HL2 - heatsink - temperature question

[Salim Hameed]

I received my HL2 towards the end of March 2021. I have been running it continuously on 4 receivers 24x7 from that point onwards (@ N8LI) using SparkSDR on Linux decoding FT8. I am truely impressed by the stability and reliability of the setup. The unit is going to be used for remote operation from a site which can be physically accessed only once in a year (the reason for testing like this).

I noticed after around 40 days that HL2 started not responding after sometime (approximetly around a few minutes after switching on - had to be switched on and off to bring it online). I found that on SparkSDR UI the temperature (?) was creeping up (to about 28C) before it shut down. I opened up the box and re-seated the stick-on heatsink (FPGA) and it went back to normal (temperature went down gradually to settle to around below 22C).

It seems to be repeating after every few weeks of continuously operating.

Is there a recommended way to fix this (like applying mechanical pressure on the heatsink by putting something in between the cover and the heatsink ) ?  

73 de salim VU2LID / N8LI

[Anthony N1IG]

is adding an internally mounted fan an option?

[Marcos X]

--
You received this message because you are subscribed to the Google Groups "Hermes-Lite" group.
To unsubscribe from this group and stop receiving emails from it, send an email to hermes-lite...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/hermes-lite/8a11365e-b96d-4959-90ed-1838c3bdd261n%40googlegroups.com.

[Marcos X]

[Alan Hopper]

Hi Salim,

28C is not very high and I doubt it is the cause for the failure unless the actual temperature gets very much higher.  I'm intrigued by the once per year access and am keen to make Spark support this sort of operation, maybe as a fall back spark could trigger a relay to power the radio down and up.  What is the network setup?

73 Alan M0NNB

[ron.ni...@gmail.com]

Several vendors make ethernet-controlled power switches, some with built-in network keep alive monitors and daily power-cycle timers.  I have one that I plan to use to allow me to reboot a remote Raspberry Pi setup.

73,

Ron

n6ywu

[Salim Hameed]

The main issue I am trying to solve is - it looks like - in my case - the stick on heatsink is coming loose (?) after a long duration of continuous operation.

If the heating conducting remains as good as how it was  initially it will continue working (since after putting pressure on and reaseating the heatsink the temperature went down).

On Tuesday, 25 May 2021 at 11:25:30 UTC-4 Anthony N1IG wrote:

is adding an internally mounted fan an option?

[Salim Hameed]

Hi Alan, Around 28C is what I noticed last on the UI (the ambient temperature, where it is being tested right now is around 16C). Since others have reported much higher temperatures I am suspecting that the actual temperature may be a different/higher value.

Inside the shack internal networking has gigabit connections through a switch. Depending on the availability (some computers in the shack may be down or unavailable - all run some version of Linux), hardware through with the XCVR is controlled may change.    The switch is connected to the outside world through a 4G modem and a tunnel through it.

It will be very useful if SparkSDR can provide a trigger like you mentioned.

73 de salim vu2lid / n8li

[Salim Hameed]

Thanks, due to the way the setup evolved over the years, right now I use more than one computer to monitor activities of others, with arrangement to reboot through connected arduinos and relays.

73 de salim vu2lid / n8li

On Tuesday, 25 May 2021 at 20:42:12 UTC-4 ron.ni wrote:

Several vendors make ethernet-controlled power switches, some with built-in network keep alive monitors and daily power-cycle timers.  I have one that I plan to use to allow me to reboot a remote Raspberry Pi setup.

73,

Ron

n6ywu

On Tuesday, May 25, 2021 at 1:22:54 PM UTC-7 ahop wrote:

Hi Salim,

28C is not very high and I doubt it is the cause for the failure unless the actual temperature gets very much higher.  I'm intrigued by the once per year access and am keen to make Spark support this sort of operation, maybe as a fall back spark could trigger a relay to power the radio down and up.  What is the network setup?

73 Alan M0NNB

[Steve Haynal]

Hi Salim,

I agree with Alan that this doesn't seem like a heat issue. What is the temperature registered right before failure? Can you still ping the unit after failure? What are you ping latencies? Does disabling the watchdog timer in SparkSDR help? Even during TX, the unit is allowed to reach 55C before shutting TX off.

The heat sinks included from Makerfabs are standard Raspberry Pi heat sinks. It should be easy and inexpensive to find a replacement if yours is coming off.

https://www.pishop.us/product/aluminum-heatsink-for-raspberry-pi-4b-3-pack/

73,

Steve

kf7o

[Ed Grafton]

For better heat sink adhesion, you can get heat conductive epoxy on Ebay & other sources. Works well. I have used it on every SDR I own with no failures.

Ed

[Duncan Clark]

I use similar.

Just don't ever try to remove the heatsink :-)

Duncan

In message <d6d72293-e5c4-449e...@googlegroups.com>, Ed
Grafton writes

>--
>You received this message because you are subscribed to the Google
>Groups "Hermes-Lite" group.
>To unsubscribe from this group and stop receiving emails from it, send
>an email to hermes-lite...@googlegroups.com.
>To view this discussion on the web visit

>https://groups.google.com/d/msgid/hermes-lite/d6d72293-e5c4-449e-9b91-badaa88d5e2dn%40googlegroups.com
>.

--
Duncan Clark
G4ELJ