IMO it's standard behaviour to blink the HDD light for any activity on the primary IDE channel. If you want per disk lights, one can use drives that have an LED header, they exist.... However, it must be taken off the interface simply somehow, as they exist on removable drive caddies without any apparent control circuitry.

Pin/line 39 on the IDE cable is an activity indicator line but since it's connected to both drives (master and slave) in parallel, they typically implement it using an open-emitter and pull-up resistor. This makes it active low, so if you want to use it to control a LED you need to connect it to the negative (cathode) side with positive voltage (normally +5V via a resistor) on the anode.

you can probably isolate this pin from cdrom side of ide cable if you want only HDD led

All good points, but my issue is that the LED seems to be triggered by the secondary channel and I don't want it to be. i.e., the HDD and CD are not sharing a cable.

jakethompson1 wrote on 2023-04-13, 17:21:

All good points, but my issue is that the LED seems to be triggered by the secondary channel and I don't want it to be. i.e., the HDD and CD are not sharing a cable.

Perhaps pin 39 is bridged between the two channels? Multimeter to the rescue?

Depending on how the circuit was implemented, i've seen on some motherboards the ide led signal is routed trough diodes from both ide channels to the led driver. If one of the diodes is missing/defective or the circuit is open (broken trace), you will get activity only from one ide channel. So look at the circuit, analyze how the activity signals come together. If you are lucky you will have to remove a diode/cut a trace if you want activity only from channel one.

The led correctly lights up as there might actually be a HDD on the secondary IDE channel, so this is the expected behavior. If you really don't want it I'd research how to modify a cable to avoid it, if it's possible at all this way.

jakethompson1 wrote on 2023-04-13, 17:21:

All good points, but my issue is that the LED seems to be triggered by the secondary channel and I don't want it to be. i.e., the HDD and CD are not sharing a cable.

Most boards provide only a single HDD LED which is triggered by any device on either channel pulling /DASP low, so this is still expect behaviour on a dual-channel IDE controller. The only way to prevent the drive active signal from the CD drive to activate the "drive active" LED of the main board is to not connect the /DASP output of the CD drive to the /DASP input of the main board (i.e. breaking pin 39 of the IDE cable).

mkarcher wrote on 2023-04-13, 20:02:

jakethompson1 wrote on 2023-04-13, 17:21:

All good points, but my issue is that the LED seems to be triggered by the secondary channel and I don't want it to be. i.e., the HDD and CD are not sharing a cable.

Most boards provide only a single HDD LED which is triggered by any device on either channel pulling /DASP low, so this is still expect behaviour on a dual-channel IDE controller. The only way to prevent the drive active signal from the CD drive to activate the "drive active" LED of the main board is to not connect the /DASP output of the CD drive to the /DASP input of the main board (i.e. breaking pin 39 of the IDE cable).

I see, but I distinctly remember the secondary channel, or at least a CD-ROM on the secondary channel, not activating the LED on multiple systems I used, including the HOT-591P.

You have to understand that the circuitry for the ide led can be different and thus when both ide channels are populated can lead to different 'blinking patterns', depending on the motherboard in that era. If in the design of the mbo let's say had a dual (or more) input gate of an logic ic unused, it could be used for the ide led, instead of adding more parts.

bogdanpaulb wrote on 2023-04-14, 09:13:

If in the design of the mbo let's say had a dual (or more) input gate of an logic ic unused, it could be used for the ide led, instead of adding more parts.

As the DASP line that is used to turn on the LED is specified as "pull-up on host, open-collector output on drives", you don't need any gates or diodes, you can just connect the primary and secondary DASP pin to obtain a combined DASP pin. This gate-free arrangement is called "active-low wired or" (see https://en.wikipedia.org/wiki/Wired_logic_connection : In "Reversing active level", it is stated that "active-low wired or" is the same as "active-high wired and"; the circuit described in "Active-high wired AND" connection is exactly what already happens between the two drives on a single channel and can be extended to all four drives on a dual-channel controller by just connecting the DASP lines.

As drives are required to be able to drive 12mA when they pull the DASP line low, an LED can be connected directly to the DASP line without any further driving electronics, just a series resistor is enough. That configuration, while easy and cheap, is not recommended, though.

mkarcher wrote on 2023-04-14, 14:30:

bogdanpaulb wrote on 2023-04-14, 09:13:

If in the design of the mbo let's say had a dual (or more) input gate of an logic ic unused, it could be used for the ide led, instead of adding more parts.

As the DASP line that is used to turn on the LED is specified as "pull-up on host, open-collector output on drives", you don't need any gates or diodes, you can just connect the primary and secondary DASP pin to obtain a combined DASP pin. This gate-free arrangement is called "active-low wired or" (see https://en.wikipedia.org/wiki/Wired_logic_connection : In "Reversing active level", it is stated that "active-low wired or" is the same as "active-high wired and"; the circuit described in "Active-high wired AND" connection is exactly what already happens between the two drives on a single channel and can be extended to all four drives on a dual-channel controller by just connecting the DASP lines.

As drives are required to be able to drive 12mA when they pull the DASP line low, an LED can be connected directly to the DASP line without any further driving electronics, just a series resistor is enough. That configuration, while easy and cheap, is not recommended, though.

I'm not talking about how it should be done OR not be done, i'm just showing the way it's done on some motherboards. I can post pictures with them and measurements if necessary.

As I am using a SATA-to-IDE converter on this drive that I don't care about, I just bent pin 39 of the IDE connector out of the way, and that gets rid of the LED on CD-ROM access.
I wonder if there was variation and if some IDE CD-ROM drives never activated ACTIVE# and that is why I feel that many period-correct PCs I had did not activate the HDD LED on CD-ROM access.

I'll have to look at some, I think I have seen a pin missing on their IDE headers before.