There was an earlier thread talking about front-end damage from nearby transmit antennas. Someone from SDR Play replied the SDR1 can briefly tolerate up to 10mW and tolerate 1mW indefinitely. I've measured over 25 mW into 50 ohms at the coax leading to my outdoor discone when I transmit on 2M or 70CM from a different antenna about 10 feet away which I guessing could be fatal and I can't easily get any more separation between the antennas.
For now I've been unscrewing the SMA from the RSP2 when transmitting but it's a hassle and sooner or later I'm going to forget and likely blow out the front end of my RSP2 as others have reported doing. I know there are protection devices that go in the antenna feed but most only are rated up to 150 Mhz or so and the ones using relays I don't trust as I doubt they can switch fast enough by the time they detect excessive RF. I could easily end up spending more on protection good to 1 GHZ than I paid for the RSP2.
Finally there are TVS ultra low capacitance protection diodes designed to be used at 2 Ghz and beyond that should clamp the voltage at the antenna terminal to about 5 volts peak which is about 1.7 volts rms or 3.5 mW into 50 ohms. Can anyone comment if such a diode would provide adequate protection for the RSP2 (assuming there's not so much power as to damage the diode itself) and no other side effects? Such a diode would also help prevent damage from nearby lightning and ESD. Or perhaps there's already such a diode in the RSP2? If not perhaps there should be? Here's an example and they're about $0.10 each in quantity:
https://www.digikey.com/product-detail/ ... ND/6576108
Thanks for any help!
Powering down the device or disconnecting it from the PC will NOT provide additional protection in the scenario where excessive RF power is applied.
All RSPs contain protection diodes. The first failure mechanism when excess RF power is applied is the destruction of the ESD protection diodes themselves at the RF input. These diodes DO clamp the input, but there is still a finite amount of power density that they can handle. Such small devices cannot handle watts as they are in fact very small PN junctions. These effectively become fused as a result of thermal dissipation. If even greater levels of power are still applied, excess heating leads to the destruction of the input amplifier and potentially the RF tuner and RF switching devices. It is impossible to say at what power levels these various failures occur.
The levels of power that we state can be applied safely are numbers that we are confident in. We regularly get asked what the ultimate limit is, but the problem is that the only way to determine this is to carry out a series of comprehensive destructive tests across different frequencies, manufacturing SKUs and temperature. As you might appreciate this would likely require the destruction of several hundred devices and take a very large amount of time and even then it is impossible to provide an absolute guarantee. We have applied much higher levels of power to devices (up to +25 dBm for short durations in our labs) and it has not destroyed the device. Does that mean that we can guarantee + 25 dBm? No it does not. We might try it tomorrow and find that a device fails. 10 dBm (10 mW) is chosen because at that point the diodes are clamping, but the short term heating effects are unlikely to cause damage. Similarly 1mW is certainly safe because the diodes will not turn on.
The protection devices integrated into the device are a compromise between providing good ESD robustness and also providing good RF performance at up to 2 GHz. As you rightly have observed, sometimes people whose only interest is HF fail to appreciate how devices that are effective at HF have catastrophic effects on the performance at L-Band or even UHF, which is why we simply cannot integrate such devices into the unit.
Reason: No reason
The math gets kind of complicated once the diodes start clamping as that, presumably, drops the input impedance well below 50 ohms. In my case my discone is resonant and has a low SWR at 50 ohms on both 2M and 70cm but, connected to a lower impedance mis-matched load (the RSP2 clamping), it presumably will be less efficient and deliver less power. So it's a rather different scenario than a constant 50 ohm signal source hard wired to the RSP2 on a lab bench. If the protection diode failing is the main issue I can also see how it doesn't matter if the device is un-powered or the input is de-selected.
Thanks for the great answer! I'll be doing my best to send more business your way.
Now I'm planning to add a HF dipole to RSP2, about 10 meters in length. I'm thinking of using a BAR14-1 PIN diode pair for protection against the same TX signal. Possible comments welcome.
Hey, is this a potential product idea for somebody: A single box providing robust RF protection for all three antenna inputs of RSP2?
Reason: No reason
@OH2BUA The capacitance of 1N4148 diodes is 4pf. A pair of them in parallel would be 8pf plus the capacitance of the leads/wiring which could easily exceed 10pf total. This will significantly load the 50 ohm input of an RSP at higher frequencies. For example 10pf at these frequencies:
440 Mhz - 36 ohms (21 ohms antenna load when in parallel with the RSP's 50 ohm input)
850 Mhz - 19 ohms (14 ohm antenna load)
2 Ghz - 8 ohms
Even at 440 Mhz the diode capacitance will "steal" over half the RF signal and attenuate the signal by around 4 dB. At higher frequencies it becomes much worse.
If you are only using your RSP below 50 Mhz 1N4148 diodes would work fairly well on the 50 ohm input and would only load the input with an extra 318 ohms at 50 Mhz dropping the input impedance to around 43 ohms. But if you're using the HiZ input as you say it's hard to say what will happen. I suspect the diode capacitance will eat up most of your signal and you're lowering the input impedance into the low hundreds of ohms.
You also need to make sure you don't turn on the Bias-T option if using 1N4148 diodes on the 50 ohm input. The DC voltage would put the diodes into conduction essentially shorting out the signal and may also excessively load the RSP.
For the engineering geeks out there I'm simplifying this and ignoring the complex impedance but the end result will be similar. Using a pair of 1N4148 diodes will significantly degrade the performance of any receiver, including the RSPs, at frequencies above around 50 Mhz and seriously degrade the performance above around 800 Mhz.
The above is why this is such a challenge. To keep the protection diode capacitance low enough at high frequencies the diodes have to be extremely tiny. But extremely tiny diodes can't dissipate much sustained energy before failing. There are more involved methods involving using an RF transformer core that saturates when overloaded but it's difficult to design an RF transformer that doesn't significantly degrade performance over the full bandwidth of the RSPs.
You can isolate the input with a relay during transmit if you have a transceiver with a "PTT" or "TX" output signal intended to key up an external power amp but only if there's enough of a delay from the signal to the start of transmission to safely close a relay. But you still have the same issue with the added capacitance and inductance of the relay seriously degrading performance at higher frequencies.
I know some put an inexpensive "sacrificial" preamp ahead of their RSP with the hope if they forget to disconnect the receive antenna while transmitting it will only destroy the preamp and not their RSP. But a preamp, and especially a cheap one, will degrade the noise figure of the RSP and may create overload issues from nearby broadcast, mobile phone, pager signals, etc.
I'll keep looking into this. I plan to put my oscilloscope on the RSP2 input and gradually increase my transmit power. I should be able to see at what point the RSP2 protection diodes start "clipping" the signal. That gives me something of a baseline for when the RSP is at risk. If nothing else I may be able to keep my transmit power to a safe level for the RSP and only increase it above that level when needed and disconnect the RSP at those times. Another option might be an external sacrificial ultra low capacitance TVS diode that will start clamping hopefully before the internal RSP diode(s) do. With any luck that might save the RSP if I forget to disconnect it and just require replacing a cheap external diode.
As @Tech_Support said, this isn't an easy problem to solve and involves trade-offs no matter what you do. It sounds like SDRPlay has done a good job of balancing performance at higher frequencies, trying to protect their RSPs, and keeping the price down. There's a reason RF spectrum analyzers with bandwidth into the Ghz are so expensive and even they have somewhat fragile front-ends (typically specified up to 20 DBm or so) and even the cheapest models are around ten times the price of the RSP.
that mitigate the destruction of the SDR device?
I wonder, because, just shorting the input (using BNC connectors and a BNC tee) I can still detect and
Reason: No reason
I ended up buying a Stridsberg Engineering MCA204M multicoupler to share my wideband discone antenna between my SDRplay and muliple other receivers and scanners. The MCA204M has a bandwidth of 25 Mhz to 1 Ghz which corresponds well to my discone antenna. And it has it's own low noise figure preamp at the input, with protection, to overcome the losses in the output port splitter and isolation section. It's rated for +24 dbM maximum input which is impressive.
I then added a tiny relatively inexpensive HF3 series RF relay to short the antenna input to the multicoupler during transmit and power off. There's room inside the multicoupler enclosure for the relay near the antenna input jack. This webpage ( https://www.w6pql.com/using_inexpensive_relays.htm ), and various data sheets, helped confirm the right relay should not cause significant degradation. The end result is a slightly inferior signal to noise ratio around 2 dB worse on very weak signals in the UHF range. The actual sensitivity is slightly better compared to connecting the antenna directly to my RSP2 as the multicoupler has a small amount of net gain but there is still the slight noise penalty. You rarely get anything for free but 98% of the time you can't tell the MCA204M and relay are there.
The only other downside is the MCA204M costs more than the RSP2 (but less than a Duo). So this isn't a cheap solution but it does let me share a single antenna among multiple receivers and, with my modification, provides transmit protection. The relay also defaults to shorted when no power is applied which provides some nearby lightning/static protection when everything is turned off. It is impressive the RSP2 has a slightly better RF front end than a device that's more expensive and is only a signal splitter with a single transistor pre-amp. Stridsberg Engineering could probably learn from SDRplay but the real world performance is generally so close as not to matter.
I should also add my solution doesn't help those using the Hi-Z antenna input for coverage below 25 Mhz. Although, in theory, you could use the same HF3 series RF relay to protect the Hi-Z input during transmit.
I have a Kenwood TS 870 and my EXT RX ANT RCA port is supposed to go to ground by the radio when I TX. I'm a CW only op, sometimes QSK at full speed. I was thinking of some front end protection like that which DX Engineering makes the "Receiver Guard Electronic RF Limiters DXE-RG5000HD" This starts to get kinda pricey for me in Australia. Does anyone else use the Receiver Guard Electronic RF Limiters DXE-RG5000HD and if so how does it work for you? Is it like a fuse where after you put RF on it it doesn't work anymore or does it work like the name suggests as a limiter that continuously works?
Cheers es 73 ..
Reason: No reason