Help with first Power Supply for LM1875

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Is the IC electrically isolated with mica or silicone washers from the heatsink?

Do you have a Boucherot cell at the output (R5 and C5 in the datasheet)?

What is your speaker impedance? Did you read pages 6 and 7 of the datasheet? What is the thermal resistance of your heatsinks + thermal grease + mica or silicone washers?
 
pacificblue said:
Is the IC electrically isolated with mica or silicone washers from the heatsink?

Do you have a Boucherot cell at the output (R5 and C5 in the datasheet)?

What is your speaker impedance? Did you read pages 6 and 7 of the datasheet? What is the thermal resistance of your heatsinks + thermal grease + mica or silicone washers?

Yes they are isolated.

Not sure to be honest but here is a link to the instructions for the kits I got. See the bottom for the schematic.
http://www.quasarelectronics.com/kit-files/electronic-kit/3050.pdf

I did read those pages of the data sheet, but being new to this a lot of it went straight over my head. My speaker impedance is 4 ohms (although I am going to purchase some 8 ohms speakers soon).
 
Your kit has a Boucherot cell aka Zobel network.

The instructions that came with your kit say in the introduction that it is supposed to work from 25 V into 8 Ohm. For that a 1,4 K/W heatsink per IC is recommended. To give you an idea of the dimensions, I have a slightly smaller1,5 K/W heatsink here. It is 75 x 100 x 40 mm.

Working with 4 Ohm speakers will bring the IC into the thermal shutdown zone even with a bigger 1 K/W heatsink.

You got three options.
  • - Buy a 2 x 15 V transformer and two 1 K/W heatsinks.
    - Add fan(s) to the heatsink(s).
    - Or buy those 8 Ohm speakers now.
While the IC is internally protected against immediate failure, driving it at its limits will shorten its lifetime significantly.
 
pacificblue said:
Your kit has a Boucherot cell aka Zobel network.

The instructions that came with your kit say in the introduction that it is supposed to work from 25 V into 8 Ohm. For that a 1,4 K/W heatsink per IC is recommended. To give you an idea of the dimensions, I have a slightly smaller1,5 K/W heatsink here. It is 75 x 100 x 40 mm.

Working with 4 Ohm speakers will bring the IC into the thermal shutdown zone even with a bigger 1 K/W heatsink.

You got three options.
  • - Buy a 2 x 15 V transformer and two 1 K/W heatsinks.
    - Add fan(s) to the heatsink(s).
    - Or buy those 8 Ohm speakers now.
While the IC is internally protected against immediate failure, driving it at its limits will shorten its lifetime significantly.

Thanks for the info. The ICs are getting very hot even with no load and no input connected. I am assuming this is very bad? What could be causing it? I read somewhere last night (can't find the link) that having the output ground connected to the ground bus on the PCBs and not straight to star ground will send the chip into oscillation, could this be the reason why they are getting so hot?
 
One step at a time. First step is to provide adequate heatsinking. Afterwards you will see, if your ICs still run hot. Until then don't bother thinking about oscillations.

The LM1875 draws 100 mA of idle current. With your ± 27 V power supply rails that means it has to dissipate 5,4 W without even driving a load.

Nice to touch is something between 30 °C and 40 °C. Somewhere between 40 °C and 60 °C people start to perceive as too hot to touch.

So let us assume you want to achieve 40 °C at the heatsink. Your room has 20 °C. The difference is 20 K. 20 K / 5,4 W = ~3,7 K/W. You need a heatsink with less than 3,7 K/W even with no load connected to perceive the temperature as comfortable.

That is not yet counting in the heat from the transformer. And you have two of those ICs in one case, which means the ambient temperature inside the case is above 20 °C. The heatsink temperature will be accordingly higher.

An IC does not need as low a temperature as 40 °C, although it doesn't hurt either. You should try to maintain it below 70-80 °C on worst case conditions. That is too hot to touch, but a good compromise between safe operation and heatsink price and size. And you won't always be listening at worst case level.

So if you connect 8 Ohm speakers, the 1,4 K/W recommendation per IC from the instruction sheet should give good results.

By the way. Did you mount some feet below the case since you posted the photos? You should at least allow fresh air to enter through the slits at the bottom and to leave at the top to improve heat disposal.
 
Thanks for your information, it is most appreciated. As this is my first build I have a steep learning curve and everyones input is helping a lot :)

Originally I was using the two aluminium sides of the casing for heatsinks (80 mm high, 230 mm deep and about 6 mm thick). Do you feel these are adequate? One side for each IC of course.

Also the IC isolator which came with the kit is a rubber pad about 2-3 mm thick. I have applied heat compound between the IC and pad then pad and heatsink. However when I feel around the area where the IC is mounted it does not seem to be getting that warm. This seems to me that the heat is not tranferring through the pad. Would you recommend a better pad? Maybe a mica one?
 
simonlarusso said:
Originally I was using the two aluminium sides of the casing for heatsinks (80 mm high, 230 mm deep and about 6 mm thick). Do you feel these are adequate? One side for each IC of course.
That is not easy to calculate, and depends also on whether it is black anodized or not and on the mounting direction. My guesstimate is they could work most of the time, but maybe not for party or live SPL.
simonlarusso said:
Also the IC isolator which came with the kit is a rubber pad about 2-3 mm thick.
According to the instruction sheet there should have been a silicon washer. It should be a tenth of the thickness you mention.
Could the rubber pad be supposed to go between screw and IC as damage prevention?

simonlarusso said:
However when I feel around the area where the IC is mounted it does not seem to be getting that warm. This seems to me that the heat is not tranferring through the pad.
Looks as if you found the culprit.

simonlarusso said:
Would you recommend a better pad? Maybe a mica one?
Yes, if the original silicon washer has disappeared.
 
I reseated the ICs on the heatsinks without an thermal compound and made sure they were good and secure. They chips are definitely cooler and I can comfortably keep my finder on them with no load.

So I think that solves my problem but I still need to get some 8 ohm speakers to test the amp fully.
 
it is an AC coupled amplifier, so It can't be DC offset problems.
No RF attenuation, could that be a problem.
The amp oscillated before and may be unstable just waiting for a high frequency whistle at it's input.
Did you hot plug it? That could have blown the fuses if it's unstable.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.