Keyboard Shortcuts
ctrl + shift + ? :
Show all keyboard shortcuts
ctrl + g :
Navigate to a group
ctrl + shift + f :
Find
ctrl + / :
Quick actions
esc to dismiss
Likes
Search
QDX Heat and Mod
|
My QDX Rev 2 works great. I have put a number of hours of constant activity with no issues. Rx sensitivity is great and transmit power is nominal.?
?
However, I am concerned about the heat generated by the finals. I built mine for 9 volts, but put a 7809 regulator inside with the case as a heatsink. This way I can power the rig from 13.8 volts, but give 9 volts to the finals and the input to the P-channel MOSFET.?
?
I noticed the case got warm on prolonged transmit activity. So at 21 C ambient, I checked temperatures with my IR thermometer. The case was 40 C after a half hour of constant FT8 activity (~45% Tx duty cycle). On opening the case, the finals bolt was 77 C and T1 was 60 C. While I believe the junction can take up to 150 C temp, I also know the TO92 case does not conduct heat very well. Leaving the case open only lowered the temps by about 4-5 degrees. A fan lowered temps by 20-30 degrees.?
?
These seem to be excessive temps and would only get worse with high ambient temps in a field environment, such as SOTA or POTA, which is where I usually use this rig. So I mounted a small 12 volt CPU cooler on the case with holes drilled to allow air to enter the case from the fan and exit through the end plate. I powered the fan from the 9 volt 7809. This gave marked improvement in temperatures. Fan noise is low and power draw at 9 volts is only a trivial 25 mA additional. Additional case height is only about 12 mm more.?
?
Temperatures after the mod are 29 C on case and 32 C at the finals bolt and T1 (case closed) after prolonged FT8 transmit activity. Again, this was at 21 C ambient. This is 35 degrees less than before! Pics are below. My write up and more pics will also be on Skyview Radio Society¡¯s ¡°Q5er¡± newsletter ( www.skyviewradio.net/newsletter/ ) for December 2022. I am hoping this will prolong the life of my finals.
-- 73, Dan? NM3A |
|
Nice! How much heat is generated by the 7809 regulator? Would it make sense to give it it's own cooling instead of dumping that heat into the QDX case? Of course, the fan sounds like the best solution.
toggle quoted message
Show quoted text
Please let us know when the ¡°Q5er¡± newsletter is out, as this group doesn't support pictures. 73, Willie N1JBJ On Nov 14, 2022, at 11:44 AM, Daniel Walter via groups.io <nm3a@...> wrote: |
|
Doh. Wrong again, must be another group I was thinking of.
toggle quoted message
Show quoted text
Nice pictures! 73, Willie N1JBJ On Nov 14, 2022, at 11:50 AM, William Smith <w_smith@...> wrote: |
¿ªÔÆÌåÓý
?Manual says 1.1A TX current at 9V.? With 13.8V in, 4.8V must be lost for 9V out and
7809 is linear, so 4.8 x 1.1 = 5.28W
You can work out the temp rise from the 7809 data sheet and this is left as an exercise
for the reader. ?
Andy
|
|
For what it's worth I measured the temperature of the four FET finals on my just-built QDX.? I wound T1 for 12V (two turns on the secondary), and am powering the QDX with 12.3V.? Transmitting on 30 meters for two minutes into a dummy load, the transistor case temperature rose to 51 deg C (about 30 deg aver ambient).? This is pretty reasonable.? T1 was a bit hotter, and the filter-switching diodes were a bit warmer still.?
The diode temperature surprised me at first, since the switched-on diode is only dissipating about 7 milliwatts DC, but I suppose RF through the diode resistance is causing it to heat up. When biased at 10 mA the 1N4007 apparently has a series resistance of about 1 Ohm, which at 5W RF power would result in about 300 mW of diode dissipation. More QDX measurements on my blog:? -- Paul Elliott - WB6CXC |
|
Andy,
|
|
Paul, Neat picture. I don't have any sophisticated measuring equipment, just a cheap Harbor Freight IR thermometer, but it does correlate well with other temperature measurements.?
|
|
On Mon, Nov 14, 2022 at 10:07 AM, Daniel Walter wrote:
I am surprised that your finals temperature stabilized that quicklyMe too.? When I started the transmit (WSJTX "Tune") I watched the temperature rise, and after two minutes it had essentially stabilized.? It's possible that the transistors aren't contacting the circuit board as well as they should, but while that will affect the thermal time-constant, poor contact certainly won't make things any worse...? I'm using a little FLIR camera that plugs into my android phone, and it's been pretty accurate.? Like all IR measurements it's a bit sensitive to color / emissivity, but the black TO-92 transistor cases are a decent target. ? -- Paul Elliott - WB6CXC |
|
On Mon, Nov 14, 2022 at 10:53 AM, Paul WB6CXC (wb6cxc.com) wrote:
poor contact certainly won't make things any worseWhat I was trying to say is that the measurement accuracy wouldn't be affected.? Of course poor heatsink contact (and so reduced radiating area) will make the transistor run hotter. ? -- Paul Elliott - WB6CXC |
¿ªÔÆÌåÓýPaul,Very interested in your FLIR camera that plugs into your Android phone. Can you tell us more? ?Model, where you got it etc.? Thanks very much.? Alan, W3AL? On Nov 14, 2022, at 13:53, Paul WB6CXC (wb6cxc.com) <paul@...> wrote:
|
|
On Mon, Nov 14, 2022 at 04:54 PM, Al Sines wrote:
your FLIR camera that plugs into your Android phoneI also have the FLIR One Pro, and like Donald I am satisfied with the performance.? It has (I think) two ranges, the shot I posted is using the lower-temp range and it is extremely sensitive.? I originally talked my wife into letting me get mine so we could troubleshoot our in-floor heating system that had some stuck valves -- even small temperature differences show up quite clearly.? The higher-temp range is useful for looking at reflow soldering ovens, etc.? There may be newer / better / cheaper versions now available, but the one I have serves me well. ? -- Paul Elliott - WB6CXC |
|
On Tue, 15 Nov 2022 at 13:35, Donald S Brant Jr <dsbrantjr@...> wrote: I have the FLIR One Pro, which has a USB-C connector, for Android phones:??? I have had it since December 2019 and I have been satisfied with its performance.? It is sensitive enough to use to track down high-current paths/short circuits by following the warm PC board trace. I became a little excited by the idea of a thermal imaging camera that I could just run with my android phone. "Wow," I thought, "the phone can supply power and processing and display, so this will be a really cheap way into a FLIR camera". So the ~AUD400 asking price was a bit of a shock. I guess I was thinking more along the lines of other optical products and maybe something like a ~AUD100 cost which might almost make it an impulse purchase. Not this time. -- MC VK1MC |
|
Hi Paul, Thanks for sharing your measurements. I also have some thermal measurements (look after my previous messages on this group) that shows T1 and diodes as a principal heating source on transmit mode. On RX and TX the most heat generator is 7805 5V linear regulator. I replaced it with a SIP3 SMPS module with good result. On the other hand I highly appreciate you FSK tone generation analysis. I never though about this having in mind that there were no reported issues on the FSK quality except the one Hans fixed in the past. I'm interested what was the FW you used when you rung those measurements. I think Hans fix this issue in the last one (1.05 I believe). So please check. Good luck with the little RIG! 73 de YO3FFF Cristi
|
|
On Mon, Nov 14, 2022 at 10:43 PM, cris blak wrote:
I'm interested what was the FW you used when you rung those measurements. I think Hans fix this issue in the last one (1.05 I believe). So please check.My QDX firmware is v1_05, and the board is PCB Rev 4.? AFAIK this is all the latest, I ordered it in late October and received it from Hans the beginning of November.? I built it on Nov 5. About the FSK issue, I would like to try repeating the tests using a longer QDX frequency measurement period than the default 100 per second.? With the WSPR symbol rate being so slow there's no reason that it couldn't measure more audio cycles and get a more precise frequency measurement.? That way we could tell if the error is due to the measurement, or to the S15351 divider setting algorithm.? (Although I believe that Hans already knows which is the case, likely the divider setting).? Certainly the default frequency measurement method *should* provide far better accuracy that my measurements were showing. ? -- Paul Elliott - WB6CXC |
|
On Mon, Nov 14, 2022 at 10:12 PM, mike.carden wrote:
So the ~AUD400 asking price was a bit of a shock. I guess I was thinking more along the lines of other optical products and maybe something like a ~AUD100 cost which might almost make it an impulse purchase. Not this time.Yes, the FLIR camera isn't cheap.? That's why I had to find an "around the house" way to use the camera, to avoid overly-annoying my wife! (although in fairness, she really doesn't complain much about my expensive hobby.) But I do see some US$50-$100 IR cameras on aliexpress, and even a $150 one on Amazon.? I have no idea if they work as advertised or are complete crap. ? -- Paul Elliott - WB6CXC |
|
Hi Paul, Thank for the feedback. That means the error still present! I have ver3 with 1_05 but never run the test (by the way I have also an SA44 analyzer). When you have time to redo the measurements it will be nice if you can quantize the errors due to this bad frequency shift! If it say 1% than not bother about it. Interesting subject. Maybe Hans could say more if it is interesting for him?! 73 de YO3FFF Cristi
Pe mar?i, 15 noiembrie 2022, 18:40:10 EET, Paul WB6CXC (wb6cxc.com) <paul@...> a scris:
On Mon, Nov 14, 2022 at 10:43 PM, cris blak wrote: I'm interested what was the FW you used when you rung those measurements. I think Hans fix this issue in the last one (1.05 I believe). So please check. About the FSK issue, I would like to try repeating the tests using a longer QDX frequency measurement period than the default 100 per second.? With the WSPR symbol rate being so slow there's no reason that it couldn't measure more audio cycles and get a more precise frequency measurement.? That way we could tell if the error is due to the measurement, or to the S15351 divider setting algorithm.? (Although I believe that Hans already knows which is the case, likely the divider setting).? Certainly the default frequency measurement method *should* provide far better accuracy that my measurements were showing. ? -- Paul Elliott - WB6CXC |
Donald S Brant Jr
to navigate to use esc to dismiss