Posts Tagged ‘repair’

Getting off Windows XP

May 18th, 2014

With the impending end-of-life of Windows XP a few weeks ago, I found myself with a bit of a predicament:  My shack PC is a dual-core 64-bit system (3.0-GHz Pentium-D) so it’s not really a slouch performer.  But, it had only 1 GB of RAM, which was fine for XP.  I could upgrade my RAM and upgrade Windows, or I could upgrade the RAM and dispense with Windows altogether.

I opted to drop Windows.

My almost-3-year-old son helped me install 4 GB of RAM last weekend.  (Actually, he spent more time asking about the capacitors on the motherboard. Kids these days.)

I’m a long-time GNU/Linux user (just over 15 years, actually) and fan.  However, as I wrote once in the past, I’m also fan of getting things done.  So, I kept using Windows 95 / MS-DOS for many years on my hamshack computer, with a brief several-year foray into using Windows XP when I moved to TR4W logging software.  Linux finally became ready for prime time in my hamshack when Kevin, W9CF, ported TR-Log over a couple of years ago.  After ensuring it wasn’t a fad, I was ready to jump.  Over the past week or so, I’ve been tailoring the setup of everything to bring back the functionality I had previously.

So, what broke?

  • I have an nVidia graphics card for my second monitor. Failure. Not sure how much I care. Maybe I’ll look for an old ATI card or something else with drivers?  Anybody have a plain PCI video card with VGA or HDMI kicking around?
  • Windows-specific software that I’ll probably just quit using:  AVR Studio and SH5. Didn’t really use these much anyway.
  • Now, for the truly bizarre:  My Elecraft K2/100 throws an ERR 080 on power-up if I disrupt communication between it and TR-Linux (does not matter if the K2 goes off first or TRLinux quits first).  I have to disconnect the RS-232 cable at the back of the K2 for a while and then it comes back OK.
  • ARRL’s Trusted QSL software version 2.0.1 didn’t compile immediately without dependency problems. Further study indicated.

And, what works?

  • TRLinux talks to both the K3/100 and the K2/100, as well as the YCCC SO2R+ controller.
  • Made some QSOs and uploaded the signed log to LoTW.
  • Pretty much everything else…

W3APL/B 903-MHz beacon

May 7th, 2013

Late last Summer, it came to my attention that the 903-MHz W3APL beacon had gone off-line.  The failure was intermittent and seemed to resolve itself after power was reset.  Several efforts to troubleshoot it were undertaken by myself and others, including running it at high duty into a dummy load over a period of days.  I was unable to get the problem to manifest itself on my bench.

A synthesized source (Analog Devices demo board) was offered by a friend of the Club, however it did not produce the desired output (or any output at all).  It’s not clear whether this was the fault of the synthesizer or the user (me).  The notional plan was to replace the beacon, which consists of a 75-MHz crystal oscillator followed by 12x of multiplication and a small RF power module, with the synthesizer and a new RF power module.  The project languished, as they often do in my hands.  But, two weeks ago I picked up the task again and made some real headway.

Really, the failure had to be one of a couple of things:  1.  Intermittent connection exacerbated by thermal cycling.  2. Oscillator “unlock” due to component aging and thermal cycling. I reasoned that as long as we could eliminate #1, the multiplier chain and amplifier should be fine.  The behavior seemed to point toward #2 or perhaps a combination of #1 and #2.  I came across a forlorn Programmed Test Sources PTS-040 that I had rescued from another group’s surplus heap to put in my lab.  I hadn’t used it in the two years that it was in my possession, so it seemed logical to provide it to the Club on a long-term loan.  The problem was that it didn’t go up to 75-MHz.  So, I cooked up a little multiplier chain.  My “good” HP spectrum analyzer is on-loan to a paying program so I had to make do with the FFT function on the fastest Tektronix portable scope I had in the lab.

My initial effort at the multiplier chain was to build a 2N3904 amplifier that swung way into saturation producing a signal rich in harmonics.  I went straight away for the 903-MHz signal but I couldn’t get a good enough lumped-element filter to eliminate the adjacent harmonics.  So, I tried for the 75-MHz injection.  This demanded a buffer amplifier so I lazily reached for the MMIC drawer in and retrieved one of the plentiful MAR-8s.  Plenty of gain…and, as I would find out in a moment…conditionally stable!  To exercise the eloquent euphemism of Ben, N3UM, the MMIC “burst into song” at about 63 MHz.

Back to the drawing board.  I knew that I had something that would work, so I redesigned the deadbug layout on an SMD protoboard (the kind with all the pads in a grid).  I replaced the discrete 2N3904 and MAR-8 MMIC amps with SGA-4586Z MMICs (which are a little too nice for this service, but I have a ton of them).  Viola!

w3apl_b_903

It’s the little board on the far wall of the diecast box with the SMA connector on the left and two toroids.  37-MHz RF comes in from the PTS-040 through the BNC jack in the wall.  It’s multiplied up to 75 MHz on the new board and piped down to the remaining 12x multiplication and amplification stages before going to the little brick PA in the lower left (not visible).

So far, it sounds good.  I was able to monitor it with my W1GHZ transverter strapped to the IC-290A in my car and using a WA5VJB cheap Yagi tossed in the back seat.  I lost the signal about 5 miles away with that setup, which is really pretty decent all things considered at that frequency, etc, etc.  Nominally, the frequency should be 903.054 MHz.  I found it at about 903.048 MHz on the lash-up.  Brian, ND3F (aka N3IQ/R) reported that he found it at 903.046 MHz with KA3EJJ’s setup.  If you’re in the vicinity of FM19ne and are setup on 902/903, we’d appreciate a report.  The big thing is the long-term stability.  So, we’ll continue to monitor it.

Now…to get back to that 930 on my bench…

(No?) Nonsense Radio

November 10th, 2011

The November 2011 issue of QST contains an Op-Ed that really left me shaking my head more than normal.  The author bemoans the complexity and feature sets of newer handheld radios and pines for the days of his IC-02AT.  He goes on at length about the “unnecessary” receive capabilities (NOAA weather broadcasts, AM/FM radio, etc) and how he has to search for the manual every time he wants to program a repeater offset.

Well, as someone who recently upgraded from a radio just slightly newer than the IC-02AT to a “modern” HT, he’s wrong on nearly every account (except the micro-/mini-USB port, which I would wholeheartedly support for charging purposes).

  • Eliminate extraneous features.  Too bad we all have different definitions of this.  I think scanning is a worthless feature, but like NOAA/NWS weather broadcasts.  In fact, my wife is delighted that we now have a battery-powered AM/FM+NOAA/NWS radio again that I will always be able to find and will guarantee that it works.  Did you hear that, guys?  My non-ham wife actually likes my HT and uses it to listen to FM radio!
  • Eliminate multilevel menu trees.   I’m just dying to replace my cell-phone-sized VX-3r with a knob-covered brick.  I’m sure you are too.  It’ll look great in my shirt pocket.
  • Eliminate the proprietary programming cables.  Maybe I’m not a typical ham, but I only have about ten memory channels programmed into my VHF/UHF FM radios and they took about 10 minutes to program through the front panel (my bad, menus).  The mini-/micro-USB port is a good idea for charging, though.
  • Allow for a battery pack that uses disposable batteries.  Last time I checked, most radios have this option.  Did I miss something?
  • Create an inter-vendor standard for user interface.  What if they standardize on Icom?!?!  The last Icom VHF/UHF FM radio I used received a “grade of S, for ‘stupid'” from its owner.  That was in 1993.  All of the Japanese manufacturers will be put out of business by the factory owned by the Chinese military that produces their products before this happens.

He should buy another IC-02AT if he liked them so much.  I bet for a Jackson or two, you could have a nice one…complete with the 6x AA battery holder.  Heck, buy two or three for spare parts.  I think I have the Service Manual around here somewhere if I didn’t already sell it.

On a more serious note, there are lots of no-frills radios available out there, even brand new ones with factory warranties.  Until recently, at least, the money in VHF FM radios was in two-way, government, and public safety, not amateur.  There are a lot of amateur rigs at the “low end” of the market that share a lot in common with their commercial counterparts.  And, of course, you can always buy used Motorola gear on eBay if you desire ultimate performance and ruggedness.

Amateur Wiring

October 13th, 2011

I’ve picked up a couple of Mirage “brick” amplifiers over the past few years.   I’ve also rewired the DC cable on each and every one.  Here’s the latest.  AWG 10 to AWG 16 transition?  Seriously?  I know it technically doesn’t matter too much for a short run, but these amplifiers suck down a lot of current and I just don’t see cutting corners on that.  Maybe I’m a purist…

A deaf VX-3R

October 12th, 2011

I do not spend a lot of time on VHF/UHF FM and have not for many years.  My first radio in 1993 was an Alinco DJ-580T handheld and I’ve thought at various times that a new handheld might suit me well, especially for receiving.  A few weeks ago, as I tend to, I came across a Yaesu VX-3R offered “for parts or repair” on attractive terms and so acquired it—my second handheld ever.  The problem seemed straightforward enough:  full TX power, but no RX on amateur bands.  RX on FM broadcast OK.  The seller assured me, after I had agreed to purchase based on his description, that “somebody” had told him that this problem occurs when “only one component is bad.”

Based on the seller’s description of the problem and a thorough reading of the Technical Supplement, I developed a short list of candidate failure modes, components, and sources.  Fortunately, all of the components could be sourced easily from the usual sources.

When the radio arrived, I gave it a functional check and it indeed exhibited the problem that the seller had advertised.  I quickly popped it open and noticed a small red sticker in the lower right-hand (when facing the device like you would operate it) corner of the Main Unit (Side A, per the Technical Supplement’s notation).  A neatly hand-drawn arrow pointed to component Q1025, which upon closer inspection, was clearly damaged.

So, I quickly set about identifying the component and procuring a replacement.  It was a NJU7007F3 operational amplifier.  Huh…it did not contribute to any of the failure modes I had initially suspected.  However, a careful reading of the Technical Supplement indicated that this op amp drives varactor diodes in a tracking preselector—so, if it does not work, the radio will exhibit varying degrees of deafness in the amateur bands.  I found out later that the “somebody” was actually a Yaesu technician who had seen the radio for repair at the behest of a previous owner and left the sticker for me.  At any rate, this component clearly would have to be replaced if I were to fix the radio, so I set about looking for a source.

Mouser listed it in their catalog but wanted me to buy a reel of 3000.  No thanks.  At least they had it listed.  So, I did the next most logical thing—I made a list of other parts that I needed and called Yaesu.  The part was back-ordered to Japan for 4-6 weeks, but only cost 0.42 USD.  I bought three.

After however many weeks it has been, a package from Yaesu showed up on my doorstep tonight.  After repairing a damaged PCB trace (non trivial on something this small), I was able to replace it.  The little black speck in the middle of this photograph is the removed component.  For my non-US readers, the US 0.01 USD coin (“Penny”) is about 19 mm in diameter.

The radio fired right up and received NOAA/NWS right away.  The entire repair once I had the parts was about 30 minutes.  There are still two unbuilt SoftRock kits, an IC-290A with an unlocking PLL, and W1GHZ transverters for 903 and 1296 to be worked on…maybe tomorrow…

SoftRock v9.0 Lite+USB Xtall

June 25th, 2011

Brian, ND3F (aka N3IQ/R), gifted me a partially-complete SoftRock kit a few months ago on the condition that I put it on the air.  I’m making some progress on that.  I took this photograph to show the SparkFun USB break-out board installed in the Bud CU-124 enclosure.  The whole thing is assembled now, but there is probably a solder bridge somewhere.  This kit is going to be a lot of fun because it’s actually more flexible for experimenting than the present Ensemble II RX kit.

More on IC-290A disappearing display

May 6th, 2011

See original post here.  Convinced that I had it working, I put the IC-290A aside and worked on some other projects.  But, when I came back to it a few weeks later, the display was still disappearing.  So, armed with the schematic and the block diagram, I began narrowing it down.

The display digits and tuning A/D converter are both driven off the same bus.  But, since the disappearance of the display did not bring loss of receive, that gave me some confidence that it was not affecting the synthesizer.  In order to simplify the analysis, I elected to assume that it was a single component (almost certainly a cold solder joint or an electrolytic capacitor) at fault.

Another clue was that the VFO dial ceased to operate once the display started acting up.  Taking the single-point-of-failure assumption into account, this strongly implicated the microprocessor on the LOGIC board as a starting point.  The only suspicious device that I could find on the schematic was C9, a 1000-uF, 6.3-volt electrolytic capacitor on the Vcc line of the microprocessor.  So, I replaced it with a similar value unit from my junk box.

And, the radio seems to be working…for now.   I did park it on WA1ZMS/B overnight and it is still running.

IC-290A disappearing display

March 15th, 2011

Regular readers know that I enjoy a good deal on a good project and just such an opportunity presented itself again recently, this time in the form of an Icom IC-290A, which is a 10-watt 144-MHz all-mode transceiver whose age rivals my own.  It was from a large estate of much more valuable (think solid-state HF kilowatt, etc) items and the sellers were probably just happy to unload it since the display would come up at first but then disappear after a few minutes.  After cleaning a film of nicotine residue off of case and knobs, I set to work diagnosing the problem.

I tuned into the two beacons that I can always hear:  W3APL/B on 144.295 and WA1ZMS/B on 144.285.  They sounded just as good as they did on the TS-700, the IC-290H, and the borrowed FT-736R.  Then, the display went blank.  But, I was tuned into WA1ZMS and I could still hear it just fine.  Turning the main dial no longer changed the frequency.  When I turned the radio off and back on, a clicking noise emanated from the speaker above the background hiss, but there was no receive and no display.  Disconnecting the power supply and allowing the radio to sit for a few minutes restored normal operation…until the display blanked again.  This seemed to me to be a thermal problem, perhaps a bad capacitor or two as I had found in a IC-290H obtained on similar terms.

I popped the cover off of the radio and immediately noticed a few dry/cold/oxidized solder joints on the “Sensor Unit” board that sits on the top of the radio behind the front panel.

In each case, the bad joints were associated with electrolytic capacitors.  I desoldered all of them, cleaned the pads, and replaced the capacitors with new units.  It was somewhat of a miracle that I had all of them in my junkbox.

The problem persisted and for a while I thought it was getting worse.  For good measure, I went through all of the boards looking for leaky capacitors and cold solder joints and found nothing.  The “Troubleshooting” section of the Service Manual was an abject waste of pulp, but I did go through it.  The “PLL” (synthesizer) alignment section yielded no clues except that the reference oscillator (6.070 MHz nominally) level was a tad bit low in amplitude.  I heeded the warning of the manual that it had been carefully adjusted at the factory and should not need adjustment.  When the radio arrived, it was missing a number of case screws, suggesting that someone else had been inside.  So, who knows what the Golden Screwdriver hath wrought!

After reseating the row of connectors at the front of the “Logic/PLL” board just behind the radio’s front panel, I noticed that the problem seemed to have vanished.  It has not failed yet since doing that and replacing the Sensor board capacitors.  I hooked the radio up to a Bird Termaline wattmeter borrowed from W3APL and it made 10 watts on FM and CW.  The needle barely moved on a 10-watt full scale in USB.  So, I need to verify if it’s the microphone or the radio.  My new plan is to use this radio as the microwave IF and keep the more-powerful IC-290H for liaison and other portable purposes.  So, USB operation is critical.

So, like I always say, it seems that there are a lot of repairs that can be done just by cleaning things up.  I’m not totally satisfied with the technical explanation of why reseating those connectors seemed to have fixed the problem.  But, that is what happened and it is consistent with a lack of communication between the encoder/display unit in the front of the radio and the microprocessor in the logic unit.  It seems to be working for now!

Note (May 6th 2011):  See update here.

Recent tinkerings (9 Jan 2011 edition)

January 9th, 2011

Several people have commented over the years that I should “write more” on the blog.  I usually respond that I could spend my free time tinkering/hamming or blogging, but not both.  Here are a month’s worth of blog posts as freestyle poetry:

  • A section on my workshop has been added to k8gu.com under Engineering.
  • Discovered that although the SoftRock VHF Ensemble II won’t fit (barely) into the Bud CU-473 diecast box I bought for it, it will fit into an extruded enclosure that housed an ancient X-band radar detector I picked up at Dayton in 2002 in hopes of stripping the Gunn diode assembly and getting on 10 GHz.  Bonus points for thriftiness.  Photos will follow once I finish the project.
  • Did not observe Quadrantid meteor pings with the SoftRock VHF Ensemble II, but did notice something interesting about the W3APL beacon.  Need to investigate.
  • Have more CE/K8GU QSL cards again, finally.  Tonight, I might finish the bureau (and, ashamedly one direct) cards languishing.  Some K8GU (and AA8UP, no kidding) bureau cards are sitting here staring at me, too.  Not a big project, though.
  • Operated the NAQP CW on Saturday (8 January 2011) for four hours and twenty minutes and made 318 QSOs x 128 multipliers for 40,740 points before log-checking discounts.  This total is pleasing to me especially considering that it was almost all leap-frogging SO2R search-and-pounce, which can be very fatiguing.  It’s fun to see the rate meter stay over 100 when you’re S&Ping.  Good Sprint practice.
  • Still the best 30 minutes on the radio every week:  I operated the NS Ladder on Thursday (6 January) night and made my customary 30 QSOs x 24 multipliers for 720 points.  Hopefully, adding 160 will give me some momentum to increase this score back over 1000.
  • We had a spell of 50F (10C) weather on New Years’ Eve.  So, I put the 2-meter beam back up on the chimney.  It was formerly mounted on a steel mast that was ratchet-strapped to the chimney.  A strong wind (>50 mph gusts) before Christmas bent the mast (actually a fence top-rail) and I had removed it.  I cut the bent lower portion of the mast off and attached the remaining top portion with rotator to a “girder” constructed from two pieces of treated 2″ x 4″ x 10′ lumber joined with a half-dozen lag screws.  Again, I ratchet-strapped the entire assembly to the chimney.  The present configuration is much stronger and less prone to damage.  The 3-element 50-MHz Yagi is still on the ground until I actually get the transverter finished, which should be soon (as it has been for 12 months now).
  • Repaired a SoftRock v6.2 downconverter for WF1L and learned that you can solder leads back onto SOIC packages if you’re careful.
  • Have had delightful exchanges with KN6X and ZL1CDP about repairing the TS-930S.  Some of these discussions (and their fruits) may make it onto the site at some point.
  • Back in December, I started integrating the W1GHZ transverters using UT-141 semirigid jumpers with pre-installed SMA connectors obtained on attractive terms from Max-Gain Systems.  Mitsubishi RA18H1213G (1296 MHz) and RA30H0608M (50 MHz) modules arrived from RF Parts.  May have a lead on something less expensive with more gain for 903 MHz via HA1AG.  The big remaining tasks in all three transverter projects are the sequencer and IF interfacing.
  • Also in December, I had dinner with NS Ladder father Bill, N6ZFO, in San Francisco at the Hyde Street Seafood House & Raw Bar, which is a favorite of NA Sprint father, Rusty, W6OAT.  Yes, I did feel that I was in the presence of greatness.  (I had their excellent pork chops since I’m not into seafood, especially raw seafood.)  Like most contesters, Bill’s a super, fascinating guy apart from his radio contesting interests.

Pin 1 and a Pound of Ferrite

September 19th, 2010

The commonly-held wisdom goes that an “ounce of prevention is worth a pound of cure.”  As things typically go around the K8GU station, 50% of problems are solved by prevention and 50% by cure.  Today, it was a cure.  Readers of the blog are no doubt aware that I have a recently-discovered problem with QRM between my radios when they were on the 20- and 40-meter bands.  (I didn’t notice this problem until I got both of the TS-930’s up and going again; so, it’s probably the result of the “new” station location and arrangement.)

I pulled the 20- and 40-meter W3NQN filters out of the circuit and measured them.  They benefited from a little tweaking, but nothing that would have caused the problem.  I put them back in and was rewarded with much lower VSWR in across both bands.

It didn’t matter which of the radios was on 20 and which was on 40, the interference, a popping hash that followed the leading edge of my keying, was there.  The next step was to put a dummy load on the output of one radio, then both radios.  Even when transmitting into the dummy load, the interference remained.  This suggested to me that the problem was very close by.  All of my gear is bonded together with heavy straps.  So, I wasn’t too worried that it would be difficult to find.  I noticed that the 40-meter LED flickered on my KK1L band decoder box following my keying.  It became the prime suspect.

I found some split beads that AD8P and I bought from AA1K in the Dayton flea market a few years ago.  I’m not sure what mix they are, probably 43 or possibly 77.  Winding a few turns of the band data lines from each radio through the beads knocked down the interference a good bit.  I slapped some more on the audio lines going to the Heil HCS that I use for SO2R audio and the interference disappeared.  But, I was using the dummy load on one of the radios.  As soon as I went back to the external antennas, the inteference came back.

Next, I dug out a couple of the 2.4-inch diamater mix-31 toroids that I purchased in the first K9YC “group buy” of these parts back in 2005 or 2006.  I wrapped the DC power supply line to the KK1L box and the AC supply to the HCS.  And, the problems pretty much disappeared.  At least one of these is due to what K9YC calles “the pin 1 problem.”  Basically, if the shield is brought through the metal enclosure to the circuit board, it conducts (noise) current that’s riding on it into the enclosure.  So, I need to dissect the HCS and KK1L boxes to see if I can find a deeper fix than just slapping ferrite on the outside.

But, there’s hope for SO2R in this week’s NSL!