F4GFT Amateur Radio Station

My first QSO

I have always been interested in amateur radio, but living in rented appartments in big cities and/or having other objectives in life to pursue, it took me more than 25 years from the first time I felt the desire to hold a ham licence to the day I actually got one. An article about stealth antennas and magnetic loops in particular got me motivated enough to pick up the excellent F6KGL/F5KFF course pdf document, and to fix an appointment with the ANFR for a class 1 test 3 months later. My initial idea was to attempt the full HAREC licence including morse, but that was overly optimistic.

I passed my class 2 licence on december 14 2009. Four days later the callsign F4GFT was assigned to me, and I had my first QSO on december 19 at 1252UTC as DL/F4GFT/p with DJ4EW, while on vacation in Germany, using a modified Bosch KF161 and a sleeve antenna ("Sperrtopf"), on 2m via the DG0XO repeater at Bergheim.

My radio shack

During the first years of my ham radio activity in France I worked from a cramped space right under the roof, with hardly enough space to make a logbook entry on some scratch paper, and the only advantage that I was very close to the antenna. I am currently building a more comfortable shack covering this time the enormous surface of 2 square meters. The picture shows the shack in May 2015, a few minutes after the transceivers were reinstalled.

My DL/F4GFT shack

The first picture shows my secondary shack in Germany as of december 2009, right after I got my ticket, with a mobile Bosch KF161, modified by DL2ECT to operate on 2m. The X30 antenna shown to the right was added in late 2010, and is currently used with a FT-60 handheld transceiver, when I'm in Germany. I abandoned the idea to do something about the feedline losses; the RG58 will have to do the job.

This is what my DL shack looks like these days. The KF161 is gone, the FT-60 is my only VHF/UHF device, and a Yaesu FT-950 drives a temporarily installed 40m inverted-V (notice the PL239 socket at the window). The entire setup is temporary, in fact, as I only operate during my brief visits. (You notice that I was a phone operator in 2011 when the picture was taken - not that there's anything wrong with that!)

QSL

After my first VHF contacts, I was eager to send out QSL cards, and produced a small batch with my trusty old HP LaserJet II and its single page feeder. The second batch, created for my first short wave contacts, used the same, slightly embellished layout printed on a commercial color copier with single page feeder. I sent out almost 2000 cards of my third batch, printed on glossy 200g papier with a background image evoking my city.

By the end of 2012, when it became clear that I would never have an F8 callsign prefix (reserved for class 1 licences with CW proficiency), since the code requirement was dropped in France, I ordered professionally printed customized colour cards. Having waited three years, I knew exactly what information I wanted on the back side of my card, in particular three QSO lines and a checkbox for my frequent operations as DL/F4GFT on my secondary QTH. Finding a motive is not difficult in Paris, there are so many photo opportunities. Mine is the pedestrian passage ("passerelle") Simone de Beauvoir, right by the new national library. I took this picture one evening when I passed there by bicycle and immediately liked it. I have chosen it for several reasons: I live nearby, it is not the typical Paris cliché, and most of all, it looks better than a picture of me in my shack.

My handheld transceiver: The Yaesu FT-60

The Yaesu FT-60 replaced the FT-50 in 2004, and has earned itself quite a reputation as reliable, no nonsense dual band handheld transceiver. It has none of the features of more advanced HTs, such as Bluetooth or APRS, but packs all the essentials in a sturdy package. I like it so much that I got myself a second one before the FT-60 will be discontinued.

Being more than a decade old, the FT-60 comes with NiMH batteries. These batteries have the advantage of being cheaper than LiIon and not requiring a smart charger, but they need a bit more attention: Charging time with the provided dumb charger is 10h if the battery is completely discharged (avoid deep discharge though), and should be respected, otherwise the battery becomes hot and is damaged. A smart charger avoiding all these considerations is available from Yaesu, but is rather expensive.

The stock rubber antenna is actually quite good. I replaced it with a Diamond SRH-771 until it broke, and now have a Smiley 270A.

What I really like about the FT-60 is the robust package. It travels with me on bike trips, ski holidays, whatever. I wish there was a leather protection case, but even without, the unit handles rain, extreme temperatures and shocks very well.

There are really only two things I disklike about the fT-60. The plug of the optional speaker- mike occasionally gets loose when I use it on my bike, and when it does, it triggers the PTT. A rubberband helps, but a threaded connector like the ones found on more expensive professional radios would be better. I also would like to be able to disable the 200 and 900 MHz bands; I have no need for them, and having to toggle over them when switching from VHF to UHF is just a nuisance.

I noticed that many handhelds get alarmingly hot when used on full power output for longer periods. Not so with the FT-60, it gets warm but temperatures stay reasonable. I attribute this to the metal frame. However, when used on 13.8V external power, it gets hot even on 0.5W. Since the FT-60 runs on a native 7.8 V battery, it has to convert the excess voltage into heat. In my stationary setup I solved this problem with a 2A 78S08 voltage regulator.

The FT-60 is often mocked for its WIRES capability. WIRES is Yaesu's proprietary Echolink-like Internet-based network protocol. Yaesu put the WIRES functionality prominently on the 0-button, which means it can easily be turned on accidentally. If done, it manifests itself by emitting a DTMF tone at the beginning of each transmission. Since WIRES appears to be used almosts exclusively in Japan, it is quite useless, and Yaesu would have better done to assign the functionality to a secondary function of a button or the menu. However, it IS possible to configure the ft-60 to "other internet connectivity systems" (menu 21), transforming the handheld into a quite capable Echolink transceiver. In this mode of operation, the FT-60 emits one out of 10 pre-programmed DTMF sequences with up to 6 digits, perfect for storing Echolink node numbers. My only and very minor complaint is that they cannot be text labeled.

My first base station: The Kenwood TS-450S

The Kenwood TS-4xx series of HF transceivers can be found at very reasonable rates on the used market, and are an excellent choice as starter rigs. All TS-4xx rigs have 100W RF output in SSB and CW, none of them has a built-in keyer, and all cover the WARC bands.

The TS440 seems to be the most popular model, and comes with RF power adjustment, 10m FM and SWR meter, which were lacking or only available as an option with the TS430. It can be modified to work with CAT, although CAT is rather limited in these rigs.

The TS450 is my favorite. Like the 440, it comes in two versions with (SAT) and without (S) tuner. It has everything the 440 has, plus a CAT plug, lacking only a level converter. With the matching power supply, it's a small and neat base station.

The 440 is known to have key bounce problems with the numerical keypad, but since this feature is rarely used, most users seem to agree that it's not a big deal. The 450 runs the risk of backup battery leakage, but the replacement is simple. The integrated tuner in both rigs is fast and has an impressive matching range. These are non-DSP rigs, but first and second IF filters can be added easily and do the job quite well.

My first contest

In my first months on HF, I accidentally got into a strange activity with brief and rapid exchanges. I did not know what a contest was (it's not part of the exam question pool in France), answered the CQ, was unable to provide the requested number, and was told to shut off if I don't know what's going on. So much for a warm welcome on the short waves!

I looked things up on the internet, and decided to try my newly gained knowledge during the next contest that would come around, which happened to be the SP DX contest during the first weekend of april. I had three 20m contacts with my indoor Bazooka, and was quite surprised when I received a certificate for being the first french station in the SO 20m Phone category. The fact that I was the only participant helped a lot, of course.

My first DX

Late at night on April 27 2010 (so late that technically it was April 28) I had my first DX contacts, with W2YP and AA2WC on 20m, both located in the state of New York. My short inverted V got me there, but I believe the contact was more likely due to excellent propagation conditions.

My secondary base station: The Yaesu FT-950

I was tempted by the FT-950 for quite some time, and eventually got one in spring 2011. The 950 is a beautiful rig with excellent ergonomics, and a highly informative vacuum fluorescence display (In 2011, Yaesu is the only company to use VFDs in their base stations) which I prefer over LCD or TFT any time. Yaesu left out the RF power adjustment knob, probably to distinguish it better from the FT-2000, which also comes with a second receiver and a built-in power supply, but since there is a short cut button to call any of the more than 100 menu items directly, this is not really an issue for me. I do believe the FT-950 was the best buy in its price category (it has been replaced in 2014 by the FT-1200). It does have a few quirks, though:

Compared to my Kenwood TS-450S, the FT-950 makes a lot of relay noise on various occasions, not only when pushing the PTT. There's even a relay to cut audio when turning down the volume beyond a certain level.

But the one thing that bothers me most is the poor antenna tuner. Where the Kenwood TS-450SAT uses servo-driven capacitors to tune almost instantly and within a wide range, the FT-950 uses relays to activate Ls and Cs. Not only is the range surprisingly poor, but the algorithm also is rather on the slow side. The tuning process takes quite some time and makes a horrendous noise. The narrow range is the reason why I use the Ft-950 at my second QTH and keep using the TS-450SAT at my main QTH, since it allows me to work 10-80m on my inverted V dipole. The FT-950 requires a resonant antenna and can only make minor adjustments.

My first PSK31 QSO

Over christmas 2011, I hooked my FT-950 up to my personal computer. I had read lots of warnings about the risk to hook up a transceiver to a sound card without isolating audio transformers, but since I had none at hand, I wanted to give it a shot anyway. The FT-950 has a dedicated Packet/RTTY 6 pin Mini DIN socket at the back, which makes this easy. The Soundcard input is straightforward, the output however requires a voltage divider to bring the 1V output down to the 50mV the FT-950 wants to see. As for the PTT, I simply used an npn transistor wired to DTR to pull down the PTT input. The entire circuit worked surprisingly well, and I even had 2 RTTY QSOs in AFSK mode. I then discovered that the FT-950 has a data vox facility and does not even need the PTT switch: After setting menu 114 TGEN to "data", the PTT circuit was obsolete. It is also noteworthy that the ft-950 comes with a complete set of menu items to adjust input and output level, allows true FSK via a dedicated pin on the Mini DIN socket, and also provides a Squelch output, which could be used by more sophisticated decoding software to distinguish an unmodulated carrier from no signal at all. It is clear for me after this experience that I will not buy a commercial interface for my FT-950.

My first Hellschreiber QSO

With an AFSK interface and a program like DM-780, almost any digital mode except ATV is possible. In March 2012 I stumbled over a weird signal, which I was able to identify as a Hellschreiber QSO thanks to a digital mode field guide I found on the net. At the end of their QSO, they were listening to other Hell OMs, and I had a short contact with a swiss operator who was actually working with a real vintage Hellschreiber from WWII. His machine worked better than my personal computer, which is actually too slow to update the screen and generate the Hell signal at the same time.

2m SSB with the Kenwood TR-9000

When I came across a mint Kenwood TR-9000, complete with matching PS-20 power supply and BO-9 system base, I could not resist. The system base cures three shortcomings of the TR-9000, and turns it into a real base station: It connects to the TR-9000 memory backup socket and delivers the backup voltage for the otherwise volatile 5 frequency memory, adds a speaker and a phones plug, and also provides a MOX switch necessary for CW operation without awkward PTT fumbling.

The TR-9000 comes in several versions. Mine is the european model, with a 1750 Hz TONE button instead of a Hi/Lo output power toggle switch. The TR-9000 does SPLIT, has two VFOs, 12.5kHz steps in FM, and scans fast. Now I need a rotor and a 2m Yagi.

My first 80m contact

On October 12 2012 late at night, I had my first 80m contacts with DH4SC and DJ5YV. A few weeks before, I had replaced my monoband-dipole with a Kelemen 4-band trap dipole covering 10m, 15m, 20m, and 40m, and noticed that I was able to tune the antenna to 80m and that the noise level on 80m had gone up considerably. This encouraged me to give it a try, and on my third or fourth attempt, I was able to establish a contact. My signal was weak, and it is certainly due to the efforts and patience of the other operators that it actually happened.

Since these first contacts, I regularly make 80m contacts in CW. Apparently my less than appropriate antenna, while not sufficient for 80m phone ragchews, does the job for CW signals.

My first 6m contact

Since the 7th of march 2013, operation on the 6m band is authorized in France between 50 and 52MHz, with a max. output power of 120W. I had anticipated this decision with the acquisition of a 50MHz module for my FT-736R, which according to the manual is able to provide up to 15W output power. I also laid out a second line of H155 coaxial cable between my new shack and the TV mast on the roof of our building. On the 13th of june I took a day off, finished the assembly on my homemade 6m stiff dipole antenna, climbed the roof and installed the antenna. The first SWR verification showed 1:1.4 on 50.150MHz, upon which I declared the installation terminated and began operation. I immediately found myself in a massive band opening, and had my first contact on 6m with Walter, IK5ZWU, in JN53VB. Altogether I made about 20 contacts, mostly with I and EA, but also with CT and even one with LZ, and a local contact with another station from Paris. I had read a few articles on operation on 6m and knew that exchanges on this band are short and include the locator. I was very happy with my first day on this new band, but a few months later, when I did some stats on my log, I could not help but notice that band openings are rare. A dedicated transceiver is definitely required to constantly observe the band and be ready when the opening occurs.

My first CW contact

On december 19, 2013, exactly four years after my very first QSO (on a 2m FM repeater) I had my first QSO in CW as DL/F4GFT with DK1QG (technically I had a really first CW attempt with IZ5OVP on december 8, but since I used the PTT of my microphone, I do not know if that counts). I was hoping that this would be the end of a 4 month journey to CW proficiency, but it wasn't: I immediately realized that I was far from being operational. I made almost 40 QSOs during a 10 day holiday, and decided to pursue my learning campaign on LCWO.net. The thing about the Farnsworth method is that one gets accustomed to the pause after each character, which in a real life QSO just isn't there. Worse, if I shout QRS QRS! people slow down altogether, making characters learned at 18wpm hard to decode. My admiration for those who just smoothly stroll along at 35wpm just keeps on growing.

The Yaesu FT-950 on CW

I wrote here that the FT-950 was a bit disappointing on SSB compared to my trusty old TS-450SAT. On CW it's quite a different story! First of all, the FT-950 has a keyer built in, with 5 memory slots and contest counting facility (did not use that yet). But there's more: There is a "narrow" button which modifies the "width" DSP filter, which now ranges from 100 to 500Hz. And it really works, I am able to eliminate a signal at 300Hz above or below. There also is an audio peak filter, although nothing on the front panel of the FT-950 says so. It is activated with a 2 second push of the contour button. The CONTOUR indicator in the VFD then disappears, and the indication shown in the picture appears. The effect is impressive - the noise goes down, the signal clearly stands out. I also like the zero-beat indicator. As a newby, my intuitive tuning more often than not does not zero-beat the signal exactly. With this indicator, it's a breeze. If you look at the picture you'll notice that IPO is on (turning off pre-amplifiers on 40m), and the attenuator is switched to -6dB, to further reduce noise. The FT-905 IS a quite capable CW rig.

The Kenwood TS-930S

The latest addition to the shack is the Kenwood TS-930S. This transceiver was Kenwood's flagship model from 1980 to 1985, and I must have seen it first around 1984 by the time when I considered becoming a radio amateur for the first time. I could never have afforded the rig at that time, but apparently it made a sufficiently strong impression on me that I took the opportunity when an almost mint one with a 5xxx serial number came along 30 years later.

I had read the usual reviews and knew about the shortcomings and required modifications. Early models suffer from through-hole failures on the digital controller board; a problem which was addressed by an after-market replacement board from Piexx, which adds CAT, non-volatile memories and a few other features. The biggest shortcoming of the 930, however, is the power supply. The transceiver uses 28V for the final PA, 21V for the tuner from a second transformer tap, and generates auxiliary voltages for the controller boards using resistors and zener diodes with considerable heat dissipation. There are basically two approaches to circumvent the problem, the most radical being a complete replacement of the transformer and power supply by a modern switching PSU. In my unit, the W6NL-Modifications had already been carried out, which was quite a relief when I discovered it. Another weak point of the TS-930 is the mode switch, which put my transceiver out of service for almost a year.

The TS-930S is famous for its very selective receiver in CW, which is the reason why some contest operators still use it today. This performance is largely due to an INRAD roofing filter, an adjustable audio peak filter, and a feature called variable bandwidth tuning, which is the analogue equivalent of today's DSP transceiver's bandwidth control. I put the TS-930S into operation during the 2014 IARU HF Championship contest, and was able to cross-compare with my trusty TS-450S. In narrow band mode, the 930S is capable to completely isolate signals which on the 450S cannot be separated, even with CW filters in both IFs activated. At the same time, the sound is clear and undistorted, and should not induce operator fatigue during long contest hours (The latter does not really matter for me, since I am a casual operator). I understand that the 450S receiver is not really a reference (neither is the Yaesu FT-950), but I have no contemporary high end transceiver to compare the 930S with.

I was tempted to activate the 10Hz digital frequency readout, but since my TS-930 needs a realignment on some bands, where it is up to 200Hz off, I figured that would make no sense. One nice feature of the TS-930 are the phone patch Cynch sockets on the rear side - I had my PC soundcard connected in no time.

LOTW

According to the eham reviews, the ARRL logbook of the world (LOTW) is a mixed bag, people either hate it or love it. My experience was quite smooth: I installed the software, generated a .tq5 file, sent it to ARRL, and sent my licence and a copy of my id document by mail to the ARRL headquarters. A few days later, I received a password and a .tq6 file, and I was ready to go.

The only real obstacle in my case was the conversion of my Excel-based log to an ADIF file. I tried to do it within Excel, which is possible and has in fact been done by some, but since my log contains a number of inconsistencies which need to be rectified in the process (such as frequency written sometimes in MHz, sometimes in kHz, or european vs. american date format) I figured it would be easier to write a small C program to do the job and carry out the necessary corrections on the fly. I also added an option to ignore 2m and 70cm FM contacts, 99% of which are repeater contacts, and another option to write out only the data necessary for LOTW. In fact, LOTW does not require frequency or signal reports, all that is needed are date, time, band and mode (and the callsign of course).

My first upload consisted of about 4000 SW QSOs made in the 4 and a half years since I obtained my licence. The process was painless, and an instant later, about 25% of my QSOs were confirmed, covering 78 DXCC entities (56 CW, 67 SSB, 78 mixed). The log upload also showed me a few errors in my log, where band and frequency did not match. But there's more: Since LOTW uses a blind check and signatures, its results are trustworthy, and therefore accepted by the DCL (that would be the DARC contest log). The import on DCL showed me that I did qualify for WAE (Worked all Europe) in SSB and CW.

One thing I noticed after a few months of LOTW use was that there are times when the ARRL server requires significant time to process logs. I think this is where there is some room for improvement - the TQSL software confirms that the log was successfully submitted and is queued for processing, but after that, there is no way of knowing that the log is in the queue, since the web site only displays the history of already processed logs. Since I am rather the patient type, and DXCC is a life long achievement, it does not really matter to me.

I have come to the conclusion that LOTW and QSL cards serve different purposes. I will continue to send out cards for the sheer pleasure of it, and rely on LOTW for DXCC, and via DCL for DARC diplomas.

My first SSTV contact

Ever since I built my PC interface in late 2011 I kept looking for new digital modes besides the ubiquitous PSK31 and RTTY signals, and SSTV was on top of my list. I eventually managed to find SSTV signals in january 2015 on 14230, and had my first QSO on january 28 with F5SIZ. But one thing after another:

The first image I captured was the severly slanted test picture on the left. I grew up at a time when television stations shut down at midnight and resumed operation in the afternoon, and transmitted test pictures during downtime, so I was familiar with TV test images. After I found and activated the "automatic slant correction" checkbox in MMSSTV, I received the image shown on the right. Bingo!

Of course, it is almost impossible to stumble over a SSTV signal by accident. I had read an SSTV primer, which gave me three important informations: most activity takes place on 14.230MHz (although there are other designated SSTV frequencies), the modulation is always USB, and the signal contains a sync pulse on 1200Hz and the image information between 1500 and 2300Hz. MMSSTV shows a spectrum, which facilitates tuning in case the signal is not located on the 14230 KHz calling frequency.

Now it was time to try to send. Up to this time, I was receiving in SSB mode. I had to switch to AFSK mode to enable transmitting using my homebrew interface. My first QSO was not a total success, because I had made a mistake - in AFSK mode, my rig has default DSP filter settings with 2000 Hz bandwidth, cutting off part of the useful bandwidth. This is what I got.

The next day I understood and corrected this error, and had three QSOs. I think SSTV is a fun mode, probably worthy to find its way into my shack permanently.

DXCC and WAE I

DXCC is the most sought after amateur diploma, and after I got LOTW set up and saw my score, I decided to go for it. I knew it would be difficult in SSB, but in CW I was able to reach the required 100 confirmed contacts within 16 months. I had to combine 93 LOTW confirmations and 7 cards, which were verified by a card checker in DL within 48 hours. I thought I had understood the application process: Complete the online form on the ARRL site with a list of QSOs for which QSLs had to be sent to the card checker, then complete another form to use the LOTW QSLs to complete the application. However, it was only after I had linked my DXCC account to the LOTW account (it's a simple process requiring the press of a button and the entry of the callsigns used for the LOTW application), the manually checked QSLs showed up on LOTW. Now I understand the meaning of the column labeled "DXCC credits awarded).

I thought that DXCC would be tougher than WAE I, but I was wrong. Worked all Europe Level I requires 60 DXCC entities in Europe, and 200 contacts on different bands, with up to 5 band points per DXCC entity. The real hard part are the band points - all of a sudden, entities like PA or ON which are trivial on 40m become really difficult on 30m, and almost impossible on 20m. After I got stuck on 170 band points, I made up a shortlist of prefix/band combinations I needed to complete WAE I, and obtained WAE I after 21 months of activity in CW.

My first /p QSO

In the summer of 2015 I bought a friend's FT-817, a rig I was fascinated of for some time. It must be the longest running HF rig in Yaesu's current line-up, and should accompany me on my trips abroad.

To complete the setup, I bought a third party NiMH battery pack with higher capacity than Yaesu's original, and a tiny CW paddle from OM2JK. My first opportunity to give it a try came in late october 2015 during a trip to Berlin.

Once there, I noticed that it is not an easy task to mount a dipole. I had brought a shortened trap dipole of 2x6m length, the same type of dipole I use at my stationary setup. I had to try several configurations in my city appartment until I found one (with one leg of the dipole hanging down the balcony and the other one being inside) which gave me spots on the reverse beacon network in DL and HA on 40m, a band my dipole was not supposed to work on! The first picture show me in my working position, with the balun behind me. In the city I tried several locations in Tiergarten (a famous centrally located park), but I just could not find a means to get the dipole up to a significant height. Anyway, it was a start, and since then, I've been /p a number of times.

My first SOTA activation

In june 2017 I found myself in a hotel on the bottom of Arthur's seat in Edinburgh. Arthur's seat is the highest point in Edinburgh with 251m above sea level, and can be climbed from Holyrood Castle, with three different paths available, with different difficulty level. Legend has it that King Arthur found the famous sword Camelot up there.

Arthur's seat is registerd as SOTA peak GM/SS-272, and being so prominently visible from my hotel, I decided to make it become my first SOTA activation. On the day of my departure, I did some research on the internet and learned that written authorization is required to play radio on Arthur's seat. I applied for a permit via the internet site of the Historic Scotland Ranger Service, and received my autorizing letter less than 24h later, accompanied by a friendly letter, in the finest tradition of scottish hospitality.

I went up on June 14 in the afternoon, after a visit to Holyrood Palace. The path up there starts smoothly, and gets considerably steeper towards the end, without serious difficulty though. I reached the peak about an hour or so later (I did not really keep track of time), and put up my antenna on a plateau somewhat below, the actual summit being extremely crowded. I about an hour, I made 3 contacts with my FT-817 at 2.5W output and a small vertical antenna. I was a bit disappointed that I was unable to make the 4th contact necessary to claim my bit of SOTA fame, but as the wind got heavier and blew my antenna down, I felt it was time to call it a day. It would have been helpful if I had put a spot on the SOTA reflector - I will do that next time. I also learned on that occasion that a radio activity up in the wild attracts the curiosity of quite a few bypassers, and on my next occasion, I will bring SOTA leaflets to hand out. Here's my short but respectable SOTA log for GM/SS-272 as GM/F4GFT/p:

My first FT8 contact

In the late evening on January 1st 2018 I had my first QSO in FT8, using the prototype of my newly built audio interface. As opposed to the interface I had built 6 years ago, this one uses a pair of P1200 transformers for full galvanic isolation between transceiver and computer, and allows switching between two transceivers. Anyway, this rant is about FT8 and not my interface: Encouraged by a fellow ham, I installed and started WSJT without reading the user manual, and immediately had the screen fill up with stations calling CQ. I then enabled my transmitter, and noticed the software to send CQ automatically.

With auto sequencing enabled, an entire QSO took place automatically! I then double clicked on a station calling CQ, and again an entire QSO sequence took place. All I had to do was complete my log and fill out the QSL card. I'm sure there's a way to include this in the automatic sequence as well. I'm not yet sure what to think about this new mode, but chances are my DXCC digital count will go up rapidly.