This feels like it should be simple, but my google-fu and electronics-fu are
failing today.

I need a switch-mode current /sink/, something that looks like this:

Vin o-------
|
---
| L |
| o |
| a |
| d |
---
|
v Ireg
|
-------
| |
| ??? |
-------
|
Gnd o-------

Vin is +24V, Ireg variable up to ~1A (assumed possible with somewhere below
24V across the load). Load is resistive.

For reasons, the load must be connected to Vin, not Gnd (there are several
loads, each requiring separate control, with a common +ve that's non-negotiable)

Any pointers to a suitable topology / part?

(I could probably work out the combination of FET, inductor and diode that works,
but worry that wiring it "upside-down" may cause stability problems if I just use
a standard buck controller...)

--
Ian

"Tamahome!!!" - "Miaka!!!"

torsdag den 9. december 2021 kl. 12.48.49 UTC+1 skrev Ian:
> This feels like it should be simple, but my google-fu and electronics-fu are
> failing today.
>
> I need a switch-mode current /sink/, something that looks like this:
>
> Vin o-------
> |
> ---
> | L |
> | o |
> | a |
> | d |
> ---
> |
> v Ireg
> |
> -------
> | |
> | ??? |
> -------
> |
> Gnd o-------
>
> Vin is +24V, Ireg variable up to ~1A (assumed possible with somewhere below
> 24V across the load). Load is resistive.
>
> For reasons, the load must be connected to Vin, not Gnd (there are several
> loads, each requiring separate control, with a common +ve that's non-negotiable)
>
> Any pointers to a suitable topology / part?
>
> (I could probably work out the combination of FET, inductor and diode that works,
> but worry that wiring it "upside-down" may cause stability problems if I just use
> a standard buck controller...)
>

look at LED drivers, quite few of those that are constant current "upside down" buck

On 2021-12-09, Lasse Langwadt Christensen <langwadt@fonz.dk> wrote:
> torsdag den 9. december 2021 kl. 12.48.49 UTC+1 skrev Ian:
>>
>> I need a switch-mode current /sink/, something that looks like this:
>>
>
> look at LED drivers, quite few of those that are constant current "upside down" buck

Great call, thanks. Sounds like that could be just what I need.

--
Ian

"Tamahome!!!" - "Miaka!!!"

On 2021-12-09, Ian <${send-direct-email-to-news1021-at-jusme-dot-com-if-you-must}@jusme.com> wrote:
> On 2021-12-09, Lasse Langwadt Christensen <langwadt@fonz.dk> wrote:
>> torsdag den 9. december 2021 kl. 12.48.49 UTC+1 skrev Ian:
>>>
>>> I need a switch-mode current /sink/, something that looks like this:
>>>
>>
>> look at LED drivers, quite few of those that are constant current "upside down" buck
>
> Great call, thanks. Sounds like that could be just what I need.
>

This might do, if it wasn't made from enriched unobtainium:

https://uk.farnell.com/microchip/hv9911ng-g/led-driver-40-to-85deg-c/dp/3580669?ost=hv9911

("Back order today stock arriving week commencing 03/02/23")

But it's a starting point, and the datasheet has the topology I was looking for.

--
Ian

"Tamahome!!!" - "Miaka!!!"

torsdag den 9. december 2021 kl. 13.48.44 UTC+1 skrev Ian:
> On 2021-12-09, Ian <${send-direct-email-to-news1021-at-jusme-dot-com-if-you-must}@jusme.com> wrote:
> > On 2021-12-09, Lasse Langwadt Christensen <lang...@fonz.dk> wrote:
> >> torsdag den 9. december 2021 kl. 12.48.49 UTC+1 skrev Ian:
> >>>
> >>> I need a switch-mode current /sink/, something that looks like this:
> >>>
> >>
> >> look at LED drivers, quite few of those that are constant current "upside down" buck
> >
> > Great call, thanks. Sounds like that could be just what I need.
> >
> This might do, if it wasn't made from enriched unobtainium:
>
> https://uk.farnell.com/microchip/hv9911ng-g/led-driver-40-to-85deg-c/dp/3580669?ost=hv9911
>
> ("Back order today stock arriving week commencing 03/02/23")
>
> But it's a starting point, and the datasheet has the topology I was looking for.

https://www.aliexpress.com/item/32820709549.html

On Thu, 9 Dec 2021 12:48:37 -0000 (UTC), Ian
<${send-direct-email-to-news1021-at-jusme-dot-com-if-you-must}@jusme.com>
wrote:

>On 2021-12-09, Ian <${send-direct-email-to-news1021-at-jusme-dot-com-if-you-must}@jusme.com> wrote:
>> On 2021-12-09, Lasse Langwadt Christensen <langwadt@fonz.dk> wrote:
>>> torsdag den 9. december 2021 kl. 12.48.49 UTC+1 skrev Ian:
>>>>
>>>> I need a switch-mode current /sink/, something that looks like this:
>>>>
>>>
>>> look at LED drivers, quite few of those that are constant current "upside down" buck
>>
>> Great call, thanks. Sounds like that could be just what I need.
>>
>
>This might do, if it wasn't made from enriched unobtainium:
>
>https://uk.farnell.com/microchip/hv9911ng-g/led-driver-40-to-85deg-c/dp/3580669?ost=hv9911
>
>("Back order today stock arriving week commencing 03/02/23")
>
>But it's a starting point, and the datasheet has the topology I was looking for.

A buck-boost is difficult to turn into a constant current regulator
without added control features (as with all flybacks) - hence the
addition of HV7800 in the led driver application diagram.

RL

Ian wrote:
> Ian wrote:
>> Lasse Langwadt Christensen wrote:
>>> Ian wrote:
>>>>
>>>> I need a switch-mode current /sink/, something that looks like this:
>>>>
>>>
>>> look at LED drivers, quite few of those that are constant current "upside down" buck
>>
>> Great call, thanks. Sounds like that could be just what I need.
>>
>
> This might do, if it wasn't made from enriched unobtainium:
>
> https://uk.farnell.com/microchip/hv9911ng-g/led-driver-40-to-85deg-c/dp/3580669?ost=hv9911
>
> ("Back order today stock arriving week commencing 03/02/23")
>
> But it's a starting point, and the datasheet has the topology I was looking for.

There's some stock available at mouser:

https://www.mouser.com/c/?q=hv9911

Danke,

--
Don, KB7RPU, https://www.qsl.net/kb7rpu
There was a young lady named Bright Whose speed was far faster than light;
She set out one day In a relative way And returned on the previous night.

09.12.21 12:48, Ian wrote:
>This feels like it should be simple, but my google-fu and electronics-fu are
>failing today.
>
>I need a switch-mode current /sink/, something that looks like this:
>
> Vin o-------
> |
> ---
> | L |
> | o |
> | a |
> | d |
> ---
> |
> v Ireg
> |
> -------
> | |
> | ??? |
> -------
> |
> Gnd o-------
>
>Vin is +24V, Ireg variable up to ~1A (assumed possible with somewhere below
>24V across the load). Load is resistive.
>
>For reasons, the load must be connected to Vin, not Gnd (there are several
>loads, each requiring separate control, with a common +ve that's non-negotiable)
>
>Any pointers to a suitable topology / part?
>
>(I could probably work out the combination of FET, inductor and diode that works,
>but worry that wiring it "upside-down" may cause stability problems if I just use
>a standard buck controller...)
>
A buck converter just set for lower output voltage than your high side. Synchronous half bridge. Current sense in the return path

--
Klaus

On Thursday, December 9, 2021 at 4:48:44 AM UTC-8, Ian wrote:
> On 2021-12-09, Ian <${send-direct-email-to-news1021-at-jusme-dot-com-if-you-must}@jusme.com> wrote:
> > On 2021-12-09, Lasse Langwadt Christensen <lang...@fonz.dk> wrote:
> >> torsdag den 9. december 2021 kl. 12.48.49 UTC+1 skrev Ian:
> >>>
> >>> I need a switch-mode current /sink/, something that looks like this:
> >>>
> >>
> >> look at LED drivers, quite few of those that are constant current "upside down" buck
> >
> > Great call, thanks. Sounds like that could be just what I need.
> >
> This might do, if it wasn't made from enriched unobtainium:
>
> https://uk.farnell.com/microchip/hv9911ng-g/led-driver-40-to-85deg-c/dp/3580669?ost=hv9911
>
> ("Back order today stock arriving week commencing 03/02/23")
>
> But it's a starting point, and the datasheet has the topology I was looking for.
> --
> Ian
>
> "Tamahome!!!" - "Miaka!!!"

Vin of 24V. What is Vload?

On 09/12/2021 11:48, Ian wrote:
> This feels like it should be simple, but my google-fu and electronics-fu are
> failing today.
>
> I need a switch-mode current /sink/, something that looks like this:
>
> Vin o-------
> |
> ---
> | L |
> | o |
> | a |
> | d |
> ---
> |
> v Ireg
> |
> -------
> | |
> | ??? |
> -------
> |
> Gnd o-------
>
> Vin is +24V, Ireg variable up to ~1A (assumed possible with somewhere below
> 24V across the load). Load is resistive.
>
> For reasons, the load must be connected to Vin, not Gnd (there are several
> loads, each requiring separate control, with a common +ve that's non-negotiable)
>
> Any pointers to a suitable topology / part?
>
> (I could probably work out the combination of FET, inductor and diode that works,
> but worry that wiring it "upside-down" may cause stability problems if I just use
> a standard buck controller...)
>

Can't resist a challenge! Here is an untried discrete hysteretic buck
sink. Q1-3-4 are the switcher, Iload is set by Q1vbe and R10. To get
lower load currents Q2 pulls current thru R9 to cheat the feedback. At
some values tempco of Q2 may approximately cancel tempco Q1?

www.dropbox.com/s/6ekvkgopqtrbrfv/20211209_BuckSink_SED_Ian.pdf?raw=1

piglet

On Thursday, December 9, 2021 at 10:34:16 AM UTC-8, erichp...@hotmail.com wrote:
> On 09/12/2021 11:48, Ian wrote:
> > This feels like it should be simple, but my google-fu and electronics-fu are
> > failing today.
> >
> > I need a switch-mode current /sink/, something that looks like this:
> >
> > Vin o-------
> > |
> > ---
> > | L |
> > | o |
> > | a |
> > | d |
> > ---
> > |
> > v Ireg
> > |
> > -------
> > | |
> > | ??? |
> > -------
> > |
> > Gnd o-------
> >
> > Vin is +24V, Ireg variable up to ~1A (assumed possible with somewhere below
> > 24V across the load). Load is resistive.
> >
> > For reasons, the load must be connected to Vin, not Gnd (there are several
> > loads, each requiring separate control, with a common +ve that's non-negotiable)
> >
> > Any pointers to a suitable topology / part?
> >
> > (I could probably work out the combination of FET, inductor and diode that works,
> > but worry that wiring it "upside-down" may cause stability problems if I just use
> > a standard buck controller...)
> >
> Can't resist a challenge! Here is an untried discrete hysteretic buck
> sink. Q1-3-4 are the switcher, Iload is set by Q1vbe and R10. To get
> lower load currents Q2 pulls current thru R9 to cheat the feedback. At
> some values tempco of Q2 may approximately cancel tempco Q1?
>
> www.dropbox.com/s/6ekvkgopqtrbrfv/20211209_BuckSink_SED_Ian.pdf?raw=1
>
> piglet

I would use a shunt regulator, but pending Vload answer from the OP.

torsdag den 9. december 2021 kl. 19.39.24 UTC+1 skrev Ed Lee:
> On Thursday, December 9, 2021 at 10:34:16 AM UTC-8, erichp...@hotmail.com wrote:
> > On 09/12/2021 11:48, Ian wrote:
> > > This feels like it should be simple, but my google-fu and electronics-fu are
> > > failing today.
> > >
> > > I need a switch-mode current /sink/, something that looks like this:
> > >
> > > Vin o-------
> > > |
> > > ---
> > > | L |
> > > | o |
> > > | a |
> > > | d |
> > > ---
> > > |
> > > v Ireg
> > > |
> > > -------
> > > | |
> > > | ??? |
> > > -------
> > > |
> > > Gnd o-------
> > >
> > > Vin is +24V, Ireg variable up to ~1A (assumed possible with somewhere below
> > > 24V across the load). Load is resistive.
> > >
> > > For reasons, the load must be connected to Vin, not Gnd (there are several
> > > loads, each requiring separate control, with a common +ve that's non-negotiable)
> > >
> > > Any pointers to a suitable topology / part?
> > >
> > > (I could probably work out the combination of FET, inductor and diode that works,
> > > but worry that wiring it "upside-down" may cause stability problems if I just use
> > > a standard buck controller...)
> > >
> > Can't resist a challenge! Here is an untried discrete hysteretic buck
> > sink. Q1-3-4 are the switcher, Iload is set by Q1vbe and R10. To get
> > lower load currents Q2 pulls current thru R9 to cheat the feedback. At
> > some values tempco of Q2 may approximately cancel tempco Q1?
> >
> > www.dropbox.com/s/6ekvkgopqtrbrfv/20211209_BuckSink_SED_Ian.pdf?raw=1
> >
> > piglet
> I would use a shunt regulator, but pending Vload answer from the OP.

it is a shunt regulator, it is just a switching rather than linear

On Thursday, December 9, 2021 at 10:42:57 AM UTC-8, lang...@fonz.dk wrote:
> torsdag den 9. december 2021 kl. 19.39.24 UTC+1 skrev Ed Lee:
> > On Thursday, December 9, 2021 at 10:34:16 AM UTC-8, erichp...@hotmail.com wrote:
> > > On 09/12/2021 11:48, Ian wrote:
> > > > This feels like it should be simple, but my google-fu and electronics-fu are
> > > > failing today.
> > > >
> > > > I need a switch-mode current /sink/, something that looks like this:
> > > >
> > > > Vin o-------
> > > > |
> > > > ---
> > > > | L |
> > > > | o |
> > > > | a |
> > > > | d |
> > > > ---
> > > > |
> > > > v Ireg
> > > > |
> > > > -------
> > > > | |
> > > > | ??? |
> > > > -------
> > > > |
> > > > Gnd o-------
> > > >
> > > > Vin is +24V, Ireg variable up to ~1A (assumed possible with somewhere below
> > > > 24V across the load). Load is resistive.
> > > >
> > > > For reasons, the load must be connected to Vin, not Gnd (there are several
> > > > loads, each requiring separate control, with a common +ve that's non-negotiable)
> > > >
> > > > Any pointers to a suitable topology / part?
> > > >
> > > > (I could probably work out the combination of FET, inductor and diode that works,
> > > > but worry that wiring it "upside-down" may cause stability problems if I just use
> > > > a standard buck controller...)
> > > >
> > > Can't resist a challenge! Here is an untried discrete hysteretic buck
> > > sink. Q1-3-4 are the switcher, Iload is set by Q1vbe and R10. To get
> > > lower load currents Q2 pulls current thru R9 to cheat the feedback. At
> > > some values tempco of Q2 may approximately cancel tempco Q1?
> > >
> > > www.dropbox.com/s/6ekvkgopqtrbrfv/20211209_BuckSink_SED_Ian.pdf?raw=1
> > >
> > > piglet
> > I would use a shunt regulator, but pending Vload answer from the OP.
> it is a shunt regulator, it is just a switching rather than linear

Yes, we need more info from OP. Is Vin stable? If not, what range. Is Vload resistive and/or stable. etc.

torsdag den 9. december 2021 kl. 19.46.57 UTC+1 skrev Ed Lee:
> On Thursday, December 9, 2021 at 10:42:57 AM UTC-8, lang...@fonz.dk wrote:
> > torsdag den 9. december 2021 kl. 19.39.24 UTC+1 skrev Ed Lee:
> > > On Thursday, December 9, 2021 at 10:34:16 AM UTC-8, erichp...@hotmail.com wrote:
> > > > On 09/12/2021 11:48, Ian wrote:
> > > > > This feels like it should be simple, but my google-fu and electronics-fu are
> > > > > failing today.
> > > > >
> > > > > I need a switch-mode current /sink/, something that looks like this:
> > > > >
> > > > > Vin o-------
> > > > > |
> > > > > ---
> > > > > | L |
> > > > > | o |
> > > > > | a |
> > > > > | d |
> > > > > ---
> > > > > |
> > > > > v Ireg
> > > > > |
> > > > > -------
> > > > > | |
> > > > > | ??? |
> > > > > -------
> > > > > |
> > > > > Gnd o-------
> > > > >
> > > > > Vin is +24V, Ireg variable up to ~1A (assumed possible with somewhere below
> > > > > 24V across the load). Load is resistive.
> > > > >
> > > > > For reasons, the load must be connected to Vin, not Gnd (there are several
> > > > > loads, each requiring separate control, with a common +ve that's non-negotiable)
> > > > >
> > > > > Any pointers to a suitable topology / part?
> > > > >
> > > > > (I could probably work out the combination of FET, inductor and diode that works,
> > > > > but worry that wiring it "upside-down" may cause stability problems if I just use
> > > > > a standard buck controller...)
> > > > >
> > > > Can't resist a challenge! Here is an untried discrete hysteretic buck
> > > > sink. Q1-3-4 are the switcher, Iload is set by Q1vbe and R10. To get
> > > > lower load currents Q2 pulls current thru R9 to cheat the feedback. At
> > > > some values tempco of Q2 may approximately cancel tempco Q1?
> > > >
> > > > www.dropbox.com/s/6ekvkgopqtrbrfv/20211209_BuckSink_SED_Ian.pdf?raw=1
> > > >
> > > > piglet
> > > I would use a shunt regulator, but pending Vload answer from the OP.
> > it is a shunt regulator, it is just a switching rather than linear
> Yes, we need more info from OP. Is Vin stable? If not, what range. Is Vload resistive and/or stable. etc."

" Vin is +24V, Ireg variable up to ~1A (assumed possible with somewhere below 24V across the load). Load is resistive."

On 09/12/2021 18:42, Lasse Langwadt Christensen wrote:
> torsdag den 9. december 2021 kl. 19.39.24 UTC+1 skrev Ed Lee:
>> On Thursday, December 9, 2021 at 10:34:16 AM UTC-8, erichp...@hotmail.com wrote:
>>> On 09/12/2021 11:48, Ian wrote:
>>>> This feels like it should be simple, but my google-fu and electronics-fu are
>>>> failing today.
>>>>
>>>> I need a switch-mode current /sink/, something that looks like this:
>>>>
>>>> Vin o-------
>>>> |
>>>> ---
>>>> | L |
>>>> | o |
>>>> | a |
>>>> | d |
>>>> ---
>>>> |
>>>> v Ireg
>>>> |
>>>> -------
>>>> | |
>>>> | ??? |
>>>> -------
>>>> |
>>>> Gnd o-------
>>>>
>>>> Vin is +24V, Ireg variable up to ~1A (assumed possible with somewhere below
>>>> 24V across the load). Load is resistive.
>>>>
>>>> For reasons, the load must be connected to Vin, not Gnd (there are several
>>>> loads, each requiring separate control, with a common +ve that's non-negotiable)
>>>>
>>>> Any pointers to a suitable topology / part?
>>>>
>>>> (I could probably work out the combination of FET, inductor and diode that works,
>>>> but worry that wiring it "upside-down" may cause stability problems if I just use
>>>> a standard buck controller...)
>>>>
>>> Can't resist a challenge! Here is an untried discrete hysteretic buck
>>> sink. Q1-3-4 are the switcher, Iload is set by Q1vbe and R10. To get
>>> lower load currents Q2 pulls current thru R9 to cheat the feedback. At
>>> some values tempco of Q2 may approximately cancel tempco Q1?
>>>
>>> www.dropbox.com/s/6ekvkgopqtrbrfv/20211209_BuckSink_SED_Ian.pdf?raw=1
>>>
>>> piglet
>> I would use a shunt regulator, but pending Vload answer from the OP.
>
> it is a shunt regulator, it is just a switching rather than linear
>
>

Not shunt - it is a series pass switching, buck topology.

piglet

On 2021-12-09, piglet <erichpwagner@hotmail.com> wrote:
> On 09/12/2021 11:48, Ian wrote:
>> This feels like it should be simple, but my google-fu and electronics-fu are
>> failing today.
>>
>> I need a switch-mode current /sink/, something that looks like this:
>>
>> Vin o-------
>> |
>> ---
>> | L |
>> | o |
>> | a |
>> | d |
>> ---
>> |
>> v Ireg
>> |
>> -------
>> | |
>> | ??? |
>> -------
>> |
>> Gnd o-------
>>
>> Vin is +24V, Ireg variable up to ~1A (assumed possible with somewhere below
>> 24V across the load). Load is resistive.
>>
>> For reasons, the load must be connected to Vin, not Gnd (there are several
>> loads, each requiring separate control, with a common +ve that's non-negotiable)
>>
>> Any pointers to a suitable topology / part?
>>
>> (I could probably work out the combination of FET, inductor and diode that works,
>> but worry that wiring it "upside-down" may cause stability problems if I just use
>> a standard buck controller...)
>>
>
> Can't resist a challenge! Here is an untried discrete hysteretic buck
> sink. Q1-3-4 are the switcher, Iload is set by Q1vbe and R10. To get
> lower load currents Q2 pulls current thru R9 to cheat the feedback. At
> some values tempco of Q2 may approximately cancel tempco Q1?
>
> www.dropbox.com/s/6ekvkgopqtrbrfv/20211209_BuckSink_SED_Ian.pdf?raw=1
>
> piglet

Interesting, thanks for the input, appreciated!

--
Ian

"Tamahome!!!" - "Miaka!!!"

On 2021-12-09, Ed Lee <edward.ming.lee@gmail.com> wrote:
> On Thursday, December 9, 2021 at 10:34:16 AM UTC-8, erichp...@hotmail.com wrote:
>> On 09/12/2021 11:48, Ian wrote:
>> > This feels like it should be simple, but my google-fu and electronics-fu are
>> > failing today.
>> >
>> > I need a switch-mode current /sink/, something that looks like this:
>> >
>> > Vin o-------
>> > |
>> > ---
>> > | L |
>> > | o |
>> > | a |
>> > | d |
>> > ---
>> > |
>> > v Ireg
>> > |
>> > -------
>> > | |
>> > | ??? |
>> > -------
>> > |
>> > Gnd o-------
>> >
>> > Vin is +24V, Ireg variable up to ~1A (assumed possible with somewhere below
>> > 24V across the load). Load is resistive.
>> >
>> > For reasons, the load must be connected to Vin, not Gnd (there are several
>> > loads, each requiring separate control, with a common +ve that's non-negotiable)
>> >
>> > Any pointers to a suitable topology / part?
>> >
>> > (I could probably work out the combination of FET, inductor and diode that works,
>> > but worry that wiring it "upside-down" may cause stability problems if I just use
>> > a standard buck controller...)
>> >
>> Can't resist a challenge! Here is an untried discrete hysteretic buck
>> sink. Q1-3-4 are the switcher, Iload is set by Q1vbe and R10. To get
>> lower load currents Q2 pulls current thru R9 to cheat the feedback. At
>> some values tempco of Q2 may approximately cancel tempco Q1?
>>
>> www.dropbox.com/s/6ekvkgopqtrbrfv/20211209_BuckSink_SED_Ian.pdf?raw=1
>>
>> piglet
>
> I would use a shunt regulator, but pending Vload answer from the OP.

Vload will be between ~10 and ~20V for the intersting range of Ireg.

Vin isn't regulated.

--
Ian

"Tamahome!!!" - "Miaka!!!"

On 2021-12-09, Klaus Kragelund <klauskvik@hotmail.com> wrote:
> 09.12.21 12:48, Ian wrote:
>>This feels like it should be simple, but my google-fu and electronics-fu are
>>failing today.
>>
>>I need a switch-mode current /sink/, something that looks like this:
>>
>> Vin o-------
>> |
>> ---
>> | L |
>> | o |
>> | a |
>> | d |
>> ---
>> |
>> v Ireg
>> |
>> -------
>> | |
>> | ??? |
>> -------
>> |
>> Gnd o-------
>>
>>Vin is +24V, Ireg variable up to ~1A (assumed possible with somewhere below
>>24V across the load). Load is resistive.
>>
>>For reasons, the load must be connected to Vin, not Gnd (there are several
>>loads, each requiring separate control, with a common +ve that's non-negotiable)
>>
>>Any pointers to a suitable topology / part?
>>
>>(I could probably work out the combination of FET, inductor and diode that works,
>>but worry that wiring it "upside-down" may cause stability problems if I just use
>>a standard buck controller...)
>>

> A buck converter just set for lower output voltage than your high side. Synchronous half bridge. Current sense in the return path

If that works I could use standard parts/design. Breadboard time...

--
Ian

"Tamahome!!!" - "Miaka!!!"