Matttcattt said:
No volume control on the schematic, but I don't believe he is listening without any form of volume control...
Or is he always listening at max volume?
Peter Daniel said:I always thought that volume control is a device that prevents from pushing it hard.
I suppose it was clear that with "pushing hard" I meant listening at loud volumes.
Hey, it's happening again and a beginner could be misled here
IF you are using the 22K resistor from pin 8 (inverting input) to ground, then the resistance from pin 7 (non-inverting input) should be covered in the range 0-50K.
If you are not using another resistor between non-inverting input and ground, then the resistance required to ground from the inverting input will be approxmately the same as the feedback resistor. As most people are using 200K-300K for feedback, then you would need at least a 300K trimmer to get zero DC offset!
For clarity, please always state if you are using the 'one resistor' or 'two resistor' method of getting 0mV offset.
Nuuk said:
if the goal is to minimize DC offset which is caused by the bias current on the input, wouldn't it be natural to ground the non-inverting input with a resistor of the same value as you have on the inverting input?
That way, both transistors on the two inputs receive the same abount of current -> no DC offset.
Nuuk said:
I am using the 22k resistor from pin 8 to ground.
And, as I reported, if you put a high value resistor from the non-inv. input to ground, you'll have noise (not humm, hiss).
You may not notice, but it's there.
I reported that on my tests with the LM3886.
And I showed the paragraph of the AN-1192 app note from National where they say te same (you should have a low value resistor there or you'll increase noise).
So, if using the 22k resistor from pin 8 to ground (as I do) allows you to pick a smaller value resistor for the non-inv. input to ground, then that's the way to go.
I use a 300K feedback res. and what you are stating is true ,except we have tolerences on resistors.After trying to match resistors (feedback and from pin 7) i found it was easier to just use a 500k pot (its what i had handy).As i stated it does drift when a greater volume is applied but i have found the drift to be =< .001 v.This is measured on my old fluke DVOM.
ron
ron
fender4 said:I built a pair of LM3875 "Gainclones" over 1 year ago based on the Thorsten schematic, and I have enjoyed it since day one. Just out of curiousity, I built a new channel up today...a stripped-down version. The schematic is below. Other than the cheapo 12-0-12 Radio Shack tranny and bridge, it only has 4 parts: 2X1000uf supply caps, a 220k resistor between pins 3 and 8, and a 10k resistor from 8 to input. No volume pot right now. All grounds meet at a "star" ground.
I didn't expect anything much, and a certainly didn't expect any stability. To my amazement, the "Gainclone Lite" has been running with CD music source for about 4 hours now with clean sound, NO noise, and 0.01mv DC offset.
I don't have access to a scope, but if this is any indication of stability--it's been running all this time without a heatsink! That's not intentional...I was just testing it initially, but it never really got that hot and the protection never kicked in. BTW, the chip is the LM3875TF, isolated package. I haven't really assessed the quality of the sound because it is currently hooked up to my test speakers (junk).
Is this just a fluke? I'm really enjoying the novelty of the minimalist approach, but this is almost TOO easy
-f4
This is the way we normally use an opamp, no? Since this chip is just an opamp with a little extra muscle, why should you be surprised that it works? It was designed to do so!
Jan Didden
ron clarke said:
If you measure the pot what resistance do you have?
I noticed hiss with the LM3886 and a 500k mt-pot which was tuned for 0mv DC offset, and a final resistance of 216k.
But the LM3886 is a different animal...
Anyway, even National says you shouldn't use a high value resistor, or you'll increase noise.
If you have a value of under 100k across the pot you may not notice any difference in noise.
Yes, that's exactly what I have done. But Ron Clarke chose to use a trimmer instead. Carlos questioned using a 500K and I was just pointing out, that is what is required when using the single resistor to ground instead of resistors to ground from both inputs.
The point is, anyone new to building a GC who reads that you can use a 50K trimmer to get zero offset, and who has not used a resistor from non-inverting input (pin 8) to ground, will not be able to achieve 0mV.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.