Hi Folks,
I recently upgraded from 100D to 7000DLE. The 7000 came up easy and I am a full-time LAN remote operator. Recently, I completed the audio setup and have spent a good bit of screen time during those adjustments. During this screen time is when I noticed an anomaly where the last 5-25 ms of the xmit audio reappears on the receive screen. The speech spurt is also heard. I've made a video of the panadapter keying the transceiver without audio input (mic disconnected) and with normal audio preceding the release of PTT.
Here is the Google Drive link to the video: https://drive.google.com/file/d/1GX_9ZG ... p=drivesdk
Thoughts Welcome,
dave
wa3gin
Odd XMIT Artifact after PTT released
-
- Posts: 23
- Joined: Sun Apr 09, 2017 6:37 pm
- Contact:
Re: Odd XMIT Artifact after PTT released
Dave,
You need to post a public link to your video, or make the link you already posted publicly accessible, we can't see it.
73,
Scott
You need to post a public link to your video, or make the link you already posted publicly accessible, we can't see it.
73,
Scott
Re: Odd XMIT Artifact after PTT released
can't access the video either, but what are your PTT Delays set to under General > Options?
Gary NC3Z
Re: Odd XMIT Artifact after PTT released
Hi Dave,
I can see the video now, not sure if others can see it. What you are seeing seems to be an affect caused by latency (delay) in the signal processing chain, as you suspect. However, it is most unusual.
It appears that you are using MOX for PTT control, is that correct?
The artifact is almost certainly caused by latency on the receive side. On the transmit side de-asserting PTT will cause a near immediate cessation in the TX state (there is a 25ms slew down factor in the code), but RX will not start until the slew is finished and both MOX delay and RF delay factors elapse. Even if you are using a long (large) linear delay TX filter any information left in that filter will simply be lost. Hence RF is truly off at the antenna when RX starts.
On the receive side I would not expect to see this effect unless you are using a long, linear phase RX filter. The time it takes to traverse the linear phase filter as implemented in PowerSDR is filter size divided by 48000. For example, if you are using linear phase filter of size (length) 4096, that would equate to .0853 seconds, or 85.3ms. Since the radio is full duplex, you will hear whatever is left in that filter when switching back to RX with an RX delay setting of 0ms, which is the default.
Post screenshots of the following settings tabs so we can see what your current settings are: Setup > General > Options and Setup > DSP > Options.
73,
Scott
P.S. You should upgrade to PowerSDR 3.4.9, and use the built in audio processing features instead of that IHY box.
I can see the video now, not sure if others can see it. What you are seeing seems to be an affect caused by latency (delay) in the signal processing chain, as you suspect. However, it is most unusual.
It appears that you are using MOX for PTT control, is that correct?
The artifact is almost certainly caused by latency on the receive side. On the transmit side de-asserting PTT will cause a near immediate cessation in the TX state (there is a 25ms slew down factor in the code), but RX will not start until the slew is finished and both MOX delay and RF delay factors elapse. Even if you are using a long (large) linear delay TX filter any information left in that filter will simply be lost. Hence RF is truly off at the antenna when RX starts.
On the receive side I would not expect to see this effect unless you are using a long, linear phase RX filter. The time it takes to traverse the linear phase filter as implemented in PowerSDR is filter size divided by 48000. For example, if you are using linear phase filter of size (length) 4096, that would equate to .0853 seconds, or 85.3ms. Since the radio is full duplex, you will hear whatever is left in that filter when switching back to RX with an RX delay setting of 0ms, which is the default.
Post screenshots of the following settings tabs so we can see what your current settings are: Setup > General > Options and Setup > DSP > Options.
73,
Scott
P.S. You should upgrade to PowerSDR 3.4.9, and use the built in audio processing features instead of that IHY box.
-
- Posts: 23
- Joined: Sun Apr 09, 2017 6:37 pm
- Contact:
Re: Odd XMIT Artifact after PTT released
This is a temporary ops position. There is construction is taking place where the primary ops position is located. The mic cord of the Shure 55S is too short to reach PC so used the W2IHY box for RCA jack and output to PC with compatible 1/8" stereo jack.
Here are the screen shots:
Thanks
Here are the screen shots:
Thanks
Re: Odd XMIT Artifact after PTT released
Well, you can't do any better than the low latency filter setting on phone. That exhibits approx. 20mS of delay at any length setting (because it is a minimum phase FIR filter, not a linear phase FIR filter).
The defaults for RX, MOX, RF and PTT delay are 0, 10, 30 and 0, in that order. Although there doesn't seem to be anything wrong with your values you might try going back to defaults and see if that removes the burp.
Just for grins, what is the ping delay from the RF hardware to the PC? It can't be more than 1mS, but I suppose you never know.
The defaults for RX, MOX, RF and PTT delay are 0, 10, 30 and 0, in that order. Although there doesn't seem to be anything wrong with your values you might try going back to defaults and see if that removes the burp.
Just for grins, what is the ping delay from the RF hardware to the PC? It can't be more than 1mS, but I suppose you never know.
-
- Posts: 23
- Joined: Sun Apr 09, 2017 6:37 pm
- Contact:
Re: Odd XMIT Artifact after PTT released
Hmmm,
For some reason I couldn't snip the command prompt window but the results are not great:
round trip Minimum = 4ms, MAX = 4ms, Average = 4ms
Sometimes average = 3ms other times 6ms
thanks
For some reason I couldn't snip the command prompt window but the results are not great:
round trip Minimum = 4ms, MAX = 4ms, Average = 4ms
Sometimes average = 3ms other times 6ms
thanks
Re: Odd XMIT Artifact after PTT released
Sounds like you have a lousy switch or maybe bad cables?
Do you know that everything is running at Gigabit speeds?
People should be considering Cat 8 cables too since they are all shielded and can help all sorts of things (better speed and reduced noise injection into your shack).
Mike
Do you know that everything is running at Gigabit speeds?
People should be considering Cat 8 cables too since they are all shielded and can help all sorts of things (better speed and reduced noise injection into your shack).
Mike
Mike W9MDB
-
- Posts: 23
- Joined: Sun Apr 09, 2017 6:37 pm
- Contact:
Re: Odd XMIT Artifact after PTT released
Thanks Mike,
I'm using multi-mode fiber. Have a 3' CAT cable between fiber to ethernet converter to router. Same on the host side. I'm not seeing throughput demands anywhere near 100 Mbps. I suspect the latency is in the converter or router or switch or all three combined.
73
I'm using multi-mode fiber. Have a 3' CAT cable between fiber to ethernet converter to router. Same on the host side. I'm not seeing throughput demands anywhere near 100 Mbps. I suspect the latency is in the converter or router or switch or all three combined.
73
-
- Posts: 23
- Joined: Sun Apr 09, 2017 6:37 pm
- Contact:
Re: Odd XMIT Artifact after PTT released
Update on the network performance. I have been using an excellent little USB TP-link wireless dongle for the remote PC. It provided the same throughput as wired Ethernet. However, after performing a ping test between the dongle and the hardwire... I could see at times where the latency on the wireless dongle was up to 40+ms while the Ethernet was 0ms to max of 2ms with average =0 So, I will run a dedicated Ethernet to the remote PC. The switch over to Ethernet did not have any effect upon the artifact. That I believe is just audio buffers that have not been completely emptied... I suppose I could spend hours tinkering with buffer size settings but since no artifact is being transmitted and it is a 99% visual artifact on receive. I don't see an issue. 73