Re: K4 Latency Time
Hi Wayne,toggle quoted messageShow quoted text
I use the same pulse generator that I use to test how a receiver's AGC handles impulse noise. The K3 is one of the few rigs where the AGC doesn't get captured by an impulse, like an electric fence. (This assumes the operator has this AGC feature turned on in the menu.)
The IC-756 Pro III was the last Icom rig to handle impulse noise properly, and no current Japanese rig from the big three OEMs works properly. No analog rig has this problem at all. I believe the Apache radios also handle impulse noise properly from an AGC standpoint. Now the latency info you wanted!
Unfortunate I don't currently have any K3 latency data. I need to correct this oversight.
I have both an HP pulse generator and a Tektronix, either set for a 10ns rise time. The duration of the impulse doesn't matter much, 500ns to 1ms is OK. The output is set for full output (10 to 16 volts) followed by a 20 dB pad. This provides the best signal to noise of the generator output stage vs. the pulse. This is more important when making the RX AGC test than making the latency test. It is convenient if you can set the rep rate to once every 5 or 10 seconds so the scope can be set to one-shot and ignore the following impulses.
The output of the 20 dB pad feeds a hybrid splitter, such as a Mini-Circuits ZSC-2-1. One output goes to the receiver antenna input and the other output goes to the external trigger of the digital storage scope. The audio output of the receiver goes to the vertical input of the scope.
You want to measure the time delta between the trigger point and the audio pulse, which will have ringing. I pick the first full amplitude point, as per the PDF on my website. If you look at the first picture in the PDF for the IC-7300 at 2400 Hz bandwidth, the cursor for the trigger point is shown on the left, and the audio pulse on the right is where the other digital cursor is set. The time delta in this case is 3.66ms. If on the other hand you look at the Flex 6600M set for the best filter shape factor (highest latency), the time delta is 170ms.
For a frame of reference, my pure analog IC-781 has a latency with a 2700 Hz bandwidth of 1.18ms and on CW with a 500 Hz bandwidth latency is 5.9ms. The IC-7610 with a 2400 Hz bandwidth has a latency of 7.96ms and on CW with a 500 Hz bandwidth the latency is 12.65ms.
What gets really interesting is using a radio over the internet with its added latency. Let's say two brand X radios had a latency of 100ms, and were working each other in a contest. The total latency could well be over 300ms. A contest over a year ago, where all the operators were remote, resulted in the US station consistently calling when the DX was transmitting.
From: Elecraft-K3@groups.io [mailto:Elecraft-K3@groups.io] On Behalf Of N6KR
Sent: Tuesday, February 04, 2020 1:15 PM
Subject: Re: [Elecraft-K3] K4 Latency Time
What are you using as the stimulus for these tests? We'd like to reproduce this in our lab.
On Feb 4, 2020, at 12:07 PM, Rob Sherwood <rob@...> wrote: