.
+5 will drive the ultimate customer output pulses. And wobble on those
chips directly modulates output amplitide, which looks like jitter.
And the chip logic thresholds will change too. More jitter. The +5 and
+4 will be super home-made LDOs, the ones I posted Spice models of.
Must the voltages be exactly regulated (with feedback regulation) or
can there be some tolerances (open loop control) ?
We\'ll go for super stable amd low noise; exact voltage can be a few
per cent accurate.
How about a transformer coupled power supply with a center taped
primary fed by a push-pull switcher and multiple secondary windings
for various output voltages ?
Custom magnetics are a nuisance. Buck switchers followed by some
linear regs look reasonable.
For those voltages that allow some variation, just rectify one of the
secondary voltages and apply some filtering. For better regulation add
some regulator possibly followed by a capacitance multiplier to reduce
noise.
Having multiple secondary windings makes it possible to keep the
individual grounds separated as possible and only connect the grounds
together at a single point at the load, thus reducing the risk for
various ground noises.
We\'ll probably use a 6-layer board with two solid ground planes.
Single-point grounding is impractical in complex high-speed products.
Here\'s the gadget.
http://www.highlandtechnology.com/DSS/T560DS.shtml
It\'s time to redesign and improve. It has a 68332 cpu programmed in
assembly, and an FPGA that\'s long EOL.
What\'s bizarre is that we did the fast signal path in a Spartan 3
FPGA, and no modern chips are that fast pin-to-pin. FPGA cores are
getting faster and general i/o is getting slower. So the fast signal
path will be all discrete logic chips.
maybe one of these would work, of the old 68332 dinosaur was enough a cortex M3 should be plenty
https://www.microsemi.com/product-directory/fpga-soc/1639-soc-fpgas
https://wiki.sipeed.com/hardware/en/tang/Tang-Nano-4K/Nano-4K.html
What\'s also bizarre is that you can still buy 68332, which has
outlasted at least 5 generations of ARM chips.
it is probably used in cars
The TPU time processor was obviously designed to run an engine.