tube preamp to mate with Gainclone

[JAZZ2250]

I have built couple of solid state amplifiers. Recently, I'm much into the sound of so-called Gainclone which is based on LM3875 (National Semiconductors Power OP-Amp). Many people said that its sound was pretty much close to that of tubes. I don't know if it's true or not because I have no experience with tube stuffs. So, I decided to make a tube preamp to mate with LM3875.
The tube preamp is supposed to do all the voltage amplification and power amp based on LM3875 will do the current amplification. Once I thought to build Foreplay, but I changed my mind to go with a complete DIY. I would like to get some recommendation from the tube guys here which project would be good for my plan. So, please feel free post what you guys think. Gain of 20-25 dB should be high enough for me. Thanks.

[Steve Eddy]

Originally posted by JAZZ2250
The tube preamp is supposed to do all the voltage amplification and power amp based on LM3875 will do the current amplification.

That won't work I'm afraid.

The LM3875 isn't unity gain stable. It should be operated at a minimum gain of 10.

Might want to consider something like this:

<a href="http://sound.westhost.com/project83.htm">MOSFET Power Follower</a>

se

[JAZZ2250]

Originally posted by Steve Eddy

That won't work I'm afraid.
The LM3875 isn't unity gain stable. It should be operated at a minimum gain of 10.

Thanks for your comment. But...
It works, actually, without any problem.
I already built a Gainclone with unity gain.
Please check the following link if you're interested.

Click here


Now, what about the tube preamp?
Any recommendation, please?

[Steve Eddy]

Originally posted by JAZZ2250
Thanks for your comment. But...
It works, actually, without any problem.
I already built a Gainclone with unity gain.
Please check the following link if you're interested.

I read it. And the LM-3875 isn't operating at unity gain. It's operating at a gain of about 11. In other words, it's applying voltage gain. It SEEMS as if it's operating at unity gain because your circuit is attenuating the input signal by the same factor as the voltage gain of the 3875. That's not the same thing as if the 3875 were operating at unity gain. With the 3875 operating with a voltage gain of 10 and attenuating the input signal by the same amount, all you're doing is increasing the signal-to-noise ratio.

If that works for you, that's great. But it's not doing what you indicated you want it to do (i.e. the preamp handing voltage gain with the power amp only applying current gain). Both your preamp and your power amp are applying voltage gain. You've just got a resistive attenuator between the two that makes it only seem that that's the case.

It's kind of like how Congress only made it seem like we had a budget surplus.

If you want the real deal, you'd use a follower.

se

[analog_sa]

SE

I read it. And the LM-3875 isn't operating at unity gain. It's operating at a gain of about 11. In other words, it's applying voltage gain. It SEEMS as if it's operating at unity gain because your circuit is attenuating the input signal by the same factor as the voltage gain of the 3875. That's not the same thing as if the 3875 were operating at unity gain. With the 3875 operating with a voltage gain of 10 and attenuating the input signal by the same amount, all you're doing is increasing the signal-to-noise ratio.

I also read it and what you say appear to be wrong. Which circuit are you refering to? If feedback resistor and input resistor are the same the gain is clearly 1.

[Steve Eddy]

Originally posted by analog_sa
I also read it and what you say appear to be wrong. Which circuit are you refering to? If feedback resistor and input resistor are the same the gain is clearly 1.

Apparently he's using this circuit:

<center>
<img src="http://www.diyaudio.com/forums/attachment.php?s=&postid=112202">
</center>

Although the input resistor is the same value as the feedback resistor (220k), it's actually in parallel with the 22k shunt resistor at the inverting input. So you have 220k||22k, which comes to 20k. So the voltage gain is 220k/20k which comes to 11 (about 21dB)

And that 220k/22k combination also forms a 1:10 voltage divider so that the input signal is attenuated by a factor of 10, or 21dB.

But the fact remains that the LM-3875 is functioning <b>EXACTLY</b> as if you were to remove the 22k resistor and replaced the 220k input resistor with a 20k resistor.

Again, it only SEEMS as if it's operating at unity gain because the signal's being attenuated by the same amount as the voltage gain. You've got 21dB of signal attenuation combined with 21dB of voltage gain.

Smoke and mirrors.

mc2 laid it all out very nicely in this post here:

http://www.diyaudio.com/forums/showt...587#post129587

se

[analog_sa]

Oh, of course you're right. I never saw much sense in Kuei's suggestion to introduce the 22k resistor. I find inverting amps stable at any gain even if the non-inverting configuration has a minimum gain for stability. After all how is the opamp to know what is your source resistance?

[JAZZ2250]

Originally posted by Steve Eddy

If you want the real deal, you'd use a follower.
se

So, are you saying that I should use LM3875 in a follower configuration? Would it be stable without oscillation? It seems my quqestion here should be posted in the solid state forum. But, I started here anyway, so I'll just keep going here.

And, I'm still waiting for a tube preamp recommendation. Please...

[Steve Eddy]

Originally posted by JAZZ2250
So, are you saying that I should use LM3875 in a follower configuration? Would it be stable without oscillation?

You can't. It's an opamp. There is no "follower configuration" per se.

A follower is basically just a tube or BJT or FET configured as a common-anode/collector/drain amplifier. What are also referred to as cathode/emitter/source followers. Their output voltage follows the input voltage with no voltage gain (in the real world, the voltage gain is going to be slightly less than 1).

An example of a follower (in this case a source follower) is that MOSFET power follower circuit I referenced you to earlier.

At best all you can do with an opamp is configure it as a unity gain buffer. But the LM-3875 isn't stable at unity gain. That's why you're using that circuit you're using now, so it's voltage gain remains at about 10 or above.

If you're happy with what you have now, then by all means stick with it. I'm not saying there's anything wrong with it. I'm simply pointing out that it's not giving you that preamp for voltage gain only/power amp for current gain only configuration you said you wanted.

But if that's what you'd rather have, then you'll either need to find a power opamp that's unity gain stable, or use a follower circuit.

It seems my quqestion here should be posted in the solid state forum. But, I started here anyway, so I'll just keep going here.

Fine with me 'til the moderators say otherwise.

And, I'm still waiting for a tube preamp recommendation. Please...

'Fraid I can't be of much help to you there. Tubes aren't my forte.

se

[Steve Eddy]

Originally posted by analog_sa
Oh, of course you're right. I never saw much sense in Kuei's suggestion to introduce the 22k resistor. I find inverting amps stable at any gain even if the non-inverting configuration has a minimum gain for stability.

Great. Unfortunately JAZZ2250 hasn't found the LM-3875 to be stable at unity gain.

After all how is the opamp to know what is your source resistance?

It doesn't. It relies on a reasonably competent designer to read the data sheets and take the appropriate measures to accommodate the range of source impedance values the circuit's likely to encounter.

se