PA Chime Generator

Upon replacement of the audio system in 2020, some individuals remarked that they missed the four-tone chime of the prior system.

A (for-fun) project has been kicked off to see if we can bring this element of nostalgia to the new system!

Background

The prior PA unit was an Inkel PA-9100RC Public Address Cassette Receiver, produced by the Korean Inter-M corporation. Unfortunately, as the unit was in disrepair, it was discarded after an extended effort to sell it on eBay. In hindsight, the chime circuit could have probably been recovered.

Research

The schematic in the PA-9100RC/9200RC Service Manual shows that the chime tone is generated by IC101, a BXA4240. No datasheet could be found for this part. Fortunately, a datasheet for a pin-compatible equivalent (the IK Semiconductor BT8040N) has been found here.

The BT8040 is a "melody generator with accompaniment". Some key observations of this schematic:

• The resistor-capacitor (RC) circuits attached to ENV1 and ENV2 appear to provide the audible decay (envelope) of the the tones.
  • Attached to each pin is a 4.7 uF (50 V) electrolytic capacitor and a 100 kohm resistor.
  • The same values are used in the interM PAM-240/360 and PAM-510/520 service manual schematics (these schematics use the BXA-4230K and BXA-4240 generators, respectively).
• The SEL1 pin (#6) is pulled HIGH (to VDD) while SEL2 (#5) is left floating.
  • This appears to select the "do-mi-sol-do" melody, according to the BT8040N datasheet, which is consistent with a spectral/audible analysis of the PA-9100RC.
  • The interM PAM-240/360 schematic, which utilizes the BXA4230K, has SEL2 tied to GND. However, the interM PAM-510/520, which utilizes the BXA4240 (same as the PA-9100RC), shows SEL2 floating. It's not entirely clear whether this is intentional or an oversight; further bench investigation may be needed.
• It's not clear whether the musical key is affected by the 1 Megaohm resistor between OSC1 and OSC2, or if this just sets the speed (tempo). According to the BT8040N datasheet, this resistor can span from 500 kohm to 1.5 Megaohm. 1 Megaohm will be used to match the PA-9100RC.
• The MT pin (#4) is used to trigger the tone sequence.
• The driver circuit (P1, OUT1 and OUT2) is not used. In the PA-9100RC and interM PAM-240/360 schematics, pin #12 (OUT2) is tied to GND.
• Audio is provided via the MO pin (#12). In the PA-9100RC circuit, there is AC coupling (C1), an application of DC bias (R1/R2?). The purpose of D1 is not yet understood. This circuit will require further analysis.

Project Status

February 2021

A small quantity of the BXA4240 has been sourced from a surplus dealer in China. When (or if) these parts ever arrive is "TBD". Once they do, experiments will be carried out to see if the tones can be replicated on the bench.

A schematic is being developed. The chime circuit will nearly duplicate that of the PA-9100RC. A mixer circuit, along with a mic pre-amp, will also be included. Details to follow.

March 2021

A preliminary schematic has been developed.

Key features:

• Implementation of the BXA4240 (consider "BT8040x" to be used interchangeably) in a manner similar to that shown in the interM schematics.
• The bandpass filter on the output of the BXA4240 is still TBD. The gain and corner frequencies may need to be adjusted based upon bench measurements. The datasheet does not provide clear guidance on the output signal of this device.
• A P-channel JFET based soft mute circuit has been added. The intent is to eliminate the use of the switch on the microphone (which creates an audible click/pop over the PA). This will also prevent passing audio from the microphone until the tone progression (chime) has completed.
• A single-ended output (RCA connector) will be used to connect to the JBL mixer-amplifier.

July 2021

This project has been on hold due to other priorities. At this point, it's possible that its introduction may be delayed to the 2022 season.

August 2021

A switch failure on the PA microphone prompted re-visiting this project. The switch was fixed - but the chime generator was finished (in "beta" form, anyhow) as well.

Some updates:

• The BXA4240 sourced from China in February was a bust. The transaction turned out a scam (received a random phone case instead of the IC). While AliExpress finally issued a full refund, it certainly doesn't reflect well upon the integrity of its sellers.
• In April, a BT8040N was extracted from an eBay-purchased InterM PA-1000BR amplifier. While functional - the sound wasn't quite right (more on this below).
• A Texas Instruments C2000 evaluation board was selected as the "brains" per familiarity in the designer's day job. Over the course of months, code was written to control the chime, microphone mute, button LED and a small LCD display via button trigger. The code functionality is described in more detail in a separate section.
• In August, a BXA4230K came onto the market via eBay, extracted from an Inkel PA-9100RC (the exact model previously owned by Fairway). This chip sounds exactly right.

Design

The preliminary schematic shown in March is mostly accurate. Modification include (but are not limited to):

• The addition of a C2000 microcontroller board (EVM), LCD, button and indicator LED (embedded in the button).
• The addition of a bandpass filter and DC biasing, as is present in the Inkel PA-9100RC (as originally mentioned in February).
• Power supplies to produce +5 V and +3.3 V for the analog and digital supplies, respectively.

When and if time permits, a final schematic will be produced to show the as-built hardware in its entirety.

BT8040N versus BXA4230K

While these parts are pin-for-pin compatible, and seemingly listed as drop-in replacements, there is no doubt that the BXA4230K has a distinctly different sound than the BT8040N. The tone progression interval, pacing and decay (with external circuits) are very similar. The BX4230K, however, has lower fundamental tones, and additional harmonics and/or sustained tones. I don't have captured audio samples - but I can say that if you've heard the PA-9100RC chime, you won't be satisfied with the BT8040N. I wasn't.

Future

The LCD display was primarily added to provide a visual level cue to the announcer. The PA microphone is very sensitive to proximity and volume; too far away or too quiet and the PA announcement isn't sufficiently audible. This seemed like a larger issue in 2020; as of 2021, though, the staff is doing quite well making clear announcements. The development semi-extensive code to add a graphical Vu meter (with louder/software cues) to the LCD seems unnecessary at this time.

The hardware is mostly ready for feeding the microphone signal into a low-passed ADC input for microphone monitoring when and if this feature is ever to be developed.