Reply by John Larkin February 10, 20152015-02-10
On Mon, 9 Feb 2015 17:43:08 -0800 (PST), Simon S Aysdie
<gwhite@ti.com> wrote:

>On Saturday, February 7, 2015 at 9:22:38 AM UTC-8, John Larkin wrote: >> Who makes decent SPI programmable attenuators? I need DC to, say, 300 >> MHz or better. > >Peregrine. For example, the PE43703. Power handling goes down as the frequency gets low. The negative charge pump can be bypassed, if you don't want the low level noise they make. The 50 mdB settling is much faster than GaAs because GaAs has "gate lag." > >Also, they have announced the PE42020, UltraCMOS&#2013266094; True DC RF Switch, which goes to DC with better power handling than anything preceding it. It is a matter of time until we see that in DSAs.
Thanks. I'll check Peregrine. -- John Larkin Highland Technology, Inc picosecond timing laser drivers and controllers jlarkin att highlandtechnology dott com http://www.highlandtechnology.com
Reply by Simon S Aysdie February 9, 20152015-02-09
On Saturday, February 7, 2015 at 9:22:38 AM UTC-8, John Larkin wrote:
> Who makes decent SPI programmable attenuators? I need DC to, say, 300 > MHz or better.
Peregrine. For example, the PE43703. Power handling goes down as the frequency gets low. The negative charge pump can be bypassed, if you don't want the low level noise they make. The 50 mdB settling is much faster than GaAs because GaAs has "gate lag." Also, they have announced the PE42020, UltraCMOS&#2013266094; True DC RF Switch, which goes to DC with better power handling than anything preceding it. It is a matter of time until we see that in DSAs.
Reply by John Miles, KE5FX February 7, 20152015-02-07
On Saturday, February 7, 2015 at 12:32:45 PM UTC-8, John Larkin wrote:
> Good grief, does no one at MiniCircuits proofread their data sheets? > It's a mess.
As I understand it, their main business is packaging other peoples' dies. So they may just post whatever specs they get from the OEM. Hardly anybody in the MMIC business publishes decent data sheets.
> The requirement that the inputs be AC coupled, or held at ground, and > that the max input limit is -0.3 volts, is interesting. That conflicts > with the 24 dBm rating. Typical RF specmanship.
I wonder if there's a charge pump in there that they're not mentioning? It's supposed to be CMOS, not GaAs, and I've never caught it making racket. One reason I like the DAT-15R5 is that it doesn't contribute any additive AM or PM noise beyond its own loss. Now that I'm looking at it, the DAT-15R5-SP+ is NRND and the data sheet for their recommended replacement part has a rather interesting new graph. I used the old part in a box rated for +20 dBm max input from 0.5-30 MHz, and figure 1 for the new DAT-15R5A-SP+ says that it should more like +10 dBm near 1 MHz. That's much more specific (and disturbing) than the original DAT-15R5 data sheet, which just says "Input IP3 and P1dB degrades below 1 MHz." So caveat emptor, I guess. -- john, KE5FX
Reply by February 7, 20152015-02-07
Try Hittite, now owned by Analog Devices
Mark
Reply by Lasse Langwadt Christensen February 7, 20152015-02-07
Den l&#2013266168;rdag den 7. februar 2015 kl. 21.29.37 UTC+1 skrev Phil Hobbs:
> On 2/7/2015 3:20 PM, Lasse Langwadt Christensen wrote: > > Den l&#2013266168;rdag den 7. februar 2015 kl. 21.11.08 UTC+1 skrev Phil Hobbs: > >> On 2/7/2015 3:00 PM, John Larkin wrote: > >>> On Sat, 07 Feb 2015 14:34:35 -0500, Phil Hobbs > >>> <hobbs@electrooptical.net> wrote: > >>> > >>>> On 2/7/2015 12:28 PM, John Larkin wrote: > >>>>> On Sat, 07 Feb 2015 09:22:37 -0800, John Larkin > >>>>> <jlarkin@highlandtechnology.com> wrote: > >>>>> > >>>>>> Who makes decent SPI programmable attenuators? I need DC to, say, 300 > >>>>>> MHz or better. > >>>>> > >>>>> Well, I did find this... > >>>>> > >>>>> http://www.triquint.com/products/d/DOC-A-00000718 > >>>>> > >>>>> It's sure easy to program. > >>>>> > >>>>> > >>>> ;) > >>>> > >>>> These RF bods can't be expected to use zero-ohm resistors like the rest > >>>> of us. > >>> > >>> Nor can they be bothered to reveal the DC behavior of their parts. > >>> > >>>> > >>>> I've recently started putting zero-ohm resistors between linear > >>>> regulators and their output reservoir caps. One recent test board had > >>>> all the inverse-parallel diodes in backwards, leading to putting +15.5 V > >>>> on the MCU's VDD pin. Too bad, so sad. > >>> > >>> I never put reverse diodes across voltage regulators, and haven't > >>> suffered. > >>> > >>> I go try to remember to put 0R or milliohm resistors in series with > >>> regulators, so we can measure the currents. Sometimes it's good to > >>> know that stuff, like FPGA core power, for future reference. > >>> > >>> > >>>> > >>>> The other good thing about that is that if the regulator has any > >>>> stability issues (e.g. an LM1117 with a big polymer aluminum output cap) > >>>> I can just put N milliohms in series and fix it. > >>> > >>> I generally use tantalums on an 1117 output, voltage derated 2:1 or > >>> so. That seems to be very stable. We avoid wet aluminum lytics, > >>> because the ESR skyrockets below 0C. > >> > >> Yup. I've been using AlPo caps there lately, because it helps prevent > >> low-frequency crosstalk between different things hung off the same > >> supply rail. > >> > >>> > >>>> > >>>> In hand-wired protos, I always wire and test the supplies first, then > >>>> finish up. > >>> > >>> Yeah, my protos sometimes include regulators, so that I don't need six > >>> bench supplies. > >> > >> My standard is a 16-19V laptop power brick, with LM78XXes and LM2594s > >> applied as necessary. > >> > > > > for protos and one-ofs i sue something like this: > > > > http://www.ebay.co.uk/itm/281474876609 > > > > hard to beat the price > > That's because they left out the inductor. ;) > > I've used similar things in the past. Most of the Chinese ones work > okay if you derate them by a factor of 2 or 3 in current. (They always > cheap out on the inductor like that.) >
yeh I wouldn't bet on 3A. but you don't need much at 1MHz, a similar sized coilcraft of that value is rated for 5A The markings on the diode on the ones I have leads to a datasheet that says 3A average The IC is an MP1584, rated at 4A So it doesn't look totally hopeless -Lasse
Reply by John Larkin February 7, 20152015-02-07
On Sat, 7 Feb 2015 10:27:19 -0800 (PST), "John Miles, KE5FX"
<jmiles@gmail.com> wrote:

>On Saturday, February 7, 2015 at 9:22:38 AM UTC-8, John Larkin wrote: >> Who makes decent SPI programmable attenuators? I need DC to, say, 300 >> MHz or better. > >My jellybean SPI attenuator of choice is the Mini-Circuits DAT-15R5-SP+. As usual, it's specified to work down to DC as long as your definition of DC is 0V. :( Also as usual, the data sheet tells you nothing about how it actually works.
Good grief, does no one at MiniCircuits proofread their data sheets? It's a mess. The requirement that the inputs be AC coupled, or held at ground, and that the max input limit is -0.3 volts, is interesting. That conflicts with the 24 dBm rating. Typical RF specmanship.
> >It'd be nice if someone built a digital pot like the AD5260 with response into the VHF range. I'm not sure why those have such a low 3 dB bandwidth. > >-- john, KE5FX
I think most of those are CMOS string dac architectures, nice and linear and monotonic but fundamentally slow. An n-bit unit has 2^n resistors and switches. -- John Larkin Highland Technology, Inc picosecond timing laser drivers and controllers jlarkin att highlandtechnology dott com http://www.highlandtechnology.com
Reply by Phil Hobbs February 7, 20152015-02-07
On 2/7/2015 3:20 PM, Lasse Langwadt Christensen wrote:
> Den l&#2013266168;rdag den 7. februar 2015 kl. 21.11.08 UTC+1 skrev Phil Hobbs: >> On 2/7/2015 3:00 PM, John Larkin wrote: >>> On Sat, 07 Feb 2015 14:34:35 -0500, Phil Hobbs >>> <hobbs@electrooptical.net> wrote: >>> >>>> On 2/7/2015 12:28 PM, John Larkin wrote: >>>>> On Sat, 07 Feb 2015 09:22:37 -0800, John Larkin >>>>> <jlarkin@highlandtechnology.com> wrote: >>>>> >>>>>> Who makes decent SPI programmable attenuators? I need DC to, say, 300 >>>>>> MHz or better. >>>>> >>>>> Well, I did find this... >>>>> >>>>> http://www.triquint.com/products/d/DOC-A-00000718 >>>>> >>>>> It's sure easy to program. >>>>> >>>>> >>>> ;) >>>> >>>> These RF bods can't be expected to use zero-ohm resistors like the rest >>>> of us. >>> >>> Nor can they be bothered to reveal the DC behavior of their parts. >>> >>>> >>>> I've recently started putting zero-ohm resistors between linear >>>> regulators and their output reservoir caps. One recent test board had >>>> all the inverse-parallel diodes in backwards, leading to putting +15.5 V >>>> on the MCU's VDD pin. Too bad, so sad. >>> >>> I never put reverse diodes across voltage regulators, and haven't >>> suffered. >>> >>> I go try to remember to put 0R or milliohm resistors in series with >>> regulators, so we can measure the currents. Sometimes it's good to >>> know that stuff, like FPGA core power, for future reference. >>> >>> >>>> >>>> The other good thing about that is that if the regulator has any >>>> stability issues (e.g. an LM1117 with a big polymer aluminum output cap) >>>> I can just put N milliohms in series and fix it. >>> >>> I generally use tantalums on an 1117 output, voltage derated 2:1 or >>> so. That seems to be very stable. We avoid wet aluminum lytics, >>> because the ESR skyrockets below 0C. >> >> Yup. I've been using AlPo caps there lately, because it helps prevent >> low-frequency crosstalk between different things hung off the same >> supply rail. >> >>> >>>> >>>> In hand-wired protos, I always wire and test the supplies first, then >>>> finish up. >>> >>> Yeah, my protos sometimes include regulators, so that I don't need six >>> bench supplies. >> >> My standard is a 16-19V laptop power brick, with LM78XXes and LM2594s >> applied as necessary. >> > > for protos and one-ofs i sue something like this: > > http://www.ebay.co.uk/itm/281474876609 > > hard to beat the price
That's because they left out the inductor. ;) I've used similar things in the past. Most of the Chinese ones work okay if you derate them by a factor of 2 or 3 in current. (They always cheap out on the inductor like that.) My stuff usually has a lot of sensitive analogue sections, so I use a lot of toroids, especially in protos. Cheers Phil Hobbs -- Dr Philip C D Hobbs Principal Consultant ElectroOptical Innovations LLC Optics, Electro-optics, Photonics, Analog Electronics 160 North State Road #203 Briarcliff Manor NY 10510 hobbs at electrooptical dot net http://electrooptical.net
Reply by Lasse Langwadt Christensen February 7, 20152015-02-07
Den l&#2013266168;rdag den 7. februar 2015 kl. 21.11.08 UTC+1 skrev Phil Hobbs:
> On 2/7/2015 3:00 PM, John Larkin wrote: > > On Sat, 07 Feb 2015 14:34:35 -0500, Phil Hobbs > > <hobbs@electrooptical.net> wrote: > > > >> On 2/7/2015 12:28 PM, John Larkin wrote: > >>> On Sat, 07 Feb 2015 09:22:37 -0800, John Larkin > >>> <jlarkin@highlandtechnology.com> wrote: > >>> > >>>> Who makes decent SPI programmable attenuators? I need DC to, say, 300 > >>>> MHz or better. > >>> > >>> Well, I did find this... > >>> > >>> http://www.triquint.com/products/d/DOC-A-00000718 > >>> > >>> It's sure easy to program. > >>> > >>> > >> ;) > >> > >> These RF bods can't be expected to use zero-ohm resistors like the rest > >> of us. > > > > Nor can they be bothered to reveal the DC behavior of their parts. > > > >> > >> I've recently started putting zero-ohm resistors between linear > >> regulators and their output reservoir caps. One recent test board had > >> all the inverse-parallel diodes in backwards, leading to putting +15.5 V > >> on the MCU's VDD pin. Too bad, so sad. > > > > I never put reverse diodes across voltage regulators, and haven't > > suffered. > > > > I go try to remember to put 0R or milliohm resistors in series with > > regulators, so we can measure the currents. Sometimes it's good to > > know that stuff, like FPGA core power, for future reference. > > > > > >> > >> The other good thing about that is that if the regulator has any > >> stability issues (e.g. an LM1117 with a big polymer aluminum output cap) > >> I can just put N milliohms in series and fix it. > > > > I generally use tantalums on an 1117 output, voltage derated 2:1 or > > so. That seems to be very stable. We avoid wet aluminum lytics, > > because the ESR skyrockets below 0C. > > Yup. I've been using AlPo caps there lately, because it helps prevent > low-frequency crosstalk between different things hung off the same > supply rail. > > > > >> > >> In hand-wired protos, I always wire and test the supplies first, then > >> finish up. > > > > Yeah, my protos sometimes include regulators, so that I don't need six > > bench supplies. > > My standard is a 16-19V laptop power brick, with LM78XXes and LM2594s > applied as necessary. >
for protos and one-ofs i sue something like this: http://www.ebay.co.uk/itm/281474876609 hard to beat the price -Lasse
Reply by Phil Hobbs February 7, 20152015-02-07
On 2/7/2015 3:00 PM, John Larkin wrote:
> On Sat, 07 Feb 2015 14:34:35 -0500, Phil Hobbs > <hobbs@electrooptical.net> wrote: > >> On 2/7/2015 12:28 PM, John Larkin wrote: >>> On Sat, 07 Feb 2015 09:22:37 -0800, John Larkin >>> <jlarkin@highlandtechnology.com> wrote: >>> >>>> Who makes decent SPI programmable attenuators? I need DC to, say, 300 >>>> MHz or better. >>> >>> Well, I did find this... >>> >>> http://www.triquint.com/products/d/DOC-A-00000718 >>> >>> It's sure easy to program. >>> >>> >> ;) >> >> These RF bods can't be expected to use zero-ohm resistors like the rest >> of us. > > Nor can they be bothered to reveal the DC behavior of their parts. > >> >> I've recently started putting zero-ohm resistors between linear >> regulators and their output reservoir caps. One recent test board had >> all the inverse-parallel diodes in backwards, leading to putting +15.5 V >> on the MCU's VDD pin. Too bad, so sad. > > I never put reverse diodes across voltage regulators, and haven't > suffered. > > I go try to remember to put 0R or milliohm resistors in series with > regulators, so we can measure the currents. Sometimes it's good to > know that stuff, like FPGA core power, for future reference. > > >> >> The other good thing about that is that if the regulator has any >> stability issues (e.g. an LM1117 with a big polymer aluminum output cap) >> I can just put N milliohms in series and fix it. > > I generally use tantalums on an 1117 output, voltage derated 2:1 or > so. That seems to be very stable. We avoid wet aluminum lytics, > because the ESR skyrockets below 0C.
Yup. I've been using AlPo caps there lately, because it helps prevent low-frequency crosstalk between different things hung off the same supply rail.
> >> >> In hand-wired protos, I always wire and test the supplies first, then >> finish up. > > Yeah, my protos sometimes include regulators, so that I don't need six > bench supplies.
My standard is a 16-19V laptop power brick, with LM78XXes and LM2594s applied as necessary. Cheers Phil Hobbs -- Dr Philip C D Hobbs Principal Consultant ElectroOptical Innovations LLC Optics, Electro-optics, Photonics, Analog Electronics 160 North State Road #203 Briarcliff Manor NY 10510 hobbs at electrooptical dot net http://electrooptical.net
Reply by John Larkin February 7, 20152015-02-07
On Sat, 07 Feb 2015 14:34:35 -0500, Phil Hobbs
<hobbs@electrooptical.net> wrote:

>On 2/7/2015 12:28 PM, John Larkin wrote: >> On Sat, 07 Feb 2015 09:22:37 -0800, John Larkin >> <jlarkin@highlandtechnology.com> wrote: >> >>> Who makes decent SPI programmable attenuators? I need DC to, say, 300 >>> MHz or better. >> >> Well, I did find this... >> >> http://www.triquint.com/products/d/DOC-A-00000718 >> >> It's sure easy to program. >> >> >;) > >These RF bods can't be expected to use zero-ohm resistors like the rest >of us.
Nor can they be bothered to reveal the DC behavior of their parts.
> >I've recently started putting zero-ohm resistors between linear >regulators and their output reservoir caps. One recent test board had >all the inverse-parallel diodes in backwards, leading to putting +15.5 V >on the MCU's VDD pin. Too bad, so sad.
I never put reverse diodes across voltage regulators, and haven't suffered. I go try to remember to put 0R or milliohm resistors in series with regulators, so we can measure the currents. Sometimes it's good to know that stuff, like FPGA core power, for future reference.
> >The other good thing about that is that if the regulator has any >stability issues (e.g. an LM1117 with a big polymer aluminum output cap) >I can just put N milliohms in series and fix it.
I generally use tantalums on an 1117 output, voltage derated 2:1 or so. That seems to be very stable. We avoid wet aluminum lytics, because the ESR skyrockets below 0C.
> >In hand-wired protos, I always wire and test the supplies first, then >finish up.
Yeah, my protos sometimes include regulators, so that I don't need six bench supplies. -- John Larkin Highland Technology, Inc picosecond timing laser drivers and controllers jlarkin att highlandtechnology dott com http://www.highlandtechnology.com