Topics

uBITX finals get hot very fast #ubitx #ubitx-help #v5

Rob French (KC4UPR)
 

I've got a new uBITX v5.  I've only been using it for CW so far (haven't gotten around to wiring up the mic yet), but I've noticed that, after calling CQ a half dozen times, the finals get way hot... too hot to hold my finger on for more than a second or so.  

As far as I can tell, my antenna is good to go... according to my antenna analyzer, < 2:1 SWR across 40 meters, which is where I'm using it currently.  I guess I could try doing the same thing into a dummy load to verify that it's not due to an SWR problem.  

Other than loading the KD8CEC firmware and making some firmware configuration settings, I have not deviated from the stock configuration--same finals, same heatsinks, no change to PA bias settings.  

I'm assuming this is not normal.  Any thoughts on resolution?  I'm using a metal case that doesn't quite let me mount the board close enough to back to bolt the finals to the back panel, but I've thought of making a hole in the back panel to accomodate bolting on a larger heatsink.  But I'm also thinking that there may be bigger underlying problem that I address, regardless of whether or not I add a bigger heatsink, fan, etc.

Things I'm thinking I should do:
- Recreate my CQ calling into a dummy load, see how hot they got (does anyone have any metrics for key-down time versus temperature?)
- Measure the current draw on transmit, and/or perform the PA bias calibration from the HF Signals web site
- Build one of those LED SWR meters to back-up my antenna analyzer
- Any other thoughts?

Clark Martin
 

What is the idle current into the PA with the mic PTT pressed (or the equivalent)?  It should be 400 mA max.  That’s measuring the power just into the PA section.  The radio as a whole should be 500mA max.

How much power out are you getting on CW?



Clark Martin
KK6ISP

On Aug 3, 2019, at 7:50 AM, Rob French (KC4UPR) <robert@...> wrote:

I've got a new uBITX v5.  I've only been using it for CW so far (haven't gotten around to wiring up the mic yet), but I've noticed that, after calling CQ a half dozen times, the finals get way hot... too hot to hold my finger on for more than a second or so.  

As far as I can tell, my antenna is good to go... according to my antenna analyzer, < 2:1 SWR across 40 meters, which is where I'm using it currently.  I guess I could try doing the same thing into a dummy load to verify that it's not due to an SWR problem.  

Other than loading the KD8CEC firmware and making some firmware configuration settings, I have not deviated from the stock configuration--same finals, same heatsinks, no change to PA bias settings.  

I'm assuming this is not normal.  Any thoughts on resolution?  I'm using a metal case that doesn't quite let me mount the board close enough to back to bolt the finals to the back panel, but I've thought of making a hole in the back panel to accomodate bolting on a larger heatsink.  But I'm also thinking that there may be bigger underlying problem that I address, regardless of whether or not I add a bigger heatsink, fan, etc.

Things I'm thinking I should do:
- Recreate my CQ calling into a dummy load, see how hot they got (does anyone have any metrics for key-down time versus temperature?)
- Measure the current draw on transmit, and/or perform the PA bias calibration from the HF Signals web site
- Build one of those LED SWR meters to back-up my antenna analyzer
- Any other thoughts?

Curt
 

Numerous opinions here, I run my v4 on cw or ssb without updates. I will add a small fan before doing digital modes, whether its needed or not.

Wait until you discover the hot regulator on the raduino. If your PA heatsink is less warm, I suggest things are okay. If it is warmer, then check out bias current, carefully follow instructions.

Curt

Tom, wb6b
 

Hi,

I found when running digital modes, particularly WSPR, that a fan was a necessity. 

Tom, wb6b

Rob French (KC4UPR)
 

All good questions!  So, I don't have a power meter, but I do have a QRP Labs dummy load that has a diode for rudimentary RF power measurements with a voltmeter.  I'll try with that, which will also help me verify that the PA's heat up similarly with the dummy load.

I'll measure the overall current consumption of the transceiver as well as the PA section.  

However, one thing I've discovered... I'm using an old Micronta 12V power supply (that's what the front panel says).  When I checked the back panel, it says 13.8VDC output on the spec label, and measured, it's actually like 14.4V.  (I'll double-check that under load.)  So I imagine the higher supply voltage could also be contributing? 

I'm going to wire up a couple diodes in series with the power supply to bring the supply voltage down to ~12V, and see how that affects things.  Then I'll look at the PA section.

Thanks!
-Rob

 

Hi Rob,

I don't think anything unusual is happening in your case.  Just making some general observations:

The heatsinks provided are probably good for about 20C/W dissipation; meaning the heatsink temperature will rise 20 degrees C above ambient temperature per W dissipated.

Ambient in these situations (a hot day, no fan, heatsink in enclosed or semi-enclosed area) is best taken at 50C.

At 10W out and the PA 50% efficient, each transistor would have 5W to get rid of on constant keydown or about 2.5W during morse sending.

So that makes 20C/W x 2.5W + 50C  =  100C (burn your finger temperature for sure)

This is a worst case example as ambient is high (you may have better air circulation around the heatsinks and a cooler room), the power may be lower, and the PA may be more efficient.  But it gives you an idea of how hot the heatsinks may get.  With no fan, the stock heatsinks are not good for more than a watt or so if you want the transistor and heatsink to stay at "just warm" temperatures (less than 60C).  And no way would they be good enough for digital without fan help.  Bigger heatsinks will allow for cooler tempertures, or you can run a fan on them as I think many do here.

Remember the transistor tabs need to be insulated from grounding through the heatsinks/housing.

73,


Mark.

Rob French (KC4UPR)
 

Wow, you're right, that regulator does get pretty hot!  I've got a TO-220 heatsink that I will bolt to that.

That said, I would say that the heatsinks for the IRF510's has gotten at least as hot as the tab for the regulator, so it seems like I should continue to investigate this.  

I do know that my power supply is running high, so I'm planning on dropping it down ~1.5V with a couple of diodes.  I have a handful of these 6A02 power rectifiers in my junk box, does anyone know if these would be suitable to put in series with the incoming power lead to drop the voltage?  Probably overkill (6 A!), but I have them on hand...

Thanks
Rob

Evan Hand
 

Rob,

Don't forget the thermal compound on the regulator. and verify there is some on the finals.

As to current draw, I have my 2 uBitx fused for 3 amps, and the only time I popped one was when I shorted out one of the final heat sinks when trying to measure a voltage.  I have measured the output of the rigs at 14 watts CW with 13.8 volts to the finals and 2 2n4001 diodes in series to drop the voltage to the rest of the board (4001s are rated for 1 amp average, not recommended to drop the voltage to the finals).   I did not run the rig for long, so cannot say if I have a thermal problem.  I am traveling on business full time now, so have not been able to go back and do any more playing with the uBitx.

I have read a number of posts that indicate digital modes will heat up the finals, unless you lower the drive.  If it were me, I would be looking at installing a fan to blow on the final heat sinks, drop the voltage to the board, and keep the finals supplied direct from the supply,  If you still have a problem with heat, then lower the drive with RV1.  I would also be sure that the finals were biased correctly per the procedure on the HF Signals web site.

Above are my thoughts, yours may be different, and mine do not come with any guarantee.  In other words, use at your own risk.
73
Evan
AC9TU

Randy.AB9GO
 

If your RadioShack power supply is running over 13.8V, check the pass transistor for leakage or for a bad zener.  I had one that the pass transistor shorted and it was sourcing over 18 Volts.  

Mark Hatch
 

Coming very soon...  will make it available on Thingiverse

73
Mark
AJ6CU

Jim Sheldon
 

That will be nice - I have a 3D printer - LOL.  Just got done upgrading my Ender 3 to a direct drive extruder this afternoon.  Used stock extruder motor and an Aluminum top part.  First print after getting it all together was a short filament guide to keep the filament from getting tangled on the spool and drops right down to the extruder input.

I was going to design something similar for a slightly smaller fan, but it looks like you have the right combination there.  That should keep the finals cool for sure.  You might consider another air duct that can be capped off or directed at things like the Raduino's 7805 or Jack's JackAl board's regulator heat sinks as well.  That would be the best of all worlds.

Jim - W0EB

------ Original Message ------
From: "Mark Hatch" <mark2382@...>
Sent: 8/4/2019 8:12:59 PM
Subject: Re: [BITX20] uBITX finals get hot very fast

Coming very soon...  will make it available on Thingiverse

73
Mark
AJ6CU

Rob French (KC4UPR)
 

Thanks... what exactly would I be looking for to check this?  (Feel free to point me at a resource, I can read...) Is this something I'd need to desolder stuff for?

I may go ahead and invest in another supply, anyhow... this one has a two prong plug, and given that I'm running into an attic antenna (albeit w/ great SWR!), I currently have no ground connection, which probably isn't awesome...

Curt
 

Rob

Check your power supply voltage key down to see if the voltage droops any. Otherwise some might find 14 volts as a feature. You may be fine on ssb and cw. You can back off a little on drive level, and also get a cleaner signal.

Farhan has commented on heatsink temps. I would not let this keep you from enjoying the rig. If you must some rainy day add heatsink compound, but keep it nice and thin.

Enjoy the nice rig.

Curt

Mark Hatch
 

See https://www.thingiverse.com/thing:3790450 for the fan shroud. (Although it doesn't seem to be relevant to this problem...)

73
Mark