A backup of the now-defunct ferider.smugmug.com page on collimating (focus-calibrating) the Konica Hexar AF, plus the source material it was built from: the legendary photo.net Hexar thread, the Silvergrain Wiki page, Frank van de Wiel's Hexar reference site, and a 2017 Facebook group discussion with real-world calibration values from several cameras.
Everything here is mirrored so the information survives. Original PDF captures of
each page live in sources/; the images from the ferider article are in
images/.
This page is an archive, not a rewrite. Part I reproduces ferider's article word for word, exactly as it appeared, with footnotes linking to supporting material. Part II transcribes the documents the article points at. Part III is the only editorial section, and it is marked as such.
⚠️ Do this at your own risk. You are working inside the camera's service ROM via an undocumented button sequence. Every author of this material says the same thing: it is possible to ruin your camera, or at minimum to wipe factory AF values you cannot recover. Write down your camera's existing values before you change anything. Nobody here accepts responsibility for a bricked Hexar.
Part I — The original article (ferider, verbatim)
Part II — The source material (transcribed)
- II.1 — Entering the ROM (patrick van den hurk, photo.net, 2003)
- II.2 — Picture A: the silent mode table (Charles Clemens, photo.net, 2003)
- II.3 — Picture B: the ROM functions (Michael Koch-Schulte, photo.net, 2006)
- II.4 — Picture D: the tables (ferider)
- II.5 — The Facebook group thread (2017)
- II.6 — Model and hyperfocal data (Frank van de Wiel, tedm.com)
- II.7 — Known hardware faults (Silvergrain Wiki)
- II.8 — The Chinese-language origin (色影无忌, 2003)
Part III — Compiled notes (editorial)
- Reference values from known cameras
- Procedure summary
- Before you calibrate: is it actually the AF?
- A note on what could not be archived
Repository contents · Credits · Notes
Reproduced verbatim from
ferider.smugmug.com/Technical/Other-Technicalities/Hexar-AF-Collimation, captured 14 September 2024. The page is no longer online. Nothing below has been reworded, shortened, or corrected — spelling and punctuation are ferider's own. The only additions are the footnote markers, which link to transcribed source material further down this page.A plain-text copy with no footnote markers at all is at
sources/Hexar-AF-Collimation-ferider-article.txt, and the PDF capture of the live page is atsources/Hexar-AF-Collimation-ferider-smugmug.pdf.
Extending the remarks in one of the longest photo net threads http://photo.net/equipment/point-and-shoot/konica-hexar),[^pnet] here is how I focus-adjusted a Hexar AF. The good news is: you need no tools - it's a pure software problem. The bad news is you'll need at least two test rolls. Here goes:
See in Picture A how Charles Clemens describes re-programming of the silver AF to introduce silent mode.1 Later (Picture B), Michael Koch-Schulte finds out that while doing the above adjustment, the camera's ROM also can be used to test the IR autofocus sensor at 0.6m, 0.8m, 3.8m and 7.6m, and possibly to adjust the camera's focus at 0.6m, 0.8m, 3m and infinity.2
After a first test roll which was quite out of whack, I did two test settings and film tests (called Scenario A and B below); then I used the results to finally converge to what I consider a good setting of the camera.
When you measure the camera's sensor and look at the focus settings described on PNET, you see that both sensor outputs and focus settings are almost linear with changing 1/(focus distance). The sensor output moves between 0 (infinity) and roughly 100 (0.6m), the focus can be set in roughly 300 steps (290 according to Stephen, Dante and others), where the lowest number is for infinity focus.3
Keep in mind that the AF's autofocus goes only upto 20m. Afterwards the camera considers everything to be at infinity. The hyperfocal distance of the 35/2 lens wide open happens to be 20m.4 How you set it will depend a bit on you prefering to use hyperfocus or wanting to have infinity "dead-on". I prefer the latter.
Here are some analysis results I collected on the way, and test shots to illustrate the procedure: a diagram (Picture C) showing focus sensor measurements, PNET focal settings, and my final camera settings as function of focal distance. Note the almost linearity. The same in numbers (Picture D), on top the sensor measurements, below the PNET focus settings, the two Scenarios I tried and my final result.5
Pictures E,F,G show test roll output collected along the way (forgive the post-processing, I did not invest a lot since I cared mostly about resolution checks). Close-up focus (Picture E), Infinity focus (Picture F); (Picture G) is my favorite test target, a GAS lamp outside (note the beautiful bokeh of the Nikkor 35/1.8 derivative in the background - better than the 35/2 pre-asph Summicron if you ask me ....)
Finally, Picture H shows the working camera.
— end of the original article —
The eight images below are the ones the article refers to as Pictures A through H, reproduced at full resolution. Filenames in this repository preserve the mapping.
Charles Clemens, photo.net, 21 December 2003. Transcribed in II.2.
Michael Koch-Schulte, photo.net, 5 September 2006. Transcribed in II.3. The red and blue annotations are ferider's.
Focus sensor measurements, PNET focal settings, and ferider's final camera settings as a function of focal distance, plotted against 1/distance.
Transcriptions of the documents ferider's article refers to, plus the later discussion that grew around it. Quoted passages are reproduced as written, including the original spelling. Editorial framing is kept to a minimum and set outside the quotes.
The button sequence that opens the service ROM was first published in English on photo.net on 14 December 2003, sourced from a Hong Kong photography site. It is the first half of the silent-mode upgrade; for collimation you stop before changing anything. In the poster's own words:
Okay, this how it goes: First make sure your camera is powered off. Then: 1: Turn the aperture dial to F22 2: Turn the camera on and press SELECT. The display will show +/-0.0 3: Press SELECT again. The display will show [0] 4: Turn the camera off, then on. Do not press any buttons. Display shows [0] 5: Enter ROM wit the exposure compensation buttons: Press right button 1 time Press left button 2 times Press right button 3 times Press left button 4 times Press right button 3 times Press left button 2 times Press right button 1 time The camera will rattle and buzz and the display will show the flashing number '232'
The remaining steps — turning to f/13 and setting A[00], turning to f/16 and setting
D[FA], powering off, then holding the shutter release for two seconds — are what
write the silent-mode feature flag. Robert Edwards later reported (photo.net,
6 January 2004) that using D[FE] in place of D[FA] additionally enables AE
bracketing and the other Hexar Classic features, at the reported cost of infrared film
setting on some cameras.
The same poster's closing caution is worth repeating:
remember that I can't be held responsible for camera's that stop working!
Charles Clemens tabulated the above a week later, on 21 December 2003, prefacing it:
I concur with the above. In fact I also went to the chinese site, and combined with other sites used the data to make this little table. Use at your own risk, however. I will accept no responsibility for problems. It might be possible to ruin your camera.
The table as it appears in Picture A:
Adding Silent Mode to the Hexar AF Silver
| # | Action | While pressing | LCD |
|---|---|---|---|
| 1 | Set aperature at F22 | Aperture Dial | |
| 2 | Turn Camera On | Select | (+/-0.0) |
| 3 | Press Select | Select | 0 |
| 4 | Turn Camera off then on | Do Not Press any Buttons | 0 |
| Enter ROM | |||
| 5 | Up/Down Buttons | 0 |
|
| Press Right Up/Down button 1 time | R1 |
||
| Press Left Up/Down button 2 times | L2 |
||
| Press Right Up/Down button 3 times | R3 |
||
| Press Left Up/Down button 4 times | L4 |
||
| Press Right Up/Down button 3 times | R3 |
||
| Press Left Up/Down button 2 times | L2 |
||
| Press Right Up/Down button 1 time | R1 |
||
232 |
|||
| 6 | Turn from f22 to f13 | Aperture dial | A(...) — opinions vary as to what may appear within brackets: 83 or 38; it probably doesn't matter |
Set Display on LCD to A(00) — NOTE: May take lots of button pushes |
Up/Down Buttons | A(00) |
|
| 7 | Turn dial to f16 | D(...) — opinions vary as to what may show in brackets: "22" or "FB"; again it probably doesn't matter |
|
| 8 | Set display to D[FA] — NOTE: May take lots of button pushes |
Up/Down Buttons | D(FA) |
| 9 | Power OFF | F16 |
|
| 10 | Press shutter release button for 2 secs | Shutter Button | F16 disappears |
To utilize Silent Mode: press & hold "MF" button while turning on camera. An "L" will appear next to the frame number in the LCD panel.
This is the post that turned the silent-mode trick into a calibration procedure. Posted
5 September 2006. ferider's annotations on Picture B
bracket the f/11–f/6.7 group in red as "RF Check" and the f/5.6–f/3.4 group in blue
as "Collimation", and circle the four values 300, 222, 055, 013.
Koch-Schulte's introduction:
I recently performed the upgrade to my "original" black Hexar AF. I played around with the aperture wheel a little bit and found out there is more diagnostic information and possibly different modes available on the hexar. Use at your own risk. The following is a description of what I observed after entering the ROM sequence.
His observations, aperture by aperture, quoted as written:
| Aperture | Observation |
|---|---|
| F22 | "the number '232' flashing (I have no idea, possibly a version number)." |
| F19 | "T (the letter 'T' is displayed -- again -- no idea)." |
| F16 | "d[16] is displayed (One of the two programming hex registers on the chip, changable using the UP/DN buttons.)." |
| F13 | "A[3F] (The second of the two programming hex registers on the chip.)." |
| F11 🔴 | "###/0.6 (Nothing happens except a flashing number followed by 0.6 on the LCD screen, probably a focus sensor. 0.6 is the closest the camera can focus coincidentally. Point the camera at infinity now and the screen reads 000, at very close range the read out say 160.) I'm guessing that this and the following seven menus have to do with distance sensor calibration used by Konica technicians." |
| F9.5 🔴 | "###/0.8 (Same as above except flashes 0.8 and next readable distance on the hexar's lens. Again read out is affected by pointing the camera at different distances.)." |
| F8 🔴 | "###/3.8 (Same as above except flashed 3.8 ie. about 12.5 feet.)." |
| F6.7 🔴 | "###/7.6 (Same as above except flashed 7.6 ie. about 25 feet.)" |
| F5.6 🔵 | "This is interesting. The lens jumps (and you get a little bit scared) to its closest focus .6 m and the LCD read 300." |
| F4.8 🔵 | "Same as previous, except the lens sets itself to .8 and the LCD reads 222." |
| F4 🔵 | "Same. Except lens focuses at 3 metres and LCD reads 055 (not so scared anymore)." |
| F3.4 🔵 | "Same. Except lens focuses at infinity and LCD reads 013." |
| F2.8 | "Something else happens this time. The lens doesn't move but the screen reads 28.4 and changes a little. Not sure what causes the change. Light? Power?." |
🔴 = RF Check (read-only sensor diagnostics) · 🔵 = Collimation (editable focus values), per ferider's annotation.
The two tables in Picture D, transcribed.
Top table — sensor measurements. Column headings: to meas · m · ft · in · Meas.:
| to meas | m | ft | in | Meas.: |
|---|---|---|---|---|
| f11 | 0.6 | 1 | 11.62 | 105 |
| f9.5 | 0.8 | 2 | 7.50 | 76 |
| 3 | 9 | 10.11 | 19 | |
| f8 | 3.8 | 12 | 5.61 | 16 |
| f6.7 | 7.6 | 24 | 11.21 | 8.5 |
| inf | 0 |
Bottom table — focus settings. Scenario A is headed "inf.def: 20"; Scenario B is headed "inf. def: 10 / min def: 307". Bold figures are the values actually entered; the unbolded figures are the interpolated intermediate points.
| to set | PNET | Sc. A | Sc. B | FINAL | ||
|---|---|---|---|---|---|---|
| f5.6 | 300 | 314.0 | 314 | 307.0 | 307 | 306 |
| f4.8 | 222 | 232.8 | 233 | 225.0 | 225 | 225 |
| f4 | 55 | 73.2 | 73 | 63.7 | 64 | 64 |
| 64.8 | 55.3 | |||||
| 43.8 | 34.0 | |||||
| f3.4 | 13 | 20.0 | 20 | 10.0 | 10 | 16 |
The two unlabelled rows are the interpolated values for the 3.8 m and 7.6 m points — present to demonstrate the linearity, not separately settable.
A thread in the Konica Hexar AF Facebook group, started 29 November 2017 by Adria
Weatherbee, is the most substantial later discussion of ferider's method, and the
only known source of comparison values from multiple cameras. Full capture:
sources/Konica-Hexar-AF-Facebook-group-thread-2017.pdf.
Ali Bosworth introduced the method to the thread, and is the person who obtained ferider's documentation directly from the author:
One thing to consider is that the hexar can have its AF system re-calibrated via the same menu that you use to enable "silent" mode, auto-bracketing etc. There is very little documentation about this but I came across some old message board posts and reached out to the author a while ago who pass on his own documentation that he had used to recalibrate his hexar
His summary of what the menu contains, and his warning:
Essentially there are 6 AF readings where it shows you the values that the system is reading (so you need to point the camera at known distances), then 6 values that you can set which correspond to those same distance readings with regard to how the system focuses the lens itself. I would be VERY careful with this though, because if it is an unrelated issue you could wipe out your factory set AF system values. If you are interested I would start by going through the menu carefully and recording your current calibration data, but not changing anything.
On how to read the values without altering them:
simply enter the weird mode via the first part of the silent-mode instructions (don't change anything), then note the values shown on the display while in "P" at f/3.4, f/4, f/4.8, and f/5.6.
And, asked how one knows the right values — the central unresolved question:
yes, that's the problem, there is no way to know if those values are correct or not… We wont really be able to tell anything simply from these values unless they are WAY off.
The camera in the thread turned out not to be a calibration case: its values were in range, and the fault lay in the shutter release switch. That diagnostic thread is summarised in Part III.
Barna Deli reported the only unambiguous success in the thread:
My Af was not good.. I calibrate it with the help of the link... Working
Frank van de Wiel's Konica Hexar reference site — compiled, in his words, from "the
manual, on the net and in magazines" — is the best single source on which Hexar model
has which functions, and provides an independent check on ferider's hyperfocal claim.
Full capture: sources/Konica-Hexar-tedm-com-Frank-van-de-Wiel.pdf.
The six models, per that site:
| Model | Notes |
|---|---|
| Standard (Black) | The original. Has silent mode. Upgradeable by Konica service for ~$30 to add multiple exposures, one-step manual exposure setting, guide number setting, and infrared film setting. |
| Date | Standard model with date back. |
| Classic (Chrome) | 1993 anniversary model, 2000 made. Everything an upgraded Standard has, plus automatic bracketing (3 exposures, ±2 EV in 1/3 EV steps). |
| Gold | Collector's anniversary edition. |
| Rhodium (Pink) | Functions identical to an upgraded Standard. |
| Silver | Functions identical to a Classic, plus a date back. No silent mode as shipped. |
This is why the ROM procedure exists in the form it does: the Silver shipped without silent mode, and the button sequence in II.1 was found as a way to enable it without sending the camera to Konica. The collimation menu was discovered incidentally, on a black camera (II.3), which suggests the ROM menu is present across models.
Hyperfocal distances for the Hexanon 35/2, calculated with H = F²/(f×c) and a circle of confusion c = 0.033 mm:
| Aperture | 2 | 2.8 | 4.0 | 5.6 | 8.0 | 11 | 16 | 22 |
|---|---|---|---|---|---|---|---|---|
| Hyperfocal distance (m) | 18.56 | 13.26 | 9.28 | 6.63 | 4.64 | 3.37 | 2.32 | 1.69 |
| Set manual focus to (m) | 20 | 20 | 10 | 7 | 5 | 3.5 | 2.4 | 1.7 |
| Sharp from (m) to infinity | 10 | 10 | 5 | 3.5 | 2.5 | 1.75 | 1.2 | 0.85 |
This corroborates ferider's remark that the hyperfocal distance wide open "happens to be 20m" — the computed figure is 18.56 m, and the site's own advice is to set 20 m. The coincidence with the AF's 20 m cut-off is what makes the infinity setting a genuine choice rather than a single correct answer.
Two other items from the same source bear on focus testing:
- Fixed focus distances available in MF mode: 0.6, 0.7, 0.8, 0.9, 1.0, 1.1, 1.2, 1.3, 1.4, 1.5, 1.6, 1.7, 1.8, 1.9, 2.0, 2.2, 2.4, 2.6, 2.8, 3.0, 3.5, 4.0, 5.0, 7.0, 10, 20, 999 (infinity) m. Useful for shooting a manual-focus control series against the AF.
- Filters and the
EEEerror. A mounted filter can obstruct the lens at 0.6 m focus, producing a persistentEEEon the LCD that survives switching off; only removing the battery clears it. Worth knowing before you run close-focus tests with a filter fitted.
The Silvergrain Wiki page on the Konica Hexar documents the lens design and, more
usefully here, the two common hardware faults with disassembly procedures:
"Shutter release (erratic pre-focus) problem" and "Sticky buttons (usually MF,
Select or both)". Full capture:
sources/Konica-Hexar-Silvergrain-Wiki.pdf.
On the AF system itself, that page notes the Hexar uses centre-only active AF with three front windows — a centre infrared emitter and left and right photodiode sensors — an arrangement chosen to cancel certain focusing errors. It also records that the Hexanon 35/2's spherical aberration is deliberately undercorrected, and that the camera "can calculate the focus offset for each aperture value and correct for the focusing error during exposure." That per-aperture offset sits downstream of the four values you edit, which is one reason a calibration that looks right at f/8 can still be wrong wide open — and why test rolls should be shot at f/2.
Charles Clemens' remark in II.2 —
"I also went to the chinese site, and combined with other sites used the data to make
this little table" — points at this thread. It is the earliest dated posting of the
silent-mode sequence found in this whole archive, predating patrick van den hurk's
English-language photo.net post by six weeks. Full capture (Wayback Machine, crawled
19 June 2008): sources/Konica-Hexar-Chinese-Forum-cameraunion-net-2003.pdf.
The forum is 色影无忌 ("Xitek"), a well-known Chinese photography community; this
particular capture is served from the domain cameraunion.net, which the page's own
Wayback header records — whether that domain was a mirror, an aggregator, or the same
site under a different name at the time is not established here, and it is not
necessarily "the Hong Kong photography site" patrick van den hurk separately
referred to. Treat it as an independent early source, not a confirmed identification of his.
Thread title: 银巧思改静音模式的方法-不必开机壳 — "Method to modify the Silver Qiaosi6 to silent mode, without opening the case." Posted by user aaa, 2 November 2003, 23:32. Translated in full:
Silver Qiaosi is better than Black Qiaosi in almost every way. These special button operations let you modify the firmware inside the Silver Qiaosi. I've tested it myself and everything is OK.
Note: some steps require the display to be set to a specific value — do not set it wrong under any circumstances. I don't know what will happen if the firmware ends up corrupted.
The table posted with it is the same sequence transcribed in II.1
and II.2 —
aperture to f/22, power on, SELECT, SELECT, power off/on, the seven-press ROM entry
(r1, l2, r3, l4, r3, l2, r1) to 232, f/13 to A[00], f/16 to D[FA], power off,
hold shutter two seconds to L[0] — column-for-column identical to Clemens' English
table three weeks later, right down to the ambiguous bracket values.
That ambiguity shows up here too, independently, in the replies:
| Poster | Date | Step 16 shows | Step 18 shows |
|---|---|---|---|
| aaa (original table) | 2003-11-02 | A(00)/232 (entry) |
D[22] |
| FATTER | 2003-11-06 | A[83] |
D[FB] |
| 一小口 | 2003-11-09 | A[3F] |
D[F8] |
| 菜丝头 (quoting FATTER) | 2003-11-07 | A[3F] |
D[F8] |
This is the same split later reported independently on photo.net — Koch-Schulte's
A[3F] in II.3
matches the second row here — which is good evidence the value genuinely varies by
camera (firmware revision, or a per-unit ID/feature register) rather than being a
transcription error on either forum. One poster in this thread guesses exactly that:
Maybe every unit's displayed values at these two steps are different? Perhaps the Qiaosi's internal firmware version differs?
— FATTER, 6 November 2003
Two other details from this thread are worth carrying forward. First, a warning about the aperture ring during the button sequence, from DONDM (3 November 2003):
When adjusting, be sure to hold the camera flat and level — otherwise it will go wrong.
Second, hexarfan's theory (5 November 2003) for why the Silver has these dormant
registers at all — that Konica dropped the Black model under patent pressure and shipped
the Silver instead, but its firmware retained the Classic model's features, gated behind
undocumented D[XX]/A[XX] register values, "sleeping" until someone found them. This
lines up with the model history in II.6:
the Silver had no silent mode as shipped, but clearly the hardware and firmware to run it
were already there.
The thread also links a follow-up (cameraunion.net/forum/showthread.php?threadid=194492,
uncaptured — see below) about extending the same
technique to auto-bracketing, consistent with Robert Edwards' D[FE] variant reported
independently on photo.net (II.1).
No part of this thread discusses the RF-check or collimation menu — that discovery, per every source in this archive, dates to Koch-Schulte's 2006 post (II.3). This thread is silent-mode only.
Everything in this section is editorial. It is compiled from the sources above by the maintainer of this repository, not taken from any single one of them. Where it makes a claim the sources do not, that is flagged. If you want only the primary material, stop at the end of Part II.
Every documented set of collimation values, gathered from the sources above. Order is f/5.6 (0.6 m) / f/4.8 (0.8 m) / f/4 (3 m) / f/3.4 (∞).
| Camera | 0.6 m | 0.8 m | 3 m | ∞ | Source |
|---|---|---|---|---|---|
| PNET published values | 300 | 222 | 055 | 013 | II.3 — black Hexar AF |
| Ali Bosworth — newer Hexar | 301 | 223 | 054 | 012 | II.5 |
| Ali Bosworth — older Hexar | 313 | 233 | 062 | 019 | II.5 |
| ferider — before calibration | 314 | 233 | 073 | 020 | II.4 (Sc. A) |
| ferider — after calibration | 306 | 225 | 064 | 016 | II.4 (FINAL) |
| Adria Weatherbee — late Silver, ~2002 | 308 | 229 | 058 | 015 | II.5 |
Every known camera sits within roughly ±8 steps of the others at each distance. If your values fall inside that spread, your ROM calibration is probably intact and the fault is elsewhere. There is no published factory-correct value per serial number: every camera was individually calibrated at the factory, so the only ground truth is a test roll from your own camera.
A condensed reading of Parts I and II, for use at the camera. The worksheet at
docs/worksheet.md is a printable version.
1. Record your factory values first. Enter the ROM (II.1),
stop at the flashing 232, then turn the aperture dial to f/5.6, f/4.8, f/4 and f/3.4
in turn and write down each number. Do not press the up/down buttons. This is the only
way back.
2. Check the sensor (optional, read-only). At f/11, f/9.5, f/8 and f/6.7 the display shows a live sensor reading; point the camera at measured targets at 0.6, 0.8, 3.8 and 7.6 m — measured from the film plane — and compare against ferider's 105 / 76 / 16 / 8.5 (II.4). Erratic or stuck readings mean a hardware problem that calibration will not fix.
3. Shoot a baseline roll. Wide open at f/2, close / mid / infinity targets, every frame logged. Depth of field at smaller apertures will hide the very error you are looking for.
4. Enter Scenario A, shoot the same set. Enter Scenario B, shoot it again. Then interpolate toward whichever direction improved — this is exactly how ferider arrived at 306 / 225 / 64 / 16 from 314 / 233 / 73 / 20 and 307 / 225 / 64 / 10.
5. Save and verify. Turn the camera off, then hold the shutter release for about two seconds — the commit step from II.2. Re-enter the ROM and read your values back before burning another roll.
Because the response is linear in 1/distance, only the two endpoints really need solving; the middle two follow:
value(d) = value(inf) + [ value(0.6) − value(inf) ] × ( 0.6 / d )
Higher number = closer focus. A frame that back-focuses needs a higher value at that distance. (This sign convention is inferred from ferider's data and from Koch-Schulte's observation that the lowest number corresponds to infinity; no source states it outright.)
The 2017 Facebook thread (II.5) is a cautionary tale: in-range calibration values, and the real fault somewhere else entirely. Rule these out first.
The two-stage shutter button fault. The classic Hexar failure — the microswitch below the shutter button deforms or oxidises, so the camera focuses on half press and unfocuses on full press. Ali Bosworth described it in the thread as affecting the later gold "crab" contacts too, not just the earlier type, and reported that his own took several attempts:
Finally with careful changing of the bends of the crab legs it was fixed for good. Also cleaned with a qtip and rubbing alcohol but i believe you can't see the dirt/oxidation that causes the poor connection.
Disassembly is documented in the Silvergrain Wiki capture (II.7).
Test for it with the self-timer. The SELF button fires the camera in place of the shutter release, bypassing the switch entirely — established in the thread by both participants checking their cameras. If self-timer frames are sharp and shutter-button frames are not, the switch is your problem. Note that in that thread even the self-timer frames came out soft, which is what pointed to a second, separate fault.
Dirty or hazed AF windows. Three on the front — centre emitter, left and right photodiodes (II.7). Contamination on any of them makes the AF wander, and no ROM value compensates for it.
Sticky MF or SELECT buttons. Also documented in the Silvergrain Wiki. A stuck MF button changes focus behaviour outright.
Rule of thumb: a camera consistently off by a fixed amount at every distance is a calibration candidate. A camera that is sometimes right and sometimes racks to infinity is a hardware fault, and calibrating it will only destroy the factory values.
Two leads did not pan out, recorded here so nobody repeats the same dead ends.
Roland R's rangefinderforum.com thread. A comment on the photo.net thread — Roland
R, 1 March 2009 — points to a thread with practical advice on using this data to adjust
focus: rangefinderforum.com/forums/showthread.php?threadid=70654. That vBulletin URL
scheme is long gone; the site now serves rangefinderforum.com/threads/<slug>.<id>/, and
thread IDs appear to be assigned sequentially across the whole forum regardless of
subforum — consistent with the site's own numbering, where nearby subforum threads run
from the high 68,000s (December 2008) to the high 70,000s (January 2009), bracketing
Roland's March 2009 reference.
Checked and ruled out:
- The Wayback Machine never crawled thread 70654 under any URL form — old
(
showthread.php?t=70654) or new (/threads/70654/,/threads/-.70654/). Its CDX index has real neighbours on both sides (t=70646, a thread titled "Kowa SW 28mm? Is that the viewfinder?", andt=70665) but nothing at 70654 itself, at any point between 2003 and today. - The live site's legacy-URL redirect doesn't work at all, for this thread or any
other — confirmed by testing it against
t=70646andt=68838, both threads that demonstrably still exist under the new URL scheme, and both 404. So a 404 for 70654 proves nothing either way. - No search engine result or forum cross-post mentions thread ID 70654 by number.
Bottom line: thread 70654 cannot be located or ruled out by automated means from
here. The one thing that can resolve it is a real browser session, which passes
Cloudflare's check where a script can't: try https://www.rangefinderforum.com/threads/70654/
directly, or use the forum's own search for Hexar/collimation posts from
January–February 2009. If the thread was deleted rather than merely un-archived, no
amount of searching will bring it back — but that hasn't been established either way.
archive.ph/iHkUB. Every archive.today mirror (.ph, .today, .is, .li)
returns a hard bot-detection challenge to every automated request tried here — direct
fetch, a plain-text reader proxy, different user agents — never the archived page
itself. This is a genuine CAPTCHA wall, not a fixable request problem. If this page
matters for the archive, the only way to preserve it is the same way the Facebook and
Chinese-forum threads were: open it in a real browser, save or print it, and pass the
PDF over directly.
├── README.md ← this document
├── NOTICE.md ← preservation and attribution notice
├── docs/
│ ├── ferider-article.md ← the original article text, standalone
│ └── worksheet.md ← blank worksheet for recording your values
├── images/ ← Pictures A–H from the ferider article
│ ├── A-adding-silent-mode-table.jpg
│ ├── B-rom-modes-rf-check-and-collimation.jpg
│ ├── C-focus-curves.jpg
│ ├── D-measurement-and-setting-tables.jpg
│ ├── E-test-closeup-focus.jpg
│ ├── F-test-infinity-focus.jpg
│ ├── G-test-gas-lamp.jpg
│ ├── H-the-camera.jpg
│ └── hexar-af-quick-reference.jpg
└── sources/ ← original page captures
├── Hexar-AF-Collimation-ferider-article.txt ← original article text
├── Hexar-AF-Collimation-ferider-smugmug.pdf ← page capture (now offline)
├── Konica-Hexar-photo-net-thread.pdf ← the 72-page photo.net thread
├── Konica-Hexar-Silvergrain-Wiki.pdf ← lens design + repair procedures
├── Konica-Hexar-tedm-com-Frank-van-de-Wiel.pdf ← models, tables, hyperfocal data
├── Konica-Hexar-AF-Facebook-group-thread-2017.pdf ← real-world values & diagnosis
└── Konica-Hexar-Chinese-Forum-cameraunion-net-2003.pdf ← earliest dated source, 2003
None of this is my work. I am mirroring it because the original page went offline and the information deserves to survive.
- ferider — the collimation article, the analysis, the test rolls, and the
interpolation method. Originally at
ferider.smugmug.com/Technical/Other-Technicalities/Hexar-AF-Collimation. - patrick van den hurk — published the ROM entry sequence (photo.net, 14 December 2003), sourced from a Hong Kong photography site.
- Charles Clemens — tabulated the silent-mode procedure (photo.net, 21 December 2003).
- Michael Koch-Schulte — mapped the ROM aperture positions to the RF-check and collimation functions (photo.net, 5 September 2006). Without this post there is no procedure.
- Robert Edwards — the
D[FE]variant enabling AE bracketing (photo.net, 6 January 2004). - Ali Bosworth, Adria Weatherbee, Kosma Moczek, Barna Deli — the 2017 Konica Hexar AF Facebook group discussion, and the comparison values from multiple cameras.
- Frank van de Wiel — the Konica Hexar reference site at tedm.com: model differences, function tables, hyperfocal data, error codes.
- Silvergrain Wiki contributors — lens design notes and the shutter-button and sticky-button repair procedures.
- Peter van de Haar — the silent photography and "S-mode" notes on van de Wiel's site.
- aaa (色影无忌 / cameraunion.net) — posted the earliest dated version of the silent-mode sequence found in this archive (2 November 2003), six weeks ahead of the English-language photo.net posts.
- DONDM, FATTER, 一小口, hexarfan and the rest of that thread's posters — testing, corroborating register values, and the theory of why the Silver's firmware carried dormant features at all.
- The photo.net Konica Hexar thread — one of the longest in that forum's history, archived here via the Wayback Machine.
Archived material is reproduced for preservation and non-commercial reference. See
NOTICE.md. If you are an original author or rights holder and want
material removed or credited differently, please open an issue.
Footnotes
-
The full button sequence is transcribed at II.1 and Clemens' table at II.2; an independently posted Chinese-language version of the same sequence, predating both, is at II.8. For collimation you only need the first half — enter the ROM, stop at the flashing
232, and leave theA[00]/D[FA]registers alone. Those write the feature flags, not the focus values. The Silver shipped without silent mode while the Black had it; see the model table at II.6. ↩ -
Koch-Schulte's post is transcribed in full at II.3. Note his hedging — "possibly", "I'm guessing that this and the following seven menus have to do with distance sensor calibration used by Konica technicians". ferider's article is what confirmed the guess by test roll. Note also that Koch-Schulte was working on a black Hexar AF, not the Silver of the original silent-mode instructions, which indicates the ROM menu is common across models. ↩
-
"Stephen, Dante and others" are other participants in the photo.net thread. The 290-vs-300 discrepancy is unresolved in the sources and does not matter in practice: what you are setting is a step index, and the useful range is bounded by the values in II.4 — roughly 010 at infinity to roughly 315 at 0.6 m across every documented camera. ↩
-
Independently corroborated. Frank van de Wiel's table (II.6) computes 18.56 m at f/2 for the 35 mm lens using a 0.033 mm circle of confusion, and advises setting manual focus to 20 m at f/2 — so ferider's "happens to be 20m" is sound. The consequence is real: because the AF treats everything past 20 m as infinity, setting the f/3.4 value low ("infinity dead-on", ferider's choice at 016) versus high (hyperfocal) is a genuine preference, not a right-or-wrong. Scenario A used 020 and Scenario B used 010; the final 016 sits between them. ↩
-
Both tables are transcribed at II.4. The "PNET" column is Koch-Schulte's observed values from II.3 — 300 / 222 / 055 / 013 — which are that camera's factory values, not a specification. Treat them as one more data point, not a target; see Reference values. ↩
-
巧思 ("Qiǎosī") is the Chinese nickname for the Hexar used throughout this thread — not a brand name, and not to be confused with any Konica model designation. It appears in the thread title, in every poster's replies, and in the follow-up thread title quoted below. Translations in II.8 render it as "Hexar" or "the Hexar" for readability rather than transliterating it each time. ↩