Uniden BCT15X - Best way to discriminator tap for DSDPlus? Tech question

[zerosix]

I’ve seen many threads online about making a discriminator tap. There are several versions floating around, most of them use a resistor + tantalum capacitor from the already available DISC point in serie to a RCA/3.5mm mono jack.

I get that and i used to make taps on my old scanner. From reading a lot online and also reading the tutorials from Bill Cheek, it seems that many of the old recommendations are no longer valid because people are connecting their discriminator output directly to their microphone/line in. From reading it seems there is often voltage coming from the PC since most microphone in lines power microphones this way.

from reading multiple threads it seems that a higher value tantalum capacitor of 10uF is best for a clear signal. Resistor values and suggestions vary wildly. I found people recommending between 1K and 12K Resistors. The guy from this thread: Discriminator Tap
says a lower value of around 1K should be used because it can interfere with the decoding.

this Is all getting over my head.. is there someone with true technical knowledge that can help me understand which values are best used.. thanks a lot

 

[zerosix]

Update 1: This guy had a similar question; BCT15 discriminator tap “how too”!

i would really like to know if it is true that a 5 or 10K resistor would decrease the audio signal too much and cut off a part of the audio frequency range, also it seems that higher capacity capacitors prevent the signal from being altered. Is that statement true? I know it’s a bit technical but I want to know the best way to approach this. From an electronical perspective I would assume that the ‘wrong’ values can indeed impact decoding performance on programs like DSD because they would act like a filter.

 

[jonwienke]

A 5-10KΩ resistor in series won't decrease the audio significantly, the input impedance of your external sound card should be at least 10KΩ, maybe higher, like 100KΩ. A 1μF nonpolar capacitor in series with the resistor will block any DC voltages, but pass low frequencies acceptably.

 

[zerosix]

Dear jonwienke, thanks for replying. This guy; Pro92b said the following:

”
To transfer a wideband signal to the sound card the tap resistor needs to be reduced to 1K or less. If a cap is to be placed in the tap feed, at 1uF the low cut off frequency is about 12 Hz. This is as high as I would like to see, and I would probably recommend 2.2uF to push the cut off lower. The capacitor should be non polar if there is concern about 12 volts accidently being connected to the tap point. The Nichicon ES series is bipolar and would be suitable.

Putting a resistor in the tap feed limits the current into or out of the IF chip and it is good practice to include one. The cap prevents steady state DC current in both directions but does not block transient current that might arise when something is connected to the tap. The resistor limits transient current in this case.

Bottom line: for sound card connections use 1K and 2.2uF in series from the tap point. Do not place a cap from the discriminator jack to ground. There probably already is a cap in the sound card input circuit.”

source: Discriminator Tap

is he wrong???

also I’ve noticed that many people with BCT15X scanners say the disc audio output signal is very low when using 10K resistors; wouldn’t a lower value increase the audio level to a more appropriate level?

thanks again, im looking forward to get more replies and learn more about this.

 

[jonwienke]

It depends on th input impedance of the sound card. If the input impedance is 1KΩ, than a 10kΩ resistor will reduce voltage by about 90% and yes, the sound level will be low. If the input impedance is 10KΩ, a 10KΩ resistor will reduce voltage by 50%--noticeable, but not an issue. But if the input impedance is 100KΩ or higher, than the voltage reduction will be less than 10%--negligible. If your limiter resistor and sound card input impedance is too low, then connecting the sound card to the tap can drop the signal level at the tap inside the scanner, and potentially cause the scanner to malfunction.

Check the specs on your sound card.

 

[zerosix]

It depends on th input impedance of the sound card. If the input impedance is 1KΩ, than a 10kΩ resistor will reduce voltage by about 90% and yes, the sound level will be low. If the input impedance is 10KΩ, a 10KΩ resistor will reduce voltage by 50%--noticeable, but not an issue. But if the input impedance is 100KΩ or higher, than the voltage reduction will be less than 10%--negligible. If your limiter resistor and sound card input impedance is too low, then connecting the sound card to the tap can drop the signal level at the tap inside the scanner, and potentially cause the scanner to malfunction.

Check the specs on your sound card.

thanks for this great information, I will try to determine the input impedance of my soundcard. Your comment was only about the resistor value, right? What about his comment that a 2.2 uF capacitor would be better because a lower value would cut off a part of the frequency response?

regards

 

[jonwienke]

The exact cutoff frequency will depend on the sum of the impedance of the limiter resistor and the sound card input. 12Hz is lower than strictly necessary (the lowest standard CTCSS tone is 67Hz) but going to 2.2μF won't hurt anything. It will just push the low frequency cutoff lower, like 5Hz or something.

 

[slicerwizard]

this Is all getting over my head.. is there someone with true technical knowledge that can help me understand which values are best used.. thanks a lot

The cap is simple - too small is bad, so use 10 uF. They're not expensive or bulky.

The best resistor values, yes values, because you should be creating a voltage divider, depend on the characteristics of your sound input. This means experimentation on your part, but I doubt you'll do it.

 

[wbswetnam]

I tapped a Whistler 1025 one time, super easy to do since it had a tapping point. I used a 5K resister in series with a 0 to 5K variable resister mounted on the side of the radio so that I could adjust the signal as necessary.

 

[Ubbe]

The input of soundcards probably lies in the 5-10kOhm range but has an impedance, a capcitance, that together with a resistor works as a low pass filter and a square wave will be filtered to a more rounded shape and makes decoding digitial signals harder. Using a 1k-2k resistor will act as a protection for short circuit and will not distort the signal too much.

But if the mic jack are used on a sound card that might have a voltage feed for a capacitor mic, then a bipolar capacitor might be necessary as it might be more positive voltage from the mic jack than from the discriminator tap, or the reverse. It's difficult to know without measuring the voltages.

I once made a discriminator tap using a 1uF capacitor. It was too small for some digitial signals and a 4,7uF cap where perfect for me in a Pro2042 scanner together with a 2,2k resistor.

/Ubbe

 

[zerosix]

It depends on th input impedance of the sound card. If the input impedance is 1KΩ, than a 10kΩ resistor will reduce voltage by about 90% and yes, the sound level will be low. If the input impedance is 10KΩ, a 10KΩ resistor will reduce voltage by 50%--noticeable, but not an issue. But if the input impedance is 100KΩ or higher, than the voltage reduction will be less than 10%--negligible. If your limiter resistor and sound card input impedance is too low, then connecting the sound card to the tap can drop the signal level at the tap inside the scanner, and potentially cause the scanner to malfunction.

Check the specs on your sound card.

It seems I have a combined microphone/headphone jack on my asus laptop. Can't find the specific impedance. I've also noticed that I cannot configure the duo port as microphone using only a mono 3.5mm plug. It will just not accept it.. probably has to do with the number of rings.. From what I've been reading a better option should be an external line-in input vs microphone, do you agree that a line input is superior to a microphone input for digital decoding? your help is appreciated!

The cap is simple - too small is bad, so use 10 uF. They're not expensive or bulky.

The best resistor values, yes values, because you should be creating a voltage divider, depend on the characteristics of your sound input. This means experimentation on your part, but I doubt you'll do it.

Thanks for your valuable contribution, I figured a larger capacitor would be better. About the second part you wrote, I am open to learning and experimenting but I'm quite new to technical modifications, so it will take me some time to understand it all. My background is in software development and so far I've only covered the basics of electronics when I received my general HAM license.

I tapped a Whistler 1025 one time, super easy to do since it had a tapping point. I used a 5K resister in series with a 0 to 5K variable resister mounted on the side of the radio so that I could adjust the signal as necessary.

That could be a handy solution for sure! I'll consider it, did you ever need to adjust it because of signal problems?

The input of soundcards probably lies in the 5-10kOhm range but has an impedance, a capcitance, that together with a resistor works as a low pass filter and a square wave will be filtered to a more rounded shape and makes decoding digitial signals harder. Using a 1k-2k resistor will act as a protection for short circuit and will not distort the signal too much.

But if the mic jack are used on a sound card that might have a voltage feed for a capacitor mic, then a bipolar capacitor might be necessary as it might be more positive voltage from the mic jack than from the discriminator tap, or the reverse. It's difficult to know without measuring the voltages.

I once made a discriminator tap using a 1uF capacitor. It was too small for some digitial signals and a 4,7uF cap where perfect for me in a Pro2042 scanner together with a 2,2k resistor.

/Ubbe

Thanks Ubbe, this makes a lot of sense and confirms my thinking about the resistor values. From what I've been reading a dedicated line-in seems to be preferred over a microphone input. What is your experience on this? thanks again!

 

[Ubbe]

Line in are for sure prefered as there's no voltage coming out of that and mic signals have some added amplifier and could also have some other negative effects to the signal. But sometimes the discriminator tap has a too low audio for the line jack to be used. On laptops it often are a single input that combines both line and mic and are set in the software if mic amplification should be used or not.

/Ubbe

 

[slicerwizard]

From what I've been reading a better option should be an external line-in input vs microphone, do you agree that a line input is superior to a microphone input for digital decoding?

No, it's not, despite what others are telling you. A line input has an impedance of 600 ohms or so. A discriminator circuit has an impedance of roughly 10k ohms. This means that to drive a line input, you have to excessively load down the scanner's discriminator circuit. The better approach is to use a mic input, which has a nice high input impedance, and then use a voltage divider (two resistors) to reduce the magnitude of the discriminator signal to match what the mic circuit can handle. Find an inexpensive USB sound device that has a mic input. They're dirt cheap. Here's one:

Aluminum USB Audio Adapter External Stereo Sound Card With 3.5mm Headphone And Microphone Jack

I like this from DX. Find the cool gadgets at a incredibly low price with worldwide free shipping here.

www.dx.com

About the second part you wrote, I am open to learning and experimenting but I'm quite new to technical modifications, so it will take me some time to understand it all. My background is in software development and so far I've only covered the basics of electronics when I received my general HAM license.

wbswetnam took the right approach - with an adjustable voltage divider, one can eyeball digital waveforms as well as monitor BER and dial it in.

 

[zerosix]

Line in are for sure prefered as there's no voltage coming out of that and mic signals have some added amplifier and could also have some other negative effects to the signal. But sometimes the discriminator tap has a too low audio for the line jack to be used. On laptops it often are a single input that combines both line and mic and are set in the software if mic amplification should be used or not.

/Ubbe

I can see from a logical standpoint why the microphone line would not be best since the amplification could alter the signal, and of course there is some voltage on it. However, the right resistor and cap would make sure the signal is not clipping/too strong if my understanding is correct. I also understand that line inputs often require more volume, the BCT15X might not provide adequate power to drive the line in, this might explain the people who complained their audio signal was extremely low with this scanner. This however might be fixable by using a lower capacity resistor if I am understanding correctly.

No, it's not, despite what others are telling you. A line input has an impedance of 600 ohms or so. A discriminator circuit has an impedance of roughly 10k ohms. This means that to drive a line input, you have to excessively load down the scanner's discriminator circuit. The better approach is to use a mic input, which has a nice high input impedance, and then use a voltage divider (two resistors) to reduce the magnitude of the discriminator signal to match what the mic circuit can handle. Find an inexpensive USB sound device that has a mic input. They're dirt cheap. Here's one:

Aluminum USB Audio Adapter External Stereo Sound Card With 3.5mm Headphone And Microphone Jack

I like this from DX. Find the cool gadgets at a incredibly low price with worldwide free shipping here.

www.dx.com

wbswetnam took the right approach - with an adjustable voltage divider, one can eyeball digital waveforms as well as monitor BER and dial it in.

I love how many here have different (conflicting) opinions on this, I am not judging in any way, but it's very interesting!

I can see why you would say the line input is not the best since the scanners discriminator circuit does not amplify the signal, the needed signal might not be there, putting a strain on the circuit. I also understand that you could lower the signal volume with resistors to a more appropriate level for the microphone input, but wouldn't the higher amplification/voltage on the microphone input distort the signal, even slightly?

It seems the general consensus from reading the replies in this thread and everywhere else on the internet is this:
(but please, anyone correct me if I'm wrong)

1. Values of the capacitor and resistor(s) vary based on input type used; line-in vs mic. in
2. Values also change based on the specific sound card input used (impedance)
3. There are those who say the microphone input will 'soften' or otherwise distort the sharp digital waveform because of the microphone amplification. Those people also say the microphone input is a higher potential danger to ruining the scanner circuit, however, from my understanding this could be mitigated by the resistor and cap.
4. There is another group of people who say line-in is better because it does not amplify/modify the signal as much as the mic. input but the overall signal volume is much lower, this means that the scanner might not have the necessary audio levels to drive the line input, this might result in some unwanted outcome. Maybe this could be overcome by using a 1K resistor instead of a 10K resistor, but i might be wrong here..

Conclusion so far;

I might be completely overthinking this all.. I just want to create the cleanest and best usable signal as possible for DSD. Especially with NXDN/Fusion the right signal quality is of great importance. DMR seems to be more forgiving in my own experience.

If anyone else has anything to add why I should/shouldn't use a certain input type, resistor/cap value I am all ears!

 

[slicerwizard]

With line in, you get weak audio that doesn't make full use of the sound input's ADC. Instead of 16 bit audio, you effectively get 8-10 bit audio. That doesn't help decoding. If you try to compensate by using lower resistor values to increase the volume, you load down the discriminator circuit. This is easily demonstrated with a Uniden scanner that exposes its "window voltage" through the serial programming API - when the scanner is properly centered on an RF signal, the window voltage value is 100, but if you then connect a line in discriminator tap that loads down the scanner's circuitry, the window voltage value shifts. It's a very clear sign that you're doing something wrong.

Mic inputs don't distort audio unless you do something wrong, like overdrive them or introduce excessive phase shifts or undesireable frequency responses with your component choices.


As far as caps go, I've already told you - bigger is better. Check out this audio waveform from a C4FM P25 LDU1 frame. It has a lot of consecutive zero bits, interrupted by only a few 1 bits. Looks to be about 100 symbols, so 200 bits, almost all set to zero. Those symbols are represented by a fairly constant voltage in the FM demodulated audio coming from my tapped 246T. Over time, that audio voltage is dropping slowly towards zero. Why is that? Because my tap has a large series capacitor and that long stream of zeroes is slowly charging the capacitor. That means the sound card sees the discriminator voltage minus the capacitor's slowly increasing voltage. With 2 level signals, this isn't a huge deal, as state transitions are easy to spot and the bits tend to stay on their own side of the zero line. Not so with the 4 level signals used by NXDN, DMR, P25, etc. You introduce any significant error voltage from an undersized cap and your decoding rate goes straight into the crapper.

A smaller capacitor charges faster and introduces larger errors. A line in pulls far more current than a mic input and more current charges the cap faster and trashes the signal. So, 100 symbols at 4800 baud is 21 milliseconds; in that time, a 24 Hz sine wave goes through a half cycle, i.e. from zero to maximum voltage and back to zero. So a tap circuit that had good frequency response down to 24 Hz could pass said sine wave unmolested. It couldn't even begin to pass that string of zeros. The comment above that says "12Hz is lower than strictly necessary (the lowest standard CTCSS tone is 67Hz)"? Not even close. Digital data needs audio paths that have good/flat frequency response down to a couple of Hz. The ideal is a true DC response, but we get by with what we can manage.

So build your tap and record the raw audio of some clear (non-encrypted) C4FM-modulated P25 Phase I voice calls and locate the string of zero bits at the start of the HDU frames; there should be one HDU frame at the start of every call. See if your tap passes the HDU test...

 

[zerosix]

With line in, you get weak audio that doesn't make full use of the sound input's ADC. Instead of 16 bit audio, you effectively get 8-10 bit audio. That doesn't help decoding. If you try to compensate by using lower resistor values to increase the volume, you load down the discriminator circuit. This is easily demonstrated with a Uniden scanner that exposes its "window voltage" through the serial programming API - when the scanner is properly centered on an RF signal, the window voltage value is 100, but if you then connect a line in discriminator tap that loads down the scanner's circuitry, the window voltage value shifts. It's a very clear sign that you're doing something wrong.

Mic inputs don't distort audio unless you do something wrong, like overdrive them or introduce excessive phase shifts or undesireable frequency responses with your component choices.


As far as caps go, I've already told you - bigger is better. Check out this audio waveform from a C4FM P25 LDU1 frame. It has a lot of consecutive zero bits, interrupted by only a few 1 bits. Looks to be about 100 symbols, so 200 bits, almost all set to zero. Those symbols are represented by a fairly constant voltage in the FM demodulated audio coming from my tapped 246T. Over time, that audio voltage is dropping slowly towards zero. Why is that? Because my tap has a large series capacitor and that long stream of zeroes is slowly charging the capacitor. That means the sound card sees the discriminator voltage minus the capacitor's slowly increasing voltage. With 2 level signals, this isn't a huge deal, as state transitions are easy to spot and the bits tend to stay on their own side of the zero line. Not so with the 4 level signals used by NXDN, DMR, P25, etc. You introduce any significant error voltage from an undersized cap and your decoding rate goes straight into the crapper.

A smaller capacitor charges faster and introduces larger errors. A line in pulls far more current than a mic input and more current charges the cap faster and trashes the signal. So, 100 symbols at 4800 baud is 21 milliseconds; in that time, a 24 Hz sine wave goes through a half cycle, i.e. from zero to maximum voltage and back to zero. So a tap circuit that had good frequency response down to 24 Hz could pass said sine wave unmolested. It couldn't even begin to pass that string of zeros. The comment above that says "12Hz is lower than strictly necessary (the lowest standard CTCSS tone is 67Hz)"? Not even close. Digital data needs audio paths that have good/flat frequency response down to a couple of Hz. The ideal is a true DC response, but we get by with what we can manage.

So build your tap and record the raw audio of some clear (non-encrypted) C4FM-modulated P25 Phase I voice calls and locate the string of zero bits at the start of the HDU frames; there should be one HDU frame at the start of every call. See if your tap passes the HDU test...

View attachment 80935

Thanks slicerwizard for your in-depth information. I really enjoy your explanation! I bought a high-end soundcard earlier today but from your earlier comment I could have gone with a cheap usb interface with a mic input.

since you’ve explained that a bigger cap is better, ill go with a 10uF, from what I’ve read I might want to use a bipolar one. I will drive to an electronics store tomorrow or the day after tomorrow to get some parts, including a variable resistor. i also got a new multimeter today so I can do some better measurements in the future. One last thing; when I use the microphone input and make a voltage divider (1 fixed + 1 variable resistor) that wouldn’t interfere with the audio quality right? Reason im asking is because of this statement;
Pro92b said the following:

”To transfer a wideband signal to the sound card the tap resistor needs to be reduced to 1K or less.}”

ive also seen this suggestion in various other places. Does it mean the resistors will cut-off certain frequencies or are they only saying that resistors should be used to get the audio volume to the optimal level for the mic input? (I think the latter is correct)

I will defintely do the test you suggested, although I don’t have P25 in my area, but many other digital modes are present.

 

[slicerwizard]

One last thing; when I use the microphone input and make a voltage divider (1 fixed + 1 variable resistor) that wouldn’t interfere with the audio quality right? Reason im asking is because of this statement;
Pro92b said the following:

”To transfer a wideband signal to the sound card the tap resistor needs to be reduced to 1K or less.}”

ive also seen this suggestion in various other places. Does it mean the resistors will cut-off certain frequencies or are they only saying that resistors should be used to get the audio volume to the optimal level for the mic input? (I think the latter is correct)

I think they're saying that a line input needs a small resistor (which can overload the discriminator circuit and definitely reduces the DC response), otherwise the audio getting to the PC is very quiet. It's a can of worms.

 

[zerosix]

I tapped a Whistler 1025 one time, super easy to do since it had a tapping point. I used a 5K resister in series with a 0 to 5K variable resister mounted on the side of the radio so that I could adjust the signal as necessary.

so, a potentiometer can work as a rheostat (variable resistor) or as a voltage divider; did you wire it with two pins (variable resistor)or three pins (voltage divider)? which is the correct way? And couldn’t I just use a 0 to 10K potentiometer? From reading your old post you even used a 100K one and it worked.

I think they're saying that a line input needs a small resistor (which can overload the discriminator circuit and definitely reduces the DC response), otherwise the audio getting to the PC is very quiet. It's a can of worms.

Got it, thanks! Also same question to you; a potentiometer can work as a rheostat (variable resistor) or as a voltage divider; do I need to wire it with two pins (variable resistor) or three pins (voltage divider)? which is the correct way? And couldn’t I just use a10K potentiometer instead of a 5K resistor + 5K potentiometer? Do you recommend a 10K potmeter or would higher be better?

 

[jonwienke]

If you're going into the mic input, 5+5 would be better to limit the signal going to the mic.

 

[slicerwizard]

Also same question to you; a potentiometer can work as a rheostat (variable resistor) or as a voltage divider; do I need to wire it with two pins (variable resistor) or three pins (voltage divider)? which is the correct way? And couldn’t I just use a10K potentiometer instead of a 5K resistor + 5K potentiometer? Do you recommend a 10K potmeter or would higher be better?

2 pins. I used two pots to experiment and find good values for both resistors, then substituted fixed 1/8w resistors. The top of the divider will be a high value (maybe 20k to 50k) while the bottom will be much smaller.