Page 7 (Fig 3.) of datasheet ( http://www.ti.com/lit/ds/symlink/lm3875.pdf )shows single supply cct.
Add inverting opamp in front of it and reduce signal for overall gain of -1.
Done.
Add inverting opamp in front of it and reduce signal for overall gain of -1.
Done.
Input gets cap in series with R2 (Xc= R2/10 at lowest F)
R1 gets a Cap to gnd (Xc= R1/10 at lowest F) (Some opamps don't need R1)
R3 goes to Vcc/2 with bypass cap at Vcc/2 to gnd
Output gets coupling cap (Xc= loadR/10 at lowest F)
Negative opamp supply goes to gnd.
Lowest F is your desired 3dB point...just a guideline.
R1 gets a Cap to gnd (Xc= R1/10 at lowest F) (Some opamps don't need R1)
R3 goes to Vcc/2 with bypass cap at Vcc/2 to gnd
Output gets coupling cap (Xc= loadR/10 at lowest F)
Negative opamp supply goes to gnd.
Lowest F is your desired 3dB point...just a guideline.
You can't use it unity gain, it is not compensated as such and will oscillate like crazy.
Minimum recommended gais is 26dB, so if you are trying to use it, as, say, the second "arm" of a bridged amp you'll have to attenuate anyway.
From the datasheet:
If you want an inverting amp, build the regular non onverting version, ground the + input and feed the - input with a 20K resistor from the other amp's output ( or the same R value you used as Rf).
No need for an external inverter.
"The LM3875 is designed to be stable when operated at a closed-loop gain of 10 or greater, "
So, a gain of -1 is greater than 10?
Just asking.
Might work. I've thought about this solution but I've not tried it so I can't say one way or the other.
Let us know of the results.
P.S. Don't ground the + input...not on a single supply amp. If you change the position of the feedback shunt resistor (low value R) and its' cap to gnd then you can connect the 20K input R to the top of the Resistor and save a coupling cap in the circuit.
Do you know what I am trying to say?
So, a gain of -1 is greater than 10?
Just asking.
Might work. I've thought about this solution but I've not tried it so I can't say one way or the other.
Let us know of the results.
P.S. Don't ground the + input...not on a single supply amp. If you change the position of the feedback shunt resistor (low value R) and its' cap to gnd then you can connect the 20K input R to the top of the Resistor and save a coupling cap in the circuit.
Do you know what I am trying to say?
Well, if it's a single supply amp, you still need to ground it (you need a ground reference) but, of course, through a capacitor.
10 to 100uF will be fine.
And yes, running the 20K resistor straight to the first amp output can be done, no capacitors involved, if both are biased from the same 1/2V+ voltage divider.
lm3875 etc have been made stable at, in inverted configuration, at unity gain. This has been much posted about in this forum, so I won't repeat it here. Point is, can it be done from a single rail.
Last edited:
It just dawned on me that ThorstenL's circuit IS 20dB of attenuation followed by 20dB of amplification.Thanks, but attenuate to then amplify is what I'm trying to avoid.
I've been looking at ThorstenL's unity gain chip amp and am trying to picture it on single rail.
"ThorstenL's circuit IS 20dB of attenuation followed by 20dB of amplification"
I'm not sure about that since with an inverting amp the - input node stays at 0V, there is no current in R1. (Or very, very little)
I've seen caps in that position to swamp input capacitance for stability issues...maybe the R is doing the same type of job.
It might be balancing DC offset factors.
Probably need full Pspice analysis to see for sure.
Or just try it.
I'm not sure about that since with an inverting amp the - input node stays at 0V, there is no current in R1. (Or very, very little)
I've seen caps in that position to swamp input capacitance for stability issues...maybe the R is doing the same type of job.
It might be balancing DC offset factors.
Probably need full Pspice analysis to see for sure.
Or just try it.
Yes. I'm full of flu and not thinking straight.
The point is to have it after a tube preamp. ECC88s at about 70-80V, which the 3875 can also use, thus simplifying the PSU, though the triodes will need extra smoothing. It will be into 16 ohm speakers so the high voltage for the chip amp should be fine.
The other alternative is running the triodes off the +/- rails of a balanced supply. I've no idea if this is a good idea.
The point is to have it after a tube preamp. ECC88s at about 70-80V, which the 3875 can also use, thus simplifying the PSU, though the triodes will need extra smoothing. It will be into 16 ohm speakers so the high voltage for the chip amp should be fine.
The other alternative is running the triodes off the +/- rails of a balanced supply. I've no idea if this is a good idea.
http://www.ti.com/lit/ds/symlink/opa445.pdf
I built an amp with this...with some current buffering of course (30A+ of it)
It was a bass guitar amp so did not really check high F
200W 2R
I built an amp with this...with some current buffering of course (30A+ of it)
It was a bass guitar amp so did not really check high F
200W 2R
That's exactly what I was going to write about the "unity gain inverters" you mentioned.
Glad you found it on your own.
ADD:
Ah!, but there *is* current from the input into R1 , just that the Op Amp tries hard to keep voltage difference between both inputs as "0", so the Op Amp injects, through the NFb 20 K resistor , the exact same value, in opposite phase, to cancel it.
And in doing so, will apply at the output whatever voltage is needed.
That's exactly what Op Amps were invented for.
Then, is it the same to use R1 (500 ohms) or not?
It depends:
1) as of total gain, it changes nothing, you will still have unity gain, end to end
2) as to loop gain, the Op Amp "sees" 20X gain, couldn't care less as to what you are connecting to "the other end" of the 20K input resistor.
And there lies the difference.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.