- ICF-7600D and ICF-7600DS a.k.a. ICF-2002 and ICF-2003
- A few 7600 series receivers compared
- Receiver utilities
- Digital shortwave portable (dual conversion)
- Coverage: 153kHz - 30MHz (German version: up to 26.1 MHz only)
- 5 kHz steps on SW; LW: 3 kHz, MW: 9/10 kHz; FM: 100 kHz
- 10 presets
- SSB reception (not in Middle East version)
- Fine tuning for AM and SSB (+/- 5kHz, 55.395 MHz osc)
- External antenna jack (not in German version)
- Rec out jack
- Alarm and sleep functions
- 1st IF: 55.845 MHz
- 2nd IF: 450 kHz
AM IF filter (earlier models): Murata CFW450H ceramic filter with 6 elements,
probable specs (may not be entirely accurate!):
- -6 dB bandwidth: +/- 3 kHz (6 kHz spec'd is a tad wide, this usually means way beyond 7 kHz real)
- Selectivity (@ -?? dB): +/- 9 kHz
- Ultimate rejection: 55 dB
- Insertion loss (max): 6 dB
- I/O impedance: 2 kOhm
- AM IF filter (later models): Murata SFR450I 6-element ceramic filter, specs and performance likely to be the same as for the SFR455I except for center freq (i.e. 4 kHz spec'd bandwidth, more like 7 kHz with a shape factor of ~2 real). Additionally there's a tuned IFT plus one tuned IFT each in front of the AM and SSB detectors.
- FM IF filter: Murata SFE10.7MA5-A 2 pcs., 280 kHz (yup, the 7600G still uses basically the same type! Check there for replacement types)
- 7600D/2002 and 7600DS/2003 mostly differ in the colors; a number of internal changes had already taken place before that transition (the filter change from CFW450H to SFR450I, for example). Comfusingly the 7600D also existed with a dark case, but can be distinguished from a 7600DS by the color of the panel with the LCDs and the buttons therein (namely, those are colored like on any other 7600D, while the background is darker on the 7600DS and the ON, SLEEP and OFF buttons are black instead of white there).
- Frequency synthesizer / microcontroller: µPD1706G + µPB556C prescaler
- Spec'd battery life: 12 hours with Sony SUM-3 (NS) "heavy duty" (zinc chloride) batteries
- Output: 400 mW max @ 10% distortion (UK model: 200 mW)
- Dimensions/weight: 180 x 120 x 35 mm, 650 g (incl. batteries)
- Produced: 1983 ... 1986 / 1987 ... 1989
- The ICF-7600D cost DM 648,- in Germany. In North America (source ), the ICF-2002 cost (US-)$210-250. The ICF-7600DS originally cost some DM 550,-, and in the end only DM 448,-. In the US, the ICF-2003 cost $210-270, in the end it could already be had for under $200.
- Typical defects: Dead clock module (display?), back of cabinet cracked around telescopic antenna base (spareparts = unobtainium), occasional dead 'lytics
- First digital receiver in this size
- Good performance for its size in 1983 (it was slightly superior to the ICF-2001)
- No battery indicator
- No stand on the back (unlike later models)
- Believe it or not, the loose contact problem (with the power supply connector) of the ICF-SW7600G already existed with that model! It can be fixed the same way.
- Selectivity (at least on earlier models) isn't that great by today's standards, neither is the noise level or susceptibility to overload – but then the set was introduced in 1983... It's still a nice SWL receiver for travellers, but certainly not a DX machine.
A manual in Dutch, German and Swedish is available at RigPix.
My take on the ICF-7600D
You can read about my findings while briefly checking out an ICF-7600D of the newer variant (with narrower IF filter) on my receiver review page.
- 7600GR – you name it.
- 7600group – originally created with ICF-SW7600G in mind, abandoned and plagued by spam nowadays
Specified power consumption
Total current spec from the schematic (with minimum output, I suppose):
- FM: 72 mA
- SW: 75 mA
- MW/LW: 77 mA
The DRM mod document talks about a current draw of 45 mA in FM and 50 mA in AM (55 mA in AM with DRM downmixer installed). Maybe later models were improved in this regard?
Bernhard Weiskopf kindly scanned and converted the schematic of the ICF-7600DS (int'l version), as of 1987.11.
You can also find a complete ICF-7600D service manual online these days.
If you need a service manual in paper form, there are various sources that sell 'em. I have mentioned a local one in the misc links section.
Opening the 7600D
Opening a 7600D has proven to be not entirely trivial. Here's the trick, as posted on 7600group:
From: h van vliet <vanvliethc@y...>
Date: Mon Feb 21, 2005 10:14 am
Subject: Re: [7600group] Re: IDF-7600D - How do I open it?
To open my ICF-7600D I loose (and remove) the two black screws on the L and R top and the two brass screws (one in each battery compartment). Then, because of a lip on the right lower side, I open the backside turning it like a page from left to right. No problem. Furthermore, like with laptops a strong thumbnail is an indispensable asset for loosening latches.
Hein van Vliet
From: "crabtreejr" <crabtreejr@...>
Date: Fri Jun 10, 2005 4:11 am
Subject: ICF-2002 circuit changes
I have recently obtained an original of the ICF-2002 service manual (9-951-280-11, 1983) and a photocopy of supplement No. 1 (9-951-280- 81, October 1984). This supplement was issued to cover two updates: additional color variation, and circuit variations to reduce drift of the second oscillator. I believe that this supplement covered the infamous warranty changes which Sony made to the ICF-2002 starting late in 1984 or early in 1985. If anybody has further information, please let me know.
The ICF-2002 and 7600D are considered to be identical, as are the updated versions, ie the ICF-2003 and 7600DS.
1. Additional color variation.
The original model was issued in a silver case. A black case was introduced. The following parts were added (silver part no. then black part no):
15 Knob (fine) 3-891-803-00 3-891-803-11 20 Button(E) 3-891-808-00 3-891-808-11 25 Strap, Hand 3-891-813-00 3-891-813-11 29 Lid, battery 3-891-817-00 3-891-817-11 30 Knob (vol) 3-891-818-00 3-891-818-11 31 Knob (tone) 3-891-819-00 3-891-819-11 32 Knob (mode) 3-891-820-00 3-891-820-11 33 Knob (sens) 3-891-821-00 3-891-821-11 42 Chassis 3-891-843-00 3-891-843-11 46 Holder 3-891-872-00 3-891-872-11 48 Lid (rear cabinet) 3-891-874-00 3-891-874-11 56 Screw 7-685-152-14 7-685-152-19 60 Front panel X-3891-809-1 X-3891-831-1 63 Label 3-891-871-00 3-891-871-11
2. Circuit changes
To quote the Sony supplement: "The circuit has been changed to reduce the drift of the 2nd local oscillator and location of some parts on printed circuit board has been changed."
The ICF-2002 circuit diagram was re-issued in the supplement. In the original the horizontal position was indicated letter, and the vertical position by number (as in an Excel spreadsheet). In the supplement and for the for the ICF-7600DS the notation was reversed.
The component changes were:
(schematic location original / supplement)
- C3 0.003uF to 0.0022uF C2/3B
- C58 47uF to 3.3uF D5/4E
- L1 3.9uH to 5.6uH C2/3B
- R4 82 to 120 C2/3B
- R5 2.2k to 1.5k C2/3B
- R39 added 47 /4C in series with C44
- R63 100k deleted H4/ gate cct of Q11 was between collector of Q8 and L19/C52
- R65 180k deleted I4/ gate cct of Q11 was across C52
- R67 1.5k to 100 I3/9D drain cct of Q11
- R168 33 to 220 B2/2B
Five of the ten components which were changed were on the antenna board. The easiest way to tell if one has an original or updated ICF- 2002 is to look at R168. It is on the antenna board just to the left of the top left hand corner of the external antenna socket. It should be possible to tell if it is 33 ohms (orange-orange-black-xxx) or 220 ohms (red-red-brown-xxx).
There appear to be some discrepancies in the manual and supplement which I have for the ICF-2002:
The original circuit diagram shows R66 connected between the source of Q11 and ground. The original PCB layout shows it connected between the source of Q11 R64/C52. This is the way in which appears in the cct diagram in the supplement. I checked my non-early ICF- 7600D and it is wired as per the supplement. Interestingly the manual for the ICF-7600DS (9-952-946-12) shows the circuit as per the ICF-2002 original manual, but the PCB layout (different from the ICF-2002) is for the circuit in the ICF-2002 supplement no. 1.
This component is in the drain circuit Q7, the AM rf amp The cct diagram for the original ICF-2002 manual and supplement no. 1 show C43 connected between cathode of D16 and drain of Q7. However the PCB layout shows C43 connected between the cathode of D16 and connection of C48 to C49.
The cct diagram in the ICG-7600DS manual shows the latter configuration.
Looking at the PCB layout, it is not clear to me why this component exists, as it connects between two points on the same PCB track. D209 is not shown on the parts list or cct diagram for the ICF- 7600DS, but still appears on the PCB layout.
I looked at the keyboard PCB on my ICF-7600D. C205, C207 and C217 were mounted on the track side of the PCB, not as per the ICF-2002 manual. What was marked as C222 on the PCB appeared to be C224.
Having documentation with apparent discrepancies makes tracking the history of the models 'challenging'.
The component changes shown in section 2 appear to have been carried over to the ICF-7600DS along with a small number of additional component changes I shall write a separate note at a later date describing these. An updated ICF-2002 and ICF-7600DS are essentially the same radio.
Thanks to Rick in Atlanta for the ICF-7600DS manual.
I would welcome any comments.
73 John KC0GGH
crabtreejr at aol dot com
On excessive battery drain
If you have an ICF-7600D(S)/2002/2003 draining the batteries even when off, this might help (again, from 7600group):
From: "Bernhard Weiskopf" <bweiskopf@g...>
Date: Thu Oct 2, 2003 10:17 pm
Subject: Re: ICF7600D/ICF2002 question
I have a service manual of the ICF-7600DS. Due to its age it is in paper shape.
In off mode there is working a 3 V linear regulator.
The main IC µPD1706G-511 or -519 (from NEC) contains the 10 preset memories which are buffered by the 3 V "computer battery" as well as by the 6 V "radio battery" (via the a. m. 3 V regulator). So you can change one of these batteries without loss of memory.
The 3 V regulator draws continuously about 30 µA out of the 6 V "radio battery", not more.
The clock module is powered only by one cell (1.5 V) of the "computer battery".
While the off-state current is about 30 µA (@ 6 V), the on-state current is 70...80 mA (@ 6 V and low volume).
He measured 10.7 milliamps with the radio turned off and of course 75 milliamps when it is turned on.
10.7 mA off-state current is too much. 75 mA in on-state is ok.
May be one of the switching transistors is damaged so it cannot switch-off anymore.
Looking at the circuit diagram I assume as critical part the swichting transistor for the AF output power amp (µPC1212C).
If a short circuit to ground occurs at the earphone jack, I guess an overload condition at the switching transistor.
Switch off the radio (tone switch to music) and hold your ear directly at the loud speker. Can you hear some low noise?
If yes, change the damaged Q32 (2SB1013).
(Other test: You must hear a noticeably pop-up noise during switching on (volume = min.). In case of a damaged Q32 there will be no pop-up noise.)
Audio frequency response
Via PM, Bernhard told me that he had, among other things, changed the
frequency response of the audio amp because originally it was too muffled.
A circuit simulation via LTspice had shown the following -6dB points:
- MUSIC: 40 Hz ... 2 kHz (after obtaining the circuit file and running the simulation with regard to the frequency response at the speaker, I'd rather say 70 Hz ... 4 kHz, with very little further decay up to some 10 kHz)
- NEWS: 150 Hz ... 2 kHz (120 Hz ... 1.4 kHz) with stronger attenuation of bass and highs
The lowest frequency the speaker can reproduce is about 250 Hz.
The front end
The ICF-7600DS/ICF-2003 uses a front end similar to later digital models (ICF-SW7600 upwards), but with 2SK152-2 FETs.
- Analog shortwave portable (dual conversion)
- Coverage: MW, 49m, 41m, 31m, 25m, 19m, 16m, 13m, FM (76 - 108 MHz on the international model ICF-7600A, 87.5 - 108 MHz on the ICF-7600AW).
- No external antenna jack
- Rec out jack
- 1st IF 10.7 MHz, 2nd IF 455 kHz
- IF filters: 2x Murata SFE10.7?? for FM, 1x Murata CFU455? + a tuned IFT for AM
- Dimensions/weight: 179 x 117 x 31 mm, 600g incl. batteries (which might work out to 640 g with alkalines or rechargeables)
- Spec'd battery life is approx. 21 hours with Sony SUM-3(NS) batteries (heavy duty = carbon zinc)
- Produced: (late) 1982 ... 1987
- It cost DM 398 in Germany, apparently about $200 ($150?) in the US.
- Typical defects: Nothing out of the ordinary for a set of its age, really. Scratchy volume pot, switch array with contact problems, in need of lubing or worn out, sloppy dial cord, dead electrolytics.
- Being a very sensitive and quite selective dual conversion set, the RX was superior to most other travel portables on SW available at that time – in fact it's far from bad today.
- Batteries lasted forever – about 40 hours.
- Fairly good dial accuracy (about as good as on the ICF-7601 on my ICF-7600A), even though on some bands there are larger unmarked ranges where you have to extrapolate.
- Good craftsmanship.
- Good AM audio with little excessive hiss, good speaker sound.
- Decent FM sensitivity and selectivity.
- Limited band coverage (no 22m, no 60m, top 300 kHz of 19m band missing).
- Although better than the ICF-7601, still rather susceptible to intermod due to the wide 1st IF on SW (a weak spot of the concept).
- No stand on the back, which would have been very advisable since the acoustic design requires that the speaker be able to "breathe" on the back. (Speaker guys would be referring to this as an "open baffle" or a dipole.)
- Some tuning backlash.
- Mediumwave shows a few spurious shortwave stations.
- The speaker is easily damaged when you open the radio, being glued to the front of the case – so beware. (My uninformed guess was that this was done for better sound, and indeed this is one of my best-sounding units with a 3" speaker.)
My take on the ICF-7600A
You can read about my findings with an ICF-7600A / AC-456C combo on my receiver review page.
A while back I scanned the relevant pages from the ICF-7600AW service manual. You can now download the result here.
- In case you should be in need of a new speaker for the ICF-7600A(W): The speaker is still available, for US$ 20.04 plus US$ 6.95 shipping. The phone number in the US is 1-800-488-7669 (SONY parts). You can supposedly get the speaker cheaper here: 1-800-248-0898
- If contact cleaner doesn't work too well on the switches, use a drop of liquid graphite (graphite powder) – of course, this should work on other radios, too.
- Analog shortwave travel portable (single conversion)
- Coverage: MW, 75m, 49m, 31m, 25m, 19m, FM (76 - 108 MHz on the international model ICF-7600, 87.5 - 108 MHz on the ICF-7600W).
- No external antenna jack
- Rec out jack
- Dimensions/weight: 180 x 120 x 35 mm, 640 g (incl. batteries)
- Produced: 1978 (1977?) ... 1981
- It cost DM 278 in Germany.
- Typical defects: Set entirely dead.
- Good selectivity, sensitivity and S/N on FM
- Batteries lasted quite long (current draw: about 35 mA; specified at 20 mA for minimum output)
- The selectivity in the AM ranges isn't great (6.1 kHz / 17 kHz @ -6 / -60 dB), and there are lots of IM products (it's just a single conversion receiver).
- Only five SW ranges are available.
- Test des ICF-7600W von Rainer Lichte (measurements are at the end of the text)
ICF-7600 schematic – as one can see, this
1977 design is much different from later models, only sporting a total of two
early ICs mainly providing FM and FM IF amplification and AGC. All the rest is
entirely discrete, with 13 transistors, 1 FET and a bunch of diodes.
While having two ceramic FM IF filters and thus good FM selectivity for its day and class, for AM IF there is no more than a CFT class filter (1 ceramic element with two IFTs) plus another tuned IFT to be found, making it the equivalent of a 6-circuit receiver in classic radio nomenclature (middle of the road level basically and comparable to the later Grundig Yacht Boys or Philips D1835, but certainly quite good for its size back then).
Looking out for potential trouble spots, this is what caught my attention:
- The set employs a number of Sanyo 2SC930 transistors, which are notorious for beta reduction and failing in other sets of similar vintage, like the ICF-6800(A)W. While the effect for FM IF amp Q3 or LO buffer Q6 would arguably not be as drastic, it's a different story for oscillator transistors Q4 and Q5. If the LOs are dead because transistor gain is not sufficient to maintain oscillation, all you're going to get out of the set is some faint amplifier hiss. This may very well explain the "set entirely dead" fault.
- As typical for oldschool receivers, there are a lot of switch contacts that must work correctly, including in the oscillator circuits. Degraded Q is not going to make problems in that area any better.
- There also are quite a few small electrolytics in the audio filtering, which may be dried out (it's been close to 40 years by now). Some values may have to be replaced with film caps nowadays, as I don't think you can still get .22µ and such.
- The audio amplifier sports one of these dreaded VDxxxx (here: VD1220) varistor diodes which have proven troublesome in early/mid '70s audio gear, causing all kinds of cracking and other funny noises. Two small-signal silicon diodes (e.g. 1N4148) in series should make a half-decent replacement if necessary.
For a complete restauration, I would suggest a workshop not afraid of working on old radios, or a vintage radio enthusiast if you aren't one yourself. Working with dial mechanics and the like in a very compact set isn't everyone's cup of tea (and the speaker may be glued to the front in this one as well). What I would suggest to have done is:
- Replace all 2SC930s (Q3..Q6). Typically suggested substitutes seem to be 2SC2786, 2SC2999 and similar.
- Replace all electrolytic capacitors by quality Japanese ones (Panasonic, Elna, Nippon Chemi-Con) – same capacitance, voltage rating chosen such that dimensions match the old part (they have gotten a lot smaller). For very small values like 0.22 µF, use film caps instead.
- Clean and seal all switch contacts, also volume pot if needed (maybe Caig Faderlube or the like?).
- Realign all oscillators at upper end of the band (CT1-4 for MW, CT2 for FM, CT3 for SW2). That's going to be quite unavoidable if Q4 and Q5 get changed, unfortunately. Touch up tracking if needed (aging etc.).
Whether a basic set like this is worth all the effort is quite another matter, of course. If you want something to actually use on shortwave, I would rather invest some efforts into, say, a well-kept ICF-SW7600 with the usual issues instead.
A few 7600 series receivers compared
|Min. tuning steps (SW)||5 kHz||5 kHz||1 kHz|
|Fine tuning AM||yes||yes||no|
|Fine tuning SSB||yes||yes||yes|
|Sideband selection in SSB mode||no||yes||yes|
|Stand on back||no||yes||yes|
|German version: full coverage?||no||yes||international version, so the problem is solved|
|Steps on LW||3 kHz||3 kHz||9kHz; 1 kHz|
|Steps on MW||9 kHz / 10 kHz||9 kHz / 10 kHz||9 kHz / 10 kHz; 1 kHz|
|Steps on SW||5 kHz||5 kHz||5 kHz; 1 kHz|
|Steps on FM||100 kHz||50 kHz||50 kHz|
Gain distribution is an important topic in receivers, as it affects both
obtainable sensitivity and dynamic range. If you have a noisy mixer without any
preamplification and feed this with the modest signal levels of a telescopic
antenna, the receiver won't be very sensitive. However, if you stuff a lot of
gain into the RF preamp, the poor mixer may overload far too quickly. (Some
noise figure calculations are helpful in striking the right balance. You need
to know the antenna's noise figure, of course – a big antenna will always catch
a lot more atmospheric noise than something as mismatched as a telescopic
The bandwidth of the various stages also plays a role, as it not only affects sum signal levels but also decides on which signals can still generate unwanted intermodulation products in following stages. The signal handling capability of the mixers also plays in – normally the requirements are less strict down the line in spite of some gain, as bandwidth is strongly reduced, but this is not always the case, as the analog models (ICF-7600A/7601) show.
(Weaker 2nd mixers aren't possible either if high close-in blocking / dynamic range is required, something typical for ham radio field days where another OM only a short distance away might be operating close to the desired reception frequency. For these applications, an equally good or even better 2nd mixer may be warranted, and/or a narrow – read <1 kHz – roofing filter on the 1st IF for narrowband modes, e.g. CW.)
Proper gain distribution in battery operated receivers is not at all easy, as not only signal levels (on the whip in particular) may differ quite a bit between battery and mains operation, but also dropping battery voltage may lead to reduced gain. Few receivers offer the flexibility needed to cope with such strongly varying input levels.
Where AGC (either as variable attenuation or direct gain control of an amplification stage) is applied affects both quieting characteristics and strong-signal handling. In the kind of receivers we're looking at here, there is only one AGC loop whose control signal is derived from demodulator output or pre-detector 2nd IF level. Now consider two scenarios:
- Rx tuned to empty channel on an otherwise busy band
- Rx tuned to strong station on a busy band
While it is clear that frontend gain should not drop too early (which would reduce SNR and thus lead to less steep quieting), no AGC in front of the first mixer in particular may mean that reception of stronger stations is still disturbed by intermodulation products, forcing the user to activate the attenuator. In the first scenario, only gain distribution and mixer signal handling capability will determine rx behavior.
The following table illustrating gain and AGC aspects of the various 7600s was split for space reasons. The first part covers the various "digital" models.
|Type||Dbl. bal. FET||Dbl. bal. FET||Dbl. bal. FET||Dbl. bal. FET|
|Gain stage||Yes |
Post xtal fltr
Post xtal fltr
Post xtal fltr
(post 2nd IF fltr)
All of these use double balanced mixers, and for good reason – when using a wideband frontend, their 2nd order intermod rejection and generally better strong signal handling when compared to unbalanced mixers are really needed.
The second part covers analogs and unique models. These all use band filters and therefore have to make do with unbalanced first mixers.
|AGC||Pre-Amp||Pre-Amp + |
|Type||Unbal. FET||Unbal. FET||Unbal. FET|
The analogs have a somewhat critical combination of high RF plus 1st mixer gain with a wideband 1st IF and less well-performing second mixer. (As explained in the '7601 section, the wideband 1st IF was deliberately chosen to improve frequency readout accuracy and stability.) Only the '7600DA uses a narrow crystal filter but still has high RF gain.
In January of 2004, I finally obtained a (used) AN-LP1 antenna, a project that had once been terminated after the decision to get an AR7030. Now the AOR isn't exactly a portable, so... My AN-LP1 (JFTR: S/N 15243) had been modified quite strongly by the previous owner who, among other things, removed the cloth (for better handling, I suppose), wrapping something around the wire and frame - well, a "naked" AN-LP1 doesn't look exactly pretty, but reception is unaltered. So, what about a little review?
Handling: I got the folding twist right virtually on the first try (after use, the antenna has to be folded together again, which had been reported as being a bit tricky until you get the hang of it), though it's quite possible that it would be more difficult with a stock antenna frame. Operation is quite simple: Unfold antenna module and attach it somewhere (that's a bit of an unsolved problem on this antenna, you can clip it somewhere and attach it to windows and such, but that's about it), connect antenna controller to both antenna and receiver (if you were wondering what the somewhat strange-looking longish part is, that's a filter into which you plug in the cable from the antenna controller on one end and connect the thing to the antenna frame with the other end), put in two AA batteries into the antenna controller (I used rechargeables, since I didn't have anything else at hand, and these work pretty well, with the switchoff voltage being roughly 2.3 V) if not already present, and off you go. At least if you have a receiver with a "hot" EXT ANT socket, like the ICF-SW7600G(R), ICF-SW100 or ICF-SW1000T. On others, it's necessary to explicitly turn the antenna controller on with the respective slider switch. While receiving, it's necessary to select the correct frequency range for best reception.
Reception: That's what we bought this thing for, right? And so far, the AN-LP1 has not disappointed me. One must know, of course, that it's only good for the frequency range from about 80 meters to 13 meters, with a preference for the lower frequencies. There the antenna really shines, with signals that are a lot stronger – say, 20 dB – than just with the whip (even when using an antenna tuner, [ADDX-PRE-1]), and usually cleaner than the signals of a tuned whip + AN-71 combo. Also, the preselection helps a lot to minimise IM (like on 80 and 22 meters, where at the time of testing the wire overloaded the receiver on the respective frequencies), and with the apparently rather sharp 5 MHz range, images from 49m on 60m (a nuisance of the 7600G) are considerably attenuated. (19m images on 20m are strong as ever, but you can't have everything.) Minimising interference by rotating the loop is also nice. Performance, while still good on 16m, is no longer stellar on 13m, where the difference to the whip + tuner is no longer that huge.
All in all: Recommended. If you need better reception "on the go", the AN-LP1 makes a good travel companion. (And not only that – it'll also deliver good performance at home and with the receiver used on a wall wart. I've retired the AN-71 for now.) Unfortunately, its availability has never been too great (apparently it's only being made once in a while, in small batches), and the price may also seem a bit deterring, given it costs about half as much as an ICF-SW7600G(R) when new.
Oh boy, I certainly have come a long way in terms of headphones since I wrote the paragraph below. Today I'd prefer 600 ohm retro cans with slightly forward upper mids, like my trusty Sennheiser HD420SLs. Those may not be a good choice for radios that are overly rolled off in the highs (e.g. ICF-SW30, DT231s are a better match here), but they do well on my array of '7600' models, and there are no problems with amp hiss, frequency response warping due to output impedance or bass rolloff due to smallish coupling capacitors to be feared.
Since the sound via the speaker isn't too great (*grmbl*), I mostly use my
7600G with phones. At home, that's the Sennheiser HD 590 (very good sound
IMHO, though the comparatively high impedance isn't ideal for this radio since
you easily get amp hiss), while "on the go" I used to have a Koss KSC50 and am
currently using a Beyerdynamic DT231 (with foam modded earcups), albeit
not with the trusty 7600G but a (more compact and economic) Panasonic RF-B11.
Generally (when listening to FM and other stuff via my home hi-fi setup) I find the HD 590 to give the best sound of those three (not surprisingly), followed by the DT231 and then the KSC50. The Koss part should have been better (it uses the same drivers as the PortaPro after all), but it seems I'm not the only one who couldn't persuade them to fit tightly against the ear without becoming uncomfortable. (The older KSC35 and current KSC75 apparently fit noticeably better. These are generally considered excellent budget choices.) My main complaints were that percussion would sound "paperish" at times (the fit issue certainly didn't help here), and then storing them quickly and safely at the same time wasn't an easy task. The DT231 may not be all too comfy if you've got big ears, but provided it fits it's quite acceptable in the comfort department (and you don't get sweaty ears due to the velour earpads) and provides some (though not outstanding) isolation from outside noises. It's not foldable or anything, but reasonably small nonetheless. (For something foldable, the Sennheiser PX100 or a PortaPro may be an option to be considered. Both of these are totally open, however.) Lately I happened to plug the DT231 into a relatively newly acquired ICF-SW30 and was pleasantly surprised about the sound on the short waves, which turned out to be very smooth. The reason for this probably is the dip in the frequency response (about 6..7 dB down) which extends from a bit above 2 kHz to about 4.5 kHz. Not too bad a characteristic in this case, as it gets rid of a good part of the distortion typically associated with AM diode demodulation without muffling things all too much.
For more headphone suggestions, I'd suggest turning to Head-Fi. Frequency response measurements for various headphones can be found on Headroom