On 30/12/2019 16:52, Klaus Kragelund wrote:> AFAIR the step down cap supplies is not that bad > > 7 stages or more sounds like a nightmare, but since individual voltage is low, the process can be low voltage and thus cheap > > That said, I am yet to see one that can compete with inductor step down > > Cheers > > Klaus >The advantage is at low power. I built a 6 stage capacitive divider 48V to 5V at max 10 microamperes input (on-hook POTS 48V). Microamp inductor based buck converters do not appear to be easy. piglet
high voltage charge pump
Started by ●December 30, 2019
Reply by ●December 31, 20192019-12-31
Reply by ●December 31, 20192019-12-31
On 30/12/2019 13:52, Hul Tytus wrote:> Anyone familiar with a high voltage charge pump designed as a step down > power supply? The voltage is between 175 & 200. Numerous high voltage > versions exist using inductors, (Onsemi has one that is "self supplied" > which is attractive) but not a one that works with just a capacitor. Are > there any out there I've missed? > > Hul >There was the Harris HV2405 in the late 1980s but that was AC input - it charged the capacitor on the leading portion of the ac line waveform and disconnected when it exceeded the output voltage. Won't work from a DC supply. Did you specify if the HV input is AC or DC and how much current is demanded at 12V? piglet
Reply by ●December 31, 20192019-12-31
Klaus, if you come across one spec'd to the higher voltage, let me know. Hul Klaus Kragelund <klauskvik@hotmail.com> wrote:> On Monday, December 30, 2019 at 10:00:46 PM UTC+1, Hul Tytus wrote: > > You're going the wrong way, Klaus. Take a look at the message to win. > >> I don't think so> See example:> https://www.analog.com/media/en/technical-documentation/data-sheets/7820fc.pdf> Page 25, 48V to 12V switched capacitor high efficiency solution.> A 200V DC type might exist
Reply by ●December 31, 20192019-12-31
On Tue, 31 Dec 2019 11:04:03 +0000, Piglet <erichpwagner@hotmail.com> wrote:>On 30/12/2019 13:52, Hul Tytus wrote: >> Anyone familiar with a high voltage charge pump designed as a step down >> power supply? The voltage is between 175 & 200. Numerous high voltage >> versions exist using inductors, (Onsemi has one that is "self supplied" >> which is attractive) but not a one that works with just a capacitor. Are >> there any out there I've missed? >> >> Hul >> > >There was the Harris HV2405 in the late 1980s but that was AC input - it >charged the capacitor on the leading portion of the ac line waveform and >disconnected when it exceeded the output voltage. Won't work from a DC >supply. > >Did you specify if the HV input is AC or DC and how much current is >demanded at 12V? > >pigletRight, the problem is underspecified. -- John Larkin Highland Technology, Inc lunatic fringe electronics
Reply by ●December 31, 20192019-12-31
On Tue, 31 Dec 2019 10:32:52 +0000, Piglet <erichpwagner@hotmail.com> wrote:>On 30/12/2019 16:52, Klaus Kragelund wrote: >> AFAIR the step down cap supplies is not that bad >> >> 7 stages or more sounds like a nightmare, but since individual voltage is low, the process can be low voltage and thus cheap >> >> That said, I am yet to see one that can compete with inductor step down >> >> Cheers >> >> Klaus >> > >The advantage is at low power. I built a 6 stage capacitive divider 48V >to 5V at max 10 microamperes input (on-hook POTS 48V). Microamp inductor >based buck converters do not appear to be easy. > >pigletEasy! Just run them discontinuous at low pulse rate. Many switcher chips will automatically go into burst mode at low load, too. -- John Larkin Highland Technology, Inc lunatic fringe electronics
Reply by ●December 31, 20192019-12-31
Input varies between 150 & 200 volts dc. Current is at the 12 volt output is 60 ma. There will probably be a 6 volt negative output also but that will cause little added circuitry. Simply stated, 2 npn's, 1 pnp, 6 resisters & 2 1nf capacitors comprise the device. That''s doable but cumbersome in pcb real estate and design time. An integrated package would be a neater fix. Hul Piglet <erichpwagner@hotmail.com> wrote:> On 30/12/2019 13:52, Hul Tytus wrote: > > Anyone familiar with a high voltage charge pump designed as a step down > > power supply? The voltage is between 175 & 200. Numerous high voltage > > versions exist using inductors, (Onsemi has one that is "self supplied" > > which is attractive) but not a one that works with just a capacitor. Are > > there any out there I've missed? > > > > Hul > >> There was the Harris HV2405 in the late 1980s but that was AC input - it > charged the capacitor on the leading portion of the ac line waveform and > disconnected when it exceeded the output voltage. Won't work from a DC > supply.> Did you specify if the HV input is AC or DC and how much current is > demanded at 12V?> piglet
Reply by ●December 31, 20192019-12-31
On Tue, 31 Dec 2019 16:28:03 +0000 (UTC), Hul Tytus <ht@panix.com> wrote:>Input varies between 150 & 200 volts dc. Current is at the 12 volt output is >60 ma. There will probably be a 6 volt negative output also but that will >cause little added circuitry. Simply stated, 2 npn's, 1 pnp, 6 resisters & >2 1nf capacitors comprise the device. That''s doable but cumbersome in pcb >real estate and design time. An integrated package would be a neater fix. > >HulSchematic? A charge pump may be no more efficient than resistors. People usually bottom-post on usenet. Mixed posting gets confusing. -- John Larkin Highland Technology, Inc lunatic fringe electronics
Reply by ●December 31, 20192019-12-31
On 12/31/19 12:12 PM, jlarkin@highlandsniptechnology.com wrote:> On Tue, 31 Dec 2019 16:28:03 +0000 (UTC), Hul Tytus <ht@panix.com> > wrote: > >> Input varies between 150 & 200 volts dc. Current is at the 12 volt output is >> 60 ma. There will probably be a 6 volt negative output also but that will >> cause little added circuitry. Simply stated, 2 npn's, 1 pnp, 6 resisters & >> 2 1nf capacitors comprise the device. That''s doable but cumbersome in pcb >> real estate and design time. An integrated package would be a neater fix. >> >> Hul > > Schematic? A charge pump may be no more efficient than resistors. > > People usually bottom-post on usenet. Mixed posting gets confusing. > >in my experience the only types of charge pumps which aren't dreadful are CW multipliers to make high voltages at very small currents, and the types that come in IC packages that can use techniques like fractional conversion to improve their efficiency/regulation and can use internal switches that have had their parameters tuned to suit the chip application. Every other time I thought I'd use a roll-your-own charge pump for something I've ended up going inductor-based because naive charge pumps made with logic or discretes suck harder IRL than the by-the-book calculations would lead you to believe
Reply by ●December 31, 20192019-12-31
On 12/31/19 12:40 PM, bitrex wrote:> On 12/31/19 12:12 PM, jlarkin@highlandsniptechnology.com wrote: >> On Tue, 31 Dec 2019 16:28:03 +0000 (UTC), Hul Tytus <ht@panix.com> >> wrote: >> >>> Input varies between 150 & 200 volts dc. Current is at the 12 volt >>> output is >>> 60 ma. There will probably be a 6 volt negative output also but that >>> will >>> cause little added circuitry. Simply stated, 2 npn's, 1 pnp, 6 >>> resisters & >>> 2 1nf capacitors comprise the device. That''s doable but cumbersome >>> in pcb >>> real estate and design time. An integrated package would be a neater >>> fix. >>> >>> Hul >> >> Schematic? A charge pump may be no more efficient than resistors. >> >> People usually bottom-post on usenet. Mixed posting gets confusing. >> >> > > in my experience the only types of charge pumps which aren't dreadful > are CW multipliers to make high voltages at very small currents, and the > types that come in IC packages that can use techniques like fractional > conversion to improve their efficiency/regulation and can use internal > switches that have had their parameters tuned to suit the chip application. > > Every other time I thought I'd use a roll-your-own charge pump for > something I've ended up going inductor-based because naive charge pumps > made with logic or discretes suck harder IRL than the by-the-book > calculations would lead you to believe >Using one capacitor and one diode to boost e.g. 3.3V logic output square wave to ~4.something to reliably drive a white or blue LED at a few mA is OK too but it's not really high-investment charge pump design here
Reply by ●December 31, 20192019-12-31
On Tue, 31 Dec 2019 12:44:13 -0500, bitrex <user@example.net> wrote:>On 12/31/19 12:40 PM, bitrex wrote: >> On 12/31/19 12:12 PM, jlarkin@highlandsniptechnology.com wrote: >>> On Tue, 31 Dec 2019 16:28:03 +0000 (UTC), Hul Tytus <ht@panix.com> >>> wrote: >>> >>>> Input varies between 150 & 200 volts dc. Current is at the 12 volt >>>> output is >>>> 60 ma. There will probably be a 6 volt negative output also but that >>>> will >>>> cause little added circuitry. Simply stated, 2 npn's, 1 pnp, 6 >>>> resisters & >>>> 2 1nf capacitors comprise the device. That''s doable but cumbersome >>>> in pcb >>>> real estate and design time. An integrated package would be a neater >>>> fix. >>>> >>>> Hul >>> >>> Schematic? A charge pump may be no more efficient than resistors. >>> >>> People usually bottom-post on usenet. Mixed posting gets confusing. >>> >>> >> >> in my experience the only types of charge pumps which aren't dreadful >> are CW multipliers to make high voltages at very small currents, and the >> types that come in IC packages that can use techniques like fractional >> conversion to improve their efficiency/regulation and can use internal >> switches that have had their parameters tuned to suit the chip application. >> >> Every other time I thought I'd use a roll-your-own charge pump for >> something I've ended up going inductor-based because naive charge pumps >> made with logic or discretes suck harder IRL than the by-the-book >> calculations would lead you to believe >> > >Using one capacitor and one diode to boost e.g. 3.3V logic output square >wave to ~4.something to reliably drive a white or blue LED at a few mA >is OK too but it's not really high-investment charge pump design hereSomething like this works... https://www.dropbox.com/s/azsc1vigh8vepoq/Led_Pump.JPG?raw=1 but the boost ratio is small and efficiency doesn't matter much. -- John Larkin Highland Technology, Inc lunatic fringe electronics
