Reply by February 1, 20192019-02-01
Gerhard Hoffmann wrote
>Am 31.01.19 um 14:15 schrieb 698839253X6D445TD@nospam.org: > >>> >>> In this app note, they mention that they have pretty accurate models >>> >>> < >>> >>> https://www.wolfspeed.com/downloads/dl/file/id/847/product/159/load_pull_validation_of_large_signal_cree_gan_field_effect_transis >>> tor_fet_model.pdf >>> > >>> >>> but it takes a harmonic balance simulator, so el cheapo LTspice won't >>> cut it. Buy a copy of ADS, for 100K +-. >>> >>> regards, Gerhard >> >> Yea, >> well I thought : >> 'Must be extremely nonlinear or large production spread else they would publish it'. >> > >No, just the opposite. The only important market for these is cell phone >base stations and to avoid spectral regrowth and for efficiency reasons >they need at least Doherty amplifiers, preferably with a monitoring >receiver and active predistortion. Just take a look at ADI's fastest >ADCs with GHz sample rates and what they are made for.
OK, but I wrote that because you never know if they read this group and would then publish it ;-)
>You cannot make that with LTspice, it takes ADS or Microwave office. >The data sheets are just teasers, you don't get far without their >engineering pack.
I need to get some experience with those FETS, really. There is a lot of those new parts that I have no hand on experience with yet, not needed those till now.
>I once did some design work on a base station, regretfully on it's >CPU and not the RF section. The customer made it clear that every &#4294967295; >on the BOM would have to be multiplied by 10 Meg over the product >lifetime. >Even Xilinx insisted they would never ever lose the design-in on price. >Shocking, those numbers. Even J&#4294967295;rg would have felt challenged.
Many years ago, eh 2010 to be precise, I bought a "super-conductor technologies super-filter" on ebay for a few hundred dollars from the US, it came from a cellphone tower. http://panteltje.com/pub/super_filter/super_filter_front_plate_img_2576.jpg https://www.ebay.com/itm/220707865078 It has a band filter made with a super-conductor (around 800 MHz or so it is?) cooled by a Sterling cooler, and a lot of nice RF stuff for that band, RF relays, RF preamps, attenuators, ... I bought it for the Sterling cooler, that alone is 12k $ new, to cool cameras to get lower noise. Condensation is always a problem, but also the thing is great for all sorts of experiments, bought a CSSP-YLME YBCO-123 Disk from can-superconductors.com... Make your own liquid air... http://panteltje.com/pub/cryo/ http://panteltje.com/pub/super_filter/super_filter_cryo_cooler_img_2545.jpg The actual super-conducting filter is in a large dewar, that the Sterling cooler sticks into: http://panteltje.com/pub/super_filter/super_filter_dewar_img_2557.jpg http://panteltje.com/pub/super_filter/super_filter_connection_panel_img_2522.jpg http://panteltje.com/pub/super_filter/super_filter_top_boards_img_2573.jpg
>> So I erased the pdf, after all I have this: >> http://panteltje.com/pub/spectrian_2.4-GHz_linear_top_view_IMG_4576.JPG >> 2.4 GHz 75 watt linear, was only 99$ on ebay >> https://www.ebay.com/itm/75W-Spectrian-Linear-RF-Amplifier-Board-2-3-2-35-GHz-18dBg-24-26V-/223351579252?oid=221556214361 >> the idea is to use it for ATV to the Es'hail-2 sat one of these days.. > >You could have got mine. I bought it for AO-40, the one with the strong >elliptical orbit. When it turned out that it would never fly, the >amplifier strip disappeared somewhere in the basement. >One of my worst investments.
I bought mine in 2014, the Es'hail-2 was delayed many times..... So the thing landed, after testing with a dummy load, in the storage too. Anyways cellphone tower are sort of interesting in many ways,,, Maybe even more so now 5G coming.
Reply by Gerhard Hoffmann January 31, 20192019-01-31
Am 31.01.19 um 14:15 schrieb 698839253X6D445TD@nospam.org:

>> >> In this app note, they mention that they have pretty accurate models >> >> < >> https://www.wolfspeed.com/downloads/dl/file/id/847/product/159/load_pull_validation_of_large_signal_cree_gan_field_effect_transis >> tor_fet_model.pdf >> > >> >> but it takes a harmonic balance simulator, so el cheapo LTspice won't >> cut it. Buy a copy of ADS, for 100K +-. >> >> regards, Gerhard > > Yea, > well I thought : > 'Must be extremely nonlinear or large production spread else they would publish it'. >
No, just the opposite. The only important market for these is cell phone base stations and to avoid spectral regrowth and for efficiency reasons they need at least Doherty amplifiers, preferably with a monitoring receiver and active predistortion. Just take a look at ADI's fastest ADCs with GHz sample rates and what they are made for. You cannot make that with LTspice, it takes ADS or Microwave office. The data sheets are just teasers, you don't get far without their engineering pack. I once did some design work on a base station, regretfully on it's CPU and not the RF section. The customer made it clear that every &#4294967295; on the BOM would have to be multiplied by 10 Meg over the product lifetime. Even Xilinx insisted they would never ever lose the design-in on price. Shocking, those numbers. Even J&#4294967295;rg would have felt challenged.
> So I erased the pdf, after all I have this: > http://panteltje.com/pub/spectrian_2.4-GHz_linear_top_view_IMG_4576.JPG > 2.4 GHz 75 watt linear, was only 99$ on ebay > https://www.ebay.com/itm/75W-Spectrian-Linear-RF-Amplifier-Board-2-3-2-35-GHz-18dBg-24-26V-/223351579252?oid=221556214361 > the idea is to use it for ATV to the Es'hail-2 sat one of these days..
You could have got mine. I bought it for AO-40, the one with the strong elliptical orbit. When it turned out that it would never fly, the amplifier strip disappeared somewhere in the basement. One of my worst investments. 73, Gerhard
Reply by John Larkin January 30, 20192019-01-30
On Wed, 30 Jan 2019 14:23:02 -0800 (PST), whit3rd <whit3rd@gmail.com>
wrote:

>On Wednesday, January 30, 2019 at 11:48:35 AM UTC-8, John Larkin wrote: > >> The eval boards just bring out the gate bias as a pin and assume users >> will magically know how to set that. > >Yeah, crank it up til it emits magic smoke, then twenty percent >less...
It's not obvious what polarity the bias should be. -- John Larkin Highland Technology, Inc picosecond timing precision measurement jlarkin att highlandtechnology dott com http://www.highlandtechnology.com
Reply by whit3rd January 30, 20192019-01-30
On Wednesday, January 30, 2019 at 11:48:35 AM UTC-8, John Larkin wrote:

> The eval boards just bring out the gate bias as a pin and assume users > will magically know how to set that.
Yeah, crank it up til it emits magic smoke, then twenty percent less...
Reply by John Larkin January 30, 20192019-01-30
On Tue, 29 Jan 2019 20:58:02 -0800 (PST), dagmargoodboat@yahoo.com
wrote:

>On Thursday, January 17, 2019 at 4:52:13 PM UTC-5, Gerhard Hoffmann wrote: >> Am 12.01.19 um 06:39 schrieb dagmargoodboat@yahoo.com: >> >> > One of my problems, currently, is that most of the high-voltage fancy >> > FETs are monstrous compared to my needs, so even though their figures- >> > of-merit Ron vs. Qg are worlds better than the older tech stuff, the >> > fact that they're 20x over-sized for my application makes them slower >> > than an appropriately-sized MOSFET. >> > >> > Even the most nimble GaN garbage truck FET still isn't as gamely as a >> > silicon itty bitty unicycle FET. >> >> A bit smaller than the other transistors that were mentioned, >> but nevertheless DC to 6 GHz: >> >> < https://www.mouser.de/datasheet/2/90/ghv27030s-947886.pdf > >> >> A nice driver, at least :-) >> >> Cheers, >> Gerhard > >Wow, that's one beast of a driver ;-) > >Cheers, >James Arthur
It's a typical RF data sheet: primitive DC specs and no DC curves at all. The eval boards just bring out the gate bias as a pin and assume users will magically know how to set that. -- John Larkin Highland Technology, Inc picosecond timing precision measurement jlarkin att highlandtechnology dott com http://www.highlandtechnology.com
Reply by January 30, 20192019-01-30
On Thursday, January 17, 2019 at 4:52:13 PM UTC-5, Gerhard Hoffmann wrote:
> Am 12.01.19 um 06:39 schrieb dagmargoodboat@yahoo.com: > > > One of my problems, currently, is that most of the high-voltage fancy > > FETs are monstrous compared to my needs, so even though their figures- > > of-merit Ron vs. Qg are worlds better than the older tech stuff, the > > fact that they're 20x over-sized for my application makes them slower > > than an appropriately-sized MOSFET. > > > > Even the most nimble GaN garbage truck FET still isn't as gamely as a > > silicon itty bitty unicycle FET. > > A bit smaller than the other transistors that were mentioned, > but nevertheless DC to 6 GHz: > > < https://www.mouser.de/datasheet/2/90/ghv27030s-947886.pdf > > > A nice driver, at least :-) > > Cheers, > Gerhard
Wow, that's one beast of a driver ;-) Cheers, James Arthur
Reply by Gerhard Hoffmann January 17, 20192019-01-17
Am 12.01.19 um 06:39 schrieb dagmargoodboat@yahoo.com:

> One of my problems, currently, is that most of the high-voltage fancy > FETs are monstrous compared to my needs, so even though their figures- > of-merit Ron vs. Qg are worlds better than the older tech stuff, the > fact that they're 20x over-sized for my application makes them slower > than an appropriately-sized MOSFET. > > Even the most nimble GaN garbage truck FET still isn't as gamely as a > silicon itty bitty unicycle FET.
A bit smaller than the other transistors that were mentioned, but nevertheless DC to 6 GHz: < https://www.mouser.de/datasheet/2/90/ghv27030s-947886.pdf > A nice driver, at least :-) Cheers, Gerhard
Reply by John Larkin January 13, 20192019-01-13
On Mon, 14 Jan 2019 10:13:11 +1100, Clifford Heath
<no.spam@please.net> wrote:

>On 13/1/19 6:55 am, Winfield Hill wrote: >> dagmargoodboat@yahoo.com wrote... >>> >>> Winfield Hill wrote: >>>> >>>> There's very little in the signal path. And the 20ns delay to >>>> *the MOSFET output* is about as fast as you'll find. My only >>>> complaint is that the FETs are too big, too much capacitance. >>>> When switching a 50-ohm cable-matching output resistor, a low >>>> 70-milli-ohm Ron is serious overkill. >>> >>> Hey Win, here's a smaller-sized version of the same concept, >>> driver-plus-GaN on a chip: >>> <https://media.digikey.com/pdf/Data%20Sheets/Navitas%20Semi%20PDFs/NV6113%20Datasheet%20(FINAL)%208-28-18.pdf> >> >> Wow, Navitas NV6113, GaN, 200V/ns, only $3.38 at Digi-Key. And >> Octopart doesn't even know about them yet! It looks hard to get >> heat out of the package, they say limited to 2MHz switching rate. > >50C/W - ugly. That would practically limit you to about 2 amps average, >at 300mOhm. That's just not... special.
It might make sense thermally to parallel a few smaller parts, EPC GaNs or these things. -- John Larkin Highland Technology, Inc lunatic fringe electronics
Reply by January 13, 20192019-01-13
On Sunday, January 13, 2019 at 6:13:19 PM UTC-5, Clifford Heath wrote:
> On 13/1/19 6:55 am, Winfield Hill wrote: > > dagmargoodboat@yahoo.com wrote... > >> > >> Winfield Hill wrote: > >>> > >>> There's very little in the signal path. And the 20ns delay to > >>> *the MOSFET output* is about as fast as you'll find. My only > >>> complaint is that the FETs are too big, too much capacitance. > >>> When switching a 50-ohm cable-matching output resistor, a low > >>> 70-milli-ohm Ron is serious overkill. > >> > >> Hey Win, here's a smaller-sized version of the same concept, > >> driver-plus-GaN on a chip: > >> <https://media.digikey.com/pdf/Data%20Sheets/Navitas%20Semi%20PDFs/NV6113%20Datasheet%20(FINAL)%208-28-18.pdf> > > > > Wow, Navitas NV6113, GaN, 200V/ns, only $3.38 at Digi-Key. And > > Octopart doesn't even know about them yet! It looks hard to get > > heat out of the package, they say limited to 2MHz switching rate. > > 50C/W - ugly. That would practically limit you to about 2 amps average, > at 300mOhm. That's just not... special.
Au contraire! Two amps is way more than I need, and I much appreciate the reduced capacitances. (Not all of us are trying to drive big metal all the time, or launch EMI out into space. :) There are lots of big wide-gap devices for big-power stuff. But we don't have many choices yet when it comes to making a dinky lil' signal generator. Cheers, James Arthur
Reply by Clifford Heath January 13, 20192019-01-13
On 13/1/19 6:55 am, Winfield Hill wrote:
> dagmargoodboat@yahoo.com wrote... >> >> Winfield Hill wrote: >>> >>> There's very little in the signal path. And the 20ns delay to >>> *the MOSFET output* is about as fast as you'll find. My only >>> complaint is that the FETs are too big, too much capacitance. >>> When switching a 50-ohm cable-matching output resistor, a low >>> 70-milli-ohm Ron is serious overkill. >> >> Hey Win, here's a smaller-sized version of the same concept, >> driver-plus-GaN on a chip: >> <https://media.digikey.com/pdf/Data%20Sheets/Navitas%20Semi%20PDFs/NV6113%20Datasheet%20(FINAL)%208-28-18.pdf> > > Wow, Navitas NV6113, GaN, 200V/ns, only $3.38 at Digi-Key. And > Octopart doesn't even know about them yet! It looks hard to get > heat out of the package, they say limited to 2MHz switching rate.
50C/W - ugly. That would practically limit you to about 2 amps average, at 300mOhm. That's just not... special.