Moon-net August 2019

moon-net@mailman.pe1itr.com

100 participants
139 discussions

Re: [Moon-Net] Thermal failure of solid state PA's

by charlie＠sucklingfamily.free-online.co.uk

Hi Brian, Jim and Bill Many thanks for your comments on gold embrittlement of solder. This might well be the root cause of failures some folks have seen with LDMOS (and other devices). Other articles state that some gold content is OK, others more recent suggest that any gold contamination is to be avoided. I don't have a copy of the paper in the conference proceedings that Bill referred to - this looks to be very interesting. I did find one potentially useful article here: https://cdn.macom.com/applicationnotes/AN3025.pdf This refers to the need to remove gold from RF package leads prior to soldering them into a board to avoid gold contaminating the solder joint. I am attracted to the method described there using a soldering iron to pre tin the leads and solder wick to then remove it. 73 Charlie G3WDG > FWIW......in my day-job many years ago, it was common to pre-tin the leads > of RF power devices in a solder wave process. The purpose was to remove the gold plate from the lead so as to stop tin/gold embrittlement issues in the final PCB solder joint. > > -Brian, WA1ZMS > > >> On Aug 30, 2019, at 10:26 AM, Charlie via Moon-net >> <moon-net(a)mailman.pe1itr.com> wrote: >> Hi John >> Thanks for your comments. >> They provide yet more evidence that the solder joint between a >> medium/high >> power device and the PCB is a very critical area. The joint is subjected >> simultaneously to high RF and DC currents as well as mechanical forces due >> to temp cycling. >> As you say, JT type operation over a long period is a tough environment. >> Most base station PAs, built pretty much like our are, operate in a more >> benign environment with lower average output power and are usually on 24/7. >> 73 >> Charlie >>> It is not just beacons that suffer from, non-device, thermally induced failures. I have experienced a particularly strange fault on the 144MHz >>> PA >>> in the TS2000 that I use for 144MHz EME. >>> I use a TS2000 for digital modes on VHF; JT65 and FSK441 as more conventional traditional transceivers with narrow crystal filters introduce >>> phase distortion on both RX and TX. One of my TS2000 transceivers is used >>> on >>> 144MHz EME and drives a 8877 linear amplifier via a coaxial attenuator. >>> It >>> runs at 100W (full) output to reduce the level of transmitter output overshoot. This is because I have an optional SSPA and do not want to destroy the devices with large 'spikes'. >>> A nearby amateur reported that I had what sounded like arching on my transmissions with the power output dropping in bursts. I swapped-out the >>> TS2000 and put it on the bench. During soak testing I also saw the output >>> drop off and this time it stayed down. I was fairly convinced that one of >>> the 144MHz PA transistors had died. >>> However close examination of the 144MHz PA area revealed that the PCB > > > _______________________________________________ > Moon-Net posting and subscription instructions are at > http://www.nlsa.com/nets/moon-net-help.html > >

3 years, 9 months

RWN 220 & RW 1136

by OE4WOG

need schematics, layouts, operation & service manual, all infos how to use / modify / upgrade and operate the PSU + RW1136 TWT from Siemens,  ( hints, tricks, pitpfalls, etc,.,. ) any help out there ? TNX in advance.... 73 de OE4WOG

3 years, 9 months

Re: [Moon-Net] IC-9700 transmit delay

by JAYB1943＠OPTONLINE.NET

rE: ic-9700 transmit delay. My solution was to build a breakout box in series with my Signalink to break into the PTT line (pin 3 on the ACCY) connector and drive it from a DEM sequencer. I use a separate foot + toggle switch or the Signalink PTT to trigger the sequencer which goes to a 3-way switch (band selection) which drives one of the 3 Gemini power amps (and the +12 for the remote preamps) before the PTT is initiated on the IC-9700. As a further improvement, I built another break-out box connected to the front panel Mike connector which isolates the PTT line on the mike, now also connected to the DEM sequencer. So if I hit the mike button, it triggers the sequencer, disconnects the +12v to the remote preamp, turns on the selected power amp and then finally triggers the PTT line in the IC-9700. A lot of work but FOOLPROOF ! It does get busy in a contest so I need things to be fully protected from my possible mistakes. If I select the wrong band power amp and hit the mike button the worse thing that happens is the IC-9700 transmits the driver power only...not a big deal.... I really wish that Icom had brought out separate PTT lines for each band (like my old 736r) and made the RF transmit delay longer... 73 Jay NY2NY

3 years, 9 months

Re: [Moon-Net] IC-9700 switching

by Edward R Cole

One variation that works is if the radio has Tx Inhibit control, then one can enable RF at last stage of sequencer. I use an Elecraft K3 that provides Tx inhibit. Previous to the K3 I had an FT-847 that also had Tx Inhibit. Both worked perfectly. http://www.kl7uw.com/TX-INHIBIT.htm My sequencer is an old design of W5LUA and marketed by DEMI. It provides 300ms total delay from Rx to Tx. At one time I used aux contacts on my TR relay to interrupt PTT, but in later design removed that as Tx Inhibit functions without fault. Some times too many control functions lead to more things to fail - KISS principle. 73, Ed - KL7UW At 10:48 PM 8/30/2019, Conrad PA5Y via Moon-net wrote: >Content-Language: en-US >Content-Type: multipart/alternative; > >boundary="_000_DB6P189MB0389A86A2B91DAD712CEFB8CB5BC0DB6P189MB0389EURP_" > >I measured the TX delay on PA2CHRs IC9700 and >it was 34ms, my Dow Key 402A relays take 20ms to >change state and for ground bounce to settle >BUT- why dont people just have the radio go >over last by driving it with an external >sequencer? I drive the sequencer with either a >footswitch, a PC , a microphone PTT or a >Winkey CW interface. The final 5th position on >the sequencer drives the radio PTT. The 402A >relays have auxiliary contacts and so I use >these contacts to interrupt the radio PTT. It is >impossible for the radio to go into TX if the >main LNA protection relay is not in the correct position. > >This is absolutely fool-proof and simple. > >Using the radio to start a PTT sequence is just >asking for trouble in my opinion. > >Of course you could use 2 PIN switches for RX/TX >separation if your LNA has sufficient gain to overcome the losses. > >73 > >Conrad PA5Y > >From: Moon-net ><moon-net-bounces(a)mailman.pe1itr.com> On Behalf Of Peter PA2V via Moon-net >Sent: 30 August 2019 15:57 >To: moon-net(a)mailman.pe1itr.com >Subject: [Moon-Net] IC-9700 switching > >I am playing with a better TX/RX delay on my IC-9700. >Maximum delay on the rig is 30 mSec, which is a little short. >Most coax and other relays need a longer timing. >I am afraid to kill the preamp with no other protection included. >With this in mind I did some experiments with an >(old) Mitsubishi MD001H hybrid antenna switch. >Using such a device is easy to delay TX RF >entering to the relay box and preamp. >It also create a separated RX and TX port. >I also used a circulator at the rig to prevent >high SWR on the rig itself during the time there >is no connection to the antenna while the Mitsubishi switch delays it. >It really works great, timing on a pin switch >can be done in many ways. For this experiment I >used a RC combination with an ordinary P-channel >FET to switch the bias on the Mitsubishi PIN switch. >As stated before it works and any delay can be >made this way without damaging the rig. > >After rough tests I would like to know the RF >performance of my creation. And I was in shock. >The data sheet said that the minimum isolation >between the TX and RX port is 30 dB. >But I measured 23 dB with a calibrated VNA. >Another MD001H gave exact the same performance, >and another PCB with better separated >connections did not gave better performance. >Using the 9700 as an exciter it will drive my >amps with around 7-8 Watt. Which means that my >preamp will have around 70 mW on the output >port, and kill the preamp sooner or later with >such a low isolation of the PIN switch. > >I do not go on with the MD001H, but would like >to know if there will PIN switches available >that are able to transfer up to 20 Watt and have >at least 40 dB isolation between TX and RX ports? >If so, please drop me an e-mail. > >73, Peter PA2V > >_______________________________________________ >Moon-Net posting and subscription instructions >are at http://www.nlsa.com/nets/moon-net-help.html 73, Ed - KL7UW http://www.kl7uw.com Dubus-NA Business mail: dubususa(a)gmail.com

3 years, 9 months

Re: [Moon-Net] Thermal failure of solid state PA's

by john.regnault＠btinternet.com

It is not just beacons that suffer from, non-device, thermally induced failures. I have experienced a particularly strange fault on the 144MHz PA in the TS2000 that I use for 144MHz EME. I use a TS2000 for digital modes on VHF; JT65 and FSK441 as more conventional traditional transceivers with narrow crystal filters introduce phase distortion on both RX and TX. One of my TS2000 transceivers is used on 144MHz EME and drives a 8877 linear amplifier via a coaxial attenuator. It runs at 100W (full) output to reduce the level of transmitter output overshoot. This is because I have an optional SSPA and do not want to destroy the devices with large 'spikes'. A nearby amateur reported that I had what sounded like arching on my transmissions with the power output dropping in bursts. I swapped-out the TS2000 and put it on the bench. During soak testing I also saw the output drop off and this time it stayed down. I was fairly convinced that one of the 144MHz PA transistors had died. However close examination of the 144MHz PA area revealed that the PCB track just beyond the collector tabs of the PA devices had de-laminated, the solder had re-crystallised, oxidised and was growing whiskers. In fact it showed all of the signs of long term aging at high temperatures. Replacing the collector track with some 14G copper wire and generous re-soldering solved the problem and the 144MHz output was restored. This particular transceiver had been in regular JT65 EME use for six years. 50 seconds TX at full power, then 80 seconds RX and so on. It had probably done over 2000 hours transmitting service in such a manner. As I worked in semiconductor reliability during the 1980s this should not have come as a surprise! However it does show that some parts of the TS2000 and probably most other amateur transceivers do age. Indeed one of my locals who had a Beko SSPA experienced similar issues with the PCB beyond the output device. It does seem that kit built from the amateur market was never designed for heavyweight duty cycles such as JT65 or beacon use. 73 John G4SWX --- This email has been checked for viruses by Avast antivirus software. https://www.avast.com/antivirus

3 years, 9 months

Wanted: Data sheets for Stanford Microdevices PHEMPT FETs

by Dr. David Kirkby

Sorry, this may be a bit off-topic, but I'm hoping someone can help me. About 40 years ago, as a teenager, I bought some low-noise Pseudomorphic High-electron-mobility (PHEMPT) field-effect transistors (FETs). Stanford Microdevices SPX 1476 Stanford Microdevices SPF 1076 These are microwave devices, but a friend and I designed and built a couple of pre-amps with these. Based on the tests we could do at that time (no real test equipment in our teens), they worked worked well. I know at least one of the devices is an N-channel depletion mode device. I would like to use this in a preamp, but I've lost the datasheets. Does anyone have them? I'm guessing I probably scrounged these from somewhere, as I expect these were very expensive at the time, and a friend has more than 20 of the things. Unfortunately, Stanford Microdevices no longer exists. It's been taken over or merged many times, so the latest company to actually have any connection with Stanford Microdevices is Qorvo. I think the chances of Qorvo having a data sheet for a 40-year old FET is pretty remote! I believe the acquisitions were Stanford Microdevices -> Sirenza Microdevices -> RF Micro Devices (RFMD) -> Qorvo I guess there's no harm asking Qorvo if they have datasheets for these FETs, but I would think the probability is extremely close to zero. Dave, G8WRB -- Dr David Kirkby Ph.D C.Eng MIET Kirkby Microwave Ltd Registered office: Stokes Hall Lodge, Burnham Rd, Althorne, CHELMSFORD, Essex, CM3 6DT, United Kingdom. Registered in England and Wales as company number 08914892 https://www.kirkbymicrowave.co.uk/ Tel 01621-680100 / +44 1621-680100

3 years, 9 months

Re: [Moon-Net] Moon-net Digest, Vol 432, Issue 15

by JAYB1943＠OPTONLINE.NET

3 years, 9 months

Re: [Moon-Net] Moon-net Digest, Vol 432, Issue 10

by Vlada ok1vpz

Re: thermal failure of SSPA: Let me send a one cent view: http://www.ok2kkw.com/more/impulsni_tranzistory _en.htm(http://www.ok2kkw.com/more/impulsni_tranzistory_en.htm) 73! Vlada ---------- Původní e-mail ---------- Od: moon-net-request(a)mailman.pe1itr.com Komu: moon-net(a)mailman.pe1itr.com Datum: 30. 8. 2019 12:06:55 Předmět: Moon-net Digest, Vol 432, Issue 10 "Send Moon-net mailing list submissions to moon-net(a)mailman.pe1itr.com To subscribe or unsubscribe via the World Wide Web, visit http://mailman.pe1itr.com/mailman/listinfo/moon-net or, via email, send a message with subject or body 'help' to moon-net-request(a)mailman.pe1itr.com You can reach the person managing the list at moon-net-owner(a)mailman.pe1itr.com When replying, please edit your Subject line so it is more specific than "Re: Contents of Moon-net digest..." Today's Topics: 1. Re: Thermal failure of solid state PA's (charlie(a)sucklingfamily.free-online.co.uk) ---------------------------------------------------------------------- Message: 1 Date: Thu, 29 Aug 2019 13:30:14 +0100 From: charlie(a)sucklingfamily.free-online.co.uk To: "Bob Atkins" <ka1gt(a)hotmail.com> Cc: "moon-net(a)mailman.pe1itr.com" <moon-net(a)mailman.pe1itr.com> Subject: Re: [Moon-Net] Thermal failure of solid state PA's Message-ID: <65af15ec54853e4339cf9d3c8f86ec13.squirrel(a)webmail.plus.net> Content-Type: text/plain;charset=iso-8859-1 Hi Bob Nice summary of the situation and thanks for the link. The 10GHz PA in DL0SHF is calculated have a continuous operational lifetime based on electrical stress of something like 10^7 hours. The device channel temperature can be estimated reasonably accurately and from the manufacturer's exhaustive life test data such predictions can be made with good confidence. De-rating (ie running at lower power) would increase the lifetime, but at 10^7 hours with no de-rating I don't think it would help as far as the active device itself is concerned. Indeed after the first failure the active device was still working. Running at lower power, eg 2-3dB below saturation, would result in the pcb track (which failed in the DL0SHF PA, not at a solder joint incidentally) running cooler and thus seeing a reduced temperature excursion every 1 min period. IF the cause of failure was metal fatigue of the copper track caused by temp cycling then de-rating would help. Agreed it would be a useful experiment to run on the bench, however a) I'm not sure anyone has the capability or motivation to do this, b) bench tests often don't reflect the stresses in actual operation and c) it would take a long time. Running the PAs in the beacon(s) has the advantages of the "test bed" being there already and as long as the PAs work then the beacons are fulfilling their role. Today I started some experiments looking at condensation on a chilled water cooler as I have seen some evidence of what appears to be corrosion in areas on the DL0SHF amplifier. First test was to see at what cold plate temperature condensation started to appear. In an indoor atmosphere of unknown humidity and temperature of 24C a visible moisture film appeared with the cooling plate at 14C. When the cooler temperature had reduced to 6C it was covered in large water droplets. I have found a nice calculator here http://www.dpcalc.org/ which works out the dew point as a function of temperature and humidity. If condensation was found to be the cause then it would not be too difficult with use of heaters or even perhaps running the PA continuously at quiescent bias to prevent this. 73 Charlie PS powerpoint file of notes is here. I will update it as we learn more (same URL): https://drive.google.com/file/d/1xRhtaVLH2OjFCYaXrEjcMh0CGjMgCeWn/view?usp= sharing > What we are seeing with the beacons seems to be the result of using > equipment originally designed for amateur service being pressed into > semi-continuous operation. If a beacon is operating an average of 8hrs/day > it goes through at least 480 cycles (off/on/off transitions). If an > average QSO consists of 5 transmit cycles, that's the equivalent of 96 > QSOs a day, every day, or just over 35000 QSOs a year. So a beacon > amplifier works as hard in 3 or 4 days as my amplifier does is a year (and > that's a pretty active year with 300-400 QSOs) > > There's an interesting website on reliability engineering - see > https://accendoreliability.com/basics-derating-electronic-components/ - > and derating of components, which lists a number of mechanisms that result > in higher probability of failure including Electrical stress, Thermal > stress, Chemical stress and Mechanical stress. > > It would be in interesting experiment to run one of the beacon amplifies > continuously and see how the reliability compared to one which was turned > on and off 480 times a day. No thermal cycling at all. When not sending > FSK digital or FSK CW, a carrier would be present. > > Efficient heat-sinking would be required (but not so much cooling as to > allow moisture to condense within the amplifier), rated for a 100% duty > cycle. I would also add air cooling passing over the PCB to further lower > passive component temperature. Even with an RF sealed enclosure there are > still ways to air cool the circuit boards (high pressure air through small > diameter ports). Also (though it goes against the grain of most amateur > practice!) I'd turn down the power a 2-3 dB from the nominal amplifier > rating. As long as small stations can see the carrier and decode the > digital signal, the beacon would be doing it's job of encouraging activity > and providing a signal for equipment testing and evaluation. If higher > power was desirable for very small station reception, going from reduced > power to full power once every 5 or 10 cycles would minimize stress on the > amplifier. > > One of the reason I chose to get on 1296 EME and I'm now moving toward > 10Ghz capability (Rx working fine, Tx will take a while, need bigger > dish..etc.) is that there's a beacon on both bands. The beacons do a great > job and are in invaluable resource for system anaysis and testing. I know > from personal experience that they DO encourage activity! My thanks go to > the beacon operators. > > Bob, KA1GT > ________________________________ > From: Moon-net <moon-net-bounces(a)mailman.pe1itr.com> on behalf of Charlie > via Moon-net <moon-net(a)mailman.pe1itr.com> > Sent: Wednesday, August 28, 2019 10:00 AM > To: Ben <LoWeb(a)esp-inc.net> > Cc: moon-net(a)mailman.pe1itr.com <moon-net(a)mailman.pe1itr.com> > Subject: Re: [Moon-Net] Thermal failure of solid state PA's > > Hi Ben > > Thanks for sharing your findings. It is amazing what thermal expansion > can do. > > I guess these effects are not often encountered on amplifiers built on > printed circuit boards mainly because in normal amateur operation there > are not a very high number of thermal cycles. With beacon operation > however things seem to be considerably more challenging! > > 73 > > Charlie G3WDG > >> A couple of days ago, several folks on this reflector were concerned >> about >> failure of their solid state PA???s. That caused me to think back about >> a >> problem I had a few decades ago with a 100 watt solid-state PA design. >> The amplifier would work just great for perhaps 30 minutes. Then the >> output stage device would blow. Once when this happened, I caught just >> a >> glimpse of a blue flash out of the corner of my eye as the device burned >> out. The 3-stage PA was built on a brass chassis, and the output match >> consisted of some series and parallel components. There was a leaded >> ATC >> capacitor, and this was soldered to a couple of piston variables that >> went >> to ground on each end of the leaded chip capacitor. The culprit was the >> thermal expansion of the brass chassis which was pulling the lead off of >> one end of the leaded chip capacitor and presenting essentially an open >> circuit to the device. I resolved the issue by putting ???an expansion >> loop,??? i.e. a slight V-shape in the lead on one end of the capacitor. >> This additional 2 nanohenries of inductance created another reactive >> component in the matching circuit, but I could compensate for this by >> changing the value of capacitance. >> >> >> >> Ben/K4QF >> >> >> >> _______________________________________________ >> Moon-Net posting and subscription instructions are at >> http://www.nlsa.com/nets/moon-net-help.html >> >> > > > _______________________________________________ > Moon-Net posting and subscription instructions are at > http://www.nlsa.com/nets/moon-net-help.html > > ------------------------------ Subject: Digest Footer _______________________________________________ Moon-Net posting and subscription instructions are at http://www.nlsa.com/ nets/moon-net-help.html ------------------------------ End of Moon-net Digest, Vol 432, Issue 10 ***************************************** "

3 years, 9 months

Weinheim Ukw

by sm4ive

If all goes well we plan to go to Weinheimhope to see some EME rs73 Lars SM4IVESkickat från min Samsung Galaxy-smartphone.

3 years, 9 months

Re: [Moon-Net] Moon-net Digest, Vol 432, Issue 10

by john.g0api

All interesting stuff Charlie .Perhaps , if not done now , slow power up ramping and same at Moon set , would reduce thermal stresses .Is PA biased during the Off periods ?Use 2 PA modules with dividers and run at reduced individual levels ,, within happier stress zone , if there is one found .73JOHNg0APISent from Samsung tablet. -------- Original message --------From: moon-net-request(a)mailman.pe1itr.com Date: 30/08/2019 11:00 (GMT+00:00) To: moon-net(a)mailman.pe1itr.com Subject: Moon-net Digest, Vol 432, Issue 10 Send Moon-net mailing list submissions to moon-net(a)mailman.pe1itr.comTo subscribe or unsubscribe via the World Wide Web, visit http://mailman.pe1itr.com/mailman/listinfo/moon-netor, via email, send a message with subject or body 'help' to moon-net-request(a)mailman.pe1itr.comYou can reach the person managing the list at moon-net-owner(a)mailman.pe1itr.comWhen replying, please edit your Subject line so it is more specificthan "Re: Contents of Moon-net digest..."Today's Topics: 1. Re: Thermal failure of solid state PA's (charlie@sucklingfamily.free-online.co.uk)----------------------------------------------------------------------Message: 1Date: Thu, 29 Aug 2019 13:30:14 +0100From: charlie(a)sucklingfamily.free-online.co.ukTo: "Bob Atkins" <ka1gt(a)hotmail.com>Cc: "moon-net(a)mailman.pe1itr.com" <moon-net(a)mailman.pe1itr.com>Subject: Re: [Moon-Net] Thermal failure of solid state PA'sMessage-ID: <65af15ec54853e4339cf9d3c8f86ec13.squirrel(a)webmail.plus.net>Content-Type: text/plain;charset=iso-8859-1Hi BobNice summary of the situation and thanks for the link.The 10GHz PA in DL0SHF is calculated have a continuous operationallifetime based on electrical stress of something like 10^7 hours. Thedevice channel temperature can be estimated reasonably accurately and fromthe manufacturer's exhaustive life test data such predictions can be madewith good confidence. De-rating (ie running at lower power) wouldincrease the lifetime, but at 10^7 hours with no de-rating I don't thinkit would help as far as the active device itself is concerned. Indeedafter the first failure the active device was still working.Running at lower power, eg 2-3dB below saturation, would result in the pcbtrack (which failed in the DL0SHF PA, not at a solder joint incidentally)running cooler and thus seeing a reduced temperature excursion every 1 minperiod. IF the cause of failure was metal fatigue of the copper trackcaused by temp cycling then de-rating would help.Agreed it would be a useful experiment to run on the bench, however a) I'mnot sure anyone has the capability or motivation to do this, b) benchtests often don't reflect the stresses in actual operation and c) it wouldtake a long time. Running the PAs in the beacon(s) has the advantages ofthe "test bed" being there already and as long as the PAs work then thebeacons are fulfilling their role.Today I started some experiments looking at condensation on a chilledwater cooler as I have seen some evidence of what appears to be corrosionin areas on the DL0SHF amplifier. First test was to see at what cold platetemperature condensation started to appear. In an indoor atmosphere ofunknown humidity and temperature of 24C a visible moisture film appearedwith the cooling plate at 14C. When the cooler temperature had reduced to6C it was covered in large water droplets.I have found a nice calculator here http://www.dpcalc.org/ which worksout the dew point as a function of temperature and humidity.If condensation was found to be the cause then it would not be toodifficult with use of heaters or even perhaps running the PA continuouslyat quiescent bias to prevent this.73CharliePS powerpoint file of notes is here. I will update it as we learn more(same URL):https://drive.google.com/file/d/1xRhtaVLH2OjFCYaXrEjcMh0CGjMgCeWn/view… > What we are seeing with the beacons seems to be the result of using> equipment originally designed for amateur service being pressed into> semi-continuous operation. If a beacon is operating an average of 8hrs/day> it goes through at least 480 cycles (off/on/off transitions). If an> average QSO consists of 5 transmit cycles, that's the equivalent of 96> QSOs a day, every day, or just over 35000 QSOs a year. So a beacon> amplifier works as hard in 3 or 4 days as my amplifier does is a year (and> that's a pretty active year with 300-400 QSOs)>> There's an interesting website on reliability engineering - see> https://accendoreliability.com/basics-derating-electronic-components/ -> and derating of components, which lists a number of mechanisms that result> in higher probability of failure including Electrical stress, Thermal> stress, Chemical stress and Mechanical stress.>> It would be in interesting experiment to run one of the beacon amplifies> continuously and see how the reliability compared to one which was turned> on and off 480 times a day. No thermal cycling at all. When not sending> FSK digital or FSK CW, a carrier would be present.>> Efficient heat-sinking would be required (but not so much cooling as to> allow moisture to condense within the amplifier), rated for a 100% duty> cycle. I would also add air cooling passing over the PCB to further lower> passive component temperature. Even with an RF sealed enclosure there are> still ways to air cool the circuit boards (high pressure air through small> diameter ports). Also (though it goes against the grain of most amateur> practice!) I'd turn down the power a 2-3 dB from the nominal amplifier> rating. As long as small stations can see the carrier and decode the> digital signal, the beacon would be doing it's job of encouraging activity> and providing a signal for equipment testing and evaluation. If higher> power was desirable for very small station reception, going from reduced> power to full power once every 5 or 10 cycles would minimize stress on the> amplifier.>> One of the reason I chose to get on 1296 EME and I'm now moving toward> 10Ghz capability (Rx working fine, Tx will take a while, need bigger> dish..etc.) is that there's a beacon on both bands. The beacons do a great> job and are in invaluable resource for system anaysis and testing. I know> from personal experience that they DO encourage activity! My thanks go to> the beacon operators.>> Bob, KA1GT> ________________________________> From: Moon-net <moon-net-bounces(a)mailman.pe1itr.com> on behalf of Charlie> via Moon-net <moon-net(a)mailman.pe1itr.com>> Sent: Wednesday, August 28, 2019 10:00 AM> To: Ben <LoWeb(a)esp-inc.net>> Cc: moon-net(a)mailman.pe1itr.com <moon-net(a)mailman.pe1itr.com>> Subject: Re: [Moon-Net] Thermal failure of solid state PA's>> Hi Ben>> Thanks for sharing your findings. It is amazing what thermal expansion> can do.>> I guess these effects are not often encountered on amplifiers built on> printed circuit boards mainly because in normal amateur operation there> are not a very high number of thermal cycles. With beacon operation> however things seem to be considerably more challenging!>> 73>> Charlie G3WDG>>> A couple of days ago, several folks on this reflector were concerned>> about>> failure of their solid state PA???s. That caused me to think back about>> a>> problem I had a few decades ago with a 100 watt solid-state PA design.>> The amplifier would work just great for perhaps 30 minutes. Then the>> output stage device would blow. Once when this happened, I caught just>> a>> glimpse of a blue flash out of the corner of my eye as the device burned>> out. The 3-stage PA was built on a brass chassis, and the output match>> consisted of some series and parallel components. There was a leaded>> ATC>> capacitor, and this was soldered to a couple of piston variables that>> went>> to ground on each end of the leaded chip capacitor. The culprit was the>> thermal expansion of the brass chassis which was pulling the lead off of>> one end of the leaded chip capacitor and presenting essentially an open>> circuit to the device. I resolved the issue by putting ???an expansion>> loop,??? i.e. a slight V-shape in the lead on one end of the capacitor.>> This additional 2 nanohenries of inductance created another reactive>> component in the matching circuit, but I could compensate for this by>> changing the value of capacitance.>>>>>>>> Ben/K4QF>>>>>>>> _______________________________________________>> Moon-Net posting and subscription instructions are at>> http://www.nlsa.com/nets/moon-net-help.html>>>>>>> _______________________________________________> Moon-Net posting and subscription instructions are at> http://www.nlsa.com/nets/moon-net-help.html>>------------------------------Subject: Digest Footer_______________________________________________Moon-Net posting and subscription instructions are at http://www.nlsa.com/nets/moon-net-help.html------------------------------End of Moon-net Digest, Vol 432, Issue 10*****************************************

3 years, 9 months

Jump to page:

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

Moon-net August 2019