> 48 volts is pretty common lately, PoE and such and maybe even in cars.
> I expect more HV-input switchers in the future.
I agree, at least for my car. I have 12V and 16V modules; so, the common storage is good at 48V. It's not too low for power and not too high for balancing problem.
Reply by Phil Hobbs●July 30, 20232023-07-30
Lasse Langwadt Christensen <langwadt@fonz.dk> wrote:
> søndag den 30. juli 2023 kl. 10.42.09 UTC+2 skrev Piotr Wyderski:
>> John Larkin wrote:
>>
>>> Diodes has some dynamite new switchers, the AP66200 parts, that I'd
>>> like to use.
>> What makes them dynamite as compared to, say, TPS564201?
>> Looks like a boring synchronous buck. What do I miss?
>
> maybe not much, but the AP66200 can take 60v input
>
Which is super handy for inverting bucks. Making -16 from +24 is too sporty
with a 42-volt abs max.
Phil Hobbs
--
Dr Philip C D Hobbs Principal Consultant ElectroOptical Innovations LLC /
Hobbs ElectroOptics Optics, Electro-optics, Photonics, Analog Electronics
Reply by Piotr Wyderski●July 30, 20232023-07-30
John Larkin wrote:
> I love those little TI switchers. We use tons of them. But the best
> they can do on input voltage is 28, and their current rating is
> optimistic, in my opinion.
Indeed, the 60V capability makes them worth attention. Thanks for
sharing the parts.
Best regards, Piotr
Reply by Piotr Wyderski●July 30, 20232023-07-30
Lasse Langwadt Christensen wrote:
> maybe not much, but the AP66200 can take 60v input
OK, that's something! Some of the TI parts can take 38V at most.
Best regards, Piotr
Reply by John Larkin●July 30, 20232023-07-30
On Sun, 30 Jul 2023 10:42:02 +0200, Piotr Wyderski
<bombald@protonmail.com> wrote:
>John Larkin wrote:
>
>> Diodes has some dynamite new switchers, the AP66200 parts, that I'd
>> like to use.
>
>What makes them dynamite as compared to, say, TPS564201?
>Looks like a boring synchronous buck. What do I miss?
>
> Best regards, Piotr
I love those little TI switchers. We use tons of them. But the best
they can do on input voltage is 28, and their current rating is
optimistic, in my opinion.
I have 48 volts and need to make 12 to power a brickwall of relays.
HV-input switchers are rare. I'd use the trusty old LM2576HV but I
don't have room for it and its giant catch diode and inductor and
caps. It switches at 50 KHz!
Pulse skip mode is fine for powering relays but can cause problems in
analog circuits, so we prefer to stock parts that don't do that. The
TIs do have spread-spectrum, which is cool for EMI tests and somehow
doesn't seem to get into the DC output. Clever.
SRH05 is a very cool part, but only good for 500 mA out. I guess I
could use four of them.
48 volts is pretty common lately, PoE and such and maybe even in cars.
I expect more HV-input switchers in the future.
(Being a Red-Blooded Real American, I say "the future.")
Reply by Lasse Langwadt Christensen●July 30, 20232023-07-30
søndag den 30. juli 2023 kl. 10.42.09 UTC+2 skrev Piotr Wyderski:
> John Larkin wrote:
>
> > Diodes has some dynamite new switchers, the AP66200 parts, that I'd
> > like to use.
> What makes them dynamite as compared to, say, TPS564201?
> Looks like a boring synchronous buck. What do I miss?
maybe not much, but the AP66200 can take 60v input
Reply by Piotr Wyderski●July 30, 20232023-07-30
John Larkin wrote:
> Diodes has some dynamite new switchers, the AP66200 parts, that I'd
> like to use.
What makes them dynamite as compared to, say, TPS564201?
Looks like a boring synchronous buck. What do I miss?
Best regards, Piotr
Reply by John Larkin●July 21, 20232023-07-21
On Fri, 21 Jul 2023 09:11:12 -0700 (PDT), Lasse Langwadt Christensen
<langwadt@fonz.dk> wrote:
>fredag den 21. juli 2023 kl. 17.56.49 UTC+2 skrev John Larkin:
>> On Fri, 21 Jul 2023 05:16:14 -0700 (PDT), Lasse Langwadt Christensen
>> <lang...@fonz.dk> wrote:
>>
>> >fredag den 21. juli 2023 kl. 13.49.55 UTC+2 skrev John Larkin:
>> >> On Thu, 20 Jul 2023 14:00:59 -0700 (PDT), Lasse Langwadt Christensen
>> >> <lang...@fonz.dk> wrote:
>> >>
>> >> >torsdag den 20. juli 2023 kl. 22.57.57 UTC+2 skrev John Larkin:
>> >> >> On Thu, 20 Jul 2023 12:01:35 -0700 (PDT), Lasse Langwadt Christensen
>> >> >> <lang...@fonz.dk> wrote:
>> >> >>
>> >> >> >torsdag den 20. juli 2023 kl. 19.51.24 UTC+2 skrev John Larkin:
>> >> >> >> On Thu, 20 Jul 2023 09:02:15 -0700 (PDT), Lasse Langwadt Christensen
>> >> >> >> <lang...@fonz.dk> wrote:
>> >> >> >>
>> >> >> >> >torsdag den 20. juli 2023 kl. 16.44.27 UTC+2 skrev John Larkin:
>> >> >> >> >> On Thu, 20 Jul 2023 02:54:02 -0700 (PDT), Lasse Langwadt Christensen
>> >> >> >> >> <lang...@fonz.dk> wrote:
>> >> >> >> >>
>> >> >> >> >> >torsdag den 20. juli 2023 kl. 02.53.23 UTC+2 skrev John Larkin:
>> >> >> >> >> >> On Wed, 19 Jul 2023 16:48:11 -0700 (PDT), Lasse Langwadt Christensen
>> >> >> >> >> >> <lang...@fonz.dk> wrote:
>> >> >> >> >> >>
>> >> >> >> >> >> >torsdag den 20. juli 2023 kl. 01.16.45 UTC+2 skrev John Larkin:
>> >> >> >> >> >> >> On Wed, 19 Jul 2023 12:15:17 -0700 (PDT), Eddy Lee
>> >> >> >> >> >> >> <eddy7...@gmail.com> wrote:
>> >> >> >> >> >> >> >On Wednesday, July 19, 2023 at 11:08:24?AM UTC-7, John Larkin wrote:
>> >> >> >> >> >> >> >> On Wed, 19 Jul 2023 16:31:27 -0000 (UTC), Phil Hobbs
>> >> >> >> >> >> >> >> <pcdhSpamM...@electrooptical.net> wrote:
>> >> >> >> >> >> >> >>
>> >> >> >> >> >> >> >> >John Larkin <jla...@highlandSNIPMEtechnology.com> wrote:
>> >> >> >> >> >> >> >> >> Diodes is lately a lot more than diodes. A similar case is Onsemi, a
>> >> >> >> >> >> >> >> >> spinoff of Motorola that inherited the cheap gumdrop business but does
>> >> >> >> >> >> >> >> >> a lot more now.
>> >> >> >> >> >> >> >> >>
>> >> >> >> >> >> >> >> >> Diodes has some dynamite new switchers, the AP66200 parts, that I'd
>> >> >> >> >> >> >> >> >> like to use.
>> >> >> >> >> >> >> >> >>
>> >> >> >> >> >> >> >> >> Any experience with Diodes, as regards support and especially keeping
>> >> >> >> >> >> >> >> >> parts in production? Any horror stories?
>> >> >> >> >> >> >> >> >>
>> >> >> >> >> >> >> >> >>
>> >> >> >> >> >> >> >> >
>> >> >> >> >> >> >> >> >We�ve had good luck with them. Some of their analog parts are odd, both for
>> >> >> >> >> >> >> >> >good or ill.
>> >> >> >> >> >> >> >> >
>> >> >> >> >> >> >> >> >Their TLV431 is better than TI�s original, and much much better than the
>> >> >> >> >> >> >> >> >onsemi version.
>> >> >> >> >> >> >> >> >
>> >> >> >> >> >> >> >> >Their TLC271 is, like, 30 dB noisier than TI�s.
>> >> >> >> >> >> >> >> >
>> >> >> >> >> >> >> >> >Haven�t used any of their switcher parts.
>> >> >> >> >> >> >> >> >
>> >> >> >> >> >> >> >> >Cheers
>> >> >> >> >> >> >> >> >
>> >> >> >> >> >> >> >> >Phil Hobbs
>> >> >> >> >> >> >> >> That switcher looks fabulous. I'm designing a board with maybe 125
>> >> >> >> >> >> >> >> 12-volt-coil relays and I have to switch +48 down to 12 and have about
>> >> >> >> >> >> >> >> no room to do it in. HV-input switchers are rare. I like the old
>> >> >> >> >> >> >> >> LM2576HV-ADJ but it, and its inductor and caps, are gigantic.
>> >> >> >> >> >> >> >
>> >> >> >> >> >> >> >How many of them are switched on at the same time? How do you deal with the initial surge? I am looking for a way to have a short delay to sequence on 24V relay coils.
>> >> >> >> >> >> >> Rough guess, maybe 50 max on at once. What surge do you expect?
>> >> >> >> >> >> >>
>> >> >> >> >> >> >> One trick is to bump up the coil bus voltage to 12 for a while
>> >> >> >> >> >> >> whenever any relay state is changed, and drop it down to the
>> >> >> >> >> >> >> guaranteed holding voltage after a short while. But my relays will
>> >> >> >> >> >> >> only need maybe 200 mW coil power each, under an amp for all 50 on, so
>> >> >> >> >> >> >> that's not worth the hassle here.
>> >> >> >> >> >> >>
>> >> >> >> >> >> >> I'm thinking of using an Efinix FPGA and ULN2003s as the coil drivers.
>> >> >> >> >> >> >
>> >> >> >> >> >> >this would save routing, https://www.ti.com/product/TPIC6B595
>> >> >> >> >> >> We use that part in some other places. But it's physically big and
>> >> >> >> >> >> they would actually cost a tad more than the fpga+uln drivers. I can
>> >> >> >> >> >> squeeze the ULN2003s on the bottom, between the relay pins.
>> >> >> >> >> >>
>> >> >> >> >> >> Diodes Inc does the ULN2003 too.
>> >> >> >> >> >>
>> >> >> >> >> >> A single mosfet under each relay would be cool, but I'd need a catch
>> >> >> >> >> >> diode or an avalanche-rated part.
>> >> >> >> >> >
>> >> >> >> >> >https://eu.mouser.com/ProductDetail/Toshiba/SSM3K357RLF?qs=F5EMLAvA7ICLn4Y138U4bA%3D%3D&_gl=1*1l31jut*_ga*MjE3OTg0NTgyLjE2Njg0NDMwOTE.*_ga_15W4STQT4T*MTY4OTg0NjM4MC4yOS4xLjE2ODk4NDYzOTUuMC4wLjA.*_ga_1KQLCYKRX3*MTY4OTg0NjM4MC4xNy4xLjE2ODk4NDYzOTUuNDUuMC4w
>> >> >> >> >> It avalanches but they don't specify the voltage!
>> >> >> >> >
>> >> >> >> >the drain-gate zener makes it shunt so one would assume it does that at ~60V and avalanche never happens
>> >> >> >> Abs max is 60 volts, without a clue what that zener clamps at.
>> >> >> >>
>> >> >> >> Same as avalanche-rated fets, no hint of the actual clamp voltage.
>> >> >> >> Grrrr.
>> >> >> >
>> >> >> >https://www.diodes.com/assets/Articles/Automotive-MOSFETs.pdf
>> >> >> >
>> >> >> >page 2
>> >> >> >
>> >> >> >
>> >> >> Lots of words, no numbers.
>> >> >
>> >> >atleast 60 and less than the mosfets avalance voltage ..
>> >> which is never even hinted it.
>> >
>> >"The back?to?back Zener stack, shown in figure 2, between the MOSFET�s gate and drain connections, is the key
>> >component in this low?side, active clamp configuration. The clamp voltage is set by the Zener stack voltage and
>> >is designed to be less than the avalanche breakdown voltage of the MOSFET�s drain to source junction but high
>> >enough not to be triggered in normal operation."
>> Why isn't that voltage on the data sheet?
>
>60V ...
>
No. It doesn't zener at 60. What would the flyback voltage actually
be? It might be sorta guessed from the abs max flyback energy, but I'd
like to know the decay time when driving a relay. All the curves,
including SOAR, end at 60 volts.
The pattern I've seen in discretes is breakdown voltages between 1.5
and 2.0 x specified abs max. Most go into soft conduction (like the
EPC GaN parts) and some fail and die hard, even current limited.
I want to know that.
I've seen some ICs, like LM1117, that work at 5x rated abs max. One
opamp wasn't even reliable at abs max.
Reply by Lasse Langwadt Christensen●July 21, 20232023-07-21
fredag den 21. juli 2023 kl. 17.56.49 UTC+2 skrev John Larkin:
> On Fri, 21 Jul 2023 05:16:14 -0700 (PDT), Lasse Langwadt Christensen
> <lang...@fonz.dk> wrote:
>
> >fredag den 21. juli 2023 kl. 13.49.55 UTC+2 skrev John Larkin:
> >> On Thu, 20 Jul 2023 14:00:59 -0700 (PDT), Lasse Langwadt Christensen
> >> <lang...@fonz.dk> wrote:
> >>
> >> >torsdag den 20. juli 2023 kl. 22.57.57 UTC+2 skrev John Larkin:
> >> >> On Thu, 20 Jul 2023 12:01:35 -0700 (PDT), Lasse Langwadt Christensen
> >> >> <lang...@fonz.dk> wrote:
> >> >>
> >> >> >torsdag den 20. juli 2023 kl. 19.51.24 UTC+2 skrev John Larkin:
> >> >> >> On Thu, 20 Jul 2023 09:02:15 -0700 (PDT), Lasse Langwadt Christensen
> >> >> >> <lang...@fonz.dk> wrote:
> >> >> >>
> >> >> >> >torsdag den 20. juli 2023 kl. 16.44.27 UTC+2 skrev John Larkin:
> >> >> >> >> On Thu, 20 Jul 2023 02:54:02 -0700 (PDT), Lasse Langwadt Christensen
> >> >> >> >> <lang...@fonz.dk> wrote:
> >> >> >> >>
> >> >> >> >> >torsdag den 20. juli 2023 kl. 02.53.23 UTC+2 skrev John Larkin:
> >> >> >> >> >> On Wed, 19 Jul 2023 16:48:11 -0700 (PDT), Lasse Langwadt Christensen
> >> >> >> >> >> <lang...@fonz.dk> wrote:
> >> >> >> >> >>
> >> >> >> >> >> >torsdag den 20. juli 2023 kl. 01.16.45 UTC+2 skrev John Larkin:
> >> >> >> >> >> >> On Wed, 19 Jul 2023 12:15:17 -0700 (PDT), Eddy Lee
> >> >> >> >> >> >> <eddy7...@gmail.com> wrote:
> >> >> >> >> >> >> >On Wednesday, July 19, 2023 at 11:08:24?AM UTC-7, John Larkin wrote:
> >> >> >> >> >> >> >> On Wed, 19 Jul 2023 16:31:27 -0000 (UTC), Phil Hobbs
> >> >> >> >> >> >> >> <pcdhSpamM...@electrooptical.net> wrote:
> >> >> >> >> >> >> >>
> >> >> >> >> >> >> >> >John Larkin <jla...@highlandSNIPMEtechnology.com> wrote:
> >> >> >> >> >> >> >> >> Diodes is lately a lot more than diodes. A similar case is Onsemi, a
> >> >> >> >> >> >> >> >> spinoff of Motorola that inherited the cheap gumdrop business but does
> >> >> >> >> >> >> >> >> a lot more now.
> >> >> >> >> >> >> >> >>
> >> >> >> >> >> >> >> >> Diodes has some dynamite new switchers, the AP66200 parts, that I'd
> >> >> >> >> >> >> >> >> like to use.
> >> >> >> >> >> >> >> >>
> >> >> >> >> >> >> >> >> Any experience with Diodes, as regards support and especially keeping
> >> >> >> >> >> >> >> >> parts in production? Any horror stories?
> >> >> >> >> >> >> >> >>
> >> >> >> >> >> >> >> >>
> >> >> >> >> >> >> >> >
> >> >> >> >> >> >> >> >We’ve had good luck with them. Some of their analog parts are odd, both for
> >> >> >> >> >> >> >> >good or ill.
> >> >> >> >> >> >> >> >
> >> >> >> >> >> >> >> >Their TLV431 is better than TI’s original, and much much better than the
> >> >> >> >> >> >> >> >onsemi version.
> >> >> >> >> >> >> >> >
> >> >> >> >> >> >> >> >Their TLC271 is, like, 30 dB noisier than TI’s.
> >> >> >> >> >> >> >> >
> >> >> >> >> >> >> >> >Haven’t used any of their switcher parts.
> >> >> >> >> >> >> >> >
> >> >> >> >> >> >> >> >Cheers
> >> >> >> >> >> >> >> >
> >> >> >> >> >> >> >> >Phil Hobbs
> >> >> >> >> >> >> >> That switcher looks fabulous. I'm designing a board with maybe 125
> >> >> >> >> >> >> >> 12-volt-coil relays and I have to switch +48 down to 12 and have about
> >> >> >> >> >> >> >> no room to do it in. HV-input switchers are rare. I like the old
> >> >> >> >> >> >> >> LM2576HV-ADJ but it, and its inductor and caps, are gigantic.
> >> >> >> >> >> >> >
> >> >> >> >> >> >> >How many of them are switched on at the same time? How do you deal with the initial surge? I am looking for a way to have a short delay to sequence on 24V relay coils.
> >> >> >> >> >> >> Rough guess, maybe 50 max on at once. What surge do you expect?
> >> >> >> >> >> >>
> >> >> >> >> >> >> One trick is to bump up the coil bus voltage to 12 for a while
> >> >> >> >> >> >> whenever any relay state is changed, and drop it down to the
> >> >> >> >> >> >> guaranteed holding voltage after a short while. But my relays will
> >> >> >> >> >> >> only need maybe 200 mW coil power each, under an amp for all 50 on, so
> >> >> >> >> >> >> that's not worth the hassle here.
> >> >> >> >> >> >>
> >> >> >> >> >> >> I'm thinking of using an Efinix FPGA and ULN2003s as the coil drivers.
> >> >> >> >> >> >
> >> >> >> >> >> >this would save routing, https://www.ti.com/product/TPIC6B595
> >> >> >> >> >> We use that part in some other places. But it's physically big and
> >> >> >> >> >> they would actually cost a tad more than the fpga+uln drivers. I can
> >> >> >> >> >> squeeze the ULN2003s on the bottom, between the relay pins.
> >> >> >> >> >>
> >> >> >> >> >> Diodes Inc does the ULN2003 too.
> >> >> >> >> >>
> >> >> >> >> >> A single mosfet under each relay would be cool, but I'd need a catch
> >> >> >> >> >> diode or an avalanche-rated part.
> >> >> >> >> >
> >> >> >> >> >https://eu.mouser.com/ProductDetail/Toshiba/SSM3K357RLF?qs=F5EMLAvA7ICLn4Y138U4bA%3D%3D&_gl=1*1l31jut*_ga*MjE3OTg0NTgyLjE2Njg0NDMwOTE.*_ga_15W4STQT4T*MTY4OTg0NjM4MC4yOS4xLjE2ODk4NDYzOTUuMC4wLjA.*_ga_1KQLCYKRX3*MTY4OTg0NjM4MC4xNy4xLjE2ODk4NDYzOTUuNDUuMC4w
> >> >> >> >> It avalanches but they don't specify the voltage!
> >> >> >> >
> >> >> >> >the drain-gate zener makes it shunt so one would assume it does that at ~60V and avalanche never happens
> >> >> >> Abs max is 60 volts, without a clue what that zener clamps at.
> >> >> >>
> >> >> >> Same as avalanche-rated fets, no hint of the actual clamp voltage.
> >> >> >> Grrrr.
> >> >> >
> >> >> >https://www.diodes.com/assets/Articles/Automotive-MOSFETs.pdf
> >> >> >
> >> >> >page 2
> >> >> >
> >> >> >
> >> >> Lots of words, no numbers.
> >> >
> >> >atleast 60 and less than the mosfets avalance voltage ..
> >> which is never even hinted it.
> >
> >"The back?to?back Zener stack, shown in figure 2, between the MOSFET’s gate and drain connections, is the key
> >component in this low?side, active clamp configuration. The clamp voltage is set by the Zener stack voltage and
> >is designed to be less than the avalanche breakdown voltage of the MOSFET’s drain to source junction but high
> >enough not to be triggered in normal operation."
> Why isn't that voltage on the data sheet?
60V ...
Reply by John Larkin●July 21, 20232023-07-21
On Fri, 21 Jul 2023 05:16:14 -0700 (PDT), Lasse Langwadt Christensen
<langwadt@fonz.dk> wrote:
>fredag den 21. juli 2023 kl. 13.49.55 UTC+2 skrev John Larkin:
>> On Thu, 20 Jul 2023 14:00:59 -0700 (PDT), Lasse Langwadt Christensen
>> <lang...@fonz.dk> wrote:
>>
>> >torsdag den 20. juli 2023 kl. 22.57.57 UTC+2 skrev John Larkin:
>> >> On Thu, 20 Jul 2023 12:01:35 -0700 (PDT), Lasse Langwadt Christensen
>> >> <lang...@fonz.dk> wrote:
>> >>
>> >> >torsdag den 20. juli 2023 kl. 19.51.24 UTC+2 skrev John Larkin:
>> >> >> On Thu, 20 Jul 2023 09:02:15 -0700 (PDT), Lasse Langwadt Christensen
>> >> >> <lang...@fonz.dk> wrote:
>> >> >>
>> >> >> >torsdag den 20. juli 2023 kl. 16.44.27 UTC+2 skrev John Larkin:
>> >> >> >> On Thu, 20 Jul 2023 02:54:02 -0700 (PDT), Lasse Langwadt Christensen
>> >> >> >> <lang...@fonz.dk> wrote:
>> >> >> >>
>> >> >> >> >torsdag den 20. juli 2023 kl. 02.53.23 UTC+2 skrev John Larkin:
>> >> >> >> >> On Wed, 19 Jul 2023 16:48:11 -0700 (PDT), Lasse Langwadt Christensen
>> >> >> >> >> <lang...@fonz.dk> wrote:
>> >> >> >> >>
>> >> >> >> >> >torsdag den 20. juli 2023 kl. 01.16.45 UTC+2 skrev John Larkin:
>> >> >> >> >> >> On Wed, 19 Jul 2023 12:15:17 -0700 (PDT), Eddy Lee
>> >> >> >> >> >> <eddy7...@gmail.com> wrote:
>> >> >> >> >> >> >On Wednesday, July 19, 2023 at 11:08:24?AM UTC-7, John Larkin wrote:
>> >> >> >> >> >> >> On Wed, 19 Jul 2023 16:31:27 -0000 (UTC), Phil Hobbs
>> >> >> >> >> >> >> <pcdhSpamM...@electrooptical.net> wrote:
>> >> >> >> >> >> >>
>> >> >> >> >> >> >> >John Larkin <jla...@highlandSNIPMEtechnology.com> wrote:
>> >> >> >> >> >> >> >> Diodes is lately a lot more than diodes. A similar case is Onsemi, a
>> >> >> >> >> >> >> >> spinoff of Motorola that inherited the cheap gumdrop business but does
>> >> >> >> >> >> >> >> a lot more now.
>> >> >> >> >> >> >> >>
>> >> >> >> >> >> >> >> Diodes has some dynamite new switchers, the AP66200 parts, that I'd
>> >> >> >> >> >> >> >> like to use.
>> >> >> >> >> >> >> >>
>> >> >> >> >> >> >> >> Any experience with Diodes, as regards support and especially keeping
>> >> >> >> >> >> >> >> parts in production? Any horror stories?
>> >> >> >> >> >> >> >>
>> >> >> >> >> >> >> >>
>> >> >> >> >> >> >> >
>> >> >> >> >> >> >> >We�ve had good luck with them. Some of their analog parts are odd, both for
>> >> >> >> >> >> >> >good or ill.
>> >> >> >> >> >> >> >
>> >> >> >> >> >> >> >Their TLV431 is better than TI�s original, and much much better than the
>> >> >> >> >> >> >> >onsemi version.
>> >> >> >> >> >> >> >
>> >> >> >> >> >> >> >Their TLC271 is, like, 30 dB noisier than TI�s.
>> >> >> >> >> >> >> >
>> >> >> >> >> >> >> >Haven�t used any of their switcher parts.
>> >> >> >> >> >> >> >
>> >> >> >> >> >> >> >Cheers
>> >> >> >> >> >> >> >
>> >> >> >> >> >> >> >Phil Hobbs
>> >> >> >> >> >> >> That switcher looks fabulous. I'm designing a board with maybe 125
>> >> >> >> >> >> >> 12-volt-coil relays and I have to switch +48 down to 12 and have about
>> >> >> >> >> >> >> no room to do it in. HV-input switchers are rare. I like the old
>> >> >> >> >> >> >> LM2576HV-ADJ but it, and its inductor and caps, are gigantic.
>> >> >> >> >> >> >
>> >> >> >> >> >> >How many of them are switched on at the same time? How do you deal with the initial surge? I am looking for a way to have a short delay to sequence on 24V relay coils.
>> >> >> >> >> >> Rough guess, maybe 50 max on at once. What surge do you expect?
>> >> >> >> >> >>
>> >> >> >> >> >> One trick is to bump up the coil bus voltage to 12 for a while
>> >> >> >> >> >> whenever any relay state is changed, and drop it down to the
>> >> >> >> >> >> guaranteed holding voltage after a short while. But my relays will
>> >> >> >> >> >> only need maybe 200 mW coil power each, under an amp for all 50 on, so
>> >> >> >> >> >> that's not worth the hassle here.
>> >> >> >> >> >>
>> >> >> >> >> >> I'm thinking of using an Efinix FPGA and ULN2003s as the coil drivers.
>> >> >> >> >> >
>> >> >> >> >> >this would save routing, https://www.ti.com/product/TPIC6B595
>> >> >> >> >> We use that part in some other places. But it's physically big and
>> >> >> >> >> they would actually cost a tad more than the fpga+uln drivers. I can
>> >> >> >> >> squeeze the ULN2003s on the bottom, between the relay pins.
>> >> >> >> >>
>> >> >> >> >> Diodes Inc does the ULN2003 too.
>> >> >> >> >>
>> >> >> >> >> A single mosfet under each relay would be cool, but I'd need a catch
>> >> >> >> >> diode or an avalanche-rated part.
>> >> >> >> >
>> >> >> >> >https://eu.mouser.com/ProductDetail/Toshiba/SSM3K357RLF?qs=F5EMLAvA7ICLn4Y138U4bA%3D%3D&_gl=1*1l31jut*_ga*MjE3OTg0NTgyLjE2Njg0NDMwOTE.*_ga_15W4STQT4T*MTY4OTg0NjM4MC4yOS4xLjE2ODk4NDYzOTUuMC4wLjA.*_ga_1KQLCYKRX3*MTY4OTg0NjM4MC4xNy4xLjE2ODk4NDYzOTUuNDUuMC4w
>> >> >> >> It avalanches but they don't specify the voltage!
>> >> >> >
>> >> >> >the drain-gate zener makes it shunt so one would assume it does that at ~60V and avalanche never happens
>> >> >> Abs max is 60 volts, without a clue what that zener clamps at.
>> >> >>
>> >> >> Same as avalanche-rated fets, no hint of the actual clamp voltage.
>> >> >> Grrrr.
>> >> >
>> >> >https://www.diodes.com/assets/Articles/Automotive-MOSFETs.pdf
>> >> >
>> >> >page 2
>> >> >
>> >> >
>> >> Lots of words, no numbers.
>> >
>> >atleast 60 and less than the mosfets avalance voltage ..
>> which is never even hinted it.
>
>"The back?to?back Zener stack, shown in figure 2, between the MOSFET�s gate and drain connections, is the key
>component in this low?side, active clamp configuration. The clamp voltage is set by the Zener stack voltage and
>is designed to be less than the avalanche breakdown voltage of the MOSFET�s drain to source junction but high
>enough not to be triggered in normal operation."
Why isn't that voltage on the data sheet?
>
>>
>> I buy parts and test them. Sometimes that's really interesting. Some
>> parts are fine at 5x abs max voltage, some blow up at 10% over.
>
>and sometimes the damage doesn't show up until later, it all depends on what you are doing
I tested a bunch of polymer electrolytics at reverse voltage for over
a month. That's not specified. Leakage current very gradually
declined, so they looked OK for production.
https://www.dropbox.com/s/hcoobynzsh4wonv/Polymer_Rev_Test.JPG?raw=1
I also tested some $300 distributed amps for max voltage. They started
leaking but didn't fail at 16 volts, so we run them at 12, very
current limited, in what might be called a "novel" bias circuit.