Ok, maxtherabbit, but for "ordinary" ones or not-rechargeable AAs or AAAs (these are "primary", right?) batteries I still have to use your simple modification suggestion - removed D3 diode or resistor (in series with D3). And I still prefer to try out first the non-bulky variant with 3 V coin cell 😉 Tomorrow...

Yes correct

Job's done. Everything seems Ok, CR2032 socket fits tight and tidy:

Time to fire it up... Will it work?

What voltage do you measure at the input pin to the RTC when the system is off?

The system is running fine with this battery modification:

maxtherabbit wrote on 2025-05-08, 01:37:

What voltage do you measure at the input pin to the RTC when the system is off?

Currently intentionally I'm using a button cell with the lowest voltage that I could find lying around - with 2.84 V measured. When system is powered on, there's no more +4.3 V on the battery terminals, but exactly +2.84 V (the actual CR2032 voltage), so the reverse current leakage through D2 diode really is negligibly low and I couldn't even measure it. When the system is off at pin 20 of RTC+CMOS chip I measure 2.48 V, but the chip keeps BIOS settings and its clock is running. So, the voltage drop from the battery caused by Si diode D2 is about 0.36 V, which is not perfect, but for now everything is Ok.

I may change this Si diode D2 (1N4148) for some schottky or even better - a Ge diode, if I found some suitable replacement.

Cool. I'd be curious if it keeps accurate time with that low of a standby voltage

maxtherabbit wrote on 2025-05-08, 13:01:

I'd be curious if it keeps accurate time with that low of a standby voltage

Yeah, I'll leave it disconnected from PSU for a couple of days to check this.

Update: I booted the motherboard today to check the date and time in the BIOS and got an unpleasing surprise. The date is still 08.05.2025 and of course the time is also incorrect, but the settings in CMOS are preserved and valid. It seems like the RTS needs some more "juice" to run fine.

So I really have to replace the Si diode D2 (1N4148) with some schottky or Ge diode. Any suggestions for a suitable diode with low forward voltage drop?

I can't choose the proper replacement diode that will do the job with the lowest possible forward voltage drop and I need help from someone who is more into electronic components.

I can get one of these schottky diodes, but I can't find all the proper data to compare their forward voltage drop at equal forward current:

DO15 case: SB160, SB260, SR2100 and 1N5819

DO27 case: SR540, SR550 and 1N5822

DO35 case: BAT42, BAT46 and BAT85

Any suggestion which one of these schottky diodes will be the most suitable (with lowest forward voltage drop) as replacement for the regular 1N4148?

Hey there, I suspect the current will be fairly low, so you can basically look at the forward voltage charts of each diode.

I attached a sample of such a chart.

Here are the values I could read off of the charts of the diodes you mentioned:

• SB160: 0.3V
• SB260: 0.3V
• SR2100: 0.4V
• 1N5819: 0.18V
• SR540: 0.25V
• SR550: 0.35V
• 1N5822: 0.22V
• BAT42: 0.15V
• BAT46: 0.09V
• BAT85: 0.09V

I'd say any of the BAT or 1N58** should do fine.

tehsiggi wrote on 2025-07-04, 08:46:

Hey there, I suspect the current will be fairly low, so you can basically look at the forward voltage charts of each diode. […]
Show full quote

Hey there, I suspect the current will be fairly low, so you can basically look at the forward voltage charts of each diode.

I attached a sample of such a chart.

Here are the values I could read off of the charts of the diodes you mentioned:

• SB160: 0.3V
• SB260: 0.3V
• SR2100: 0.4V
• 1N5819: 0.18V
• SR540: 0.25V
• SR550: 0.35V
• 1N5822: 0.22V
• BAT42: 0.15V
• BAT46: 0.09V
• BAT85: 0.09V

I'd say any of the BAT or 1N58** should do fine.

Thank you very much for the clarification, tehsiggi!

I had a suspicion, that those high-current schottky diodes (like SR540 or SR550) may not be a suitable choice for this particular 3V battery circuit, but it's nice to have a confirmation from someone who's into electronics repairs. So, in this case I'll go with BAT85 as it also has forward current of 200 mA just like the original 1N4148.

Just soldered BAT85 diode in place of 1N4148 (D2). Voltage drop through D2 is now reduced from 0.36 V to 0.131 V which is a big improvement - 2.75 times less.

Neat. Of course there's some variance in between datasheet ideal values and real world. But 0.13V forward voltage sounds pretty good to me.

tehsiggi wrote on 2025-07-12, 19:46:

Neat. Of course there's some variance in between datasheet ideal values and real world. But 0.13V forward voltage sounds pretty good to me.

Actually you are the only person, who gave me some good-enough info on these schottky diodes, to which I have the most quick access, so once again - thank you very much!

I didn't knew the current draw from the battery of the RTC+CMOS chip. And there's no exact info for Vf (forward voltage) in the BAT85's datasheet at If (forward current) values lower than 0.1 mA - just some "maximum value" of 240 mV, which is way better compared to regular diodes with same Vr (reverse voltage) and If (forward current) characteristics:

And from this graphic with logarithmic coordinate system we can only "judge" by eye (and a bit of imagination):

I only can suggest, that you looked the curve for working temperature = 85° C as the minimum depicted value for Vf seems like ~90 mV at If = 0.1 mA.

I measured the Vf value of 0.131 V for the D2 replaced with BAT85 at temperature of more than 20° C (probably between 25 and 35° C), so if I'm not mistaken the current drawn from the battery by the RTC+CMOS chip when the board is powered off is somewhere around 0.1 mA.

Finally, it seems like this modification is working as it should.