Squelch tail very long with KG1000G repeater

[wqpn591]

I've built a repeater using dual KG1000G units and a Sinclair duplexer.

It seems to work just fine except that most of the time (not every time) the Tx continues 15-20sec after the end of PTT transmission.

My testing is with Wouxun kg905g HT units.

Occasionally the squelch tail is just a second or two. More reasonable.

Anyone have a tip or reason on how to shorten the squelch tail?

 

 

[wqpn591]

I did follow this recommendation to eliminate the Squelch tail but no real change.

Wouxun Archives - Buy Two Way Radios

 

[OffRoaderX]

I built a KG-1000G repeater a few months ago and had the opposite issue - by default there is no squelch tail/kickback and IIRC the only way to turn it on was in the software (could not do it on the radio).  

Are you sure it's the tail that you're getting and not something else?

[mbrun]

I've built a repeater using dual KG1000G units and a Sinclair duplexer.
It seems to work just fine except that most of the time (not every time) the Tx continues 15-20sec after the end of PTT transmission.
My testing is with Wouxun kg905g HT units.
Occasionally the squelch tail is just a second or two. More reasonable.
Anyone have a tip or reason on how to shorten the squelch tail?
 
 

Squelch tail duration is adjustable on the radio, but it should not vary arbitrarily. Since it seems to be varying I suspect some other factor(s) are coming in to play. As Offroader mentioned, that setting is adjustable, but it is accessible only via software. The adjustment range is 0-5sec. The manual calls it “Repeat Hold Time”

Make sure sure you are using CTCSS or DCS on the KG1000G unit you are using as the receiver so that the transmitter will be activated only when the receiver detects them. If you do not have tones set on the receiver then interfering noise can be picked up, the squelch will open and thus cause the transmitter to continue to Tx.

Pay close attention to the Rx light on the repeater receiver when you are experiencing the long tail. Is the Rx light on the whole time, even when your radio is not transmitting? It it is, that would indicate some other signal is holding the Receiver squelch open and thus causing the Transmitter to continue longer than desired.

If you are already using CTCSS or DCS codes, consider trying different ones for Rx and Tx codes to see if it yields a different behavior. The duplex tuning could also be in question. Leakage from Tx to Rx side could perhaps manifest itself in strange ways.

Just food for thought.


Michael
WRHS965
KE8PLM

[gortex2]

I would check for desence. If low power eliminates the issue that's most likely your issue.

 

[wqpn591]

I do use CTCSS codes. Duplexer is specifically tuned to my freq of use. the test results show -82dB attenuation from Tx to Rx.

I'll try lower Tx power and see if that helps.

The Rx light does stay on the whole tail time.

 

[wqpn591]

Well guys, I think I got this figured out. With your help. I went to try two things.

Key the Tx unit and see if the Rx unit activates. And it did. Just when I was reaching over to adjust the power level down I noticed that the coax connector on the Tx unit was not fully tight. Tightened that up and bingo...........no more activating when the Tx unit is keyed up. Must have had some leakage detected by the Rx unit.

And the tail is completely gone now. Maybe because I turned RPT-TONE off

which brings up the question. what do you guys set the Repeat-hold-time to? The description in the manual is a bit confusing to me. It's unclear if the "PTT on the radio" is the kg1000g Tx unit or an external unit like an HT transmitting to the repeater. 

How do you guys interpret this?

Repeat Hold Timer (Programming Software Only)
The Repeating Hold Timer is used to prevent the PTT from being used to transmit too frequently. When the PTT on the radio is released after transmitting in repeater mode, the hold time prevents the unit from transmitting for a predermined interval while waiting for a response.  If no valid QT/DQT is detected within the hold time, the transmitter will release the hold and allow the PTT to transmit. The repeating hold timer sets the hold time for the transmitter to resume transmitting 100ms-5000ms (5 seconds) after the received QT/DQT signal disappears.  This function is only accessible and configured through the KG-1000G programming software.

 

 

[mbrun]

… I noticed that the coax connector on the Tx unit was not fully tight. Tightened that up and bingo...........no more activating when the Tx unit is keyed up. …

And the tail is completely gone now. Maybe because I turned RPT-TONE off
which brings up the question. what do you guys set the Repeat-hold-time to? The description in the manual is a bit confusing to me. It's unclear if the "PTT on the radio" is the kg1000g Tx unit or an external unit like an HT transmitting to the repeater. 
How do you guys interpret this?
Repeat Hold Timer (Programming Software Only)
The Repeating Hold Timer is used to prevent the PTT from being used to transmit too frequently. When the PTT on the radio is released after transmitting in repeater mode, the hold time prevents the unit from transmitting for a predermined interval while waiting for a response.  If no valid QT/DQT is detected within the hold time, the transmitter will release the hold and allow the PTT to transmit. The repeating hold timer sets the hold time for the transmitter to resume transmitting 100ms-5000ms (5 seconds) after the received QT/DQT signal disappears.  This function is only accessible and configured through the KG-1000G programming software.
 
 

Glad you figured out the issue.

As far as repeater tail length, you may receive a variety of personal preference opinions. As the owner that is up to you. I have worked repeaters with zero to 5-seconds and more of tail. If you include no tail then it makes for more guess work on the part of folks trying to figure out when they have had success connecting to the repeater. If you include a tail, it gives users some confidence they had enough signal to open the repeater.

For an etiquette standpoint some senior radio enthusiasts long ago advised me of the following (and I paraphrase): When two or more parties are in a radio conversation they should all let the repeater squelch tail drop before keying up so as to give opportunity for a third party break in as they need to. For this purpose I personally think 3 seconds is sufficient and more than 5 seconds annoying.

I agree that the wording in the manual is less clear than it could be. It is not 100% clear whether the setting in the Rx or Tx portions of the repeater is the one that actually gets used while in repeater mode. If you set both the same it should not make a difference. To know which is actually being used, set the value in one unit and test. Then set the value in the other unit and test. Based on the wording “If no valid QT/DQT is detected within the hold time”, I speculate it is the radio that serves as the Rx function in the repeater is the one that actually gets used in operation. But that is just speculation. After you test and confirm which one is being used, please report back.

Congrats on getting the repeater operational.


Michael
WRHS965
KE8PLM

[wqpn591]

See next comment

[wqpn591]

I got my programming cable and did change the setting in the radio. it seems to be called "Hold time of Repeater". Set it to 1000ms on the Tx unit and now I have an intentional squelch tail.  BTW, it did not do anything to also put it in the Rx unit. So only needs to be in the Tx unit.

As it turns out, when I was testing prior, I had left the Tx unit on low power and the Rx unit does not breaks squelch when the Tx unit is keyed up.

But with the Tx unit id on M or H power levels, indeed the Rx unit hears it.  I disconnected the Tx cable to the duplexer and terminated into a 50ohm load and Rx unit does not hear it. So still getting some leakage through the duplexer.

I put a preamp unit in line between the duplexer and Rx radio and left it powered down to get some insertion lost.

That seems to allow me to Tx on max power without the Rx unit hearing it. But not sure about the hit to my Rx sensitivity. have to do some field testing.

Was thinking about taking a second duplexer and tune it specifically to give me further filtering to the Rx unit. 

Thoughts?

 

[n1das]

On 9/16/2021 at 9:42 PM, wqpn591 said:

I've built a repeater using dual KG1000G units and a Sinclair duplexer.

It seems to work just fine except that most of the time (not every time) the Tx continues 15-20sec after the end of PTT transmission.

My testing is with Wouxun kg905g HT units.

Occasionally the squelch tail is just a second or two. More reasonable.

Anyone have a tip or reason on how to shorten the squelch tail?

 

 

Is the "squelch tail" you are referring to the hang time of the repeater transmitter or the receiver staying unsquelched for an extended period of time after a signal into the receiver went away?  These are 2 different things.

Hang time refers to how long the repeater's transmitter remains keyed after the receiver's squelch closes to mute the audio.  I like repeater hang times in the range of 2 to 4 seconds in length.

Squelch tail technically refers to how long the receiver's squelch remains open after a received signal goes away.  This is the short burst of white noise heard when a signal input to the receiver goes away before the squelch closes to mute the audio.  With plain old carrier squelch (CSQ), this time is typically a few 10s of milliseconds and up to about 50ms.  Time constants longer than about 50ms start to be annoying because of the longer "crash" sound heard due to the squelch staying open longer.  With a CTCSS/DCS decoder, the time constant is much longer and is spec'd at 350ms.  Motorola repeaters running CTCSS or DCS typically run the CSQ wide open and only the CSCSS/DCS decoder mutes the audio from the receiver.  This works best for handling mobile flutter and signal fading but can create an annoyingly long squelch tail because of the 350ms time constant.  Commercial radios typically transmit a "reverse burst" designed to eliminate the squelch "crash" sound aka squelch tail at the end of a transmission.  The reverse burst means "Close squelch NOW!" before the transmitter un-keys.  Radios talking through the repeater that don't generate a reverse burst at the end of their transmission leave a long squelch tail aka crash sound heard at the end of a transmission.

Are you using CTCSS/DCS on the repeater's receiver?  Are you also using CSQ or leaving it wide open like what Motorola does?  These two functions typically are AND'd together in a repeater application.  This requires CSQ to be open AND decoding a valid CTCSS tone or valid DCS code to unsquelch the receiver and key PTT on the repeater.

CTCSS/DCS access and CSQ can also be set to "OR" the two functions together to unsquelch the receiver in some radios and repeaters.  Check to make sure the two functions are AND'd together instead of OR'd together.  An easy test you can do is transmit into the repeater without a CTCSS tone or DCS code or with a different tone or code that the repeater requires and see if this keys up the repeater.  If you can key the repeater by just opening CSQ on the receiver, this may be the problem.

 

[mbrun]

I got my programming cable and did change the setting in the radio. it seems to be called "Hold time of Repeater". Set it to 1000ms on the Tx unit and now I have an intentional squelch tail.  BTW, it did not do anything to also put it in the Rx unit. So only needs to be in the Tx unit.
As it turns out, when I was testing prior, I had left the Tx unit on low power and the Rx unit does not breaks squelch when the Tx unit is keyed up.
But with the Tx unit id on M or H power levels, indeed the Rx unit hears it.  I disconnected the Tx cable to the duplexer and terminated into a 50ohm load and Rx unit does not hear it. So still getting some leakage through the duplexer.
I put a preamp unit in line between the duplexer and Rx radio and left it powered down to get some insertion lost.
That seems to allow me to Tx on max power without the Rx unit hearing it. But not sure about the hit to my Rx sensitivity. have to do some field testing.
Was thinking about taking a second duplexer and tune it specifically to give me further filtering to the Rx unit. 
Thoughts?
 

I hear forward progress, so that is good news.

Since there is 10dB difference between the low-power (5w) and high-power (50w) settings, and your testing seems to indicate lower power works and anything above that does not, it would seem to indicate you are in need of 10dB or more improvement in your Tx-Rx isolation. That improvement can come from either improved tuning of the duplexer if it is is not already tweaked to its maximum or you can switch to a high-quality duplexer. Duplexers can drift during transportation and with temperature shifts as I understand them. Perhaps you can connect with a local radio shop that can check the existing one out before you opt for a replacement. You may also need them to tweak the new one for you.

Good job BTW in your isolation test.

I believe your question regarding a second duplexer is where it would be viable to install it in the return path of between the receiver and primary duplexer so that you can take of the additional attenuation it provides? Correct? I will admit if I had one laying around I would certainly try that for experimentations sake. But if both units are lower-end units I would expect higher overall insertion losses leading to performance compromises. If I new for certain that the existing duplex was already peaked for its best performance already I believe I would cut my losses and invest in a single unit rated with the performance I needed. Perhaps others here may have some additional suggestions for you.

On a different note. N1DAS brought up squelch tail. I too have and continue to use the term ‘squelch tail’ when a transmitting radio un-keys and when my local radio mutes (squelches). Years ago I learned that early CTCSS operation was achieved in part by the use of a mechanical reed switch that physically resonated. When the transmitter un-keyed the reed-switch continued to resonate for a period during which it had the effect of keeping the receiver unmuted for a short time even though Tx had stopped. This allowed noise into the receiving radio. Because this phenomena affected both simplex and duplex communication it stuck with me and has stayed even though nowadays the hold time on the repeater is an intentional and user adjustable parameter. While an official squelch tail and hold time are two different things, they both have the net affect of keeping one’s receiver unmuted (un-squelched) for a brief time after the transmission of concern has ended. N1DAS, thanks for the clarification.


Michael
WRHS965
KE8PLM

[wqpn591]

Thanks for the clarification between repeater hang time and squelch tail. 

I am using CTCSS on the Rx and Tx devices in the repeater so CSQ should not be a factor.

Its actually performing pretty well now, as far as the repeater hang time being predictable, as long as the Rx unit does not re-key as a result of Tx signal getting back to the Rx input.  My Sinclair MR356-N-2 was tuned to my frequencies. And since my nano-VNA unit is less accurate than the equipment they probably used, cant really verify without taking to someone with good test equipment. 

The test results sheet provided with the Sinclair show 82dB of attenuation

Since even Medium power (20W (43dBm))  triggers the Rx but low (5W (37dBm)) does not, seems like only 6dB of attenuation would do the trick.
can easily get that by using my XLT duplexer tuned as a high pass filter. 
Then I may be able to turn on the 12dB preamp to improve my intended receive signal.  

The receive sensitivity of the KG1000-G in the datasheet is 0.25uV which is about -119dBm.  Not sure what it takes to break squelch. But with 82dB of attenuation of a 37dBm signal, I'm guess its North of -45dBm. if it can even be calculated this way.

 

[mbrun]

I had done some testing previously on what it takes to break squelch on a couple of different KG-1000G units. What I found was that they would break squelch with an input signal of around -125 to -123dBm. They would show all 9 bars on the meter when the signal level exceeded that break squelch point by approx 12dBm. In addition, when I performed some effectively sensitivity measurements on the radios while in my shack using two different antenna setups, I also found that 12dBm value to be the magic number for how much more the the desired signal needs to be above the noise or interference level in order for the radio to show full bars. As a result of this, I am currently associating full quieting or near full quieting with all nine bars being lit solid. Of course this is not being done with lab grade equipment so absolute values are certainly in question, but the relevant levels are useful nonetheless.


Michael
WRHS965
KE8PLM

[gortex2]

So have you tested this into a suitable dummy load on high and low power ? If your still getting desense it could be the jumpers between the duplexer and radio. If it does not do it with a dummy load check the antenna and antenna line. I have seen antenna's with high reflected cause different issues. Also how far away is your test setup from the radios ? If your using an HT in front of your bench its very possible the HT is getting into the mobile radio. This happens alot with non shielded radios. Mobiles were not designed for repeater use so not all are properly shielded for high RF around the box.

 

[wqpn591]

Still doing some debug.

Given the non-shielded radios theory, I'm going to focus on that next.

Here is a video of what the units do.   https://1drv.ms/v/s!ApcwUuoTYXaIhMB874A38r6NDNsX7w?e=rmYu3r

and a pic of the box.

[wqpn591]

Did some more experiments today.

Changed out all coax cable to better RG400 cables.  Results: no change

disconnected antenna and connected a 50ohm load. Results: no change. So its not the antenna itself.

Put ferrite suppressors on all power lines to radios and AC power to power supply. Results: no change.
but have some better material #61 ferrites on order to better suppress high freq.

Measured insertion loss of my powered off preamp in line with Rx radio. That is 26dB loss without power and 14dB gain with power. So somewhere between 0 and 26dB attenuation  stop the re-keying of the Rx unit. 

Put heavy duty aluminum foil over Tx and Rx unit. Results: no change. Not sure if those cases are metal or not.

I need to find an RF meter to measure what is coming from duplex to Rx radio. Which is on a different frequency so this whole thing is strange.

 

[gortex2]

I would remove all those adapters. I assume your tests are with the cabinet closed ? Without a real service monitor its hard to do a true test but it appears the TX is still getting in the RX. I doubt its a power issue. I assume the CAT5 is not shielded so I'd start with some ferrites on that. 

[wqpn591]

Did one more experiment. Took the receive signal from the duplexer and instead of going into my Rx KG1000G, I connected it to a HT unit tuned to receive on my Rx freq.

Sure enough, that one keys up also.

I'm about 99% convinced that I'm getting leakage through the duplexer.

[wqpn591]

23 minutes ago, kb2ztx said:

I would remove all those adapters. I assume your tests are with the cabinet closed ? Without a real service monitor its hard to do a true test but it appears the TX is still getting in the RX. I doubt its a power issue. I assume the CAT5 is not shielded so I'd start with some ferrites on that. 

Yeah that CAT5e is shielded with shielded ends. Tried a low freq ferrite on that too. No change.

[mbrun]

Did one more experiment. Took the receive signal from the duplexer and instead of going into my Rx KG1000G, I connected it to a HT unit tuned to receive on my Rx freq.
Sure enough, that one keys up also.
I'm about 99% convinced that I'm getting leakage through the duplexer.

Do you have any way to measure the signal level coming out of the duplexer destined for the Rx radio while the Tx radio is operating? I am curious to know how much of that 467 signal is actually making it into the receiver.


Michael
WRHS965
KE8PLM

[BoxCar]

1 hour ago, wqpn591 said:

Yeah that CAT5e is shielded with shielded ends. Tried a low freq ferrite on that too. No change.

Shielding the cable is worthless unless the shield is grounded. You have to have the shielded sockets and those must be connected to a ground. In LAN cabling systems the ground is through the cable to the patch panel and then through the rack which is connected to electrical ground.

[wqpn591]

1 hour ago, mbrun said:

Do you have any way to measure the signal level coming out of the duplexer destined for the Rx radio while the Tx radio is operating? I am curious to know how much of that 467 signal is actually making it into the receiver.


Michael
WRHS965
KE8PLM

I do not. Unfortunately. I have a nano-VNA but dont want to damage port 2 trying to measure it.

However, I did take my XLT duplexer and tuned Ant-High side port to filter the Tx freq. Took out my broad band attenuator and put that in.   Made a huge difference in my receive sensitivity. Instead of my receive signal getting attenuated by 26dB, its now only 2dB but the Tx freq gets attenuated 38dB.  Works like a champ. Did a little field testing and this works in places I figured it would not do to some rolling hills. Its a work around until I can do something with that Sinclair Duplexer.

 

[wqpn591]

1 hour ago, BoxCar said:

Shielding the cable is worthless unless the shield is grounded. You have to have the shielded sockets and those must be connected to a ground. In LAN cabling systems the ground is through the cable to the patch panel and then through the rack which is connected to electrical ground.

Agreed. Since the cable that came with it was shielded and had shielded RJ45, figured the socket was grounded. But did not look.