The Lathem LTR8-128 is a digital electronic master clock based on an RCA CDP1802 8-bit CMOS microprocessor clocked at 2.097152MHz. The power/output board contains twenty relays, which allows the unit to operate eight signal circuits, eight control circuits, and two dissimilar types of clock systems simultaneously.
A five-second duration signal (Bell) event can be programmed to occur at the beginning of any minute of any day of the week, and stored in any of three 128-event schedules. Each event has an absolute numbered address within its schedule, allowing empty spaces and retaining events in order as entered. Any schedule or combination of multiple schedules can be selected, such that 384 signal events may be active simultaneously.
A 32-event calendar schedule allows the signal schedule selection and/or enabled zones to be automatically switched at 12:00.00am on any numbered date of any specific year or recurring each year (years are entered in two-digit format but are not connected to days of the week, allowing this feature to be perpetually relevant). The calendar schedule only has any effect as the date changes at exactly midnight; used events do not clear themselves from the schedule, and empty spaces may also be left (the spaces must be counted manually as they are not numbered, and do not loop at the end).
Each Control circuit can be programmed to either latch or unlatch at the beginning of any minute of any day of the week. Each circuit provides both normally-open and normally-closed terminals, allowing the relays to remain latched for the shortest possible times and to default to the proper safe setting during a power outage. The Control events are stored in a single 128-event schedule (entered more-or-less the same as Bell schedules), which can be deactivated to freeze the current Control relay settings. The Control circuits may be used for many things, including controlling lights and such. I use them to control low-voltage relays that switch hallway lights and an outlet in my kitchen and the PA system power, and I also use one to silence the gakkou-no-chime at night.
The LTR8-128 is designed to operate two dissimilar types of clock systems simultaneously. It seems that this model was specifically designed for retrofit into buildings that were being expanded, when the original building had some type of clock system that was not economically (or otherwise) feasible to expand.
Of the different types of clock systems the LTR8-128 (EPROM version I-1) is supposed to be able to control, I have only personally verified four: Standard GRC synchronous, IBM/Simplex synchronous (“77-series” and Lathem SS-series), Standard AR2A impulse, and the special program for “Lathem type SS recorders and type SC1000 sub-controller”. The DPDT system relay contacts are also completely broken out into screw terminals, which allows more than one circuit of the same type of system to operate on one set of relays, as detailed below.
Standard GRC synchronous
Of course it runs my favorite type of clock system. Interestingly enough, it performs the 12-hour correction in an odd way. It’s actually better this way, under certain circumstances, allowing worn-out movements to continue operating for longer without attention. Unlike the LTRx-384 series, both system relays are inactive during normal operation—in the small-case models, the drive relay is wired normally-open, which forces those units to unnecessarily waste energy.
SET GRC system control program:
h:59.30 to h:59.59 every hour, both relays activate;
5:15.00 to 5:29.59 twice per day, reset relay activates.
(Equivalent timings to system code 03 in small-case models; newer units may use proper 5:12 to 5:28.)
» See modified page 19 diagram above, demonstrating how to wire a single set of system relays to drive two different circuits/voltages simultaneously. (This is not natively possible on small-case Lathem models, without adding an external relay.)
» Note that relay contacts that are switching AC should be protected by MOVs (included on-board only in small-case models). These should be wired in parallel with the contacts. This not only reduces contact wear but also electrical noise (transient voltage surges that create popping noise in radio receivers and such).
» To connect Remote Secondary Master relays: change 24vac (in above diagram) to the proper supply for the relays, and switch the wire from terminal #4 to #5 (so that each relay mirrors the operation of the master relays).
» To run an original GR system (without 12-hour correction): connect power to terminal #1 (or #7), run to #3 (or #9), reset to #2 (or #8), and disregard the reset relay. (This is also not natively possible on small-case Lathem models, without adding an external relay.)
» When the manual Advance Clocks mode is initiated, both relays activate for a duration of fifty seconds beginning immediately. (This mode is compatible with an original GR system, as long as you wait “at least three minutes between operations”.)
Standard AR2A minute-impulse
This is listed as just “AR-2” in the manual, but it is definitely the -A version that resets to the 60th minute. Due to the terminal breakout of the system relays, it can control AR2 and AR3 simultaneously with one set of system relays; see the wiring diagram shown below (small-case models may also be wired this way). All SET AR movements can be easily adjusted to reset to any minute.
SET AR2A system control program:
h:mm.58 to h:(m+1).00 every minute other than 60th (59min/hr), minute relay activates;
h:59.50 to (h+1):00.00 every hour, reset relay activates.
(See manual page 33. Equivalent to system code 04 in small-case models. System code 17 is the same but resets to 59th minute for native AR3 and AR2, which is not available in LTR8-128.)
» The manual Advance Clocks mode activates the minute relay 62 times at a rate of 0.5Hz over a span of two minutes time, causing the secondaries to advance exactly one hour (same operation as small-case models). The accumulator function should also work, in theory (this function remains untested by me).
Lathem SS-series time recorders (special)This program was designed to operate the “Lathem type SS recorders and type SC1000 sub-controller”—which may not have ever existed (I really want to see one if they did) based on the fact that this page was removed from the official current version of the manual (I suspect this control program may have also been removed from newer units). This timestamp would have been driven by minute impulses accompanied by IBM/Simplex synchronous correction. An unmodified scan of the missing page 25 from an older paper manual is shown to the right. The modified wiring to drive AR systems is shown below.
Lathem SS (special) system control program:
h:mm.57 to h:mm.59 every minute other than 58th (59min/hr), minute relay activates;
h:57.54 to h:58.02 every hour other than fifth, reset relay activates;
5:57.54 to 5:58.08 twice per day, reset relay activates.
(No equivalent in small-case models; only available in LTR8-128 with EPROM version I-1.)
» As shown, this program can simultaneously drive an AR system (movements adjusted to reset to 58th minute) and correct an IBM/Simplex/etc synchronous system.
» The manual Advance Clocks mode activates the reset relay for eight seconds starting at the 54th second of the following minute (same as plain Simplex control program). The impulse system must be advanced by external controls, as there is no way to provide extra minute pulses in this program. The accumulator function will not work for the impulse system.
» Note that relay contacts that are switching DC should be protected by diodes, which should be wired in parallel with the contacts, in reverse of the way that would short the signal. This not only reduces contact wear (absorbs back-EMF that causes arcing) but also electrical noise (transient voltage surges that create popping noise in radio receivers and such).
» ((Untested idea: 24v clutch-magnets may be able to be added directly across AR3 Reset line. I have not yet obtained any low-voltage clutch-magnets for testing. An additional relay can certainly be connected to switch the extra circuits.))
» For those who did not notice: this program actually allows the LTR8-128 to control three dissimilar types of clock systems simultaneously (or four, if you count non-corrective minute-impulse—see note below).
And, of course, it runs the type of movement that its own manufacturer uses. The reset relay operates the same as the special program above, except that the minute relay never does anything.
A non-corrective minute-impulse system may also be driven by either of the impulse programs listed above, by simply wiring a set of minute and reset contacts in parallel (or one after the other), thereby providing all sixty minute pulses per hour (with one having an odd duration). Manual controls and a pilot dial would certainly be necessary.
A system of 24v parallel movements may also be driven off of the AR3 system as wired above, with two diodes provided to sum the Minute and Reset lines together. This has been tested to function with AR2 movements, and is useful to perform long-term tests after rebuilding AR movements. (This should work with all movements that could otherwise be driven by only the Minute line from an original AR3 master.)
A version of the manual (newer, odd typesetting, missing page 25) is available on the Lathem website. (I intend to upload this document if the company ever stops offering it—the LTR8-128 and similar small-case models did go out-of-production at the end of the last millennium, after all. If you find it to be absent over there, please tell me about it.)
Additionally, the pages including my modified wiring diagrams are shown and linked below, following an unmodified scan of the original page 25.
Or, the file linked here (pdf, 04 pages, 1.01mb) contains all the above manual pages, packaged in a conveniently printable format that may be used as a supplement to the original manual.
(Note that my additions to the manual, as with most of the content of this site, will likely be more useful/interesting to a system clock collector than to anyone charged with using this type of equipment in-the-wild. Use at your own risk, of course, in any event.)
The picture at the top of the page and below is of my LTR8-128 that I pieced together. The CPU board and display is from my first one, which was in rather poor cosmetic condition when I bought it. The power supply board and the bezel and some other trim parts are from a newer non-working unit that apparently used to belong to the city of Phoenix.
The inside of the front panel is somewhat boring with the cover in place. The lock takes a Chicago #2387 key (and they are all keyed the same, with the small-case models having the same locks also). I should get a few more copies of that key.
The computer—the CPU chip is the second larger one from the left. I have no idea if the peeling sticker killed EPROM #3 (but I put the extra sticker on there the same day I learned what an EPROM chip is and how you go about erasing one), and I probably won’t ever find out unless I decide to try to operate a more exotic system someday. Note the timebase crystal and variable capacitor at the bottom left of the CPU board.
The power supply board, shown above connected to control both 120v and 24vac GRC systems, 120v Simplex synchronous (with an extra contact that could control some other voltage of the same), three 120v and one 12vac signal circuit, and three 12vdc Control circuits (one of which is wired normally-closed). (This is a configuration that I used some years ago.) The backbox is serial #0003 (and it has paint marks that suggest it was flush-mounted). The backbox and front panel serial numbers are supposed to match (the newer pieces from Phoenix match). Note the little squares of rubber that I glued to the corners of the backbox to keep the front panel from rattling—sometimes it’s the small things that make all the difference.
The LTR8-128 is the only Lathem model (from the pre-2000 LTRx-128 and LTRx-384 series) that is capable of running 384 active signal events simultaneously. The LTRx-384 models only allow 64 events per schedule and two simultaneously running schedules for a total of 128 simultaneously active events, and the LTR2/4/6-128 series only has one schedule that can hold 128 events.
The LTR8-128 is the only Lathem master clock model housed in the larger format case. The backbox is based on a standard 18x12x6in junction box (which would cost around $200 if you try to buy one new) with what appears to be a baked-enamel finish. I have never seen an LTR8-128 in any other configuration (e.g. surface-mount backbox), with any model name modifications (other than EPROM version), or branded under any other company name.
(Does anyone know when these models were introduced, closer than “around 1980”? I take the dates in the manuals to be the last point of revision, rather. Actually proving the LTR8-128 to be the first electronic master made by Lathem would be an interesting fact.)
We’ve all seen the LTR8-128, now let’s take it apart!
The photographs show the same LTR8-128 that is shown above. There is seemingly no good way to photograph the acrylic front panel—that’s the red panel with the window reflected on it. The front panel frame is made of the same gold-glazed aluminum channel as used on the smaller models, the pieces are just longer. Note also that the unit only contains three circuit boards; the front and back of each is shown. I replaced the battery leads on the power/relay board (from the newer unit) and soldered that Entry Code switch assembly (from the newer unit) to the CPU board (display and CPU boards from older unit)—the boards are otherwise all original.
((The parts comprising the non-functional unit may be shown in the future. The only differences, really, is that the sans-serif typesetting is in Helvetica on the newer acrylic front panel, and the labels inside are more plastic than paper on the newer unit.))
The LTR8-128 uses a 2.097152MHz quartz crystal timebase without temperature compensation, which can be tweaked by a variable capacitor. The manual states that a 1Hz error in the timebase is equivalent to one second every 22 days.
I have tweaked the timebase the old-fashioned way (as it lost several seconds a week when I bought it)—in other words, to treat the variable capacitor like the nut on a pendulum. I remember that I had it down to “only losing about 18 seconds in six months” at some point. There used to be a chart in this section, but it did not show anything useful due to too many uncontrolled variables.
I will now state my anecdotal observations:
» I believe the drift to be more-or-less smoothly linear rather than displaying choppy sharp changes, through observation in reference to radio broadcasting (it turns on the PA system and also switches the chime).
» Ambient humidity is the key to accuracy, I now believe. It lost, maybe, five minutes during a certain winter, in an occupied building with a woodstove maintaining around 50-60% ambient humidity. However, it would have lost nearly a whole minute during the next summer, in the same un-air-conditioned location with an 80% ambient humidity (I had to shut it off in the middle due to a power outage outlasting the battery). I need to get a nice controlled dehumidifier to operate next to it, is the conclusion.
(All statements regarding accuracy and adjustment of the timebase should also apply to all pre-2000 small-case Lathem models, as they contain the same parts—you just have to remove the red acrylic front panel to get to them. See: LTR6-384 ready for adjustment.)
In any event, I am perfectly happy with 18 seconds (which would average only a maximum of nine seconds off overall—if the offset were predictable). I’m not going to try to adjust the timebase again until after I construct my permanent house-building (and obtain a dehumidifier).
This clock is not particularly efficient. It is rated to use 20 watts “maximum” in both the manual and on the serial tag. One day, I decided to put my little Kill-A-Watt energy measuring toy to work on this problem. According to said toy, my LTR8-128 uses about 23w under normal operating conditions with no relays activated (the line voltage was 119v during the test). Each relay uses about one watt when latched (the specs for similar plug-in relays is 85mA at 12v, which is 1.02VA). Therefore, the unit can use as much as 33w constantly (with all controls latched) or 40w for a five-second period (with all controls latched during a signal event).
However, most of this power is actually consumed by the display. The clock only draws about 11w when parked in the menu selection mode (press [#]  and stop—it uses 14w with all the menu options lighted). All the programmed functions (system correction, etc) work while parked in the middle of the menu.
An apparent design flaw in this clock is the fact that the power supply cooks itself under normal operation. It’s really no wonder the plastic on the transformer is always brown and the voltage regulator in my first one failed so soon. (Well, actually, it had an unfortunate accident a few years earlier. In a related note, always disconnect the battery when you move the case—otherwise, you’ll get a face full of smoke on the highway and have to make new battery leads. Trust me on this.) Parking the computer in the menu selection mode also has the added benefit of allowing the power supply to run much cooler.
Is this a good idea? I don’t really know. Is it worth cutting the operating cost in half and likely exponentially increasing the lifespan of the power supply components? I am almost certainly doing it right now as you read this. I just wish I could figure out how to trick the display into power fault mode (including the ten-second timeout) under normal operation without the computer knowing about it (in so many words). Has anyone tried this?
I can’t say it enough times: I love the LTR8-128. That said, if I had designed it, I would have changed the following things:
(1) Give the Controls multiple selectable schedules (and a calendar selector), the same as the Bells. The Control circuit feature would be almost unimaginably more useful if the Controls had multiple schedules.
(2) Use higher-current-rated transformers and voltage regulator in the power supply. (I actually know how to fix this one.)
(3) Add a switch or keystroke sequence to power-off the LED display.
(4) Add provisions so that the time/date can be regularly and automatically set by an external device, such as a specially-designed radio/GPS receiver (made by me), either through a direct link or a contact closure that synchs to a certain time. (The new LTRx-512 series does have this feature, with both methods of connection available.)
(5) Program a new set of EPROM chips to control the systems exactly properly. (Something I would change now, but they could have gotten it exactly right the first time.)
Maybe someday I’ll be knowledgeable enough to make these modifications.