Topics

On TX Bandpass filter


Ashhar Farhan
 

How many here need a tx bandpass filter for the ubitx?
This is needed when operating in the neighborhood of another amatuer station. I was wondering if we should convert our tx output LPFs to banpass filters.
Opinions and advice are welcome.
- f


Lawrence Galea
 

Hi Farhan.
I think it would be better as lower garbage would also be attenuated, but it would also depend on the bandwidth you have in mind and whether the present 4 filters would be enough.
Regards
Lawrence

On Tue, Oct 6, 2020 at 8:13 AM Ashhar Farhan <farhanbox@...> wrote:
How many here need a tx bandpass filter for the ubitx?
This is needed when operating in the neighborhood of another amatuer station. I was wondering if we should convert our tx output LPFs to banpass filters.
Opinions and advice are welcome.
- f


Ashhar Farhan
 

Umm.. it will have to be one filter per band. The toroids and the C0G caps are hard to come by. Let's see how to play this.


On Tue 6 Oct, 2020, 3:28 PM Lawrence Galea, <9h1avlaw@...> wrote:
Hi Farhan.
I think it would be better as lower garbage would also be attenuated, but it would also depend on the bandwidth you have in mind and whether the present 4 filters would be enough.
Regards
Lawrence

On Tue, Oct 6, 2020 at 8:13 AM Ashhar Farhan <farhanbox@...> wrote:
How many here need a tx bandpass filter for the ubitx?
This is needed when operating in the neighborhood of another amatuer station. I was wondering if we should convert our tx output LPFs to banpass filters.
Opinions and advice are welcome.
- f


 

Remember my old suggestions of a BCD decoder to get 7/8 bands with just 3 lines.

Raj


At 06/10/2020, you wrote:
Umm.. it will have to be one filter per band. The toroids and the C0G caps are hard to come by. Let's see how to play this.

On Tue 6 Oct, 2020, 3:28 PM Lawrence Galea, <9h1avlaw@...> wrote:
Hi Farhan.
I think it would be better as lower garbage would also be attenuated, but it would also depend on the bandwidth you have in mind and whether the present 4 filters would be enough.
Regards
Lawrence

On Tue, Oct 6, 2020 at 8:13 AM Ashhar Farhan <farhanbox@...> wrote:
How many here need a tx bandpass filter for the ubitx?
This is needed when operating in the neighborhood of another amatuer station. I was wondering if we should convert our tx output LPFs to banpass filters.
Opinions and advice are welcome.
- f


iz oos
 

A TXRX bandpass filter would be interesting not just for ubitxers but wouldn't it be too lossy and costly? Using and having built some of them for RX frontends (I often use separate TX and RX) I really symphatize your idea.


Il mar 6 ott 2020 08:13 AM Ashhar Farhan <farhanbox@...> ha scritto:
How many here need a tx bandpass filter for the ubitx?
This is needed when operating in the neighborhood of another amatuer station. I was wondering if we should convert our tx output LPFs to banpass filters.
Opinions and advice are welcome.
- f


Ravi Miranda
 

Hi Farhan,

What if the shared BPF was before the Pre-drive stage, so the signal
is cleaned up before the driver and PA? Then a standard LPF could be
added?

The same BPF could be used in Rx stage?

Just my 2p.

Ravu

On Tue, 6 Oct 2020 at 11:47, Ashhar Farhan <farhanbox@...> wrote:

Umm.. it will have to be one filter per band. The toroids and the C0G caps are hard to come by. Let's see how to play this.

On Tue 6 Oct, 2020, 3:28 PM Lawrence Galea, <9h1avlaw@...> wrote:

Hi Farhan.
I think it would be better as lower garbage would also be attenuated, but it would also depend on the bandwidth you have in mind and whether the present 4 filters would be enough.
Regards
Lawrence

On Tue, Oct 6, 2020 at 8:13 AM Ashhar Farhan <farhanbox@...> wrote:

How many here need a tx bandpass filter for the ubitx?
This is needed when operating in the neighborhood of another amatuer station. I was wondering if we should convert our tx output LPFs to banpass filters.
Opinions and advice are welcome.
- f
--
I'm here to add more value to the world than I'm using up.


Lawrence Galea
 

HI Farhan
A few quick ideas. Maybe I would think of more and let you know later

Considering that BPF's are being considered and that the set easily covers 160M, I would suggest that 160M is also included with a wideband receive position for general coverage.

Use of the BPFs instead of LPF's would also provide more protection on the Rx side from out of band signals.

Since this would entail a new pcb design for the main board, a daughter board that could piggy back on the main board or fixed to the back panel of the cabinet could be considered. While at it, the Si frequencies generator could also be on the main board to shorten the rf lines to the respective mixers. Also a good AGC, s-meter circuit and an audio filter designed on the main board would be desirable additional features that could either be factory included or added by the buyer.

For capacitors sources perhaps you could get in touch with QRPLabs.

Regards
Lawrence






On Tue, Oct 6, 2020 at 1:30 PM Ravi Miranda <ravimiranda@...> wrote:
Hi Farhan,

What if the shared BPF was before the Pre-drive stage, so the signal
is cleaned up before the driver and PA? Then a standard LPF could be
added?

The same BPF could be used in Rx stage?

Just my 2p.

Ravu

On Tue, 6 Oct 2020 at 11:47, Ashhar Farhan <farhanbox@...> wrote:
>
> Umm.. it will have to be one filter per band. The toroids and the C0G caps are hard to come by. Let's see how to play this.
>
> On Tue 6 Oct, 2020, 3:28 PM Lawrence Galea, <9h1avlaw@...> wrote:
>>
>> Hi Farhan.
>> I think it would be better as lower garbage would also be attenuated, but it would also depend on the bandwidth you have in mind and whether the present 4 filters would be enough.
>> Regards
>> Lawrence
>>
>> On Tue, Oct 6, 2020 at 8:13 AM Ashhar Farhan <farhanbox@...> wrote:
>>>
>>> How many here need a tx bandpass filter for the ubitx?
>>> This is needed when operating in the neighborhood of another amatuer station. I was wondering if we should convert our tx output LPFs to banpass filters.
>>> Opinions and advice are welcome.
>>> - f
>>>
>



--
I'm here to add more value to the world than I'm using up.






Ashhar Farhan
 

Lawrance, Raj and others
We could even use a single line to select different band pass filters witha a CD4017. With each pulse, it moves a high to to the next decoded output. It provides 10 outputs .
An external bank of bandpass filters could be left between the antenna and the rig for both :Rx and Tx.
This will provide the rx with better filtering as well. It can always be switched out when not needed.
We would need 10 of them to cover all the HF bands. With three sections in each, that is 30 inductors and 90 capaictors (the current LPFs have 12 inductors and 20 capacitors).
The cost of making these will be significant. I am wondering of should or not include them in.
- f
- f


On Tue 6 Oct, 2020, 5:29 PM Lawrence Galea, <9h1avlaw@...> wrote:
HI Farhan
A few quick ideas. Maybe I would think of more and let you know later

Considering that BPF's are being considered and that the set easily covers 160M, I would suggest that 160M is also included with a wideband receive position for general coverage.

Use of the BPFs instead of LPF's would also provide more protection on the Rx side from out of band signals.

Since this would entail a new pcb design for the main board, a daughter board that could piggy back on the main board or fixed to the back panel of the cabinet could be considered. While at it, the Si frequencies generator could also be on the main board to shorten the rf lines to the respective mixers. Also a good AGC, s-meter circuit and an audio filter designed on the main board would be desirable additional features that could either be factory included or added by the buyer.

For capacitors sources perhaps you could get in touch with QRPLabs.

Regards
Lawrence






On Tue, Oct 6, 2020 at 1:30 PM Ravi Miranda <ravimiranda@...> wrote:
Hi Farhan,

What if the shared BPF was before the Pre-drive stage, so the signal
is cleaned up before the driver and PA? Then a standard LPF could be
added?

The same BPF could be used in Rx stage?

Just my 2p.

Ravu

On Tue, 6 Oct 2020 at 11:47, Ashhar Farhan <farhanbox@...> wrote:
>
> Umm.. it will have to be one filter per band. The toroids and the C0G caps are hard to come by. Let's see how to play this.
>
> On Tue 6 Oct, 2020, 3:28 PM Lawrence Galea, <9h1avlaw@...> wrote:
>>
>> Hi Farhan.
>> I think it would be better as lower garbage would also be attenuated, but it would also depend on the bandwidth you have in mind and whether the present 4 filters would be enough.
>> Regards
>> Lawrence
>>
>> On Tue, Oct 6, 2020 at 8:13 AM Ashhar Farhan <farhanbox@...> wrote:
>>>
>>> How many here need a tx bandpass filter for the ubitx?
>>> This is needed when operating in the neighborhood of another amatuer station. I was wondering if we should convert our tx output LPFs to banpass filters.
>>> Opinions and advice are welcome.
>>> - f
>>>
>



--
I'm here to add more value to the world than I'm using up.






 

Farhan,

Decade counter can go for a six due to RF and there is no way the processor would know which filter is on.

Raj

At 06/10/2020, you wrote:
Lawrance, Raj and others
We could even use a single line to select different band pass filters witha a CD4017. With each pulse, it moves a high to to the next decoded output. It provides 10 outputs .
An external bank of bandpass filters could be left between the antenna and the rig for both :Rx and Tx.
This will provide the rx with better filtering as well. It can always be switched out when not needed.
We would need 10 of them to cover all the HF bands. With three sections in each, that is 30 inductors and 90 capaictors (the current LPFs have 12 inductors and 20 capacitors).
The cost of making these will be significant. I am wondering of should or not include them in.
- f
- f


MadRadioModder
 

This radio was supposed to be low cost… keep adding bells and whistles and it will be as much as a commercial radio.  WHAT I WOULD SUGGEST is adding the traces to the circuit board for more advanced projects and not populate them on the basic radio kit… to keep the cost in check.  Then people can populate additional projects themselves.  Filters, AGC, SWR, etc.  The incremental cost for designing a new board to do this is very low.

 

MRM

 

From: BITX20@groups.io <BITX20@groups.io> On Behalf Of Ashhar Farhan
Sent: Tuesday, October 6, 2020 5:46 AM
To: BITX20@groups.io
Subject: Re: [BITX20] On TX Bandpass filter

 

Umm.. it will have to be one filter per band. The toroids and the C0G caps are hard to come by. Let's see how to play this.

 

On Tue 6 Oct, 2020, 3:28 PM Lawrence Galea, <9h1avlaw@...> wrote:

Hi Farhan.

I think it would be better as lower garbage would also be attenuated, but it would also depend on the bandwidth you have in mind and whether the present 4 filters would be enough.
Regards

Lawrence

 

On Tue, Oct 6, 2020 at 8:13 AM Ashhar Farhan <farhanbox@...> wrote:

How many here need a tx bandpass filter for the ubitx?

This is needed when operating in the neighborhood of another amatuer station. I was wondering if we should convert our tx output LPFs to banpass filters.

Opinions and advice are welcome.

- f

 


Virus-free. www.avg.com

--

…_. _._


Ravi Miranda
 

Would this work? Not sure if the file gets through.
The board accomodates all necessary sub-boards and additionally a CW filter board, a meter amplifier, an AGC board, an post IF amp, and mic compression board, the Si5351 board and a RFpreamp.
The driver + PA is not on this board.
Ravi/M0RVI

On Tue, 6 Oct 2020 at 15:23, MadRadioModder <madradiomodder@...> wrote:

This radio was supposed to be low cost… keep adding bells and whistles and it will be as much as a commercial radio.  WHAT I WOULD SUGGEST is adding the traces to the circuit board for more advanced projects and not populate them on the basic radio kit… to keep the cost in check.  Then people can populate additional projects themselves.  Filters, AGC, SWR, etc.  The incremental cost for designing a new board to do this is very low.

 

MRM

 

From: BITX20@groups.io <BITX20@groups.io> On Behalf Of Ashhar Farhan
Sent: Tuesday, October 6, 2020 5:46 AM
To: BITX20@groups.io
Subject: Re: [BITX20] On TX Bandpass filter

 

Umm.. it will have to be one filter per band. The toroids and the C0G caps are hard to come by. Let's see how to play this.

 

On Tue 6 Oct, 2020, 3:28 PM Lawrence Galea, <9h1avlaw@...> wrote:

Hi Farhan.

I think it would be better as lower garbage would also be attenuated, but it would also depend on the bandwidth you have in mind and whether the present 4 filters would be enough.
Regards

Lawrence

 

On Tue, Oct 6, 2020 at 8:13 AM Ashhar Farhan <farhanbox@...> wrote:

How many here need a tx bandpass filter for the ubitx?

This is needed when operating in the neighborhood of another amatuer station. I was wondering if we should convert our tx output LPFs to banpass filters.

Opinions and advice are welcome.

- f

 


Virus-free. www.avg.com

--

…_. _._



--
I'm here to add more value to the world than I'm using up.


Reed N
 

I second Raj's concerns about the decade counter. In addition to not having a way for the processor to validate state (without consuming more GPIO), using such a chip would require many more relay activations, since jumping bands would then require cycling every relay between the two. A binary decoder or I2C IO expander would be better suited. An advantage of the IO expander is that you could add the band pass filters in addition to the low pass filters using the existing setup.

I also agree with MRM that it may be better to design it as an easy add-on, rather than a required base component, for cost reasons. Perhaps add jumper pins to some of the base board traces, that are shipped with jumpers installed, but can be un-jumpered and attached to mods with minimal extra effort?


Reed


Ashhar Farhan
 

I am attaching a circuit diagram of the RF386 QRP linear that I had built for the Minima transceiver. Of interest is the CD4017 in the bottom.
You will see that the Reset line has a capacitor to the ground and a resistor in series with the digital "Band Select" line. 
The way it works is as follows:
1. To reset the counter, the digital line is driven to HIGH and held long enough for the capacitor to charge, activating the resetting the counter.
2. After 10 milliseconds, the line is driven LOW and the capacitor discharges rapidly through the diode. At this moment, the 0utput-0 is high and the others are low.
3.  Now, we send quick pulses (about 10 usec wide) to shift the active line from Output-0 to whichever output we want. As the quick pulses do not charge the capacitor to full value, the line is not reset until a long pulse resets it.
I had seen this scheme in the old analog radio control. Each of the CD4017's output was attached to a different servo channel. The CD4017 was driven by the output of an AM detector.

- f

On Tue, Oct 6, 2020 at 10:42 PM Reed N <greenkid336600+groupsio@...> wrote:
I second Raj's concerns about the decade counter. In addition to not having a way for the processor to validate state (without consuming more GPIO), using such a chip would require many more relay activations, since jumping bands would then require cycling every relay between the two. A binary decoder or I2C IO expander would be better suited. An advantage of the IO expander is that you could add the band pass filters in addition to the low pass filters using the existing setup.

I also agree with MRM that it may be better to design it as an easy add-on, rather than a required base component, for cost reasons. Perhaps add jumper pins to some of the base board traces, that are shipped with jumpers installed, but can be un-jumpered and attached to mods with minimal extra effort?


Reed


KA5FPT Paul
 

Would it not be easier to use external bandpass filters rather then modify the radio? Many multi-op contest stations go this way, and also large Field Day groups. Use external switching to choose the band wanted.
Paul
KA5FPT

Sent with Aqua Mail for Android
https://www.mobisystems.com/aqua-mail


Ashhar Farhan
 

The greater challenge is not the transmit side but the receive side. The ubitx input filters nothing below 30 MHz. If you are operating on 14 MHz, an AM broadcast station across the city at 800 KHz will swamp your front-end. With the bandpass filters even another ham operating a kilowatt on a different band, a few yards away will not interfere, it is useful for field days and multi-operator contesting or if you have a strong transmitter in the neighborhood, not otherwise.
- f

On Tue, Oct 6, 2020 at 11:16 PM Ashhar Farhan via groups.io <farhanbox=gmail.com@groups.io> wrote:
I am attaching a circuit diagram of the RF386 QRP linear that I had built for the Minima transceiver. Of interest is the CD4017 in the bottom.
You will see that the Reset line has a capacitor to the ground and a resistor in series with the digital "Band Select" line. 
The way it works is as follows:
1. To reset the counter, the digital line is driven to HIGH and held long enough for the capacitor to charge, activating the resetting the counter.
2. After 10 milliseconds, the line is driven LOW and the capacitor discharges rapidly through the diode. At this moment, the 0utput-0 is high and the others are low.
3.  Now, we send quick pulses (about 10 usec wide) to shift the active line from Output-0 to whichever output we want. As the quick pulses do not charge the capacitor to full value, the line is not reset until a long pulse resets it.
I had seen this scheme in the old analog radio control. Each of the CD4017's output was attached to a different servo channel. The CD4017 was driven by the output of an AM detector.

- f

On Tue, Oct 6, 2020 at 10:42 PM Reed N <greenkid336600+groupsio@...> wrote:
I second Raj's concerns about the decade counter. In addition to not having a way for the processor to validate state (without consuming more GPIO), using such a chip would require many more relay activations, since jumping bands would then require cycling every relay between the two. A binary decoder or I2C IO expander would be better suited. An advantage of the IO expander is that you could add the band pass filters in addition to the low pass filters using the existing setup.

I also agree with MRM that it may be better to design it as an easy add-on, rather than a required base component, for cost reasons. Perhaps add jumper pins to some of the base board traces, that are shipped with jumpers installed, but can be un-jumpered and attached to mods with minimal extra effort?


Reed


Clark Martin
 

With three CPU lines and shift registers you could operate as many relays as you need. An you can get a chip that is a shift register and relay driver in one, the TPIC6B595.

To get the full value of the bandpass filters you’d have to keep the relay energized during receive.


Clark Martin
KK6ISP

On Oct 6, 2020, at 5:16 AM, Ashhar Farhan <farhanbox@...> wrote:

Lawrance, Raj and others
We could even use a single line to select different band pass filters witha a CD4017. With each pulse, it moves a high to to the next decoded output. It provides 10 outputs .
An external bank of bandpass filters could be left between the antenna and the rig for both :Rx and Tx.
This will provide the rx with better filtering as well. It can always be switched out when not needed.
We would need 10 of them to cover all the HF bands. With three sections in each, that is 30 inductors and 90 capaictors (the current LPFs have 12 inductors and 20 capacitors).
The cost of making these will be significant. I am wondering of should or not include them in.


Curt
 

Bandpass filters often have more loss than low pass filters, especially when we try to make them small.  Note QRP-Labs has a nice set of both - one can easily compare the loss. 

Also. someone already caught using bandpass filters takes us from 4 filters to 8 or 9. 

Hams in multi-multi contest operations from a common site sometimes use large, expensive bandpass filters.  These can be homemade.  I am not sure there is a huge market for accessory filters for the uBitx.  Now I do have parts sitting on the bench to assemble a 160m low pass filter -- just in case my 160m vertical does not already reject the 80m harmonic enough (the local RBN says it does ....). 

73 curt


Ashhar Farhan
 

The tx band pass filters are not the same as the ex band pass filters. The QRP lab filters are designed as rx filters.
Whe. Desiging a Bandpass filters you trade off the losses with bandwidth. A 7 MHz tx filter is less bothered about a 7.4 Mhz shortwave broadcast station from overloading the receiver than a nearby 14 Mhz receiver getting swamped by the noise from your tx.
At 100 watts, a 1 db loss is losing out 20 watts to the filter. This 20 watt loss will appear as heating of the coils. Hence, the tx bandpass filters have to use very low loss and wider bandwidths. At 10 watts level, the heat is more managable.
With good quality inductors and capacitors, it should be possible to hold down the losses to 0.5db while providing a few mhz of usable bandwidth. This can prevent AM broadcasts from overloading the receiver as well.
Another approach is to use a a set of low pass filters and a set of high pass filters. By choosing a correct combination, the two can produce a usable bandpass shape without too much loss. But the circuit is elaborate. Three sections of LPF and at least two sections of HPF will make the filter module more complex than the whole trx!
These are the high water marks if performance, they can be build by those who need them. It is impossible to acheive these in low cost radios such as the ubitx or the ucx.
A back of the envelop calculation shows that the cost of such tx band pass filtet bank will be almost 100 bucks. Do we really want that for a 150 dollar radio?
- f
- f


On Wed 7 Oct, 2020, 2:55 AM Curt via groups.io, <wb8yyy=yahoo.com@groups.io> wrote:
Bandpass filters often have more loss than low pass filters, especially when we try to make them small.  Note QRP-Labs has a nice set of both - one can easily compare the loss. 

Also. someone already caught using bandpass filters takes us from 4 filters to 8 or 9. 

Hams in multi-multi contest operations from a common site sometimes use large, expensive bandpass filters.  These can be homemade.  I am not sure there is a huge market for accessory filters for the uBitx.  Now I do have parts sitting on the bench to assemble a 160m low pass filter -- just in case my 160m vertical does not already reject the 80m harmonic enough (the local RBN says it does ....). 

73 curt


iz oos
 

This is all clear to me. The way to go for an RTX would be HPF and LPF. This would not be as selective as a dedicated RX preselector such as the tunable bavarian contest club preselektor. I am not dire of It could make sense to add a tunable preselector just before the first mixer, I mean just a selectable inductors and a variabile capacitor in series. Similarly to the MFJ 956 with added 4:1 transformers to keep low the impedance and have a higher Q especially above 7Mhz.


Il mer 7 ott 2020 06:29 AM Ashhar Farhan <farhanbox@...> ha scritto:
The tx band pass filters are not the same as the ex band pass filters. The QRP lab filters are designed as rx filters.
Whe. Desiging a Bandpass filters you trade off the losses with bandwidth. A 7 MHz tx filter is less bothered about a 7.4 Mhz shortwave broadcast station from overloading the receiver than a nearby 14 Mhz receiver getting swamped by the noise from your tx.
At 100 watts, a 1 db loss is losing out 20 watts to the filter. This 20 watt loss will appear as heating of the coils. Hence, the tx bandpass filters have to use very low loss and wider bandwidths. At 10 watts level, the heat is more managable.
With good quality inductors and capacitors, it should be possible to hold down the losses to 0.5db while providing a few mhz of usable bandwidth. This can prevent AM broadcasts from overloading the receiver as well.
Another approach is to use a a set of low pass filters and a set of high pass filters. By choosing a correct combination, the two can produce a usable bandpass shape without too much loss. But the circuit is elaborate. Three sections of LPF and at least two sections of HPF will make the filter module more complex than the whole trx!
These are the high water marks if performance, they can be build by those who need them. It is impossible to acheive these in low cost radios such as the ubitx or the ucx.
A back of the envelop calculation shows that the cost of such tx band pass filtet bank will be almost 100 bucks. Do we really want that for a 150 dollar radio?
- f
- f


On Wed 7 Oct, 2020, 2:55 AM Curt via groups.io, <wb8yyy=yahoo.com@groups.io> wrote:
Bandpass filters often have more loss than low pass filters, especially when we try to make them small.  Note QRP-Labs has a nice set of both - one can easily compare the loss. 

Also. someone already caught using bandpass filters takes us from 4 filters to 8 or 9. 

Hams in multi-multi contest operations from a common site sometimes use large, expensive bandpass filters.  These can be homemade.  I am not sure there is a huge market for accessory filters for the uBitx.  Now I do have parts sitting on the bench to assemble a 160m low pass filter -- just in case my 160m vertical does not already reject the 80m harmonic enough (the local RBN says it does ....). 

73 curt


Tom, wb6b
 

Hi,

Would it make sense to keep low pass filters on the TX/RX circuit and then add the high pass filters just to the RX circuit. Then use the combined high/low pass to improve the receiver performance with strong signals such as broadcast band transmitters. Maybe could switch the high pass filters with some other method than relays, if that would save cost.

Also, there are cheap processors that don't even need an external xtal. Some might even be cheaper than a shift register IC. Maybe that would be a great way to free yourself from worrying about the limitations on the pins of just the one Nano.

An ATTiny85 is easy to use and is compatible with the Arduino IDE. There may be other cheap Arduino compatible chips with more pins. Here is a "famous" article on one dollar microprocessor chips. Could be a starting point. https://jaycarlson.net/microcontrollers/

There are some other chips (some less than 35 cents) that can be programmed in C with the SDCC open source compiler. But, Arduino compatible might be best in this case.

Tom, wb6b