So I'm in the middle of fixing up a Color Classic that I recapped and seems to have a blown PSU switching FET, and it has the same heavily charred area on the PCB under the two 18V zeners you see on most. Has anyone tried replacing that with a linear voltage regulator, and if so what were the results? I mapped out the schematic a bit around that 18V circuit and it seems like it should be pretty easy to remove the zeners, and then put a TO220 7818 with a little heatsink on it in place of the 47 ohm resistor.
-
Hello MLAers! We've re-enabled auto-approval for accounts. If you are still waiting on account approval, please check this thread for more information.
Replacing CC DL21/DL22 with a linear regulator
- Thread starter DaveW
- Start date
Replacing a switching regulator with a linear regulator is usually not a good idea.
You need to know what current and what voltage drop to expect. And from that you get the expected power dissipation on the linear regulator which in turn tells you if that's feasible at all.
What's the input voltage and what current is expected on the 18v rail?
You need to know what current and what voltage drop to expect. And from that you get the expected power dissipation on the linear regulator which in turn tells you if that's feasible at all.
What's the input voltage and what current is expected on the 18v rail?
Note I'm not replacing the switching circuit coming off the ZP1 monitor transformer. That runs 24V to a bunch of things and I'm not messing with it. I'm just talking about replacing the RL62/DL21/DL22 circuit, which regulates the 24V down to 18V for whatever it is that circuit feeds.
For current, unknown what the actual draw of that circuit is. I measure a hair over 5V across RL62 with the computer on and off though, so around 110mA through the resistor and 570mW dissipation in that resistor in all conditions. With the system and monitor off I'd assume most of that is through the zeners, in which case they're dissipating close to 2W without the fan on which explains why they torch the PCB.
The other zener regulator circuit on that 24V rail is the RF11/DF2 combination, which gives an 8.2V rail for something. I measured 8.3V there on and off, so 16.1V across the 390 ohm resistor and 41mA draw. That's 660mW dissipation in the resistor there at all times, but only worst case 336mW dissipation in the zener when it's off which explains why DF2 doesn't get nearly as toasty. I probably wouldn't bother replacing this if for no other reason than it doesn't seem to cause an issue. Replacing it would probably drop off state power dissipation by another watt though.
For current, unknown what the actual draw of that circuit is. I measure a hair over 5V across RL62 with the computer on and off though, so around 110mA through the resistor and 570mW dissipation in that resistor in all conditions. With the system and monitor off I'd assume most of that is through the zeners, in which case they're dissipating close to 2W without the fan on which explains why they torch the PCB.
The other zener regulator circuit on that 24V rail is the RF11/DF2 combination, which gives an 8.2V rail for something. I measured 8.3V there on and off, so 16.1V across the 390 ohm resistor and 41mA draw. That's 660mW dissipation in the resistor there at all times, but only worst case 336mW dissipation in the zener when it's off which explains why DF2 doesn't get nearly as toasty. I probably wouldn't bother replacing this if for no other reason than it doesn't seem to cause an issue. Replacing it would probably drop off state power dissipation by another watt though.
So I went ahead and did this today. I ended up using LM317 for both the 8.2V and 18V rail since there's not an off the shelf regulator for the 8.2 and I wanted to get the 18V one closer to the stock 19.2 I measured. It turned out pretty well I think. I also swapped RL22 for a 5W 100ohm resistor and raised it up.
Before the mod:
Measured 10W and 14VA draw with the system off, and ~98°C on both DF2 and DL21/DL22. Note I've already replaced DL21/DL22 with 5W 18V zeners, so they're probably a little cooler than stock.
With the system on just after getting to desktop the zener regulators didn't get any hotter but RL22 got up to 126°C. I just got a few measurements and turned it off, I didn't want to run with no fan very long with those temps.
The mod:
After modding:
In the off state, power dropped to 8W and 10VA. Not a huge saving there. Temps are wildly better as expected. The regulators stayed 38°C
On at desktop the regulator stabilizes at 61°C without fan, and the new RL22 only gets to 72°C even after sitting at desktop for several minutes.
Overall I'm quite happy. I don't really leave it switched on in standby when I'm not using it, but the lower standby power draw and much better temps are still nice. The lower operating temperature of RL22 is also great, it's crazy how hot that got before.
Before the mod:
Measured 10W and 14VA draw with the system off, and ~98°C on both DF2 and DL21/DL22. Note I've already replaced DL21/DL22 with 5W 18V zeners, so they're probably a little cooler than stock.
With the system on just after getting to desktop the zener regulators didn't get any hotter but RL22 got up to 126°C. I just got a few measurements and turned it off, I didn't want to run with no fan very long with those temps.
The mod:
After modding:
In the off state, power dropped to 8W and 10VA. Not a huge saving there. Temps are wildly better as expected. The regulators stayed 38°C
On at desktop the regulator stabilizes at 61°C without fan, and the new RL22 only gets to 72°C even after sitting at desktop for several minutes.
Overall I'm quite happy. I don't really leave it switched on in standby when I'm not using it, but the lower standby power draw and much better temps are still nice. The lower operating temperature of RL22 is also great, it's crazy how hot that got before.
