Beginner's Gainclone, HiFi LM1875, The Amplifier Board

Jump to Latest
A

AndrewT

R.I.P.
Joined 2004
N

Nag

Member
Joined 2015
Thank you guys for helping me on this.
Actually it is not point to point soldered build. I brought below products and EI 15V-0-15V(2A) and soldered wires accordingly. First one week worked and that later tranfo burned.
https://cdn.shopify.com/s/files/1/1038/6272/products/LM1875_01_1024x1024.JPG?v=1466597104
https://cdn.shopify.com/s/files/1/1038/6272/products/LM1875_02_1024x1024.jpg?v=1463666574
https://cdn.shopify.com/s/files/1/1...-a542-adb7ba8c3fc4_1024x1024.JPG?v=1463666576
https://cdn.shopify.com/s/files/1/1038/6272/products/EAPPS07_01_1024x1024.jpg?v=1464436425
https://cdn.shopify.com/s/files/1/1038/6272/products/EAPPS07_02_1024x1024.JPG?v=1464436425
https://cdn.shopify.com/s/files/1/1038/6272/products/EAPPS07_03_1024x1024.JPG?v=1464436425
Now I ordered 15V-0-15V 3A transformer.
Thanks,
Nag.
 
danielwritesbac

danielwritesbac

Member
Joined 2007
startfetch.com
Let's look closer at 90VA (15V-0-15V 3A): So 80 watts out of the power supply, two 60% efficiency little chips, each doing 16W to heatsink and 24W to speaker, is especially close to the right size transformer. Bearing in mind that with LM1875 chips, it is necessary to run out of power before running out of chip (no need to cook the chips!); so, we wouldn't want a transformer larger than that.

In fact, for longevity purposes, I'd like to see a smaller transformer (that 2A model), if 4 ohm speakers were also involved (and I hope that isn't the case, because heavy load is not an ideal use for LM1875).

Basically, fingernail size chip doesn't need huge transformer. :)
 
A

AndrewT

R.I.P.
Joined 2004
danielwritesbac said:
Let's look closer at 90VA (15V-0-15V 3A): So 80 watts out of the power supply, ..........
Click to expand...
Neither you nor I can get 80W out of a 90VA transformer when it is connected to a capacitor input filter.

The manufacturer tells us so.

the most we can get out of it will be around 55% of it's VA and if that were continuous, it will run hot.

60VA and 55% comes to ~33W and that is barely enough to run one channel.
90VA & 55% comes to ~50W and that is not enough for two channels.

It is normal to specify the mains transformer having one times to two times the total maximum power of the amplifier,
except when that comes to very low VA. Instead use a minimum of ~160VA to keep the transformer regulation lowish to avoid serious sagging of the power supply rails when the audio gets loud.
160VA will work well with all ClassAB upto ~50W+50W stereo amplifier.
 
Last edited:
danielwritesbac

danielwritesbac

Member
Joined 2007
startfetch.com
AndrewT said:
Neither you nor I can get 80W out of a 90VA transformer when it is connected to a capacitor input filter.

The manufacturer tells us so.

the most we can get out of it will be around 55% of it's VA and if that were continuous, it will run hot.

60VA and 55% comes to ~33W and that is barely enough to run one channel.
90VA & 55% comes to ~50W and that is not enough for two channels.

It is normal to specify the mains transformer having one times to two times the total maximum power of the amplifier,
except when that comes to very low VA. Instead use a minimum of ~160VA to keep the transformer regulation lowish to avoid serious sagging of the power supply rails when the audio gets loud.
160VA will work well with all ClassAB upto ~50W+50W stereo amplifier.
Click to expand...

I am most grateful for such a helpful and informative post.

Okay, the 90VA, 55% of that in watts at power board output, then ~50W, divided by two if two equally powered channels in stereo, then 25W each chip, 60% of which may go to the speaker for 15W. . . which is 2W short of what LM1875 could do on +-20vdc with an 8 ohm speaker (17W).
And then, by this estimate, that transformer falls short slightly.
I fail to realize how the difference in decibels output at the speaker will be noticeable.

Momentarily almost changing the topic (except that we'd have to consider both factors), for having both performance and longevity, it is vital to provide a just-right current to LM1875's signal input and power input pins as well. And, then the case remains that a slightly too small transformer is better than poor longevity for LM1875.

So, how close can we get if we chose a larger transformer, while also assuring that we don't provide more than 22W to each speaker? Quite the gamble with a stereo build? I've been a bit puzzled by that.
 
danielwritesbac

danielwritesbac

Member
Joined 2007
startfetch.com
Also, performance and longevity of the inputs.

It is also a bit difficult to figure the appropriate feedback resistor value for optimized current.
200k for single rail, a high figure to prevent breakage (in the datasheet).
22k for split rail with the 12+12vac transformer involved
100k for split rail with the 18+18vac transformer involved
SO, what's the ballpark resistor figure when you've got a 15+15vac transformer? That is not yet documented.

Too little current in the feedback resistor is basically fuzz, or too much current in the feedback resistor results in either Basset Hound Tone (instability) and/or a broken chip, probably both.

With that voltage, I guess 56k would work. But, I just don't know. What I do know is that neither 22k nor 100k would work suitably, and so the optimal figure must be somewhere in-between.
 
A

AndrewT

R.I.P.
Joined 2004
danielwritesbac said:
I am most grateful for such a helpful and informative post.

Okay, the 90VA, 55% of that in watts at power board output, then ~50W, divided by two if two equally powered channels in stereo, then 25W each chip, 60% of which may go to the speaker for 15W. . . which is 2W short of what LM1875 could do on +-20vdc with an 8 ohm speaker (17W).
And then, by this estimate, that transformer falls short slightly.
I fail to realize how the difference in decibels output at the speaker will be noticeable. ...............
Click to expand...
you are looking at the wrong parameter.

A small transformer will allow the voltage to sag a lot whenever the current demand increases.
The opposite effect happens when the current demand lowers. A small transformer will deliver a much higher voltage during quiescent conditions.
Note I said:
Instead use a minimum of ~160VA to keep the transformer regulation lowish to avoid serious sagging of the power supply rails when the audio gets loud.
Click to expand...
That has told you that the transformer regulation affects the supply rail voltage. You chose to ignore a very important aspect of reliability.
Will a too small transformer overheat the chip and reduce reliability?
A 160VA toroid could have a regulation around 8%.
A 60VA EI could have a regulation around 15%.
That could be sufficient to overvoltage the chip.
 
danielwritesbac

danielwritesbac

Member
Joined 2007
startfetch.com
AndrewT said:
I agree, you don't know.
Select a gain value and determine the NFB and input impedance to give that gain.
Click to expand...
For optimal use, the LM1875 will want a certain gain factor and feedback current, specifically suited to the conditions wherein it is used.
At outcome, the chip will choose--it is not up to me.

Other inconvenience is that it has wide range usage, which is a also a big difference in current between examples, so that only one schematic couldn't possibly suit all cases.
 
danielwritesbac

danielwritesbac

Member
Joined 2007
startfetch.com
Nag said:
Dear Andrew and Members,
As per the above comments I realized and changed my older order 90VA to150VA--15V-0-15V 5A transformer.
I will make schematic diagram and post my build pic.
Thanks for valuable inputs,
Nag.
Click to expand...

So, if there's no heavy load speakers within cable range then there's a chance that the speakers and amplifiers might not blow up. But, otherwise we need a discussion about fingernail size chips powered by huge transformers.
 
N

Nag

Member
Joined 2015
Dear Andrew T & Daniel and Members,

Here is my schematic diagram. Some capacitor and resistance values are difficult calculate. It is too hard to draw diagram from brought amp/psu.
out put speakers chr-70 8ohm. for both chips draw power from c3 and c6 are same.
PSU33.JPG
 
A

AndrewT

R.I.P.
Joined 2004
Nag said:
..................
PSU33.JPG
Click to expand...
remove S1 & S2 Fuses here need to be so big to survive start up that they offer no safety advanatge when something goes wrong. A close rated mains fuse is far better at giving the required safety.
remove C5 & C6. They are more likely to increase ringing on the supply lines.
They cannot offer lower impedance seen from the amplifier because they are too far away. Local Supply Rail Decoupling has to be LOCAL to where the currents change.This sch is completely ****** up. Is that an interpretation of what Daniel is giving you?
Go back to the datasheet from National and start with that.
Then add in filters at the input and Thiele Network at the output and proper Local Supply Rail Decoupling.
 
Last edited:
N

Nag

Member
Joined 2015
Dear Andrew,
I am completely newbie on amp/audio circuits. Some times even I don't understand technical words like feedback/zobel network/star earth ground bla bla. So as per your reply PSU is ok except c5 and c6. Can I through that amp and start build own amp form National semiconductor or form this DAN reference circuit. Which one is good ?. Speakers have efficiency 85.5dB 8ohm.
Thanks,
Nag.
 
You must log in or register to reply here.
Top Bottom