Shoutout to PIN diodes, I found out about those only a year ago and these are now my favorite type of diode. If you build RF stuff I recommend getting a selection of PIN diodes so you can build any switch you need! They switch high RF power with only a few mA of bias current. Virtually no effect on IP3
That page 9 of the datasheet could be pretty interesting to some hobbyists. Looks like they specced it as 100MHz - 6GHz because that's where the market is, but it will function just fine down to 2MHz, or lower if you're not picky. The power handling and distortion characteristics get worse below 100MHz, but the insertion loss and isolation only get better (no surprise), so if you're switching tens of milliwatts at HF or low VHF it might be handy.
The serif font on the board tells me this is designed in china. These boards can be really hit and miss. I tested a china HMC252 board. The RF traces where wider as I expected and measurements showed it was only good for like 300MHz where the chip was good for 3 GHz. Later that day I fried that board + the NANO VNA V2, because the VNA was connected via USB and the wall wart SMPS for the board had a high coupling into 230V mains. At some point while reconnecting the SMA cables, center was connected without ground. Sh*t happens.
The part looks great! I WANT one of those boards, but dunno why. How far do you run boards from the foundries to? I want to build a 10 GHz prescaler. ALSO: Where do you get those arthritic assist SMA rubber nut grips? Me need too! 73 de W3IHM
A lot of those high frequency boards are using basic FR4 so less than perfect berhaviour is expected really . Also of course these may well be not top bin parts to start with . I have these or similar switches they do work well at the frequencies I care about (2.4GHZ)!
It might be interesting as well to up to which frequency the channels can be switched. I did not find it in the data sheet. When the common output is switched between signal input and ground input a AM pattern will show up that will vanish with rising switching frequency.
@@Manf-ft6zk If you use this switch as a modulator (i.e. connecting a UART to A or B) the 150ns (On Off Time) may give you an upper limit of 6.67Mbaud. For OOSK you just switch the carrier on and off. For ASK switch between different levels of the carrier (power divider -> different attenuators -> switch). For FSK you toggle between different carriers. For PSK switch between different phases of the carrier (power divider -> cables with differences in length (fractions of lambda) -> switch).
I'd buy 1 if it included lower freq coverage as well. I prefer my devices have as much versatility as possible. I have some SPDT mechanical RF relays linked together that cover a wide freq range, but they are large and awkward compared to this.
Take a look at the datasheet, particularly pages 6 and 9. It *does* cover lower frequencies, if you derate the max power a bit. P1dB is down to around 100mW at 1MHz, but otherwise it's fine - and in fact has lower insertion loss and higher isolation than it does at the higher frequencies it's sold for. (And if you noticed the screwed-up X axes on the page 9 graphs... yeah, so did I. That makes things a little dubious. But the values at 1MHz align with a claim on page 1, and the values at 100MHz pretty much track with the bottom end of the 0.1-6GHz graphs, and the general trend is pretty much a straight line on the log-log graph, so you can draw your own conclusions.)
It is unusual that the datasheet only gives a functional diagram. These are PIN diodes, not physical switches. This leads to a minimum frequency. I don't understand how a related chip, HMC241AQS16, can function down to DC. I thought the limitation of the diode would require a minimum switching frequency.
Ah yes, the HMC241AQS16 is GaAs mmic, while the 7992 uses Si mmic. I have never seen mmic. It seems(?) pin diodes are not in ic designs. Perhaps mmic is better since it does not have the minimum effective frequency limitation of pin diodes?
Pretty neat. But I am going to rant about my favourite pet peeve. Everything in dBm and dB. But fortunately we can just put 33 dbm into google and it tells us 2.00 Watts :-)
@@WolfmanDude Hmm.. I'm too used to thinking of power in terms of watts and all the formulas like P = V x I. dBs might be useful for those who are used to it, but that is not me :-)
A nano vna with 10 inch screen and buttons and knobs is known as an HP/Agilent 8753 - preferably the C D E or ES varieties - the A is getting pretty long in the tooth these days
Part of the limitation is likely board substrate and layout. Impedance matters at these frequencies.
Shoutout to PIN diodes, I found out about those only a year ago and these are now my favorite type of diode. If you build RF stuff I recommend getting a selection of PIN diodes so you can build any switch you need! They switch high RF power with only a few mA of bias current. Virtually no effect on IP3
I'm just glad I understood like ~>50% of this content. Good job keeping it extra nerdy and yet approachable.
5:17 I like making one of the cables longer by an inch so the open probes cannot short accidentally
That page 9 of the datasheet could be pretty interesting to some hobbyists. Looks like they specced it as 100MHz - 6GHz because that's where the market is, but it will function just fine down to 2MHz, or lower if you're not picky. The power handling and distortion characteristics get worse below 100MHz, but the insertion loss and isolation only get better (no surprise), so if you're switching tens of milliwatts at HF or low VHF it might be handy.
I saw the spec and was a little disappointed. But now I maybe buy one of these anyway
The serif font on the board tells me this is designed in china. These boards can be really hit and miss. I tested a china HMC252 board. The RF traces where wider as I expected and measurements showed it was only good for like 300MHz where the chip was good for 3 GHz. Later that day I fried that board + the NANO VNA V2, because the VNA was connected via USB and the wall wart SMPS for the board had a high coupling into 230V mains. At some point while reconnecting the SMA cables, center was connected without ground. Sh*t happens.
sup Pop's!! it's non-reflective, with silly-con!! love it..
Lol, when I saw your Siglent Spectrum Analyzer, was wondering if you have unlocked it to go to 3.2GHz, but you gave the answer.
The part looks great! I WANT one of those boards, but dunno why. How far do you run boards from the foundries to? I want to build a 10 GHz prescaler. ALSO: Where do you get those arthritic assist SMA rubber nut grips? Me need too! 73 de W3IHM
www.pcbway.com/project/shareproject/SMA_Thumbwheel_Wrench_b3dc35b6.html
Yeah, I wouldn't rely on those tin-plated traces between the chip and sockets when it comes to impedance control.
Might have gotten a better response if you loaded the unused ports. Also how come you never use your big vna?
A lot of those high frequency boards are using basic FR4 so less than perfect berhaviour is expected really . Also of course these may well be not top bin parts to start with . I have these or similar switches they do work well at the frequencies I care about (2.4GHZ)!
hello, does this suitable for wifi router? i mean what does wifi are dc ir ac or what cause i don't understand what does that mean, sorry
It might be interesting as well to up to which frequency the channels can be switched. I did not find it in the data sheet. When the common output is switched between signal input and ground input a AM pattern will show up that will vanish with rising switching frequency.
It is under switching speed, somewhat 30ns to 320ns
@@TheTrueCBaerThank you, I did not see it. That will be just fast enough to see some modulation, not much more, that might be why it is a switch. 😀
@@Manf-ft6zk If you use this switch as a modulator (i.e. connecting a UART to A or B) the 150ns (On Off Time) may give you an upper limit of 6.67Mbaud.
For OOSK you just switch the carrier on and off.
For ASK switch between different levels of the carrier (power divider -> different attenuators -> switch).
For FSK you toggle between different carriers.
For PSK switch between different phases of the carrier (power divider -> cables with differences in length (fractions of lambda) -> switch).
@@TheTrueCBaerYes, thank you.
I'd buy 1 if it included lower freq coverage as well. I prefer my devices have as much versatility as possible. I have some SPDT mechanical RF relays linked together that cover a wide freq range, but they are large and awkward compared to this.
Take a look at the datasheet, particularly pages 6 and 9. It *does* cover lower frequencies, if you derate the max power a bit. P1dB is down to around 100mW at 1MHz, but otherwise it's fine - and in fact has lower insertion loss and higher isolation than it does at the higher frequencies it's sold for.
(And if you noticed the screwed-up X axes on the page 9 graphs... yeah, so did I. That makes things a little dubious. But the values at 1MHz align with a claim on page 1, and the values at 100MHz pretty much track with the bottom end of the 0.1-6GHz graphs, and the general trend is pretty much a straight line on the log-log graph, so you can draw your own conclusions.)
make more of these videos and also sp8t ones
It is unusual that the datasheet only gives a functional diagram. These are PIN diodes, not physical switches. This leads to a minimum frequency. I don't understand how a related chip, HMC241AQS16, can function down to DC. I thought the limitation of the diode would require a minimum switching frequency.
what makes you so sure they are PIN. how about FETs
Some of these chips internally generate a negative voltage rail.
Ah yes, the HMC241AQS16 is GaAs mmic, while the 7992 uses Si mmic. I have never seen mmic.
It seems(?) pin diodes are not in ic designs. Perhaps mmic is better since it does not have the minimum effective frequency limitation of pin diodes?
Or at least a chinesium simulacrum of one... I have learned the hard way to put very little trust in AliX components!
Pretty neat. But I am going to rant about my favourite pet peeve. Everything in dBm and dB. But fortunately we can just put 33 dbm into google and it tells us 2.00 Watts :-)
DONT convert to watts, calculate everything in dB and dBm. Its super useful if you get it.
@@WolfmanDude Hmm.. I'm too used to thinking of power in terms of watts and all the formulas like P = V x I. dBs might be useful for those who are used to it, but that is not me :-)
RF parts are SO weird.
I want a NanoVNA with a 10 inch screen and buttons and knobs. Touchscreens suck when you have shaky old hands.
A nano vna with 10 inch screen and buttons and knobs is known as an HP/Agilent 8753 - preferably the C D E or ES varieties - the A is getting pretty long in the tooth these days