Re: BitX40 voltage booster implementation

Gordon Gibby <ggibby@...>
 

and if the die to heat sink is truly 4 deg C per watt....then the hot point on the heat sink at max dissipation would be at about 110 deg C.    ​that's pretty hot!!!


Probably you want to keep the heat sink below 80 deg C.    (176 deg F)


Maybe attach a thermisor right to the to-220 tab or glue a disassembled cheap thermometer bulb there and mark your "limit" if you plan to push the limits!!


Gordon




From: BITX20@groups.io <BITX20@groups.io> on behalf of Gordon Gibby <ggibby@...>
Sent: Thursday, January 25, 2018 2:59 PM
To: BITX20@groups.io
Subject: Re: [BITX20] BitX40 voltage booster implementation
 

So, some estimations:


Junction to case, IRF510 is 3.5 deg C per watt

Case to greased surface is 0.5 deg C per watt typical

Those heat sinks look similar to  WV-T220-101E  so roughly 13 deg C per watt.


Total thermal resistance to ambient air:   17 deg C per watt.


Assume ambient of 25 (it might be hotter where you are!!!)

Junction spec'd to 175;  for safety figure 140 at the most...


Means maximum dissipation would be around   (140-25) deg C   /  17 deg C per watt or   6.8 watts per device.   


There have been some efficiency figures around here and they were not outstanding....so maybe figure 50% efficiency.   There are two MOSFETS on the ubitx, so you can dissipate a total of about 13.6 W (max) and with 50% effiiciency that would allow the same level of average output.


So....if you're considering uncompressed SSB with a duty cycle of maybe 30-40%, suggests you'd be OK to a peak power output of 1/0.4  x 13.6 =   34 watts  (that is NOT screaming in the mike nor using a compressor, of course)


Different digital modes use different amounts of power.   When my winlink "shifts gears" you can see the average power drop considerably --- like to 50-30% of peak -- but I've read that PSK31 is an 80% duty cycle --- so for PSK31 you'd be restricted back to 17 watts PEP.  


those are all based on available numbers, your mileage may vary!!!


http://www.arrl.org/news/10-tips-for-the-psk31-digital-mode

http://tymkrs.tumblr.com/post/11910142433/ham-radio-averaging-duty-cycle

http://www.ohmite.com/assets/docs/sink_w.pdf?r=false

https://www.digikey.com/product-detail/en/ohmite/WA-T220-101E/WA-T220-101E-ND/2202817

https://www.vishay.com/docs/91015/sihf510.pdf


And I don't know how to estimate the impact of "hot spots" on the MOSFET die.....



Gordon





From: BITX20@groups.io <BITX20@groups.io> on behalf of Vic WA4THR via Groups.Io <vhklein@...>
Sent: Thursday, January 25, 2018 2:34 PM
To: BITX20@groups.io
Subject: Re: [BITX20] BitX40 voltage booster implementation
 
Did some testing today. Running the booster at 14v input and setting the output to 23v I could detect no discernible noise change either on a dummy load or on an antenna. Transmit audio was recorded on a nearby receiver and no added hash or other artifact was noted either. Idle current with no load is perhaps 10ma, so leaving it on all the time the rig is powered is probably no issue. I was reading power into a 450w metered dummy load, so accuracy was not great, but the power level seemed to be up to around 15w (not PEP) on steady AHHH into the mic, as opposed to close to 5 with the standard configuration. Makes sense, (23/14.4)e2 times 6w says I should see 15w in this setup. So this looks like a go, giving nearly an S-unit improvement for little effort.

As an aside, i can tell almost no warming of the PA transistor heat sink during these tests, something I was concerned about. I've been struggling to find a suitable replacement heat sink, keeping with my target of minimizing purchases and using mostly junk-box parts. I do have a potential heat sink that I can use, but it was designed for a different transistor case style so the surface I would use with this TO-220 case is somewhat rough, like a cast finish, and I'm not sure that is a good idea. With the results I've seen is a bigger heat sink really needed? What power level was this one designed for? I only intend to use SSB on this particular rig, too.

=Vic=

Join BITX20@groups.io to automatically receive all group messages.