Components you wish existed

[mikeselectricstuff]

So what parts have you wanted to use, but nobody makes them.... 
Here's a start..

8 Mosfets in a single SO /TSSOP18 package, common source, similar spec to SOT23-6 duals
A 74HC138 or HC595 with 1A P-channel source drivers on the output 
Small FPGAs in 48 and 64 pin QFP or QFN
SMD DIP switches that operate from the underside through a slot in the PCB
A mid-range microcontroller with a buttload of RAM (>64K) but without the cost of a ton of expensive flash & high-end peripherals, in a 48 or lower pin count QFP/QFN (Microchip PIC32MZ looks very interesting, when it appears)

And of course a component with exactly the same characteristics as a scope probe , for those times when probing fixes  a problem

[KJDS]

I'd quite like a component with the same characteristics as a finger, as that cures so many RF woes.

[Rerouter]

just visit the local morgue, though the emi compliance guys might get a bit uppity about 5 human finger tips glued in place on your product,

[ecat]

just visit the local morgue, though the emi compliance guys might get a bit uppity about 5 human finger tips glued in place on your product,

But the Australian government would let it pass until you poked one in someone's ear
https://www.eevblog.com/forum/chat/warning-over-usb-chargers-after-woman-dies-from-apparent-electrocution/?topicseen

 
( to soon?)

All I need right now is a time machine, or something which makes money or jobs. These sound like hardware problems, come on you guys 

[amyk]

8 Mosfets in a single SO /TSSOP18 package, common source, similar spec to SOT23-6 duals

Supertex used to make them in DIP18,  the AN01xx.

[Wh1sper]

And of course a component with exactly the same characteristics as a scope probe , for those times when probing fixes  a problem

True! I think it might be possible to build this as community projekt relative easy

[HighVoltage]

Power transistor's or IGBT's that have the Emitter on the mounting / heatsink and not the Collector for better and direct cooling without a 1000 V insulation needed.

May be it is not possible to build?

[Psi]

A basic MCU similar to a AVR/PIC but which ran at an internal clock speed of 1+ghz.
16/32bit core would be better but i'd be totally happy with an 8bit core.
Price would be <$10 in qty of 10

The raw power of 1+ghz would be quite useful in cases where you would otherwise have to go FPGA or full on ARM CPU with associated parts (DDR/flash etc).


Oh, and it would be nice if you could buy realistically priced 0805 resistor arrays.
I would love to have a PCB with footprints that can take 10 individual 0805 resistors/caps OR a single 10 block array (like 10x 0805 glued together with typical per resistor spacing).

Being able to put a single 10 array resistor/cap block on the PCB would speed up hand assembling so much yet still allow for picknplace of individual parts.

0805 arrays seem very hard to find and cost $$$
0604 are a bit easier to get but still seem far too expensive.

[AlanR]

For me it would be a single chip solution for controlling high frequency design 3-phase AC induction motors that has the following features:

1) 450 volt MOSFETS that handle 10 amps continuous for each half bridge (three half-bridges for 3-phase motor) that would allow high frequency design 3-phase motors up to 1200 watts mechanical power.

2) Flash memory to program in motor characteristic parameters.

3) Rectifier bridge built in.

4) Current fold back at 50 C on chip and current fold back option for motor if motor will have RTD sensor embedded in the windings.

5) Input voltage range of 85 to 250 volts AC single phase.

6) Motor current limiting

7) Input for tachometer option for close looped feedback to give AC motor fine torque control from zero RPM to base speed RPM.

8  ) Power factor circuitry built in.

9) Input for potentiometer or hall effect sensor for speed control (programmable defined  by flash memory)

10) Filter capacitor pins to hook up filter capacitor.

11) Finally, a proper MCU that has enough compute horsepower to do the motor vector calculations for the AC motor and the options)

[daqq]

A MEMS XY mirror that does not cost an obscene amount of money.
An EINK display with a temperature range of -40 to 80.
A supercapacitor/battery with zero leakage.
A two+ output synchronous DC DC converter with a digital control interface (you set up the output voltage(s - one could be fixed to power the initial logic) via, say, I2C, within some range), that accepts 50V DC. Needn't be very high current, 500mA on each channel.
SMD geiger tube.

[peter.mitchell]

8 Mosfets in a single SO /TSSOP18 package, common source, similar spec to SOT23-6 duals
A 74HC138 or HC595 with 1A P-channel source drivers on the output 
SMD DIP switches that operate from the underside through a slot in the PCB

I'd really like the FETs (or even BJTs!) in a single package, and not even 8, maybe just 4 if you want to get the current on the pins to reasonable levels... Cmon guys!
High current shift registers would be nice too, yes.
Reverse mount dip switches would be really nice too, you can put config switches and stuff on a topside load board but have them accessible from the bottom.

[Psi]

oh, i thought of another one.

Single chip DVB-T transmitter

You feed HDMI into the chip and out of it you get DVB-T RF ready for power amp stage.

eg..
HDMI input - > low latency h264/mpeg encoder --> dvb-t wrapper --> low power RF output.

(I would say built in 400mW RF output stage, but i think that would be pushing things a bit far    )

Such a chip would REVOLUTIONIZE the fpv rc community.
We could finally fly FPV in HD with a video system built from the ground up for signal penetration through objects and to be immune to multipath interference

[mikeselectricstuff]

A basic MCU similar to a AVR/PIC but which ran at an internal clock speed of 1+ghz.
16/32bit core would be better but i'd be totally happy with an 8bit core.
Price would be <$10 in qty of 10

The raw power of 1+ghz would be quite useful in cases where you would otherwise have to go FPGA or full on ARM CPU with associated parts (DDR/flash etc).

Have a look at XMOS

Oh, and it would be nice if you could buy realistically priced 0805 resistor arrays.
I would love to have a PCB with footprints that can take 10 individual 0805 resistors/caps OR a single 10 block array (like 10x 0805 glued together with typical per resistor spacing).

Being able to put a single 10 array resistor/cap block on the PCB would speed up hand assembling so much yet still allow for picknplace of individual parts.

0805 arrays seem very hard to find and cost $$$
0604 are a bit easier to get but still seem far too expensive.

Why would you want them as big as 0805?
Plenty of cheap smaller arrays.

[mikeselectricstuff]

oh, i thought of another one.

Single chip DVB-T transmitter

You feed HDMI into the chip and out of it you get DVB-T RF ready for power amp stage.

eg..
HDMI input - > low latency h264/mpeg encoder --> dvb-t wrapper --> low power RF output.

(I would say built in 400mW RF output stage, but i think that would be pushing things a bit far    )

Such a chip would REVOLUTIONIZE the fpv rc community.
We could finally fly FPV in HD with a video system built from the ground up for signal penetration through objects and to be immune to multipath interference

Don't see why it would need to be single chip - I suspect the main issue is power due to the processing needed.
Broadcast formats are targetted at low cost decode and good compression without worrying about latency.
For FPV you could probably design a different encoding scheme that would be a much better fit - e.g. low latency, optimised for low encode effort and a graceful degrade in resolution as error level increases

[Codemonkey]

I'd quite like a component with the same characteristics as a finger, as that cures so many RF woes.

They used a Peperami on one of our products at the test house 

[Macbeth]

A flux capacitor of course! and a 1.21GW PSU as well.

[jlmoon]

A P-Channel mosfet with N-Channel electron flow characteristics (really low Rds on value) so I can switch on the + rail without having to cheat the system.


or.... The perfect transistor!

[Rigby]

old RGB video amplification and processing ICs that are only used for CRT arcade monitor repair. 

It is insane what people charge for NOS these days.  I will put myself through any education necessary, paying any amount required, and learn to design and manufacture my own circuit before I pay $80 for a 6-pin video processing IC from some sweaty fatass who is sitting on the world's stock of that particular IC laughing whenever he sells one.

[nuno]

"Component tape". You know, like glueing tape. You tweezer out an SMD component from the tape, you press it against the pads at its final PCB location and you're done!

[Rigby]

"Component tape". You know, like glueing tape. You tweezer out an SMD component from the tape, you press it against the pads at its final PCB location and you're done!

that's called Z-axis tape.  It exists, but isn't meant for production, just prototyping.  https://www.sparkfun.com/products/12042

[jahonen]

Real-world equivalent of SPICE E-primitive, ideal 4-terminal differential amplifier with infinite bandwidth, zero delay, user configurable gain without any external components and perfect isolation 

Regards,
Janne

[Precipice]

I'd like a LED that runs at a useful brightness on <1.8V, without needing to faff around with a driver transistor. SOT23 with an external power feed and maybe a constant current stage to save a resistor would do. A tiny voltage doubler (only needs to deliver 5mA or so, and could clock pretty fast) would maybe do the trick. 
Ideally, though, IR to get the bandgap down far enough to run native from 1V8 (or 1V5), with some frequency doubling magic on top to shift it into visible.

[BravoV]

My ultimate component, Alladin lamp with the genie that grants any 3 wishes, and my current wishes are :

1. Make the wishes unlimited 
2. Undo feature for any wishes that I've made except for no.1 
3. -spare wish- not going to use it yet 


C'mon guys, only components that is possible to make or manufacture with the current technologies or materials.

[nuno]

"Component tape". You know, like glueing tape. You tweezer out an SMD component from the tape, you press it against the pads at its final PCB location and you're done!

that's called Z-axis tape.  It exists, but isn't meant for production, just prototyping.  https://www.sparkfun.com/products/12042

Ah!! Almost there, now they just need to sell parts with the tape already on the "pins"

[krivx]

Real-world equivalent of SPICE E-primitive, ideal 4-terminal differential amplifier with infinite bandwidth, zero delay, user configurable gain without any external components and perfect isolation 

Regards,
Janne

Good luck finding design work after that part comes out 

[Psi]

Have a look at XMOS

Will have a look at that, thanks

Why would you want them as big as 0805?
Plenty of cheap smaller arrays.

Just because its quick/easy to hand assemble 0805.
The 0604 arrays are still quite expensive and usually use a custom footprint which doesnt match the individual components in a row.

Don't see why it would need to be single chip - I suspect the main issue is power due to the processing needed.
Broadcast formats are targetted at low cost decode and good compression without worrying about latency.
For FPV you could probably design a different encoding scheme that would be a much better fit - e.g. low latency, optimised for low encode effort and a graceful degrade in resolution as error level increases

Yeah, i wasn't being too serious.
In practice yeah, it probably needs separate  HDMI decode chip, video encoder, transport encoder, and RF amp chip

Yeah, a encoding scheme with feedback would be best. So as you fly out of range it automatically dials back the bitrate to keep an image.

However the data encoding format should really be OFDM to avoid multipath, like dvb.

[SeanB]

How about a power switch, capable of 60V, 30A and with a power dissipation of 3W max, logic level input with a current monitor output and able to operate to say 10kHz. High side switch, and in a simple pentawatt or similar TO220 like package. Oh, and priced at under $5 each in single quantities.

[mikeselectricstuff]

The 0604 arrays are still quite expensive

couple of cents @ 100x
http://www.digikey.co.uk/product-detail/en/741X083330JP/741X083330JPCT-ND/1124601

[free_electron]

So what parts have you wanted to use, but nobody makes them.... 
Here's a start..

8 Mosfets in a single SO /TSSOP18 package, common source, similar spec to SOT23-6 duals  ->

ULN2803 style ? i believe toshiba has that ...

A 74HC138 or HC595 with 1A P-channel source drivers on the output 

shift register with high side outputs : TI TISP series.

SMD DIP switches that operate from the underside through a slot in the PCB

cut an opening for the body in the board and sink the part in the board so that the gull wing pins are soldered top side. the smd body is about the same thickness as standard pcb.

A mid-range microcontroller with a buttload of RAM (>64K) but without the cost of a ton of expensive flash & high-end peripherals, in a 48 or lower pin count QFP/QFN (Microchip PIC32MZ looks very interesting, when it appears)

STM32F411CE : 128K ram in 49 pin QFN ( 48 pins + central slug ground )

And of course a component with exactly the same characteristics as a scope probe , for those times when probing fixes  a problem

i like a component that replace a wet finger .. to indicate overheating regulators...

[free_electron]

Real-world equivalent of SPICE E-primitive, ideal 4-terminal differential amplifier with infinite bandwidth, zero delay, user configurable gain without any external components and perfect isolation 

Regards,
Janne

just use a spice engine that behaves exactly like real components 

[free_electron]

telepatically configurable chips. so i can just think what i want the damn thing to do and it reconfigures itself.

[David Hess]

Tunnel diodes, step recovery diodes, and both undedicated and dedicated monolithic transistor arrays.  I miss the LM389.

[tautech]

SMD DIP switches that operate from the underside through a slot in the PCB

cut an opening for the body in the board and sink the part in the board so that the gull wing pins are soldered top side. the smd body is about the same thickness as standard pcb.

Nice. 

[mikeselectricstuff]

So what parts have you wanted to use, but nobody makes them.... 
Here's a start..

8 Mosfets in a single SO /TSSOP18 package, common source, similar spec to SOT23-6 duals  ->

ULN2803 style ? i believe toshiba has that ...

Can't see anything on Digikey, by anyone

A 74HC138 or HC595 with 1A P-channel source drivers on the output 

shift register with high side outputs : TI TISP series.

TISP? I've seen the TPIC series but they're all open drain

A mid-range microcontroller with a buttload of RAM (>64K) but without the cost of a ton of expensive flash & high-end peripherals, in a 48 or lower pin count QFP/QFN (Microchip PIC32MZ looks very interesting, when it appears)

STM32F411CE : 128K ram in 49 pin QFN ( 48 pins + central slug ground )

No stock showing anywhere at findchips.com

[c4757p]

both undedicated and dedicated monolithic transistor arrays.  I miss the LM389.

At least you can still get matched pairs cheaply. DMMT3904W, DMMT3906W, PMP4201/PMP5201, etc.

LM389 looks awesome though.

[free_electron]

Can't see anything on Digikey, by anyone

that would indeed be a no. but i am almost certain toshiba has em

TISP? I've seen the TPIC series but they're all open drain

even top side drivers will be open drain.. that is the funny thing with mosfets. a Pmos drives out of its drain. an nmos sinks into its drain...
it is always gate-source voltage that controls the mos , irrespective if it is p or n.

explain a bit more why you need open source pmos ? that doesnt make sense to me..

A mid-range microcontroller with a buttload of RAM (>64K) but without the cost of a ton of expensive flash & high-end peripherals, in a 48 or lower pin count QFP/QFN (Microchip PIC32MZ looks very interesting, when it appears)

STM32F411CE : 128K ram in 49 pin QFN ( 48 pins + central slug ground )

No stock showing anywhere at findchips.com
[/quote]
possible. the damn thing is barely a few months old... it is in mass production though. i got boards in the lab with these on em.

[SeanB]

I still have some MPQ6502 quad core drivers around. Thought they were 6502 processors, but in a 14 pin DIP I was wrong.

[VK3DRB]

The holy grail: A safe low cost button cell battery that held 1000Ah of capacity.

Battery technology is the biggest issue. Electric cars would be standard if battery technology could improve by a couple of orders of magnitude.

[Rigby]

It's getting there.  Just yesterday I heard of a new battery technology that sounded very promising.

[tautech]

It's getting there.  Just yesterday I heard of a new battery technology that sounded very promising.

Would you care to share that?

[SirNick]

A few things spring immediately to mind:

1) A good 3.3 + 5v linear reg in single package that can do 250mA or more on each channel.  Infineon made the TLC4476, which was exactly what I wanted.  I placed an order for a few just as stock dried up and the manufacturer discontinued the part.

2) Anything from E-Switch that isn't immediately discontinued by the time I order a few as samples.

3) A decent 4- / 8-gang pot that doesn't cost >$10, if it can be had in single quantities at all.  (Think: Linkwitz-Riley filters with adjustable corner frequency, like a variable crossover.)

4) To ante-up on Mike's request, I would love to see a modest microcontroller (Cortex-M would be fine) with 1MB of RAM.  C'mmmoooonnnnnn.....  I know you wanna.

[NiHaoMike]

A chip that takes DVI/HDMI or DisplayPort in at one end and outputs a HD stream (H.264 or similar) over Gigabit Ethernet (maybe 100Mbps actually used) at the other end with very low latency, plus a matching one to do the reverse. Price ideally under $50 for the set in small quantities. Bonus if there's an Android app for accessing the stream, albeit at reduced quality. More bonus if it also streams USB.

[mikeselectricstuff]

explain a bit more why you need open source pmos ? that doesnt make sense to me..

Ok strictly speaking, high-side outputs. 

[magetoo]

Small FPGAs in 48 and 64 pin QFP or QFN

MachXO2 comes in small QFP and QFN packages IIRC, but only the smaller (in logic) parts of course.

I'd very much like to see some hobbyist-friendly FPGAs with open source tools.  Older product lines perhaps could have documentation released, but I guess that's something that might have to be planned from the start.

telepatically configurable chips. so i can just think what i want the damn thing to do and it reconfigures itself.

One-time programmable generic chips would be nice.  If there was a way to connect up uncommitted transistors on a die in some way, 3D printing style perhaps, things could get pretty interesting.

Probably not very realistic either..

[Stonent]

More cool stuff in 40 pin pdip. I'm talking Arm, cpld/fpga, you name it.

Transistor arrays that aren't significantly more expensive than the individual parts.

More stuff that can drive loads directly with out additional components.

[mikeselectricstuff]

Small FPGAs in 48 and 64 pin QFP or QFN

MachXO2 comes in small QFP and QFN packages IIRC, but only the smaller (in logic) parts of course.

There's a 32QFN (with 4 bank power pins FFS!) , but the next up from that is 100QFP.

I'd very much like to see some hobbyist-friendly FPGAs with open source tools. 

Why does OS matter for tools like this?
The only barrier to hobbyist FPGAs is whne you get to the bigger ones that aren't supported by the free tools, and these devices are typically way above the price and complexity of even ambitious hobbyist projects.

One-time programmable generic chips would be nice.  If there was a way to connect up uncommitted transistors on a die in some way, 3D printing style perhaps, things could get pretty interesting.

Silego is a bit of a step in this direction, but like FPGAs, software tools will be a limitation as soon as things get more complex.

[electrophiliate]

I thought of 100-in-one ICs, but others have already suggested better versions (cannot beat a telepathically programmable IC!).

It would be nice to be able to have a generic resistor which can be easily set before placement. Then we would only need to keep a bunch of resistors for each power rating (and sometimes voltage rating) rather than for every value as well as power rating.

I guess something similar could be said for zeners, TVS diodes, and varistors. Would be pointless for capacitors of larger value, there would be wasted space if for example sacrificing a potential 1000uF electrolytic capacitor for 100uF in the same size package.

Might as well go the whole hog and demand a tissue sized box that will print out just about any component on demand.

Here is a simpler but unimportant request: a Sziklai transistor arrangement in one cheap jellybean package.

[nowlan]

i like a component that replace a wet finger .. to indicate overheating regulators...

There are thermochromatic paints and ink (hypercolor). Although painting IC would obscure the part number etc. I wonder if they could include the ink in the silk screen. This would indicate (radiant) hot areas of the board. dunno if its conductive.

[Dajgoro]

HV opamp in a single package, where HV is 1~2KV.

[Rigby]

It's getting there.  Just yesterday I heard of a new battery technology that sounded very promising.

Would you care to share that?

http://www.gizmag.com/dual-carbon-fast-charging-battery/32121/

Not sure if that's the one I read yesterday... I do so much reading that I simply lose track of when and where I read things.  This article is about a month old.  I know I read something a couple days back, but I don't remember if this is it or not...

[edit] also:
http://www.gizmag.com/structural-supercapacitors-batteries-obsolete/32246/
http://www.gizmag.com/silly-putty-battery/32089/
http://www.gizmag.com/flexible-high-performance-battery/31830/

this site is full of "could" and "might" but lots of people are clearly working on battery tech right now.

[richfiles]

I wish memristors were on the market.

Small chips containing a couple memristors, CPLD/FPGA style chips with internal arrays of memristors that can be interconnected by PC software. I'd like to see real experimentation in that new technology... Not just some closed off research in an HP lab somewhere.

[Rigby]

I wish memristors were on the market.

Small chips containing a couple memristors, CPLD/FPGA style chips with internal arrays of memristors that can be interconnected by PC software. I'd like to see real experimentation in that new technology... Not just some closed off research in an HP lab somewhere.

Couple years.  HP is gearing up for manufacture, and will start selling memristor-based memory in 2015, I believe.  There is a long video of a keynote speech they did somewhere in the last month.  Can't remember the conference for the life of me.  Lemme dig up a link...

[edit] here it is: http://youtu.be/Gxn5ru7klUQ

[SeanB]

Here is a simpler but unimportant request: a Sziklai transistor arrangement in one cheap jellybean package.

IGBT transistors come pretty close to implementing that, albiet with a PNP and a high voltage N channel mosfet in the same package.

[daqq]

HV opamp in a single package, where HV is 1~2KV.

http://www.apexanalog.com/apex-products/pa89/

Beware - it's expensive.

[Dajgoro]

HV opamp in a single package, where HV is 1~2KV.

http://www.apexanalog.com/apex-products/pa89/

Beware - it's expensive.

I should had added that it should not be made out of Unobtainium.

[T3sl4co1l]

HV opamp in a single package, where HV is 1~2KV.

http://www.apexanalog.com/apex-products/pa89/

Beware - it's expensive.

I should had added that it should not be made out of Unobtainium.

It is necessarily unobtainium, because it's a very unusual application and will always be high cost.  And high margin.  Physics, microactuators, that sort of thing.  Those few remaining applications, you roll your own and deal with it.

Tim

[T3sl4co1l]

Power transistor's or IGBT's that have the Emitter on the mounting / heatsink and not the Collector for better and direct cooling without a 1000 V insulation needed.

May be it is not possible to build?

RF transistors are often made this way; the die is mounted on a ceramic insulator.  Some parts (diodes most often, I think) are available in isolated-tab TO-220 or 247 styles, but SOT227 modules are much more common.

Expect to pay double for a module of given ratings.

Real-world equivalent of SPICE E-primitive, ideal 4-terminal differential amplifier with infinite bandwidth, zero delay, user configurable gain without any external components and perfect isolation 

Regards,
Janne

That would violate causality.

Tim

[richfiles]

I wish memristors were on the market.

Small chips containing a couple memristors, CPLD/FPGA style chips with internal arrays of memristors that can be interconnected by PC software. I'd like to see real experimentation in that new technology... Not just some closed off research in an HP lab somewhere.

Couple years.  HP is gearing up for manufacture, and will start selling memristor-based memory in 2015, I believe.  There is a long video of a keynote speech they did somewhere in the last month.  Can't remember the conference for the life of me.  Lemme dig up a link...

[edit] here it is: http://youtu.be/Gxn5ru7klUQ

That's part of the issue. Memory is a high value market, and thus they are pushing memristor based memory. I KNOW that innovation will take time and research, but right now, HP is effectively keeping a monopoly on ALL that research and development. Discrete memristors and reconfigurable memristor grid arrays are non existent, and I've heard nothing of intent from HP on such devices.

As far as memristor technology is concerned, I am EXCITED at the future prospects. I myself have played with Copper-Coper Sulfide-Aluminum Oxide-Aluminum junction memristors, and have even made a few on the lab bench. I obviously can't mimic the density of memristors on silicon, but it's intriguing to posit what this relatively new tech may do for us! I've been trying to find a means to manufacture point-contact style memristor components in a two PCB sandwich, supporting multiple devices brought out to a header or solder terminals.

The one thing that bothers me with memristor technology is the one difference that it STILL bears from all the other three fundamental components:

You can buy a discrete resistor.
You can buy a discrete capacitor.
You can buy an discrete inductor.

You can NOT buy a discrete memristor.

This is a critical failure for innovation. I get it... Modern technology relies HEAVILY on modern integration. You don't build a modern device with many discrete components. You avoid them, instead favoring integrating as many functions on as few integrated components as possible. This keeps cost and complexity down. You wouldn't build a TV or radio with discrete components the way you would have back in the 1980s and 70s and earlier. It's inefficient.

Why then can you still buy resistors and capacitors and inductors, and many other small discrete and small scale integrated components? Simple. Hobbyists, students, educators, and prototypers still use them for learning, educating, creating one off devices, and for fun!

The people that grow up and become skilled at the discrete level will carry that knowledge of HOW those parts work up the chain of industry, and eventually, the best of those will work on the integrated solutions. Education has to start with a foundation. NO ONE jumps into integrated circuit design with NO fundamentals in electronic technology!

The fact that there has been seemingly NO EFFORTS to release a small scale component that makes memristors available to the masses is a horrifyingly short sighted blunder! Tell me? How many textbooks teach memristor design principals? Not many, if any, yet... Even if they discuss them, how are students to TRULY learn with no examples to experiment with? HP seems content to develop the memristor technology internally, but in doing so, they are shooting the longterm advancement of memristors in the foot. Yes, they can make some highly profitable memory, but what other innovations will be missed with them holding this so close to their chest?

Did you know memristors can implement boolean logic elements? Did you know that memristors are useful in implementing some amazing neuron models? Who even knows what we will discover... but we NEED the option to experiment!

I would go so far as to argue, that if they wish to continue claiming the memristor to be a 4th fundamental circuit element, that they should be obligated to provide fundamental examples of the tech to educators and the public, to allow for the masses to actually learn the new tech, and to creat innovations that HP does not see themselves!

We need a discrete memristor, something in the same form factor as a small SMD resistor, or...
We need a chip style component that simply provides a few memristors to the leads.
We need a small crossbar latch in a chip that has a couple pins for programming.
And we need an FPGA like device that allows for massive memristor design.

So yeah, while I am CERTAINLY excited about the prospect of a new form of storage memory being made available, and driving capacity up and prices down... I'm more excited about what people COULD do with them that doesn't fit the mainstream view.

I think it's practically a sin for HP to be proclaiming this as the 4th fundamental circuit element, while neglecting to establish a stable foundation of usable discrete and reconfigurable integrated components for educators, experimenters, prototypers, and hobbyists, even as they push a highly integrated application already, all for the glory of the almighty dollar.

[magetoo]

MachXO2 comes in small QFP and QFN packages IIRC, but only the smaller (in logic) parts of course.

There's a 32QFN (with 4 bank power pins FFS!) , but the next up from that is 100QFP.

Huh, I could have sworn they had them in small QFPs, but I must have been thinking of something else.

I'd very much like to see some hobbyist-friendly FPGAs with open source tools. 

Why does OS matter for tools like this?
The only barrier to hobbyist FPGAs is whne you get to the bigger ones that aren't supported by the free tools, and these devices are typically way above the price and complexity of even ambitious hobbyist projects.

Well, it matters when you don't run Windows (or Linux).  It would just be convenient for me personally not having to mess around with a platform I don't otherwise use, and hopefully the community could at least manage some minor tweaks to the tools too.

[mikeselectricstuff]

The holy grail: A safe low cost button cell battery that held 1000Ah of capacity.

Battery technology is the biggest issue. Electric cars would be standard if battery technology could improve by a couple of orders of magnitude.

What would be nice are lithium coin cells in 2032 etc. sizes  with the same capacity as now, but with really low internal resistance so it can supply a couple of hundred mA. Peak current delivery is often more of a limiting factor than capacity in long-life wireless devices. Imagine if you could send an SMS from a 2032. 

[mikeselectricstuff]

Here's another : MicroSD cards in light colours so you can write on them

[HighVoltage]

Power transistor's or IGBT's that have the Emitter on the mounting / heatsink and not the Collector for better and direct cooling without a 1000 V insulation needed.
May be it is not possible to build?

RF transistors are often made this way; the die is mounted on a ceramic insulator.  Some parts (diodes most often, I think) are available in isolated-tab TO-220 or 247 styles, but SOT227 modules are much more common.

Expect to pay double for a module of given ratings.

Regards,
Janne

I am not working with RF, so I was not aware of this. Thanks for the explanation, I will look in to that.

[tautech]

Here's another : MicroSD cards in light colours so you can write on them

Man, are you on a roll... keep them coming. 

[Rasz]

Japanese hacker slave maid with big tits, to solder my crap ... oh wait, you meant electronic parts

A basic MCU similar to a AVR/PIC but which ran at an internal clock speed of 1+ghz.
16/32bit core would be better but i'd be totally happy with an 8bit core.
Price would be <$10 in qty of 10

The raw power of 1+ghz would be quite useful in cases where you would otherwise have to go FPGA or full on ARM CPU with associated parts (DDR/flash etc).

http://www.digikey.com/product-detail/en/XS1-L8A-128-QF124-I10/XS1-L8A-128-QF124-I10-ND/3906735
Its 2x 32bit, and package, QFN124, is a bit unwieldy, but otherwise what you wanted
Then you have 4 core ARMv7 at >1GHz at $5
http://www.cnx-software.com/2014/06/23/allwinner-a33-tablet-price/

HDMI input - > low latency h264/mpeg encoder --> dvb-t wrapper --> low power RF output.

half of that:
http://www.hides.com.tw/product_opencaster_eng.html
you dump mpeg_TS stream over USB, it takes care of making DVB-T out of that

A chip that takes DVI/HDMI or DisplayPort in at one end and outputs a HD stream (H.264 or similar) over Gigabit Ethernet (maybe 100Mbps actually used) at the other end with very low latency, plus a matching one to do the reverse. Price ideally under $50 for the set in small quantities. Bonus if there's an Android app for accessing the stream, albeit at reduced quality. More bonus if it also streams USB.

will mjpeg do? not a chip, but still ~$50
http://danman.eu/blog/reverse-engineering-lenkeng-hdmi-over-ip-extender/


I like this game

[Rigby]

I wish memristors were on the market.

Small chips containing a couple memristors, CPLD/FPGA style chips with internal arrays of memristors that can be interconnected by PC software. I'd like to see real experimentation in that new technology... Not just some closed off research in an HP lab somewhere.

Couple years.  HP is gearing up for manufacture, and will start selling memristor-based memory in 2015, I believe.  There is a long video of a keynote speech they did somewhere in the last month.  Can't remember the conference for the life of me.  Lemme dig up a link...

[edit] here it is: http://youtu.be/Gxn5ru7klUQ

That's part of the issue. Memory is a high value market, and thus they are pushing memristor based memory. I KNOW that innovation will take time and research, but right now, HP is effectively keeping a monopoly on ALL that research and development. Discrete memristors and reconfigurable memristor grid arrays are non existent, and I've heard nothing of intent from HP on such devices.

As far as memristor technology is concerned, I am EXCITED at the future prospects. I myself have played with Copper-Coper Sulfide-Aluminum Oxide-Aluminum junction memristors, and have even made a few on the lab bench. I obviously can't mimic the density of memristors on silicon, but it's intriguing to posit what this relatively new tech may do for us! I've been trying to find a means to manufacture point-contact style memristor components in a two PCB sandwich, supporting multiple devices brought out to a header or solder terminals.

The one thing that bothers me with memristor technology is the one difference that it STILL bears from all the other three fundamental components:

You can buy a discrete resistor.
You can buy a discrete capacitor.
You can buy an discrete inductor.

You can NOT buy a discrete memristor.

This is a critical failure for innovation. I get it... Modern technology relies HEAVILY on modern integration. You don't build a modern device with many discrete components. You avoid them, instead favoring integrating as many functions on as few integrated components as possible. This keeps cost and complexity down. You wouldn't build a TV or radio with discrete components the way you would have back in the 1980s and 70s and earlier. It's inefficient.

Why then can you still buy resistors and capacitors and inductors, and many other small discrete and small scale integrated components? Simple. Hobbyists, students, educators, and prototypers still use them for learning, educating, creating one off devices, and for fun!

The people that grow up and become skilled at the discrete level will carry that knowledge of HOW those parts work up the chain of industry, and eventually, the best of those will work on the integrated solutions. Education has to start with a foundation. NO ONE jumps into integrated circuit design with NO fundamentals in electronic technology!

The fact that there has been seemingly NO EFFORTS to release a small scale component that makes memristors available to the masses is a horrifyingly short sighted blunder! Tell me? How many textbooks teach memristor design principals? Not many, if any, yet... Even if they discuss them, how are students to TRULY learn with no examples to experiment with? HP seems content to develop the memristor technology internally, but in doing so, they are shooting the longterm advancement of memristors in the foot. Yes, they can make some highly profitable memory, but what other innovations will be missed with them holding this so close to their chest?

Did you know memristors can implement boolean logic elements? Did you know that memristors are useful in implementing some amazing neuron models? Who even knows what we will discover... but we NEED the option to experiment!

I would go so far as to argue, that if they wish to continue claiming the memristor to be a 4th fundamental circuit element, that they should be obligated to provide fundamental examples of the tech to educators and the public, to allow for the masses to actually learn the new tech, and to creat innovations that HP does not see themselves!

We need a discrete memristor, something in the same form factor as a small SMD resistor, or...
We need a chip style component that simply provides a few memristors to the leads.
We need a small crossbar latch in a chip that has a couple pins for programming.
And we need an FPGA like device that allows for massive memristor design.

So yeah, while I am CERTAINLY excited about the prospect of a new form of storage memory being made available, and driving capacity up and prices down... I'm more excited about what people COULD do with them that doesn't fit the mainstream view.

I think it's practically a sin for HP to be proclaiming this as the 4th fundamental circuit element, while neglecting to establish a stable foundation of usable discrete and reconfigurable integrated components for educators, experimenters, prototypers, and hobbyists, even as they push a highly integrated application already, all for the glory of the almighty dollar.

Well, memristors were JUST discovered in 2008... I agree with the sentiment of your post, and I think it is a bit early to call HP out when they still aren't selling memristors themselves.

[David Hess]

More cool stuff in 40 pin pdip. I'm talking Arm, cpld/fpga, you name it.

I really liked PLCC (plastic leaded chip carrier) packages because you could both solder them or use through-hole sockets and probing was easy.

[amyk]

The holy grail: A safe low cost button cell battery that held 1000Ah of capacity.

Battery technology is the biggest issue. Electric cars would be standard if battery technology could improve by a couple of orders of magnitude.

What would be nice are lithium coin cells in 2032 etc. sizes  with the same capacity as now, but with really low internal resistance so it can supply a couple of hundred mA. Peak current delivery is often more of a limiting factor than capacity in long-life wireless devices. Imagine if you could send an SMS from a 2032.

By 2032, this might be possible.

[peter.mitchell]

Here's another : MicroSD cards in light colours so you can write on them

In the same vein, microsd cards with a hole for a lanyard or so you can get some leverage on them to remove them (some go in real real tight, i have a bit of success with very very carefully shaped bits of kapton tape either side)

[David Hess]

both undedicated and dedicated monolithic transistor arrays.  I miss the LM389.

At least you can still get matched pairs cheaply. DMMT3904W, DMMT3906W, PMP4201/PMP5201, etc.

LM389 looks awesome though.

The LM389 was a way to get 3 monolithic matched transistors inexpensively although the amplifier came in handy as well.  You could use it for instance to make temperature compensated log and antilog converters.

Matched pairs are better than nothing but not as good as monolithic parts and the other problem is lack of diversity in type; 2N3904 and 2N3906 clones are not suitable for every application.

[sync]

The LM389 was a way to get 3 monolithic matched transistors inexpensively although the amplifier came in handy as well.  You could use it for instance to make temperature compensated log and antilog converters.

The transistors of a LM3046 array seems to be better matched and it's still available. Plus you can use the additional transistors for a heater circuit which stabilize the die temperature. I done this for antilog converters with good results.

[abaxas]

There should be a device that automatically nukes a major US city every-time they try and standardize something using imperial measurements.

[richfiles]

I wish memristors were on the market.

Small chips containing a couple memristors, CPLD/FPGA style chips with internal arrays of memristors that can be interconnected by PC software. I'd like to see real experimentation in that new technology... Not just some closed off research in an HP lab somewhere.

...

That's part of the issue. Memory is a high value market, and thus they are pushing memristor based memory. I KNOW that innovation will take time and research, but right now, HP is effectively keeping a monopoly on ALL that research and development. Discrete memristors and reconfigurable memristor grid arrays are non existent, and I've heard nothing of intent from HP on such devices.

...

The one thing that bothers me with memristor technology is the one difference that it STILL bears from all the other three fundamental components:

You can buy a discrete resistor.
You can buy a discrete capacitor.
You can buy an discrete inductor.

You can NOT buy a discrete memristor.

...

I would go so far as to argue, that if they wish to continue claiming the memristor to be a 4th fundamental circuit element, that they should be obligated to provide fundamental examples of the tech to educators and the public, to allow for the masses to actually learn the new tech, and to creat innovations that HP does not see themselves!

We need a discrete memristor, something in the same form factor as a small SMD resistor, or...
We need a chip style component that simply provides a few memristors to the leads.
We need a small crossbar latch in a chip that has a couple pins for programming.
And we need an FPGA like device that allows for massive memristor design.

...

I think it's practically a sin for HP to be proclaiming this as the 4th fundamental circuit element, while neglecting to establish a stable foundation of usable discrete and reconfigurable integrated components for educators, experimenters, prototypers, and hobbyists, even as they push a highly integrated application already, all for the glory of the almighty dollar.

Well, memristors were JUST discovered in 2008... I agree with the sentiment of your post, and I think it is a bit early to call HP out when they still aren't selling memristors themselves.

The point, is what HP is doing is pushing a highly integrated specialized product (memristor memory) as their very first released product, while ignoring the public's needs in regards to the fundamental steps they took to get there.

let me provide a hypothetical equivalency. It would be the equivalent of, if how back when the transistor was invented decades ago, if companies had just started releasing transistor radios, with germanium dies bonded to the PC boards, along with the point contacts, with the PC boards being fully potted, and the companies not selling individual transistors! You can't access the transistor, you can't salvage the transistor without destroying it, you can't use the transistor, you can't buy the transistor. All you'd have been able to get was an end application!

The idea seems absurd, but that is what HP proposes to do with memristor technology. They propose a highly specialized, highly integrated memory product, while providing no means to access the underlying fundamental technology... All while they tout memristors as a 4th fundamental circuit element. HP went through Steps A-E to develop that memory chip... But they only intend to release step F to the market. They are not releasing any parts related to the steps on the way to getting there! Fail

Imagine a world where all instances of one of the following components: resistors, capacitors, or inductors were just outright MISSING! And the only way to get them, was inside specialized chips? How the hell do you innovate? Create? Invent? Learn! The thought is absurd, yet here we have HP doing just that to fundamental electronic element number 4!

[Rigby]

Again, I see your point, but they're so new not even HP has any in-hand, yet, and you're complaining about a strategy that they've not announced, unless I've missed something.

Edit: besides, they invested all the time and money into the discovery and realization of the technology, shouldn't they get first crack at the market?  I think they should.

Everyone is free to evaluate memristors in software, just like everyone is free to do whatever the hell they want in software.  If you, or anyone, has the desire to fiddle with memristors, add them to a logic simulation program, or add a rudimentary component to a spice engine.  Better than nothing.

[NiHaoMike]

will mjpeg do? not a chip, but still ~$50
http://danman.eu/blog/reverse-engineering-lenkeng-hdmi-over-ip-extender/

I read about that but I have no idea of its latency. I intend to use it for gaming.

[daqq]

Oh, I forgot:

Pretty much everything Hamamatsu makes at 1/10 or less of the price 

[AndyC_772]

I'd like to see a range of P-channel power MOSFETs that include two devices back-to-back with the drains connected together.

The combination would still be a 3 terminal device, but it wouldn't conduct in either direction when switched off.

Bonus marks if it also includes two Schottky diodes and a high value resistor, so the gate is always pulled up to whichever source terminal is more positive.

[David Hess]

let me provide a hypothetical equivalency. It would be the equivalent of, if how back when the transistor was invented decades ago, if companies had just started releasing transistor radios, with germanium dies bonded to the PC boards, along with the point contacts, with the PC boards being fully potted, and the companies not selling individual transistors! You can't access the transistor, you can't salvage the transistor without destroying it, you can't use the transistor, you can't buy the transistor. All you'd have been able to get was an end application!

In a way this has always been the case with transistors though.  There no equivalents of many of the transistor types used in integrated circuits like super-beta transistors.  Actually I can think of one discrete super-beta transistor but its poor availability makes it a myth.

[David Hess]

The LM389 was a way to get 3 monolithic matched transistors inexpensively although the amplifier came in handy as well.  You could use it for instance to make temperature compensated log and antilog converters.

The transistors of a LM3046 array seems to be better matched and it's still available. Plus you can use the additional transistors for a heater circuit which stabilize the die temperature. I done this for antilog converters with good results.

I have as well with the CA3046.  There are a few others like that but availability tends to be poor and most of the parts which are not monolithic are not even matched.

I keep hoping for Linear Technology to release something since they seem to be of that mind but they never have.

[richfiles]

Again, I see your point, but they're so new not even HP has any in-hand, yet, and you're complaining about a strategy that they've not announced, unless I've missed something.

The point you are missing, is HP is talking about releasing complex integrated memristor devices in the form or memory chips in the not too distant future. They are not discussing discrete memristors, fixed configuration multi memristor integrated devices, or reconfigurable memristor array integrated devices that would allow true innovation, learning, prototyping, and experimentation by those who sorta, kinda don't have access to a chip fab.

The point is that they are seem to be focusing on one and only one particular end product, and have made little efforts to drive open evolution. Imagine a world where all the resistors, or all the capacitors, or all the inductors were taken away, and you could only get such fundamental parts as integrated into premanufactured integrated components. The concept is absurd, but that is what HP is doing!

By the time HP releases a memristor based memory, they will have proven memristors can work as a commercially viable component...

But where will the discrete memristor components be? Oh wait... They won't exist, because HP decided to keep all that internal, and aim instead for the highly profitable memory market because it provides for a potentially quick profit. They have chosen NOT to open the evolution of memristor technology, for further advancement of technology as a whole. They could release discrete devices and reconfigurable general purpose devices in parallel with their memory development. Nope! While they sit and make a memory chip, the world has no access to the "fourth fundamental circuit element", as they claim memristors to be.

It's the theoretical equivalent of a single corporate entity being the only maker of resistors, and choosing not to release them to the world, and instead only build them into finished integrated devices. It just seems so absurd!

part of me honestly wishes that the long history of memristance observations could serve as prior art to nullify some of HP's patents. Sir Humphry Davy described the first known observations of a memristive effect in 1808, and the Titanium Oxide materials were described with the effects that HP "claims" to have "discovered" back in 1968. A LOT of what HP is doing seems to be trying to corner a lucrative memory market with claims of "invention", when they weren't even first MODERN company to research resistive memory devices. Samsung, I think was looking into it back in 2003, 5 years before HP! it seems to me that HP just snagged lucrative patents and hype over their "discovery" in order to corner a new memory market. I fear they consider non memory applications entirely secondary, and are in no rush to let anyone else play with the tech if they can avoid it.

So much for fundamental fourth circuit element...

[richfiles]

let me provide a hypothetical equivalency. It would be the equivalent of, if how back when the transistor was invented decades ago, if companies had just started releasing transistor radios, with germanium dies bonded to the PC boards, along with the point contacts, with the PC boards being fully potted, and the companies not selling individual transistors! You can't access the transistor, you can't salvage the transistor without destroying it, you can't use the transistor, you can't buy the transistor. All you'd have been able to get was an end application!

In a way this has always been the case with transistors though.  There no equivalents of many of the transistor types used in integrated circuits like super-beta transistors.  Actually I can think of one discrete super-beta transistor but its poor availability makes it a myth.

"In a way this has always been the case with transistors though."
"this has always been the case"
"always been"
 

Well, my point was entirely and utterly missed... 

You COULD buy discrete transistors. You STILL can! Integrated circuits didn't even exist back then. Saying that because transistors on today's integrated circuits have different characteristics than the discrete parts they evolved from is only a characteristic of the scale and physics. That you even go so far as to say such a discrete part exists (even if utterly rare) disproves your own statement. Limited availability of one particular variant is irrelevant to the fact that it actually exists!

My POINT is arguing for a proper evolution of the technology from the bottom up, so learning, experimentation, innovation, and prototyping are possible. Your argument completely MISSED the point that transistors DID exist as discrete devices first. That integrated devices came later is great, but it also did not eliminate the discrete transistor.

I'm sorry, but I can buy a damn transistor. That it's not commonly available in every conceivable variant under the sun does not eliminate the fact that I can buy a damn transistor!

With memristors... there are NO VARIANTS available...

That would be why I'm gonna go build my own memristors, with blackjack and hookers. In fact, forget the blackjack and hookers! 

[Galaxyrise]

LT1021 hermetic with temperature sensing pin(s).  I believe hermetic parts are scheduled to return in the ceramic surface-mount package, and I suspect there's temp sensing already there, but severed after factory calibration... (There's 5 NC pins on it, after all!)  So close!

[Rigby]

Again, I see your point, but they're so new not even HP has any in-hand, yet, and you're complaining about a strategy that they've not announced, unless I've missed something.

The point you are missing, is HP is talking about releasing complex integrated memristor devices in the form or memory chips in the not too distant future. They are not discussing discrete memristors, fixed configuration multi memristor integrated devices, or reconfigurable memristor array integrated devices that would allow true innovation, learning, prototyping, and experimentation by those who sorta, kinda don't have access to a chip fab.

The point is that they are seem to be focusing on one and only one particular end product, and have made little efforts to drive open evolution. Imagine a world where all the resistors, or all the capacitors, or all the inductors were taken away, and you could only get such fundamental parts as integrated into premanufactured integrated components. The concept is absurd, but that is what HP is doing!

By the time HP releases a memristor based memory, they will have proven memristors can work as a commercially viable component...

But where will the discrete memristor components be? Oh wait... They won't exist, because HP decided to keep all that internal, and aim instead for the highly profitable memory market because it provides for a potentially quick profit. They have chosen NOT to open the evolution of memristor technology, for further advancement of technology as a whole. They could release discrete devices and reconfigurable general purpose devices in parallel with their memory development. Nope! While they sit and make a memory chip, the world has no access to the "fourth fundamental circuit element", as they claim memristors to be.

It's the theoretical equivalent of a single corporate entity being the only maker of resistors, and choosing not to release them to the world, and instead only build them into finished integrated devices. It just seems so absurd!

part of me honestly wishes that the long history of memristance observations could serve as prior art to nullify some of HP's patents. Sir Humphry Davy described the first known observations of a memristive effect in 1808, and the Titanium Oxide materials were described with the effects that HP "claims" to have "discovered" back in 1968. A LOT of what HP is doing seems to be trying to corner a lucrative memory market with claims of "invention", when they weren't even first MODERN company to research resistive memory devices. Samsung, I think was looking into it back in 2003, 5 years before HP! it seems to me that HP just snagged lucrative patents and hype over their "discovery" in order to corner a new memory market. I fear they consider non memory applications entirely secondary, and are in no rush to let anyone else play with the tech if they can avoid it.

So much for fundamental fourth circuit element...

YOU ARE NOT OWED A FUNDAMENTAL 4TH CIRCUIT ELEMENT!  A 4TH FUNDAMENTAL CIRCUIT ELEMENT IS NOT A RIGHT.

The link I shared is a video from an HP conference for people interested in server technology.  So, server topics are what is going to be covered and little else.  That conference is NOT the final word on what is going on within HP.

Holy shit, you're foaming at the mouth because of ideas that you fabricated inside your own mind based on a 30 minute video about SERVERS... 

WHERE DOES HP SAY THEY WILL NEVER EVER EVER NEVER RELEASE MEMRISTOR COMPONENTS!? 

[Zad]

Can I have a modern equivalent of the S1D13806 video generator with onboard memory? 8MB should do for up to 1920x1080, a bit more if you need alpha. It needn't do analogue, but various digital options would be nice. Oh yes, it must also be a pinned package. It needn't have ultra advanced draw hardware though.

[Tinkerer]

After reading the arguements on the memristors, I think I can say we are getting a little ahead here. On one side, there is the point that keeping things packaged together with no individual components is bad for innovation. On the other side, is the arguement that the tech is new enough that there hasnt been a chance to actually do anything.

Memristors finally came about in 2008, only 6 years ago. During this time, they have obviously been refining the approach to actually create usuable devices. It took about 7 years from demonstration to commercial device for the transistor. So the time frame isnt all that different. Obviously they want to create something that they can make money from and thats what they are doing. If another few years pass and we have yet to see any aside from HP products that include memristors, than I think the point is very valid that they are holding back innovation...who am I kidding, I would be willing to bet that the Chinese will cook up a poorly working copy within a year of them going on the market.

In all seriousness, I think we need to wait a couple more years before passing judgement on this. Could they release an individual component right now if they wanted? Probably. However, they would have to start tooling everything up for this and spend money in doing so. Instead they are tooling up to begin selling finished products, which even though contain the same part, would still need to be made differently. They stand to make more money from from a high tech assembly than they do from an individual component. This is where they are pouring their resources into regarding this. Once they start pumping out actual products, then the time will come to sit and talk on when everyone else gets to experiment with factory made parts. They are having their chance to smack the 'just discovered money tree' right now on a tech that hasnt even been in existance for a decade. If they decide to block all others access after they have had their first whack at this, then we can berate them.

Finally, why not just send them a letter asking if they plan on offering any components at some point? Wouldnt hurt, at worst you likely get a cookie cutter response.

Edit: Oops! typo!

[richfiles]

YOU ARE NOT OWED A FUNDAMENTAL 4TH CIRCUIT ELEMENT!  A 4TH FUNDAMENTAL CIRCUIT ELEMENT IS NOT A RIGHT.

The link I shared is a video from an HP conference for people interested in server technology.  So, server topics are what is going to be covered and little else.  That conference is NOT the final word on what is going on within HP.

Holy shit, you're foaming at the mouth because of ideas that you fabricated inside your own mind based on a 30 minute video about SERVERS... 

WHERE DOES HP SAY THEY WILL NEVER EVER EVER NEVER RELEASE MEMRISTOR COMPONENTS!?

Sorry, my "ideas" were not based on your video... I honestly never even noticed the link. They come from years of hearing and seeing HP do things that are, quite frankly, annoying. I'll admit, I've watched countless videos and read countless documents... I have not seen this server video, so it was never applicable to the nature of any of my replies. Only my prior knowledge.

I know for a fact that multiple hobbyists, myself as well as other EEVblog members amongst that number, have successfully created and experimented with our own memristors on the lab bench, while HP has "claimed" their technology is not "suited" to discrete components. As far as I am concerned, HP is spewing bullshit. Nyle Steiner was probably one of the first hobbyists to achieve the effect. I've been working on sandwiching multiple point contact style memristors, based on his work, between two PC boards myself. One memristor is never enough! 

I've been listening to HP's rhetoric since they first announced their "discovery". I'll be the first to admit that yes, I am pretty aggravated that they are telling people who HAVE CREATED working memristors that they can't make discrete components. It comes down to one simple fact. HP is in the computing business, and wants to hit up the lucrative memory market. They have no interest in innovation, unless it is fully internal, and will appease the almighty shareholder.

Here is a rather old (2010) video that is rather comprehensive, for reference:
11:10 Acknowledging the existence of pinched hysteresis observance prior to 1971, without giving up the claim of "inventing" it.
17:15 Claims memristance is unobservable at larger scales (true technically, but people ARE observing it with hand made components).
29:07 They admit to creating in house FPGA style memristor devices... Yet they gotta release that memory first, you know!
46:56 "I mean, right now we're just sorta looking at this as a non volatile memory system"... That says it right there. They speak of disruptive ideas, and what other things people can do... but they don't REALLY want people outside HP making those developments. Not really.

What do I take from this video? HP has not really shared with the world with this internally produced FPGA... They keep talking bout memory, which is a good thing, but is not all memristors are good for. It's just what they are most profitable for. Even discussing the FPGA, they speak of it's use in digital switching, in emulating an FPGA's digital logic functionality, but simply with better density. They fail to consider the potential for an analog FPGA style device, or at least have not made clear that the functionality is there.

That is why I am not holding my breath for HP to really follow through on any of those promises, any time soon... Except for memory...

LOL, the delays are probably cause the NSA is buying it all up for their fancy-schmancy data logging center!

[T3sl4co1l]

In a way this has always been the case with transistors though.  There no equivalents of many of the transistor types used in integrated circuits like super-beta transistors.  Actually I can think of one discrete super-beta transistor but its poor availability makes it a myth.

Matching is pretty much out, so there's that.  Also the incredibly low propagation delay between nearby components.

But most of the rest is available -- I think 2N5088 is equivalent to LM741 era ICs (~40V process), though with area (and thus current capacity) around what would be used for the output devices (not of op-amps, but more comparable to like TL431 or 78L05 pass devices, maybe LM386 too).  RF MOSFETs dating from the same period (I don't know numbers offhand) are close to e.g. CD4007, though without the crappy performance (slow -- unusually high internal resistances and capacitances).  Higher performance devices (low voltage CMOS, or BiCMOS, on up to modern SiGe processes) are reflected by various RF devices.  Usually at vastly greater current capacity (>10s mA, instead of ~uA for the smallest structures), just because of physical size and ability to drive transmission lines.

But superbeta I don't think is very common, so you've got that point.  Last of that sort I bought was 2SD1273 (60V 3A -- unusual!), which had been in stock (NRND) at Mouser.  They've long since disappeared, of course.  The unfortunate truth of it, I guess, is, when do you actually need that?  They're much slower, so you can't really use them in switching applications.  For almost everything else, you don't need the hFE, a TIP31 or whatever will work.  I've mostly used them in low bias followers for simple power supplies.

Tim

[T3sl4co1l]

You COULD buy discrete transistors. You STILL can! Integrated circuits didn't even exist back then. Saying that because transistors on today's integrated circuits have different characteristics than the discrete parts they evolved from is only a characteristic of the scale and physics. That you even go so far as to say such a discrete part exists (even if utterly rare) disproves your own statement. Limited availability of one particular variant is irrelevant to the fact that it actually exists!

My POINT is arguing for a proper evolution of the technology from the bottom up, so learning, experimentation, innovation, and prototyping are possible. Your argument completely MISSED the point that transistors DID exist as discrete devices first. That integrated devices came later is great, but it also did not eliminate the discrete transistor.

I'm sorry, but I can buy a damn transistor. That it's not commonly available in every conceivable variant under the sun does not eliminate the fact that I can buy a damn transistor!

With memristors... there are NO VARIANTS available...

That would be why I'm gonna go build my own memristors, with blackjack and hookers. In fact, forget the blackjack and hookers! 

So what?  It's a novelty component.  If you want to play with one, slap together an integrator with an analog multiplier and make one.  It's not a particularly useful component (despite academics cheering about theory and symmetry), and physical (rather than synthesized) examples, such as used in the memory, probably have awful properties (consistency, accuracy, drift, tempco), barely good enough for storing data.

That said, I should think something like an MOV or thermistor could be constructed to exhibit the same behavior in bulk, so you could have a memristor with real watts capacity and hundreds to k's of resistance range.

Tim

[David Hess]

Here is a simpler but unimportant request: a Sziklai transistor arrangement in one cheap jellybean package.

IGBT transistors come pretty close to implementing that, albiet with a PNP and a high voltage N channel mosfet in the same package.

I would not mind having complementary IGBTs for linear applications but P-channel IGBTs are another part whose availability makes it a myth.

[richfiles]

You COULD buy discrete transistors. You STILL can! Integrated circuits didn't even exist back then. Saying that because transistors on today's integrated circuits have different characteristics than the discrete parts they evolved from is only a characteristic of the scale and physics. That you even go so far as to say such a discrete part exists (even if utterly rare) disproves your own statement. Limited availability of one particular variant is irrelevant to the fact that it actually exists!

My POINT is arguing for a proper evolution of the technology from the bottom up, so learning, experimentation, innovation, and prototyping are possible. Your argument completely MISSED the point that transistors DID exist as discrete devices first. That integrated devices came later is great, but it also did not eliminate the discrete transistor.

I'm sorry, but I can buy a damn transistor. That it's not commonly available in every conceivable variant under the sun does not eliminate the fact that I can buy a damn transistor!

With memristors... there are NO VARIANTS available...

That would be why I'm gonna go build my own memristors, with blackjack and hookers. In fact, forget the blackjack and hookers! 

So what?  It's a novelty component.  If you want to play with one, slap together an integrator with an analog multiplier and make one.  It's not a particularly useful component (despite academics cheering about theory and symmetry), and physical (rather than synthesized) examples, such as used in the memory, probably have awful properties (consistency, accuracy, drift, tempco), barely good enough for storing data.

That said, I should think something like an MOV or thermistor could be constructed to exhibit the same behavior in bulk, so you could have a memristor with real watts capacity and hundreds to k's of resistance range.

Tim

Transistors were once an over expensive, terribly fragile novelty component, in an era where valves were the defacto standard technology. People understood they could mark an improvement and a fundamental shrinking of technology, with reduced power needs than valves... but yes, even in the very early days, transistor were once novelty components. They VERY quickly found their niche with the transistor radio, and then made their way into computing applications, other audio amplification devices, operational amplifiers, etc. Sound familiar? ICs were once quite frivolously novel. Chips were out for a WHILE before they were universally adopted. Between the US military buying every last one they could find, and their expense, people saw no need to integrate what could be made cheaper with a few transistors. Eventually, the market caught up, and that changed.

I hope that changes with memristors sooner than later...

My interest has always been in the neural / synaptic potential of memristors. It's a field with little research, and it pertains to my own interests. The fact that people dismiss memristors so easily is exactly WHY it's so necessary to be able to experiment with the real deal. The experimenters, the prototypers, the curious! Those are the ones who will come up with the idea, or stumble on some unexpected configuration that the next person misses! not everyone can see passed the everyday.

It just feels like wasted potential, mixed with a bit of hypocrisy. HP says they want to usher a new revolution with a "fourth fundamental circuit element",  but they expect that revolution to be on their shoulders, and that everyone must follow their path. No thanks. How bout they give me a component that I can be creative with! Revolutionary memory is great! FPGAs are great as well, but give us not just an FPGA that is enhanced through the integration of memristors to make it more dense, but give us an... I dunno, let's call it an FCMA - a Field Configurable Memristor Array, that doesn't just perform digital logic using memristors to achieve better density, but actually uses something like CMOS analog switches (as in similar to a CD4051, for example) to create a truly user accessible memristor array that an be reconfigured into custom analog circuits. Even having a small 16 pin chip containing 8 individual memristors would be great! Maybe we could have a 20 pin IC that has 16 memristor interconnect lines (connecting to 8 memristors), but features CMOS analog switches internally to switch banks, and has 2 wires to either select from a multitude of banks (each bank being 8 memristors) via serial commands, or has 4 banks selected by a 2 bit code. A 24 pin IC could easily contain 64 banks of 8 memristors (6 bit binary bank selection). 32 or 512 individual memristors might not sound like much at all, but consider that in the hands of a resourceful experimenter, they would be IDEAL learning tools, and I know I could personally use such a device myself. Another option would be small crossbar latch style chips. An 18 pin chip could easily contain 16 rows and a pair of columns. a 20 pin chip could do a 16x4 crossbar latch, as a 24 pin chip would enable a 16x8 configuration. An additional option would be to have disable lines that isolate a column's memristors from adjacent columns (using analog CMOS switches), to allow hard setting the memristor states via isolation, or alternately, enabling or disabling the memristor connection at that point ( a little CMOS logic plus an additional memristor to store the state of that matrix point). I'm rambling on about a potential entire LINE of memristor based small scale integrated chips. Kinda pointless at this point. Even just the FPGA style chip with accessibility to analog compatible I/O fed directly to the memristors would be AMAZING!

I just have issue with the fact that HP COULD create such devices, and sell them to the world so they could learn the new technology and become even more excited at the potential... But they instead just keep all of it internal, and promise memories. I can only store data with a memory. I can't really experiment with neural synaptic devices, or other people experiment with the non linear analog aspects of memristors... Not unless someone creates such parts.

The point of my entire argument, is I believe HP has completely missed the point and forgotten that they can develop all they want internally, but if they want the technology to take off, the WORLD needs to adopt it, and that means teaching the technology by doing... Letting the world get their hands dirty with memristors!

I wouldn't BE here on this blog's forum if I didn't want to get my hands dirty!

If I sound agitated, it's cause I'm tired of rhetoric, and i've been hearing it for more than half a decade! What can I say? I understand that HP wants to pay the bills and bring home a new Aston Martin for shareholder's A through V (W to Y each want a Ferrari, and Z wants a Lambo! LOL)  Still, there really isn't any reason at all that a small team could devolop their in house development silicon into real world applicable devices, such as small scale memristor arrays. Such devices would NOT compete with their memory product line, and would not threaten their internal development! Releasing small scale memristor devices would only educate the world of the potential of their new technology! I just have not seen any indication that they are actually doing anything of the sort!

HP is focusing on the evolution of existing concepts with new technology. They want to make what exists faster, denser, more unified... More evolved, but they are not thinking outside the box. They don't have that spark of innovation that asks, what can we do with this that hasn't been done EVER before! They just ask how we can make a shit ton of cash making mundane, but far superior memory.

[JFA]

I'd quite like a component with the same characteristics as a finger, as that cures so many RF woes.

You're getting into digital device design?

[Psi]

What would be nice are lithium coin cells in 2032 etc. sizes  with the same capacity as now, but with really low internal resistance so it can supply a couple of hundred mA. Peak current delivery is often more of a limiting factor than capacity in long-life wireless devices. Imagine if you could send an SMS from a 2032.

There was a company making those, but ive not been able to find their website for a few years.
I suspect they went out of business.

The high current 2032's they sold were quite expensive for no good reason, other than maybe low sales.

[Psi]

A basic MCU similar to a AVR/PIC but which ran at an internal clock speed of 1+ghz.
16/32bit core would be better but i'd be totally happy with an 8bit core.
Price would be <$10 in qty of 10

The raw power of 1+ghz would be quite useful in cases where you would otherwise have to go FPGA or full on ARM CPU with associated parts (DDR/flash etc).

http://www.digikey.com/product-detail/en/XS1-L8A-128-QF124-I10/XS1-L8A-128-QF124-I10-ND/3906735
Its 2x 32bit, and package, QFN124, is a bit unwieldy, but otherwise what you wanted
Then you have 4 core ARMv7 at >1GHz at $5
http://www.cnx-software.com/2014/06/23/allwinner-a33-tablet-price/

Yeah, mike posted a few pages back showing me XMOS.
It's not really 1ghz, the speed is the sum of all cores. So a 4 core 1000MIPS xmos chip  = 4x ~250mhz cpus.

It's still really interesting though, for the price.
I wonder what the development environment is like, should be able to code with gcc since the cores are ARM.

[T3sl4co1l]

Transistors were once an over expensive, terribly fragile novelty component, in an era where valves were the defacto standard technology. People understood they could mark an improvement and a fundamental shrinking of technology, with reduced power needs than valves... but yes, even in the very early days, transistor were once novelty components.

I disagree.  In gross terms, they operate exactly the same as the existing dominant technology of their day.  Interface and design migration was easy and rapid, limited only by availability and cost.  The advantages were (are) dramatic, and immediately apparent.

None of these is the case with the element in question.

Compare to magnetic amplifiers: they can be used in, sort of generally the same way as transistors, and have reasonable on/off ratio, linearity and gain.  Never really caught on, beyond a few peculiar uses (notable cases: core RAM, 3.3V regulator in ATX supplies).

As the rest of your post seems utterly intent on bashing HP's research model, I shall disregard.  Good day,

Tim

[richfiles]

Transistors were once an over expensive, terribly fragile novelty component, in an era where valves were the defacto standard technology. People understood they could mark an improvement and a fundamental shrinking of technology, with reduced power needs than valves... but yes, even in the very early days, transistor were once novelty components.

I disagree.  In gross terms, they operate exactly the same as the existing dominant technology of their day.  Interface and design migration was easy and rapid, limited only by availability and cost.  The advantages were (are) dramatic, and immediately apparent.

None of these is the case with the element in question.

Compare to magnetic amplifiers: they can be used in, sort of generally the same way as transistors, and have reasonable on/off ratio, linearity and gain.  Never really caught on, beyond a few peculiar uses (notable cases: core RAM, 3.3V regulator in ATX supplies).

As the rest of your post seems utterly intent on bashing HP's research model, I shall disregard.  Good day,

Tim

The duration of novelty doesn't change the fact that, yes, at one time, even transistors were an expensive novelty. Fortunately, for transistors, they VERY rapidly found viable applications and dropped in price, while gaining availability. Additionally, the first transistors to be created were of the bipolar junction variety, in 1947. It wasn't until the FET was created, around 1960 or so, that transistors truly mimicked the specific characteristics of valves more closely. 13 years passed before that happened! Before that, you had to rethink the circuit from what you would have done using valves, not that it wasn't amazing that you could. Like I said, transistors QUICKLY became mainstream, faster than a significantly landscape altering technology like memristors.

One of the areas people forget about, is the logic capacity of memristors. These are devices that switch states and then stay that way with no applied power. A transistor must be powered to be on. A memristor needs only be powered to change states. There have been theories that this would allow lower power consumption for processors and other types of logic hardware. HP also claims to have seen some efficiencies in compiling instructions to run on their unique hardware configuration. This is both an increase in speed and a reduction of power wasted as heat. Memristors have also proven to be capable of being made extraordinarily small. That adds a density bonus to their score.

My issue, again, stems only from the fact that there has been no positive discussion on the possibility of discrete memristors, by HP, and the FPGA claims are only in chip fabrication efficiencies, and not related to user accessible memristance... All memristor function on the proposed FPGA is simply to provide a more efficient chip. I'm just saying that HP seems to be dead set focused on making money with memory. There is NOTHING wrong with that, but at least put an effort to move small scale components to market for educators, experimenters, prototypers, and hobbyists.

Is that truly so hard to ask for?

No, what I see is the industry feeling they do not have to service a discrete industry at all, since the future of industry is in integration. I would say this is a very bad precedent to set. They only see the internal developments for their own short term business model, without seeing the bigger picture of where an industry with no foundation of tinkerers and experimenter will lead them. NO ONE coming into HP in the near future will have memristor experience. The company likely must train everyone new in from scratch. There is no room for innovation if the technology is hidden behind closed doors, and the only thing you get is an application specific black box that can only store digital data.

[Rasz]

xcore

Yeah, mike posted a few pages back showing me XMOS.
It's not really 1ghz, the speed is the sum of all cores. So a 4 core 1000MIPS xmos chip  = 4x ~250mhz cpus.

It's still really interesting though, for the price.
I wonder what the development environment is like, should be able to code with gcc since the cores are ARM.

its two 500MHz cpus glued internally, 4 threads each, so 1000/8.
Dev env is Eclipse, language is basically C. You can have one thread handle 100Mbit ethernet PHY, another one doing TCP/IP, another bitbanging SDcard at full 50MHz in 4bit mode.

[owiecc]

I'd like a multichannel transmitter/receiver working with a single fibre.

[HackedFridgeMagnet]

Here's another : MicroSD cards in light colours so you can write on them

I thought it was only the big ones that had the write lock.


Ok I wish I hadn't said that.

[Precipice]

I'd like a multichannel transmitter/receiver working with a single fibre.

This probably exists, depending on what you want... Not necessarily difficult or expensive, either.

[owiecc]

This probably exists, depending on what you want... Not necessarily difficult or expensive, either.

I would imagine they are common in telecommunication sector. I'd like something cheap to connect e.g. DSP with a gate driver. Send PWM/sync signal one way and send status/diagnostic data back over a single fiber. Can you buy something like this?

[Precipice]

I would imagine they are common in telecommunication sector. I'd like something cheap to connect e.g. DSP with a gate driver. Send PWM/sync signal one way and send status/diagnostic data back over a single fiber. Can you buy something like this?

Ah, 'and back over a single fiber' - harder, I've not seen such a thing retail. I use one fibre there, one fibre back.
Serialisers that let you stuff parallel data down a fibre (and reassemble it at the far end) are quite common. I'm still working my way few a few sticks of AMD TAXI-chip (search will find you stuff) that do exactly this, at 125MBaud, with embedded syncs if you want. I use them to get data in & out of EMC chambers (although I have more recently stuck  an mbed micro in, and use it to drive fibre ethernet). The TAXI chips are fast enough that I can send PWM and FM encoded analogue on individual data pins with reasonable results.

[SeanB]

Telecoms will use 2 fibres, as the optical splitters required to make a bidirectional link on a single fibre are going to be quite expensive, and a hard to align item with a high insertion loss. Multiple fibres in a single jacket are cheap, reliable and have built in redundancy as well in case you have a snapped strand along the cable run.

Short distance links you can use a LED each end and run one as a photodiode and the other as emitter, and arrange the protocol so as to do the half duplex required.

[TMM]

A 74HC4066 that doesn't have a retarded pinout - place all the enable pins next to each other please!

[mikeselectricstuff]

A 74HC4066 that doesn't have a retarded pinout - place all the enable pins next to each other please!

And make the input thresholds 1V regardless of supply.
 

[David Hess]

A 74HC4066 that doesn't have a retarded pinout - place all the enable pins next to each other please!

And make the input thresholds 1V regardless of supply.

The 74HCT4066 is not close enough?

[free_electron]

explain a bit more why you need open source pmos ? that doesnt make sense to me..

Ok strictly speaking, high-side outputs.

mc33880
VNI8200
UCN5891
UCN5890

there are manymore...

[free_electron]

Well, it matters when you don't run Windows (or Linux).  It would just be convenient for me personally not having to mess around with a platform I don't otherwise use, and hopefully the community could at least manage some minor tweaks to the tools too.

i take it you are on a MAC then ? sorry bub. wrong computing platform.
All the FPGa tools are winodws / linux. Even Apple designs their boards using windows.

[free_electron]

They are not discussing discrete memristors, fixed configuration multi memristor integrated devices,
-snip-
Imagine a world where all the resistors, or all the capacitors, or all the inductors were taken away, and you could only get such fundamental parts as integrated into premanufactured integrated components. The concept is absurd,

and what exactly would you do with a single memristor ? Besides do you have any idea about what is needed to read and write it ? the currents and voltages you are dealing with are extremely small. reading the state is destructive : state needs to be written back. it is an impractical device off-chip. i doubt that hwat you have constructed is the same as what hp is building on-chip.

as for a world without passives... coming sooner than you think. severalcompanies already offer integrated passives as in 'a single package containing all r and c and l required'. these are used for further miniaturisation and space saving. tablets and smartphones already use these.

[mikeselectricstuff]

explain a bit more why you need open source pmos ? that doesnt make sense to me..

Ok strictly speaking, high-side outputs.

mc33880

Way more than a simple source driver, lots of pins, expensive

VNI8200

36 pins for 8 channels , expensive

UCN5891
UCN5890

discontinued, unavailable

[David Hess]

They are not discussing discrete memristors, fixed configuration multi memristor integrated devices,
-snip-
Imagine a world where all the resistors, or all the capacitors, or all the inductors were taken away, and you could only get such fundamental parts as integrated into premanufactured integrated components. The concept is absurd,

and what exactly would you do with a single memristor ? Besides do you have any idea about what is needed to read and write it ? the currents and voltages you are dealing with are extremely small. reading the state is destructive : state needs to be written back. it is an impractical device off-chip. i doubt that hwat you have constructed is the same as what hp is building on-chip.

Presumably operation would be scaled up with larger devices to provide lower noise and larger signal-to-noise ratios.  I would use them for low drift or non-volatile integration.

I have to point out that this function does occasionally show up in part failures where a resistance element fails as a mechanical junction and the resistance very roughly varies by the integration of the current through it.

[free_electron]

a bunch or rejections

but i don;t quit

here's a whole table of em SPI interface. sources, sinkers, floaters and combinations (floaters gove you both source and drain , so you tie em where you want em)

http://www.infineon.com/cms/en/product/channel.html?channel=ff80808112ab681d0112ab69dfb9034d#goto_producttable

http://www.onsemi.com/PowerSolutions/product.do?id=AMIS-39101

the 5891 is still buitl by micrel :
http://www.micrel.com/index.php/products/power-management-ics/driver-arrays/latched-drivers/article/9-mic5891.html

[Precipice]

36 pins for 8 channels , expensive

It also has a 'hilarious' bug, where, at elevated (but still well within rated) temperature, commands from the SPI bus get acknowledged and can be read back, but don't get passed to the output drivers until the temperature drops, when they suddenly change to what they should have been.
Thanks, ST. As ever, designing in one of your parts turns out to be a mistake. (I'll check the actual part number tomorrow. It was, I think, the automotive rated version of that chip).

Edit: L9733XP is the (very similar) part with the mentioned bug. Grr.

[steve30]

A component I would love to see, is an op-amp, in a triangle shaped package, with the output pin on the point of the triangle, the inputs on the opposite side, and the power supply pins coming out of the other two sides.

That would be fantastic for prototyping stuff. I've wanted one like that for years .

[Stonent]

A component I would love to see, is an op-amp, in a triangle shaped package, with the output pin on the point of the triangle, the inputs on the opposite side, and the power supply pins coming out of the other two sides.

That would be fantastic for prototyping stuff. I've wanted one like that for years .

 

[T3sl4co1l]

A triangle wouldn't fit on a breadboard very well though.  They could be marked with the pinout on the DIP, at least -- or something like that.  But, who would ever want to do that...

Tim

[Druzyek]

1. A shift register like a 595 that can do 1, 0, and Z on each pin, not just Z on all pins or none.

2. A cheap TRNG chip. Measuring a transistor in reverse mode seems convoluted and according to some is really not that random.

3. A voltage regulator in TO-220 or similar that can give -27v from 3.3v for LCD contrast without needing 10+ external components.

A mid-range microcontroller with a buttload of RAM (>64K) but without the cost of a ton of expensive flash

I have also thought for a long time that 256k RAM and 8k flash makes more sense than the other way around. A few k is enough for a bootloader to load code from an SD card or EEPROM to run everything from RAM to get around flash wait states. FRAM is a neat idea too but the MSP430s TI is making with it now top out at 8MHz.

[T3sl4co1l]

1. A shift register like a 595 that can do 1, 0, and Z on each pin, not just Z on all pins or none.

Look at SPI GPIOs -- you're probably doing something like SPI to drive it, already.  If it needs to be chainable, not sure.

Tim

[Rigby]

A component I would love to see, is an op-amp, in a triangle shaped package, with the output pin on the point of the triangle, the inputs on the opposite side, and the power supply pins coming out of the other two sides.

That would be fantastic for prototyping stuff. I've wanted one like that for years .

I think a 3D printer and the application of some ingenuity could make that a reality fairly easily.

[free_electron]

A component I would love to see, is an op-amp, in a triangle shaped package, with the output pin on the point of the triangle, the inputs on the opposite side, and the power supply pins coming out of the other two sides.

That would be fantastic for prototyping stuff. I've wanted one like that for years .

I think a 3D printer and the application of some ingenuity could make that a reality fairly easily.

Repackage a sot23 opamp

[Monkeh]

FRAM is a neat idea too but the MSP430s TI is making with it now top out at 8MHz.

Except for the 23 chips which run at 16MHz, and the 10 which run at 24MHz..

[Galenbo]

Button switch signal conditioner for PIC/other embedded.

-8 channel in 1 IC.
-opto isolator per input.
-5V/12V on open contacts for light in the button.
-RC or other filtering, debounce, limiting the signal to 5? Hz.
-Schmitt trigger with 10%/90% for eliminating ingress.
-Output buffer capable of lighting a led, relay and uC.
-configurator pins for settings of filter, invertor, trigger,...

No rocket science, just an all-in-one for (long distance) buttons in your home, you robot, you machine.

[Psi]

1. A shift register like a 595 that can do 1, 0, and Z on each pin, not just Z on all pins or none.

+1 for this

[Precipice]

1. A shift register like a 595 that can do 1, 0, and Z on each pin, not just Z on all pins or none.

+1 for this

By the time you're at that level of complexity, shouldn't you just move to an IO expander chip?
I tend to use
http://www.nxp.com/products/interface_and_connectivity/i2c/i2c_general_purpose_i_o/PCA9538PW.html
but SPI versions also exist if you don't like i2c or want more speed, and there are many, many  flavours. They're cheap, too.

[jlmoon]

A component I would love to see, is an op-amp, in a triangle shaped package, with the output pin on the point of the triangle, the inputs on the opposite side, and the power supply pins coming out of the other two sides.

That would be fantastic for prototyping stuff. I've wanted one like that for years .

Cut out some triangle shaped perf-board with those solder pads on the bottom, solder your favorite op amp in the middle and fashion some stick pins off the bottom add some periph parts, power and enjoy.  Whooo Hoo!

[retrolefty]

A component I would love to see, is an op-amp, in a triangle shaped package, with the output pin on the point of the triangle, the inputs on the opposite side, and the power supply pins coming out of the other two sides.

That would be fantastic for prototyping stuff. I've wanted one like that for years .

Cut out some triangle shaped perf-board with those solder pads on the bottom, solder your favorite op amp in the middle and fashion some stick pins off the bottom add some periph parts, power and enjoy.  Whooo Hoo!

+1,
 Except it's not Whoo Hoo, it's Bob's your Uncle on this site.

[babysitter]

All chips shall have programmable multiplexers allowing to exchange any pin position. Except power rails. They need to be on adjacent pins.

[rob77]

3d printable chips just design you IC and print it in silicon
IBM made that 10nm resolution printer already - so probably in few hundred years we'll be there 

[Rigby]

few hundred years

Less.  It'll happen, but it probably won't be done in silicon.

[Zad]

All chips shall have programmable multiplexers allowing to exchange any pin position. Except power rails. They need to be on adjacent pins.

Putting power rails on adjacent pins is just asking for trouble when probing around looking for faults! I would be happy with current DIP/SOIC/TSSOP type packaging, but with 2 extra pins on each end for JTAG-like configuration signals.

Can I also have a range of nice cheap epoxy packages that look like QFP, but which I can clip a BGA onto, in order to break out the connections? I know there are various BGA breakout boards, but I'm thinking of something not just for the 400 ball+ behemoths, but for 6 ball tiny little things too.

[Psi]

1. A shift register like a 595 that can do 1, 0, and Z on each pin, not just Z on all pins or none.

+1 for this

By the time you're at that level of complexity, shouldn't you just move to an IO expander chip?
I tend to use
http://www.nxp.com/products/interface_and_connectivity/i2c/i2c_general_purpose_i_o/PCA9538PW.html
but SPI versions also exist if you don't like i2c or want more speed, and there are many, many  flavours. They're cheap, too.

The reason i wanted that, maybe him as well, was so i can parallel load a large number of bits into the shift reg and then stream them out serially as L, H, or Z for composite OSD overlay.
The video data is too fast to clock directly with absolute timing using a generic MCU i/o in software, so being able to load a large chunk of data all at once takes the processing load off the MCU. (Obviously it would uses 2 bits per 1 bit out due to tristate operation)
High Z is needed so the composite video signal can pass through unmodified if the pixel is not part of the osd black or white.
An i/o expander wouldn't work for the same reason a generic mcu i/o wouldnt work (too slow and no locked timing).

[David Hess]

If a 595 would otherwise work then use two 595s in parallel and an external tristate driver.  One 595 drives the logic input and the other drives the tristate input.

Can I also have a range of nice cheap epoxy packages that look like QFP, but which I can clip a BGA onto, in order to break out the connections? I know there are various BGA breakout boards, but I'm thinking of something not just for the 400 ball+ behemoths, but for 6 ball tiny little things too.

This is another reason I liked PLCC (plastic leaded chip carrier) packages although they were not quite this small.

[Psi]

yeah, there are ways to do it, it would just be nice if a single package existed.

[Melt-O-Tronic]

I'd like to have high voltage RF capacitors that could be installed in a circuit, then programmed with an exact value, perhaps by shouting at it through a bullhorn.   

Same for toroidal inductors.

[T3sl4co1l]

I'd like to have high voltage RF capacitors that could be installed in a circuit, then programmed with an exact value, perhaps by shouting at it through a bullhorn.   

Same for toroidal inductors.

They had those back in the day, but they were expensive back then, and aren't any better today... 

[free_electron]

• Microcontrollers with embedded bypass caps.

intel pentium Iv  , Xeons and later have had those for a long time.

[T3sl4co1l]

A microwave LED. Figure that one out, physicists!

Gunn diode.

Not the same, physically speaking, but a true "microwave LED" wouldn't operate much above absolute zero, in the same way that a red LED doesn't work over, oh, 200C or so (assuming you had a bare die to demonstrate with).  Or to put it another way, anything with that kind of bandgap is literally a metal at room temperature.

It should be considered an advantage that the Gunn diode is tunable

Tim

[Kjelt]

All microcontrollers having (at least some) non volatile FRAM inside instead of volatile SRAM 

[AndyC_772]

How about an amplifier with its gain programmed by the voltage on a dedicated pin?

I don't mean a PGA with a few discrete settings (eg. x1, x5 or x10, take it or leave it), but continuously variable gain from 0 up to some defined maximum, with either a linear or log response in between.

Along similar lines, a nice, simple filter IC with its cut-off frequency adjusted in the same way would be nice. No need for multiple, ganged external pots that all vary together.

[T3sl4co1l]

How about an amplifier with its gain programmed by the voltage on a dedicated pin?

I don't mean a PGA with a few discrete settings (eg. x1, x5 or x10, take it or leave it), but continuously variable gain from 0 up to some defined maximum, with either a linear or log response in between.

Old school: LM13700 (or even older relatives!), current controlled.  Gain is also proportional to absolute temperature.

Newer: VGAs are common for video applications (oddly enough, for the name, I guess?).  Example:
http://www.ti.com/lit/ds/symlink/lmh6505.pdf
Gain still varies with temp, possibly by the same figure.

There's also internal-ADC types like the AD603 that do it in steps, but fine enough you may not care (<1dB increment isn't much to fret over).

Along similar lines, a nice, simple filter IC with its cut-off frequency adjusted in the same way would be nice. No need for multiple, ganged external pots that all vary together.

Peruse some old school analog synth schematics.  You'll find piles of OTAs like the LM13700.   Also, JFETs used as variable resistors (usually with op-amps), or CdS photocells paired with LEDs (aka vactrols).  (Or before LEDs, neons or incandescents were used too.)  JFETs are generally hard to use at high linearity and dynamic range, and have to be sorted for matched characteristics.  CdS have weird time constants and poor repeatability.  Such circuits were huge in compensation components (thermistors and diodes and etc.) and had many trimpots to calibrate.

The better alternative for filters is a switched-capacitor type, many of which are available as ICs (I think LT and Maxim are the most important producers of these).  Cutoff is proportional to clock frequency, which can easily be varied continuously with a VCO.

Tim

[Galenbo]

How about an amplifier with its gain programmed by the voltage on a dedicated pin?

This exists, it's called a transistor.

For the purists: fet-transistor, or a normal one in a common-x circuit.

[Precipice]

DDRn with a jtag port, because testpoints really spoil a fast wide bus layout.

i2c 7-segment LEDs, because why not.

More components moulded in that orange epoxy that some tants use, that change colour when they're hot.

BGAs with location pins, like some fine pitch connectors have. (or cradles with locating pins that fit around BGAs, so you can easily line chips up on prototypes, then not use them on production)

Camera ICs with datasheets that aren't full of lies. Or, if you've got to stick 5K of DSP code down the i2c bus to stop the camera sucking, how about mentioning it and making it publicly available? Yes, Aptina, You! OVL are just as bad.

[mikeselectricstuff]

Camera ICs with datasheets that aren't full of lies. Or, if you've got to stick 5K of DSP code down the i2c bus to stop the camera sucking, how about mentioning it and making it publicly available? Yes, Aptina, You! OVL are just as bad.

..or at the very, very least, have the factory default register values produce a vaguely useable output from startup

[Psi]

BGAs with location pins

heh, that would be cool


How about micro laser-etching the actual full part number on all surface mount components.
Now that would be expensive but so useful to hobbiests

[tszaboo]

Single chip analog PID controller with resistor/I2C programmable parameters.
And "those" components in standard footprints. Every time I use a new inductor I have to draw a different footprint. How about using the same for once? Not to mention PCB mounted batteries or any stupid mechanical parts. How about making a battery in DO-214 package?

[GeorgeOfTheJungle]

Electrolytics that don't dry up. Or a syringe to wet them again when they do.

[GeorgeOfTheJungle]

And a decent gas soldering iron.

[mariush]

I was searching for a thermal wire stripper today (you know, strip insulation by placing wire between two nichrome blades) ... was surprised there's no chinese tools that would do this on eBay on a brief search, was only able to find $100-$150 American strippers. 

Digikey has something cheaper but only rated for PVC so I assume it's lower temperature.

Maybe I got the search terms wrong..

[Rigby]

Electrolytics that don't dry up. Or a syringe to wet them again when they do.

Seems like they could be refillable.

[Stonent]

Electrolytics that don't dry up. Or a syringe to wet them again when they do.

Seems like they could be refillable.

There could be a market for DIY refill kits. 

[nuno]

I have missed many times having current limited microcontroller pins, essentially for LEDs, but not only. Something like 1mA resolution up to 32mA. On all (GPIO) micro pins.

[rob77]

Electrolytics that don't dry up. Or a syringe to wet them again when they do.

Seems like they could be refillable.

There could be a market for DIY refill kits. 

just punch a hole with a needle and inject some fresh electrolyte, then simply seal with a drop of epoxy 

[Druzyek]

Except for the 23 chips which run at 16MHz, and the 10 which run at 24MHz..

Except for 0 of those can run from FRAM at those speeds. Anything above 8MHz is wait stated, so only peripherals benefit from speeds higher than that.

[magetoo]

i take it you are on a MAC then ?

BSD.

All the FPGa tools are winodws / linux.

Well, yeah.  That's why I said I want open source.

Open source that wraps some secret sauce library if it has to be that way, but there's no reason for tools that just massage data to be locked to specific platforms.

[T3sl4co1l]

I have missed many times having current limited microcontroller pins, essentially for LEDs, but not only. Something like 1mA resolution up to 32mA. On all (GPIO) micro pins.

Hmm, interesting thought.  In regards to FPGAs, the pin drivers are selectable -- though I don't know offhand that any offer such a wide range, and that's by configuration, not user control.

It's something of a special-purpose and not-very-useful feature, which is why you probably won't see it any time soon...

The output stages wouldn't take up too much extra space -- a 20mA pin already needs a pair of 20mA transistors, whereas a variable stage only needs a binary series of transistors of the same total area.  It would take extra drive circuitry to get a usefully constant current, but that can be a relatively small blob of analog stuff (assuming the designers know what they're doing, and the CMOS process is capable of doing it well enough).  Big pain is, you'd have to route data and address buses all the way around the periphery of the chip, to set the drive strength registers.  Or have those embedded with the rest of the IO registers, but send all the individual bits around to each port pin.  Either way, the logistics are messy -- possibly worse than the need for pins at all!

Hmm, if you combined that with extended voltage range, or buffering, or something, you could have pins directly compatible with an even more diverse number of standards: e.g., LVDS (+/-8mA current mode; complementary in pairs), RS232 (+/- 6V or more, 20mA limited; single pins), RS422/485 (2-5V, 50-100mA limited; complementary), etc.  Optional would be ESD resistance on such pins, but the user can always add that externally.

Would also be neat to have analog input thresholds for all pins, either against one reference voltage (internally or externally programmed), or by bank or port, or even per pin.  That wouldn't need too much more analog stuff, since a dumb LVDS input stage sort of thing could be used -- keeping good performance and relatively tight voltage thresholds.  FPGAs have this sort of thing, when they have to support SSTL (SDRAM signaling) standards -- think LVDS, but a whole bus worth of "+" lines, with all the "-" brought into one VREF pin.  There would no longer be any need for a stupid "analog comparator" element in the chip -- any pin could serve the job, and with much better speed than normal.  What's typical of these things is bad offset voltage -- LVDS receivers are only guaranteed within 100mV, but they're usually better than that, and they can be calibrated.  You could build yourself a self-calibrating ADC with this, for example.

Tim

[David Hess]

Modern IC processes used for FPGAs and other high performance logic are not going to be suitable for general purpose programmable outputs.

Drive strength for instance applies to saturated operation (except for current mode outputs) and is for controlling transmission line effects.  Output voltage ranges are very limited unless a much more expensive IC process is used and even then, they are still limited.

RS-232 and RS-422/485 signal levels are not going to be possible and even 5 volt logic families usually used external level translation for them.  USB signal levels will also be difficult for an advanced process.

[T3sl4co1l]

Modern IC processes used for FPGAs and other high performance logic are not going to be suitable for general purpose programmable outputs.

Could you explain?

I wouldn't think FPGA processes would be all that applicable to MCUs, which I guess are usually 3.3V through and through.  Then again, there are plenty available in the >100MHz class with that sort of architecture (~1V core, bank IO), so it depends what you're referring to I guess.

Ultimately I should suppose the highest-falootin' FPGAs are very similar to the highest-falootin' processors, but those are all GTL or whatever, which I expect isn't far from the stuff the core is made of.  But yeah, that's all special fancy low voltage stuff, hardly the crusty high voltage stuff we like for embedded.

Drive strength for instance applies to saturated operation (except for current mode outputs) and is for controlling transmission line effects.

Isn't that the same thing?

Usually what's done for a CMOS current sink/source is, either just running it flat out -- you get a lot of resistance near the rail, but eventually it comes out of the resistive range and becomes constant-ish current.  This usually takes more than half the supply voltage, so it's not exactly useful.  But that already helps "transmission line effects", where it goes up a little bit right away, then even more slowly as the TL charges.  Instead of jamming it over and causing big bounce, or whatever.

A real current limited output should drop the gate voltage a little, usually as a current mirror.  Which could be switched on and off, but that might be slow; or cascoded, which would take a lot more area (the CCS will already have to be bigger to get the same Rds(on) or equivalent at the reduced Vgs, plus the switch).

So it really shouldn't be too big of a deal, but it would take more area, and a lot more nuts and bolts to make it useful.  Should be the same on any process, just scaled by the volts/amps of whatever is typical of it..

Output voltage ranges are very limited unless a much more expensive IC process is used and even then, they are still limited.

Well, you put it in a 3.3V IO bank and get 3.3V ranges, or etc.  Typical of FPGAs, at least the non-high-falootin' ones.  I haven't even checked if the fanciest ones have 3.3V banks or not.. wouldn't be surprised if they don't, at least on the high speed banks or whatever.

Having a high voltage bank would definitely be hard work in most processes.  Like a 15V (or even +/-15V) section among 3.3V (or lower) logic.  The transistors would also be massive, probably easily bigger than the rest of a chip that size (referring to like 250nm scale processes).

Out of, like, a 5V chip (most ATMega/PIC stuff people dick around with..), it needn't be much harder than the rest.

Tim

[David Hess]

Modern IC processes used for FPGAs and other high performance logic are not going to be suitable for general purpose programmable outputs.

Could you explain?

I wouldn't think FPGA processes would be all that applicable to MCUs, which I guess are usually 3.3V through and through.  Then again, there are plenty available in the >100MHz class with that sort of architecture (~1V core, bank IO), so it depends what you're referring to I guess.

Ultimately I should suppose the highest-falootin' FPGAs are very similar to the highest-falootin' processors, but those are all GTL or whatever, which I expect isn't far from the stuff the core is made of.  But yeah, that's all special fancy low voltage stuff, hardly the crusty high voltage stuff we like for embedded.

I am not saying that it is a problem of economics and not what can actually be done.  As you make the IC process more complicated it takes longer to make them and yields are poorer.  Any extra features have to be economical even for the applications which do not use them.

There is also a strong motivation to use a high volume process instead of a specialized low volume one.

Drive strength for instance applies to saturated operation (except for current mode outputs) and is for controlling transmission line effects.

Isn't that the same thing?

I am not exactly sure how this is done with CMOS outputs but the result is that the output impedance varies with voltage so the transition time is slowed by applying less current.  They want minimum delay and high drive but without hard switching which would cause excessive noise.

Usually what's done for a CMOS current sink/source is, either just running it flat out -- you get a lot of resistance near the rail, but eventually it comes out of the resistive range and becomes constant-ish current.  This usually takes more than half the supply voltage, so it's not exactly useful.  But that already helps "transmission line effects", where it goes up a little bit right away, then even more slowly as the TL charges.  Instead of jamming it over and causing big bounce, or whatever.

CMOS does not saturate the way bipolar does.  The channel resistance is still low when the output is near either rail so switching tends to be hard making ground and power bounce a real problem.  Bipolar output stages have a real advantage here.

A real current limited output should drop the gate voltage a little, usually as a current mirror.  Which could be switched on and off, but that might be slow; or cascoded, which would take a lot more area (the CCS will already have to be bigger to get the same Rds(on) or equivalent at the reduced Vgs, plus the switch).

So it really shouldn't be too big of a deal, but it would take more area, and a lot more nuts and bolts to make it useful.  Should be the same on any process, just scaled by the volts/amps of whatever is typical of it..

A complementary current mirror output (transconductance) could work but would need to be designed to operate in the linear instead of saturated region.  LVDS outputs use switched currents but not this type of output stage and have a limited voltage range at least according to their specifications but maybe that is an issue with their receivers.

Output voltage ranges are very limited unless a much more expensive IC process is used and even then, they are still limited.

Well, you put it in a 3.3V IO bank and get 3.3V ranges, or etc.  Typical of FPGAs, at least the non-high-falootin' ones.  I haven't even checked if the fanciest ones have 3.3V banks or not.. wouldn't be surprised if they don't, at least on the high speed banks or whatever.

Having a high voltage bank would definitely be hard work in most processes.  Like a 15V (or even +/-15V) section among 3.3V (or lower) logic.  The transistors would also be massive, probably easily bigger than the rest of a chip that size (referring to like 250nm scale processes).

Out of, like, a 5V chip (most ATMega/PIC stuff people dick around with..), it needn't be much harder than the rest.

For CMOS it is the gate oxide thickness and maybe diffusions and not the area and feature size that are the problem.  They already use multiple gate oxide thicknesses to support 3.3 volt I/O (and separate fast and low power core transistors) and adding more steps to support a rarely needed feature is going cost even for the applications that do not need it.

High performance microcontrollers separate their I/O and core voltages just like high performance gate arrays, microprocessors, and other logic.

[nuno]

Big pain is, you'd have to route data and address buses all the way around the periphery of the chip, to set the drive strength registers.  Or have those embedded with the rest of the IO registers, but send all the individual bits around to each port pin.  Either way, the logistics are messy -- possibly worse than the need for pins at all!

Or clock that data serially around the pins. Slow current level setting is ok as long as there's still fast ON/OFF control.
Or have current DACs in the core or wherever it makes more sense, route the single pin output to the periphery and multiply the current there.

Hmm, if you combined that with extended voltage range, or buffering, or something, you could have pins directly compatible with an even more diverse number of standards: e.g., LVDS (+/-8mA current mode; complementary in pairs), RS232 (+/- 6V or more, 20mA limited; single pins), RS422/485 (2-5V, 50-100mA limited; complementary), etc.  Optional would be ESD resistance on such pins, but the user can always add that externally.

We're talking microcontrollers here, so I don't see much use of LVDS and similar "high speed" related stuff - although RSxxx would be nice, although not my preference. I would like however, open drain/colector outputs capable of tolerating up to 15V on the pin when the transistor is off.

[Artlav]

A 100V, 3000F supercapacitor the size of BCAP3000.

But seriously:
-An FPGA-like device, but with analog cells - opamps, resistors, current sources, etc.
-Wider selection of bolt-mounted parts - diodes, mosfets, capacitors, etc. Not just high power, high price ones.
-More integrated power parts, like a MOSFET H-bridge with all the driving and level-shift circuitry already in it.

More wishful thinking:
-An MWIR/LWIR photodiode with high sensitivity/shunt resistance of silicon ones.

[Precipice]

But seriously:
-An FPGA-like device, but with analog cells - opamps, resistors, current sources, etc.

Check out Cypress PSOC. Sure, it comes with a free micro, but that's not the end of the world.
You may have to apply some careful thought to make it do what you want - there's no 'just copy a schematic in from a spice simulation', but it's proper analogue on a chip.

-More integrated power parts, like a MOSFET H-bridge with all the driving and level-shift circuitry already in it.

These exist, in various sizes - from SO8 to decent size power devices. What are you after?

[AdShea]

Actual Reverse-Blocking IGBTs that you can buy and aren't 3 generations behind.

GMR+Bias Magnet+Amplifier in a CSP for high-bandwidth current sensing.

Magnetic insulators (<1 mu materials) for actual flux barriers.

A wind-to-order service for standard bobbins/cores that doesn't require a trip to China and a MOQ of 1000s.

[coppice]

Except for the 23 chips which run at 16MHz, and the 10 which run at 24MHz..

Except for 0 of those can run from FRAM at those speeds. Anything above 8MHz is wait stated, so only peripherals benefit from speeds higher than that.

Those devices will incur wait states when running at full speed, but nothing like the quantity you imply. Configured correctly, changing the core clock from 8MHz to 16MHz to 24MHz gives substantial increases in execution speed. Perhaps you have tried them with the pipelining disabled.

Few modern MCUs run entirely free of wait states. Its rather inefficient to make them do so. Whether they use flash or FRAM, the non-volatile memory is generally the speed limiter. To get the best out of the silicon you need some measure of pipelining, caching or related tricks to hide that limitation. When the tricks run out you have to insert a wait state.

[coppice]

Electrolytics that don't dry up. Or a syringe to wet them again when they do.

Are you talking about really big electrolytics, because solid electrolytics can solve that problem over quite a wide range of sizes these days. They just cost a bit more.

[David Hess]

Electrolytics that don't dry up. Or a syringe to wet them again when they do.

Are you talking about really big electrolytics, because solid electrolytics can solve that problem over quite a wide range of sizes these days. They just cost a bit more.

The specifications for the solid electrolytic capacitors do not list operating lifetimes higher than high quality liquid electrolytic capacitors though.  I have not had any solid ones long enough to see what happens when they get old.

[coppice]

Electrolytics that don't dry up. Or a syringe to wet them again when they do.

Are you talking about really big electrolytics, because solid electrolytics can solve that problem over quite a wide range of sizes these days. They just cost a bit more.

The specifications for the solid electrolytic capacitors do not list operating lifetimes higher than high quality liquid electrolytic capacitors though.  I have not had any solid ones long enough to see what happens when they get old.

Most solids don't look that spectacular at maximum temperature. A wet electrolytic might be 1000 hours, and a solid might be 2000 hours at 105C. However, the lifetimes diverge considerably at more modest temperatures. Typically wets are rated for a doubled life with every 10C drop in temperature. Solids are rated for more than 3 times the life for every 10C drop in temperature. That doesn't sound a huge difference until you work it through. A 1000 hour/105C wet electrolytic has a life of 32000 hours at 55C (3 years). A 2000 hour/105C solid has a life of about 500,000 hours at 55C (50 years).

[kony]

But seriously:
-An FPGA-like device, but with analog cells - opamps, resistors, current sources, etc.

Ever heard of FPAAs ? http://www.anadigm.com/fpaa.asp

[coppice]

GMR+Bias Magnet+Amplifier in a CSP for high-bandwidth current sensing.

Magnetic insulators (<1 mu materials) for actual flux barriers.

Something like http://www.alpsgd.com/eng/product/current.html ?