my preamp has a gain of 8x. Can someone please help on the recommended gain for the LM3886.
I searched the forums but came back empty. I know that Brian's user manual uses a default of around 30+ for gain. it is possible that this was set because the LM3886 doesn't uses an active preamp.
thank you very much for the help
I searched the forums but came back empty. I know that Brian's user manual uses a default of around 30+ for gain. it is possible that this was set because the LM3886 doesn't uses an active preamp.
thank you very much for the help
Dear Jarthel,
There is no data abgout your power needs. Anyway,
Overture Design Guide says,
With +/-35v PSU you can get ~61W at 8 ohm load.
You have 8x gain already..
If you give 22K/1K2 = 19,33 V/V gain then you will have 19,33+8 = 27,33 Total gain.
Result:
To get 61W (with 1% THD) on 8 ohm loads 200mV will be enough (as usual)..
If you use +/-35v PSU and have 8 ohm speaker then set the gain 19,33 V/V with 22K/1K2..
There is no data abgout your power needs. Anyway,
Overture Design Guide says,
With +/-35v PSU you can get ~61W at 8 ohm load.
You have 8x gain already..
If you give 22K/1K2 = 19,33 V/V gain then you will have 19,33+8 = 27,33 Total gain.
Result:
To get 61W (with 1% THD) on 8 ohm loads 200mV will be enough (as usual)..
If you use +/-35v PSU and have 8 ohm speaker then set the gain 19,33 V/V with 22K/1K2..
Because you will determine the gain anyway.
Ok if you use +-30v supply then,
You will get 44W into 8 ohm load.
To get this power with 8x pre amp and 200mV input sensivity, you must set the amplifiers gain 16V/V.
And with 18K/1K2 resistors you may get this gain. However in practise if you use more lower values on NFB line (like 1K8 / 120R) that gives your more noisefree operation, as my experiences..
Ok if you use +-30v supply then,
You will get 44W into 8 ohm load.
To get this power with 8x pre amp and 200mV input sensivity, you must set the amplifiers gain 16V/V.
And with 18K/1K2 resistors you may get this gain. However in practise if you use more lower values on NFB line (like 1K8 / 120R) that gives your more noisefree operation, as my experiences..
jarthel said:
Hi,
If you have Excel or equal go here for disign guide:
http://www.national.com/appinfo/audio/files/Overture_Design_Guide15.xls
Dxvideo said:
No,
If you have 8x gain already and set the 3886 gain to 19.33, then the total gain is 8 * 19.33 = 154.6 not 27.33.
You read that right; it is not a mistake. Gain factor is multiplied not added. Only if you express gain in dB can you add it. For example, a gain of 8 V/V is expressed as 18 dB, and a gain of 19.33 V/V is 25.7 dB, so the total gain should be 18 + 25.7 = 43.7 dB. This is confirmed: 154.6 V/V is 43.7 dB.
Remember that the 3886 requires gain to be set to 10x or higher for stability. So the minimum total gain of your system is 8 * 10 = 80 (V/V) or 38 dB. Your amp would clip with approximately 250 or 300 mV (RMS) at the input of the pre-amp. The overall gain is a little high but not too bad.
jarthel said:
I already know the formula for the gain from past experiences and reading the user manual of Brian's board. i guess the thread is a question of why such a gain/what gain to use instead how to calculate gain
Ok. You say your preamp has a gain of 8x (not 8v or 8db), that means that at ordinary line output (1.2Vrms) you will have 9.6Vrms on output.
For non clipping performance from the chip amp, gain would have to be ~10db or 3.2v/v. That is too little - gain must be 10v/v or greater for stability.
How's that?
it seems to me there are conflicting views in this thread which is getting me confused.
how did you get 16x gain as the ideal value? what datasheet/article are you looking at?
how do you know the non-clipping performance of the chip? what graph/page in the datasheet should I be looking at?
============
maybe I should lower the Vsupply?
Dxvideo said:
how did you get 16x gain as the ideal value? what datasheet/article are you looking at?
MJL21193 said:
how do you know the non-clipping performance of the chip? what graph/page in the datasheet should I be looking at?
============
maybe I should lower the Vsupply?
jarthel said:
In the following graphic from the xls worksheet you will see the box labeled "input for 1% output" it has a red arrow beside it. This label is a little bit misleading. It really should say "input of 100% output". Regardless, this is the maximum input voltage that the amp will handle before clipping.
Changing the gain of the amp will change this figure. If you lower the gain, you will increase this figure, therefore more input voltage can be dealt with before the amp clips.
Decreasing the rail voltage will decrease the maximum input voltage and maximum power output so this is not a good solution.
As you increase or decrease the gain of the amp, the power output does not change (second red arrow), only the maximum input voltage.
If you must use this preamp, you should try to find some way to reduce it's gain.
Addition of a pot before the chipamp with a standard gain of 27db will be fine for line input from most devices or you could use it to reduce the input from your preamp..
Attachments
MJL21193 said:
I played with the feedback resistors to get some idea of "input for 1% output". Even with a gain of 11x and Vsupply of +/- 30V (6ohm load), the "input for 1% output" = 2.4+Vrms.
My CD player has a 2Vrms output. I've looked around and ML's CD player (http://www.marklevinson.com/products/specs.asp?prod=no390s) has a 2.25Vrms. that's 2+Vrms without active preamplication. even with a preamp with a 2x gain (2 x 2Vrms), that's already over the 2.4+Vrms max input voltage.
Is there something wrong with the spreadsheet or a chipamp is better suited for passive preamp?
jarthel said:
Hi,
Site you linked to the ML CD player give this:
"Maximum Output (0dBFS): 4.45V balanced
2.225V single-ended "
2.25V is less than 1.6Vrms. Rms is the mean voltage, that spec is for peak voltage (the balanced is for peak to peak).
A good example is our household current. We have 120 VAC here but thats the rms value. It's actually 170 volts peak. Total voltage swing is 340 volts peak to peak, but the negitive protion of the wave is not considered for AC.
A quick way to convert peak voltage to rms is to divide by 1.414.
It's confusing, I know.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.