EMRFD Message Archive 12434

Message Date From Subject
12434 2016-02-22 19:48:39 w0pwe RF Switch ICs for Filter Selection

I'm building a multi-band transceiver and as I begin to design the bandpass filters for the receiver and low pass filters for the transmitter I hate to think about all the relays I am going to need and the space they are going to occupy.


I can't do much about the low pass filter switching but it seems that the bandpass filters could be switched with one of the many RF switch ICs that are available. Intended for GHz service, many of them are also good to a few MHz. They are inexpensive, have impressive isolation and insertion loss specs, and there are lots of options to choose from.


Has anyone had experience using these parts for an HF an application like this? Any specific part recommendations? A few I have looked at:

Skyworks AS169-73

Macom MASWSS0115

Fairchild FSA3157

Peregrine PE4250

73

Jerry - WØPWE


12435 2016-02-22 20:22:03 DuWayne Schmidlko... Re: RF Switch ICs for Filter Selection
You might want to look at the schematic of the SoftRock RX Ensemble III HF Receiver.
http://fivedash.com/index.php?main_page=product_info&cPath=1&products_id=16
They use a pair of inexpensive Quad Buss switches for the four band pass filters.  If you also use the output enable pins in your switching control, you can  add multiple banks of switches.
That should work for the receiver switching.  For the transmitter low-pass filters, you will probably want to stick to relays.
DuWayne
KV4QB

12436 2016-02-23 04:04:35 i7swx Re: RF Switch ICs for Filter Selection
Hi Jerry,

recently there was a discussion in the rsgbtech newsgroup about filters switching using relays or other switches. I commented as follows:

The Pic-A-Star is the first example of using such switches replacing the classic pin diodes. The FST3126 is enabled when High and the FST3125 is enabled when Low, this can help with the switching. 
An interesting part for the use of these ICs is that, although used in the H-Mode Mixer and other mixers,  have a limited high frequency limits for low convertion losses (ok 50 and 70 MHz, when used as commutating switches can have a bandwidth use of 200 to 300 MHz (max Vcc 7V), while the low voltage PI3C3125-6 (max Vcc 4.6V) have a BW up to 450MHz, The FSA3157, used in the PA3AKE transceiver H-Mode Mixer, are also a good selection (single dual position)  and have a BW up to 250MHz. These switches are certainly not as good as relays, having an Ron around 5 ohm, while the latter ones are under 1 ohm, but they will not need a wetting current and will not have contacts oxided in the time. These ICs will not be usable for high power.

The PE4250 seems an interesting switch to select a filter to select it and then put it to ground, this will need two switches to have a simple single signal selection for each filter input and output.

The FSA3157 can be a good solution also. Still will need two ICs for each filter but it will need a single line (High level) selecti
12438 2016-02-23 10:23:48 IEN Re: RF Switch ICs for Filter Selection
Have a look at the attachment.
73

norberto.modanesi@gmail.com
San Nicolás
----- Original Message -----
12439 2016-02-23 19:28:07 w0pwe Re: RF Switch ICs for Filter Selection
Thank you for the great info and examples. I will give this a try.
73 - W0PWE
12440 2016-02-24 04:57:24 i7swx Re: RF Switch ICs for Filter Selection
Hi,

you have to take note, about my circuit diagram, that a switch bias is on both sides, input and output. I have drawn in my previous drawing I had boxes representing filters. In case of a LPF you may use a single bias, for a high pass filter you need to have buffer capacitors and for the BPF probably the same unless the circuit has already buffer caps. Maybe I should correct the schematic with the 3 filter cases.

73

Gian
I7SWX
12442 2016-02-24 06:26:55 Steve Dick Re: RF Switch ICs for Filter Selection
Gian, have you compared the Fairchild FSA3157 to the Peregrine PE4250 used in a filter switching application with regard to strong signal performance?
 
“Digital Steve”, K1RF
 
12443 2016-02-24 06:48:29 Lasse Moell Re: RF Switch ICs for Filter Selection
Given that the PE4250 is an RF switch vs a bus switch for the Fairchild thingy, the performance should point to Peregrene. But remember that lowest frequency for the PE4250 is 10 MHz, and no warranty on what to expect at lower f.

/Lasse SM5GLC

24 februari 2016 15:26:46 +01:00, skrev Steve Dick sbdick@optonline.net [emrfd] :
 


Gian, have you compared the Fairchild FSA3157 to the Peregrine PE4250 used in a filter switching application with regard to strong signal performance?
 
“Digital Steve”, K1RF
 
12444 2016-02-25 08:48:04 i7swx Re: RF Switch ICs for Filter Selection
Hi Steve,

I have not had any chance to try the PE4250 switches. Looking at the datasheet it seems to be a very interested component as can switch frequencies up to 1GHz. As Lasse wrote we do not know what would happen below 10MHz, as datasheet reports linearity degrading, insertion loss should be within specs.

I have redrawn my FSA3257 filters switching schematic using PE4250, the circuit is simplified and, if no DC is present on the input and output connections, the decoupling capacitors may be omitted. I am loading the schematic in files

If you are going to try the PE4250 please let us know the results.

Enjoy the project

73

Gian
I7SWX


12446 2016-02-25 11:28:43 Graham / KE9H Re: RF Switch ICs for Filter Selection
The problem with the PE425x series switches is that they contain an internal bias generator, which is a square wave oscillator driving a charge pump at about 900 kHz.
This fundamental frequency and its odd harmonics bleed into the signal path, and can cause spurious signals in the low HF range. 
It is not a linearity problem in the normal sense.
This is not documented in the spec sheet. 
It is not a problem in the cellular band frequencies, which is where this switch was designed to operate.

--- Graham

==

12447 2016-02-26 04:59:30 i7swx Re: RF Switch ICs for Filter Selection

Hi Graham,

 

thank you very much the important note you reported. As you said the datasheet does not report the  PE425x series switches is that they contain an internal bias square wave generator, driving a charge
pump at about 900 kHz and generating spuries in the low HF.

 

Yesterday I requested infos on linearity degradation to Peregrine support, I will share the reply once received.

 

Due to the above problem, the easier solution is to use the FSA3157 for switching filters. The solution is quite simple.

 

Best 73 and have all a nice weekend

 

Gian

I7SWX 

12448 2016-02-26 06:43:15 Steve Dick Re: RF Switch ICs for Filter Selection
Hi Graham and Gian. I was unaware of the internal bias square wave generator.  I believe the large signal degradation itself below 10 MHz is a gradual effect and they can be used as low as 1 MHz without serious issues. Don’t know how important the charge pump noise is though.  At low frequencies, band noise, I would think, should totally overwhelm any tiny charge pump noise spurs. There are some other practical considerations with regard to produceability. I think the Perigrine parts are basically single source while the FSA3157 has more pin compatible parts from multiple vendors.  
 
“Digital Steve”, K1RF
 
12450 2016-02-26 10:00:40 i7swx Re: RF Switch ICs for Filter Selection

Hi Steve and Graham,

 

I have added a new file on files, following the information received from Peregrine Support, I modified my schematic suggested for the PE4250 referring to the PE4239 device.

 

I reported Graham comments to support and they suggest the following:

 

This is true. The charge pump fundamental is about ~-120dBm  at ~ 1Mhz ( plus harmonics) so mostly isn’t a problem except in a rx path at the front end.

You can use the 4239 which has no charge pump, or the 4257 which has  the external bias option.

Plus a curve related to linearity degradation was reported. Such curve is reported in the loaded file.

 

As Steve commented, it is right that several manufacturers provide the FSA3157.

 

I believe this has been a nice discussion and it should be interesting for many members.

 

Best 73

 

Gian

I7SWX

12457 2016-02-28 10:16:26 jorschei Re: RF Switch ICs for Filter Selection
Hi Gian and Steve/Graham,

Indeed the Peregrine RF IC'switch PExxxx  serie are interesting specialy for my focus on RX development working on a 3.3 volt Li-Ion cell batterrie.
Let me know if anyone has results on the lower frequency bands.

73'Joris  KTH rf-Design
-----------------------------------------

Hi Steve and Graham,

 

I have added a new file on files, following the information received from Peregrine Support, I modified my schematic suggested for the PE4250 referring to the PE4239 device.

 

I reported Graham comments to support and they suggest the following:

 

This is true. The charge pump fundamental is about ~-120dBm  at ~ 1Mhz ( plus harmonics) so mostly isn’t a problem except in a rx path at the front end.

You can use the 4239 which has no charge pump, or the 4257 which has  the external bias option.

Plus a curve related to linearity degradation was reported. Such curve is reported in the loaded file.

 

As Steve commented, it is right that several manufacturers provide the FSA3157.

 

I believe this has been a nice discussion and it should be interesting for many members.

 

Best 73

 

Gian

I7SWX

12465 2016-03-03 14:05:27 Steve Dick Re: RF Switch ICs for Filter Selection
Did a bit more research. I’m liking the FSA4157 as it has maximum 1.15 ohm on resistance at Vcc = 4.5V and much smaller Ron flatness of 0.3 ohms.  This should result in better IMD performance. This is also indicated by about 5X lower total harmonic distortion compared to the FSA3157.  It has somewhat less bandwidth and isolation than the FSA3157 but still fine for switching receiver filters. Has a little over 0.5dB loss up to 50 MHz.  The much better IMD performance is a big advantage.
 
 
Regards, “Digital Steve”, K1RF
 
 
12470 2016-03-04 10:43:55 i7swx Re: RF Switch ICs for Filter Selection

Hi Steve,

 

thanks for your note.

 

Yes, you are right, the FSA4157 is a good switch. It is strange but I did download the FSA4157 datasheet some years ago but I forgot about it as I am using the FSA3157 on the H-Mode Mixer and this chip has Ton 1.7nS and Toff 0.8nS while the FSA4157 has Ton 35nS and Toff 15nS.

 

The schematic I loade for the FST3157 switching is still valid for the FSA4157 as the pins have the same definition.

 

73 and have all a nice weekend

 

Gian

I7SWX

12477 2016-03-05 11:42:29 Ashhar Farhan Re: RF Switch ICs for Filter Selection
The lowly 2N7000s are quite good switches. They are also quite good as amps all the way to VHF. At times, these devices hide in plain sight labelled as something else. For the Minima, I did evaluate these as switches and the result was quite impressive. Entirely on par with the FSA3157 at the lower and mid HF ranges.

In a week or two, I will finish evaluating the 2N7000 as a low level and high level amplifier and post the results. The initial results are quite encouraging.

At times, the devices we need hide in plain sight inside our junk boxes.

- f

12481 2016-03-05 20:30:17 Mvs Sarma Re: RF Switch ICs for Filter Selection
SMD version of 2N7000 appears very cheap as compared to standard through hole version.


12483 2016-03-06 05:05:12 k1rf_digital_stev... Re: RF Switch ICs for Filter Selection
The 2N7000 may work well in many applications for switching.  However, like any FET, its on resistance varies as a function of gate to source voltage and drain current.  Many of the FSA type devices are based on transmission gates, which use a p-channel FET and an N-channel FET tied back to back. The worst condition for these tends to be near the midpoint of of the power supply voltage around which point its resistance is the highest. You will still get some nonlinearity. Figure 7 of the FSA 4157 shows the variation in switch resistance as a function of input voltage.  The curve shows two peaks, coming from each FET before the other FET starts turning on. The on resistance varies from about .85 ohms to 1.1 ohms for small signals if the switch is biased around the midpoint.  Assuming the switch is used in a 50 ohm system, this is a rather small variation implying fairly good linearity.  I have seen some designs bias the switches off the midpoint of the power supply in an attempt to get "better" switch performance. A single FET will have a much larger on resistance variation as shown in Figure 2 of the 2n7000 data sheet.  It can certainly work, but you need to take its characteristics into consideration for the intended circuit. There's more to an analog switch than its simple "on-off" functionality - namely how nonlinear is it over the signal range that will be driving it.

Regards, "Digital Steve", K1RF with an analog comment
12485 2016-03-06 07:09:26 kb1gmx Re: RF Switch ICs for Filter Selection
I've tried 2N7000 and the isolation is poor as it has high capacitance and that leads to blow through.
Its ok for a shunt switch if the node an tolerate the device capacitance.

Generally CMOS switches, Jfets, Pin diodes(try 1n4007), Junction diodes(1n4148) are how 
I go depending on levels and performance (IMD) required.


Allison
12486 2016-03-06 07:15:58 farhanbox@gmail.c... Re: RF Switch ICs for Filter Selection

you will need a shunt and a series fet for proper isolation. 

12529 2016-03-22 14:00:02 jorschei Re: RF Switch ICs for Filter Selection
Hi Graham,

Interesting info about the 900KHz charge pump oscillator in the PE425x switch.
Can you tel me where I can found this data ?

I order the PE4259  SPDT switch for testing,  should work from DC up to 3.0 GHz.

Joris,  KTH rf Design  ex  PE1KTH.


The problem with the PE425x series switches is that they contain an internal bias generator, which is a square wave oscillator driving a charge pump at about 900 kHz.
This fundamental frequency and its odd harmonics bleed into the signal path, and can cause spurious signals in the low HF range. 
It is not a linearity problem in the normal sense.
This is not documented in the spec sheet. 
It is not a problem in the cellular band frequencies, which is where this switch was designed to operate.

--- Graham
12531 2016-03-22 15:38:47 Graham / KE9H Re: RF Switch ICs for Filter Selection
Joris:

I have personal experience with the PE4259.
It will work across the frequency range you describe.
It has the 900 kHz charge pump, and if you can see down below -120 dBm, you will see it and it's harmonics up to 25 MHz or so.
And in each switch, the charge pump is on a slightly different frequency.
It is not documented in the data sheet.
The way you find out is to build a circuit, then call up Peregrine engineering and ask:
"Why are there low level 900 kHz signals and harmonics coming out of each of my PE4259 switches?"

It was designed to be a cellular radio RF switch, and it does that very well.

--- Graham / KE9H

==

12534 2016-03-23 05:39:06 J van Scheindelen Re: RF Switch ICs for Filter Selection

Graham:

 

Thanks for your experience information about the charge pump.

The problem is if used in a sub octave band filter set,   the 900 KHz will probably heard in the receiver when receiving low level signals. The -120 dB over 50 ohm is apron 0.2 microvolt.

May be an option is to install an L/C serial circuit on the filter block ..in and... output switch adjusted on 900 KHz to absorb the 900 KHz main component.

If this is sufficient the harmonics will also be suppressed and the switch can be used above 2 MHz

If so the PE4259 can be used for my wearable SW receiver project working on 3.3 volt with solar cell where space and power consumption is important.

 

Thanks for the info to be continued.

 

73" KTH rf Design

 
12536 2016-03-23 09:30:29 Graham / KE9H Re: RF Switch ICs for Filter Selection
Joris:

Suppressing the fundamental 900 kHz signal with a filter, will not result in suppressing the harmonics, since they are all real signals, generated in the charge pump.
If you have a good antenna like a dipole or better, most of the harmonics will be below the atmospheric noise level.
If you are running a low efficiency antenna, like a short loaded antenna, you might still hear them.

You might want to consider one of the Peregrine switches that does not use the charge pump, as noted in the I7SWX email of February 26.

--- Graham / KE9H

==



12537 2016-03-24 04:03:33 J van Scheindelen Re: RF Switch ICs for Filter Selection

Graham,

 

Thanks for the info.

You are right it is only possible to notch out de 800 KHz coming from the PE4259 bud the harmonics are still there and if they are strong it will be received.

The FSA4157P6X  is a interesting  replacement bud width a lower IP3 I think.

I will investigate it.

 

73" Joris
12538 2016-03-24 04:43:37 Steve Dick Re: RF Switch ICs for Filter Selection
Hi Joris.  I had high hopes for the FSA4157 but as was pointed out by Graeme in the Hermes Lite Yahoo group, its capacitances are very high. Too bad, because their IP3 should be very good. Graeme stated: "The bit that concerns me with the FSA4157 switches is the off port's 12pF capacitance. If you are switching say 6 filters these ports will be commoned at 72pF and will connect to the filter when switched. Port A is 40pF so you could have a total of 100 odd pF presented to the filter. Of course this value could be included in the filter itself and a series parallel conversion done to absorb it.
“Digital Steve”, K1RF
 
12539 2016-03-24 05:29:16 J van Scheindelen Re: RF Switch ICs for Filter Selection
Hi Steve,



Yes the FSA capacity should be in the simulation added to see how it looks like.

If capacity is a problem (may be on VHF ) you could use the PIN diode for example the BA892 , has 1 pf and low los.

Look at the Attach where I use them in my KTH-SDR filter block thad works well.

Bud with more components...



Joris

12775 2016-05-01 13:25:15 shaynal01 Re: RF Switch ICs for Filter Selection
Hi Group,

I have made some measurements of the RX noise visible when using the PE4259 discussed in this thread. You can find the Quisk RX screenshot measurements on the Hermes-Lite group, at:


Feedback is welcome and appreciated.

73,

Steve
KF7O