Problem with LM1875T amp - distortion, low volume and instability

[fjhuerta]

Hi all,

This is the first time I've built a chipamp. I thought I'd try the LM1875 due to its simplicity.

http://www.electronics123.com/s.nl/it.A/id.425/.f

I have a power supply that provides +-17V of unregulated DC power. I know it's a bit lower than optimal, but I wanted to use whatever I had on hand for this project.

Anyway, I bought two kits and built them both side by side. I powered the first one, and it worked beautifully. Then I powered the second one, and I have the following problems...

1) Sometimes, volume is too low and distorted.
2) Sometimes, volume is normal, but even more distorted.
3) Sometimes, I hear like a very fast sweep (instability?), the amp turns itself off, turns on after a while and repeats the sweep.

When I slighlty move the PCB apparently the LM1875 reacts in one of those three ways. Which is strange, since it seems to be very well soldered to the board.

I took out the amp and resoldered everything except for the LM1875T, which seemed OK. Everything seems the same.

I re-checked the components and their values. Then I followed the traces and checked for continuity. Everything is fine - the amp is identical to the one that works.

I've tested a different source. The problem is still there.

I've already checked to see if the chip is properly insulated against the heatsink. It's fine.

Today I'll try to replace the LM1875 with the one from the amp that is working, and see if this fixes the issue.

In case this doesn't help - what do you think the problem could be? I'm out of ideas by now...
thanks!

[danielwritesbac]

Originally Posted by fjhuerta

Hi all,

This is the first time I've built a chipamp. I thought I'd try the LM1875 due to its simplicity.

http://www.electronics123.com/s.nl/it.A/id.425/.f

I have a power supply that provides +-17V of unregulated DC power. I know it's a bit lower than optimal, but I wanted to use whatever I had on hand for this project.

Anyway, I bought two kits and built them both side by side. I powered the first one, and it worked beautifully. Then I powered the second one, and I have the following problems...

1) Sometimes, volume is too low and distorted.
2) Sometimes, volume is normal, but even more distorted.
3) Sometimes, I hear like a very fast sweep (instability?), the amp turns itself off, turns on after a while and repeats the sweep.

When I slighlty move the PCB apparently the LM1875 reacts in one of those three ways. Which is strange, since it seems to be very well soldered to the board.

I took out the amp and resoldered everything except for the LM1875T, which seemed OK. Everything seems the same.

I re-checked the components and their values. Then I followed the traces and checked for continuity. Everything is fine - the amp is identical to the one that works.

I've tested a different source. The problem is still there.

I've already checked to see if the chip is properly insulated against the heatsink. It's fine.

Today I'll try to replace the LM1875 with the one from the amp that is working, and see if this fixes the issue.

In case this doesn't help - what do you think the problem could be? I'm out of ideas by now...
thanks!

This sounds like a break in the trace on the PCB, or
you may need to apply flux and reheat the solder connections to stabilize them.

Also possible is a grounded heatsink (normal for a metal amp enclosure); however, when the LM1875's insulator is slipping somewhat, then that can cause partial electrical conductivity to the heatsink (but what you want is 100% stable electrical insulation between LM1875 and its heatsink).

EDIT: A fried or broken resistor or bad contact to a resistor can also cause these symptoms. The amplifier is acting like there is too much load per voltage and/or it is acting like the request (gain setting) is bizzare, so check the resistor soldering too. Check for DC offset before connecting speaker.

[91art]

1875 does not do insulation,Self-excited or bad connection.try this one will be better.8 components per channel very simple.Resistance and capacitance should also test match, the error can not be too big。transformer do not lower than the 60W

[fjhuerta]

Daniel, thank you for your excellent explanaton. I assumed many things (problematic capacitor was my first choice) until I read your post. Then I switched both LM1875ts and the problem still was with the same board. So you must be right - there's a problem with the traces or a resistor. The LM1875T is properly insulated, I did notice about V- being present at the heatsink.

I've read a couple of posts around here stating that the board I bought isn't the best design out there, and doesn't have the best quality components, so I just bought the pieces in order to duplicate 91art's design. I'll be building it tonight, without the pot (I have a passive linestage before the amp, so I won't be needing it).

Hopefully I'll have my first chipamp working very soon!

Thanks again for all your help, it's very appreciated.

[danielwritesbac]

Originally Posted by fjhuerta

Daniel, thank you for your excellent explanaton. I assumed many things (problematic capacitor was my first choice) until I read your post. Then I switched both LM1875ts and the problem still was with the same board. So you must be right - there's a problem with the traces or a resistor. The LM1875T is properly insulated, I did notice about V- being present at the heatsink.

I've read a couple of posts around here stating that the board I bought isn't the best design out there, and doesn't have the best quality components, so I just bought the pieces in order to duplicate 91art's design. I'll be building it tonight, without the pot (I have a passive linestage before the amp, so I won't be needing it).

Hopefully I'll have my first chipamp working very soon!

Thanks again for all your help, it's very appreciated.

Your K50 amplifier kits are easily hacked into stellar performers. One huge advantage to having the boards (and a desoldering iron "solder sucker) is being able to easily swap components in order to hear the various differences possible.

The largest problem with K50 is the inbuilt voicing to overcome the bum 22uF cap. You can simply install a better cap there, plus a little "bypass cap" for extra nice treble. After that you can change that bizzare 180k value down to 120k (or lower). Change its partner resistor (in the feedback voltage divider) to set your gain.

For example, if you had an mp3 player, you can divide 120k by 45 to get 2.7K which sets the LM1875 on maximum gain. In this case, the corresponding input load resistor would be 100k. Its compatible with the litespeed attenuator, tube projects, etc. . . and such a light load won't run down the batteries in your ipod.
If maximum is too much gain, you could use 3.3K instead of 2.7k (inverting input to nfb cap).

Another example is if your source device is quite stubborn it may need a stout load to get some power out of it. Creative Labs Xfi is one especially "stubborn" example. That amp setting is like 10k (metal) input load, 12k feedback resistor (metal), and 330R (carbon) from inverting input to ground. . . you can check DC offset with the NFB cap bypassed, and this particular resistor setting may provide the "under 90mv" whereby you could omit the NFB cap as is currently popular with chip amplifiers (however, don't omit the input cap).
For speaker protection, the NFB cap for such an amplifier is at least 220uF and you may prefer to use an output (series with speaker) cap as the protection means instead of using an NFB cap, as an output cap is usually easier to select for clarity.

The point of mentioning these options is that you can witness the differences and decide what you like. Its really easy to do it with the K50 board.

[fjhuerta]

Originally Posted by 91art

1875 does not do insulation,Self-excited or bad connection.try this one will be better.8 components per channel very simple.Resistance and capacitance should also test match, the error can not be too big。transformer do not lower than the 60W

I was wondering if I should add the RC circuit at the output. I read the designers' webpage, and it seems his LM1875's got too hot from the oscillation.

I'll add a 1 Ohm / 0.22 uF shunt to the circuit... and I'll use the boards from my kit later, with another amp. I think I'm getting hooked on chipamps!

[Juergen Knoop]

sorry
regards

[AndrewT]

Originally Posted by fjhuerta

I'll add a 1 Ohm / 0.22 uF shunt to the circuit...

why add another? The circuit you referred us to already has the Output Zobel fitted.

[fjhuerta]

Originally Posted by AndrewT

why add another? The circuit you referred us to already has the Output Zobel fitted.

Oh, that's because I built the circuit in the second post - which doesn't have the Zobel.

I finished it a couple of hours ago - and again, one channel worked perfectly well. But this time I know what happened - one LM1875t is damaged from removing it 3 times from the circuit. I had DC at the output, so I disconnected everything and noticed one pin was broken.

So I'll have to wait until I can buy another one....

But I'm now hooked on Chipamps. They are so simple to build!

[AndrewT]

post3 is a different circuit! Posted by someone else?