What about a sort of "reverse Marx generator"? i.e. a string of capacitors charged in series from the HV and discharged in parallel into the LV load?
I know it works for 72V to 5V but it's a lot of parts compared to a high-wattage resistor.
On Sun, 09 Mar 2014 07:56:04 -0700, Jim Thompson wrote:
The "miser" bringing back memories, I found this...
Bob Hirschfeld responds to some "miser" questions... apparently NOT a creation while he was at National... and mentions my contributions... wa-a-ay down the page (ctrl-F on "Hirschfeld" to find it).
I only recently found out that Bob had died... pursuing inquiries from the old Goodale (MIT) dormitory gang, I tracked down Bob's son Willy, who responded...
====
"Dad passed on August 5th, 2011 of, what appears to be, a massive heart attack due to dehydration on his couch facing the beach in Puerto Penasco (Rocky Point) Sonora, Mexico. Due to the fact that the cost of transporting his body to the US would be unrealistic, Serisa and I chose to have him cremated. His remains are still with me until we are able to determine whether to spread his ashes in Mexico or Big Sur, California. The week before his passing, he was diagnosed with skin cancer and was told he would need to undergo a series of radiation treatments, here, in Tucson. He told his Dr. he wanted to go home first and would come back after he allowed the reality to sink in. It appears that he may have made a deal with God and chose to stay in Mexico for the remainder of his life on this earth. Dad lived his last few years just like he did his previous 69 years.......any way he darn well liked! His character and demeanor never changed and always looked for new challenges and conflicts to address. After deciding to leave the engineering field to become an attorney he fought for fathers custody rights and became a pioneer in that arena until he made one too many judges angry and left his practice. Following his departure from the legal field, he moved to Rocky Point in the late 90's and exclaimed that "this will be the home I will die in!" Dad never took any flack from anyone and had a huge giving heart when he had anything to give. He is sorely missed."
====
The "reminiscences" of a loving son... Bob was actually disbarred for malpractice...
My wife and I interceded in that messy divorce... for which Bob never forgave us... but we acted as trustees, and put both children thru college... Serisa at ASU and Willy at UofA.
...Jim Thompson
--
| James E.Thompson | mens |
| Analog Innovations | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| San Tan Valley, AZ 85142 Skype: Contacts Only | |
| Voice:(480)460-2350 Fax: Available upon request | Brass Rat |
| E-mail Icon at http://www.analog-innovations.com | 1962 |
I love to cook with wine. Sometimes I even put it in the food.
Not directly relevant to your question because it needs external power, but an interesting way of measuring high voltages which might set people thinking:
Terman(1) describes a triode valve being used with a very high negative voltage on the anode, which generates a high potential gradient in the grid-cathode area; this means that grid current will not flow until the grid is several volts positive volts of the cathode. Because the grid positive voltage at the commencement of grid current is proportional to the anode negative voltage, it can be used as a way of measuring the high voltage whilst drawing no current from it.
(1) "Radio Engineering" Terman, F.E. Sect 36 "The inverted tube".
-- ~ Adrian Tuddenham ~ (Remove the ".invalid"s and add ".co.uk" to reply) www.poppyrecords.co.uk
Inverted triodes (and such) have been studied by some; Stephanie Bench (formerly Steve) wrote a few pages about it:
Using it as a measurement tool depends on traditional vacuum tube properties: plate resistance, mu, linearity... Although leakage current may well be better.
Tim
-- Seven Transistor Labs Electrical Eng>
This one is fun:
All of the resistor string current is dumped into the LEDs, instead of being wasted. So at 3KV, some or most of the dissipation can be in the resistors, not the fets. So, SOT23 fets maybe.
I drew two LEDs, but it should be more... ran out of paper.
It probably doesn't need the zeners, but one would have to analyze the potential for gate blowout before deleting them. A lot depends on how big the input RC can be, to limit dv/dt. The top resistors(s) can be the telecom types that open clean on huge over-voltages, like if the fets short.
One issue with zeners is that real low-current zeners above 8 volts or so have very low current below the zener knee, but their specs (and Spice models) assume a lot more leakage. That complicates the gate protection math. Not wasting the resistor string current lets the resistors be relatively small, which helps.
The dropout voltage here should be pretty low, under 5 volts, pretty much the gate threshold voltage of the top mosfet.
Vaux(s) could be tapped anywhere inside the LED string.
Circuits are neat, sort of like crossword puzzles or chess problems.
-- John Larkin Highland Technology Inc www.highlandtechnology.com jlarkin at highlandtechnology dot com Precision electronic instrumentation
I would imagine the geometry of the electrodes was the underlying factor. Not only would that control the ratio of the two voltages when the triode was 'inverted', but it would also determine all the other traditional parameters.
-- ~ Adrian Tuddenham ~ (Remove the ".invalid"s and add ".co.uk" to reply) www.poppyrecords.co.uk
-- Actually, I'm kind of confused... What I think you want is a box which - when connected to a voltage source with an output of between 45 and 3000 volts - draws no more than 2 milliamperes and causes an LED to illuminate. Is that right?
Yes it's a blinky LED, in the previous schematic i mention LED statically powered but yes the spec mention a 1Hz blinking LED, charging a capacitor for 0.5s and discharging it in 0.5s in the LED.
Oh i see variation the gap air between the two armatures of C (sinusoidally) giving dC/dt and mesuring I to find V. smart but complicated imho.
Funny indeed ... good idea !
Habib.
Not at 36kV.
Plus I had power to burn, so I used a motor-driven capacitor >:-}
"Blinking" has possibilities... saves using expensive parts.
The trick is to find something available "off-the-shelf", so you don't have to build too much at device-level. ...Jim Thompson
--
| James E.Thompson | mens |
| Analog Innovations | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| San Tan Valley, AZ 85142 Skype: Contacts Only | |
| Voice:(480)460-2350 Fax: Available upon request | Brass Rat |
| E-mail Icon at http://www.analog-innovations.com | 1962 |
I love to cook with wine. Sometimes I even put it in the food.
Here you go...
No expensive parts, or risky voltage divisions.
However... I see no way to have ON:OFF at 0.5s:0.5s, except with extremely low LED currents, given your power budget, or with more expensive parts... coming next ;-)
In my linked example LED current is ~2mA for ~3ms, which you should easily see. ...Jim Thompson
--
| James E.Thompson | mens |
| Analog Innovations | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| San Tan Valley, AZ 85142 Skype: Contacts Only | |
| Voice:(480)460-2350 Fax: Available upon request | Brass Rat |
| E-mail Icon at http://www.analog-innovations.com | 1962 |
I love to cook with wine. Sometimes I even put it in the food.
What is it that determines the current level flowing in the gate-source zeners?
Stacking the MOSFET's (DMOS) is a non-trivial issue... Infineon's models show substantial gate leakage.
[snip]Yes. ...Jim Thompson
--
| James E.Thompson | mens |
| Analog Innovations | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| San Tan Valley, AZ 85142 Skype: Contacts Only | |
| Voice:(480)460-2350 Fax: Available upon request | Brass Rat |
| E-mail Icon at http://www.analog-innovations.com | 1962 |
I love to cook with wine. Sometimes I even put it in the food.
Their I-V curve of course. One could use 15 or 20 volt zeners to protect the gates, and they might not leak much at 4 or 5 volts. But it would take some research to make sure the leakage isn't a problem.
A transistor b-e junction (probably a PNP) would be interesting, potentially a small-geometry, properly passivated, very low leakage zener. That would be interesting to characterize.
It probably doesn't need the zeners.
That's another reason to keep the resistor stack values as low as possible. Ideally, 100% of the LED current is through the resistors at
3KV.Big fets can have substantial D-S leakage, a good reason to use small fets.
Since it's just lighting up some LEDs, it doesn't have to be perfect, just pretty good.
Another interesting brute-force design would be a string of LND150s running at Idss (around 1.5 mA) with a zener or two across each one.
(current sources in series... where have I seen that before?)
What's the highest voltage zener around? Anything over 200 volts?
-- John Larkin Highland Technology, Inc jlarkin att highlandtechnology dott com http://www.highlandtechnology.com
Zener "leakage" isn't the issue... it's the bias current running in them. In your circuit, what sets that value?
The trick, which you haven't demonstrated, is to get the right current in the LED's at 50V.
The Infineon _models_ say they also have 10's of uA's GATE leakage, which is a killer for doing divider games.
Have no idea, I never use them on-chip. ...Jim Thompson
--
| James E.Thompson | mens |
| Analog Innovations | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| San Tan Valley, AZ 85142 Skype: Contacts Only | |
| Voice:(480)460-2350 Fax: Available upon request | Brass Rat |
| E-mail Icon at http://www.analog-innovations.com | 1962 |
I love to cook with wine. Sometimes I even put it in the food.
Looks like there is higher voltage in TVS diodes, 440V say.
-- John Devereux
In HV_LED_2, only the lower zener should have current. It's obviously set by the resistor string and the voltages. As noted, it needs a zener with a low sub-knee current. I've done some micropower stuff, and they exist.
Maybe at high drain voltages. At low drain voltages, I've set gate voltages, opened the gate lead, and had them not change much in a week.
I'm playing with circuits. I don't have to demonstrate anything.
This one is more interesting.
Or, maybe this is actually the best:
Why don't you design something, or contribute something?
-- John Larkin Highland Technology, Inc jlarkin att highlandtechnology dott com http://www.highlandtechnology.com
Way big improvement over that stacked MOSFET kluge... add some protection against reverse polarity and critical component failure and you're home.
Seems like a voltage dependent blink frequency would be valuable here.
Maybe it's a characteristic of HV DMOS, but Infineon's spice model shows 20uA _exiting_ the gate at VDS = 500V (a 600V enhancement device, BSP125)
But it could be they use PhD's straight-out-of-school to do their modeling... that disease seems to be going around right now >:-}
Which is probably why TI has taken to encrypting their models... which only adds to my fun... my version of the LM94022 temperature sensor (good for all Spice variants) is on its way... probably by tomorrow ;-)
Does that meet the OP's power budget?
Why don't you learn to be civil instead of being the perpetual asshole?
As for what I design, you seem to have a reading comprehension problem. I mentioned two very low power, frequency _stable_ oscillators this very day...
1kHz at 6nA1MHz at 1.3uA
Both output logic swings at 1.8V
0.18um XFAB-XP018 ProcessAnd I posted, also today, a blinking LED solution to the OP's problem, with components names and values, so that even you could build it...
No expensive parts, or risky voltage divisions.
Demonstrated to work.
Can you _prove_ your hairballs actually work... or do you just lick yours ?:-) ...Jim Thompson
--
| James E.Thompson | mens |
| Analog Innovations | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| San Tan Valley, AZ 85142 Skype: Contacts Only | |
| Voice:(480)460-2350 Fax: Available upon request | Brass Rat |
| E-mail Icon at http://www.analog-innovations.com | 1962 |
I love to cook with wine. Sometimes I even put it in the food.
~3mA total drain...
But those indicated gate currents make me suspect the Infineon Spice models. ...Jim Thompson
--
| James E.Thompson | mens |
| Analog Innovations | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| San Tan Valley, AZ 85142 Skype: Contacts Only | |
| Voice:(480)460-2350 Fax: Available upon request | Brass Rat |
| E-mail Icon at http://www.analog-innovations.com | 1962 |
I love to cook with wine. Sometimes I even put it in the food.
R5 dissipates 6 watts at 3 KV. Not 2 watts.
DZ1 needs to have sub-uA current below its 27 volt knee.
The minimum LED drive is 2 mA for 3 ms at about 0.6 Hz. That won't be visible in normal room light. It won't be visible at 3 KV, either.
A resistive-limited thing has problems. If the resistor doesn't fry at
3 KV, there's very little power available at 50 volts. All 50 volts across 1.5M is just 33 uA, not much to work with.And the OP wanted an auxiliary DC output to run some low-power stuff. This circuit doesn't do that. The series stack does.
-- John Larkin Highland Technology, Inc jlarkin att highlandtechnology dott com http://www.highlandtechnology.com
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.