hi, right now i am designing a pcb for a paralleld LM4780 for each channel.
i'm planing on using the mute function for a softstart.
i want the capacitor to discharge fast, so the softstart will occure the next time i turn on the amp even if i do it right away.
so.... this is why i'm gonna have to short the capacitor and the V- (-30V)
and have a diode there.
the thing is i know nothing about diodes and i am asking for the type of diode i have to use, and its values.
tnks for the help!
i'm planing on using the mute function for a softstart.
i want the capacitor to discharge fast, so the softstart will occure the next time i turn on the amp even if i do it right away.
so.... this is why i'm gonna have to short the capacitor and the V- (-30V)
and have a diode there.
An externally hosted image should be here but it was not working when we last tested it.
the thing is i know nothing about diodes and i am asking for the type of diode i have to use, and its values.
tnks for the help!
1N4004 OK, but your resistors are off a bit
The 1n4004 will be fine...you need some a little beefy to handle the current from the 100 uF...the 1n4004 will be fine.
However...75K plus 13k7 is too high for a 30 V negative supply rail. You won't really get the 0.5 mA the data sheet calls for...
You need the sum of these resistors (assuming you're driving a single mute pin) to be less than (30-2.6)/0.5mA=54.8K...Maybe a total of 40K would give you a safe margin...that will change your mute time constants, but at least you'll come all the way out of mute...
The 1n4004 will be fine...you need some a little beefy to handle the current from the 100 uF...the 1n4004 will be fine.
However...75K plus 13k7 is too high for a 30 V negative supply rail. You won't really get the 0.5 mA the data sheet calls for...
You need the sum of these resistors (assuming you're driving a single mute pin) to be less than (30-2.6)/0.5mA=54.8K...Maybe a total of 40K would give you a safe margin...that will change your mute time constants, but at least you'll come all the way out of mute...
hi djoffe, tnks for your comment!
the RM2 resistor is calculated for a current of 1mA for each mute pin as suggested in the datasheet (this is a LM4780 so it has 2 mute pins).
the calculation:
RM2=(30-2.6)/2mA=13.7k
the RM1 is calculated to have a 4 sec delay when the AMP starts. (i'm having another circuit with a relay that close 3 sec after the AMP starts and i don't want the closing sound heard from the speakers)
does it still look wrong?
the RM2 resistor is calculated for a current of 1mA for each mute pin as suggested in the datasheet (this is a LM4780 so it has 2 mute pins).
the calculation:
RM2=(30-2.6)/2mA=13.7k
the RM1 is calculated to have a 4 sec delay when the AMP starts. (i'm having another circuit with a relay that close 3 sec after the AMP starts and i don't want the closing sound heard from the speakers)
does it still look wrong?
The current has to flow through RM1 and RM2. Once CM is charged, you can just sum the values of RM1 and RM2. Calculate the values without the capacitor, then change the value of the capacitor to change the delay. Also, it states in the documentation that you should double the minimum required current to 2ma for both channels of a LM4780.
Last edited:
if i understood you correctly RM1+RM2=13.7k so the current will be 2mA (1mA per mute pin).
lets say RM2=3.7k and RM1=10k
according to my calculation T=7.5 (t=4sec)
T=R*C
C=T/R =7.5/10k = 0.00075F = 750uF
now i just have to make the capacitor a 1000uF and the resistor 7.5k
then RM2=13.7k-7.5k=6.2k ?
is this what you ment??
tnks for helping by the way!
lets say RM2=3.7k and RM1=10k
according to my calculation T=7.5 (t=4sec)
T=R*C
C=T/R =7.5/10k = 0.00075F = 750uF
now i just have to make the capacitor a 1000uF and the resistor 7.5k
then RM2=13.7k-7.5k=6.2k ?
is this what you ment??
tnks for helping by the way!
values...
If we accept the pessimistic data sheet declaration that they want 2 mA (1 mA per side, for a total of 2 mA), then we'd need Rm1+Rm2<(30-2.6)/0.002=13.7K.
One reasonable choice would be to make Rm1=Rm2=6850 Ohms.
C then is made as big as you need to hold the mute on for the desired time. Its exact size is a bit hard to say without a good model of the mute input pins.
If we accept the pessimistic data sheet declaration that they want 2 mA (1 mA per side, for a total of 2 mA), then we'd need Rm1+Rm2<(30-2.6)/0.002=13.7K.
One reasonable choice would be to make Rm1=Rm2=6850 Ohms.
C then is made as big as you need to hold the mute on for the desired time. Its exact size is a bit hard to say without a good model of the mute input pins.
Yes TigerGuy, that's exactly what I'm saying And really, you don't need such exactly values, you can pick others that are close if they are easier to get. The mute is not really a full on/full off type circuit, it follows a curve that I believe is shown in the datasheet. As long as you have a bit more than 0.5mA per channel, it'll be fine. 2mA total was suggested to make sure supply sag doesn't cause the output to go a bit lower.
And really, you don't need such exactly values, you can pick others that are close if they are easier to get. The mute is not really a full on/full off type circuit, it follows a curve that I believe is shown in the datasheet. As long as you have a bit more than 0.5mA per channel, it'll be fine. 2mA total was suggested to make sure supply sag doesn't cause the output to go a bit lower.
Shameless plug -- you can use my boards specifically for parallel, bridged or dual mono or Peter Daniel's which are adaptable for any situation. i think it's a lot less expensive than having boards burned on your own. the diode isn't going to be troubled by any leakage issue in this ap so anything like an 1N4002, etc should do fine.
Hi,
+IN A and +IN B share a common grounding resistor (47k) before reaching the DC blocking cap.
I do not think this is good practice. Mixing the input offset currents likes this may lead to instability of the paralleled channels.
I think you need a DC block on each +IN input.
The +IN input sees an effective 1k0 + 2*[47k] as the route to ground for it's input current. That is ~95k.
The -IN input sees 22k (21k shown) as the route to an effective ground.
This will result in a big input offset voltage and to a bigger than necessary change in input offset voltage with changes in temperature.
This will lead to a variable output offset for the two paralleled amplifiers.
I suspect this proposal to suffer similar stability problems as other paralleled chipamp implementations.
+IN A and +IN B share a common grounding resistor (47k) before reaching the DC blocking cap.
I do not think this is good practice. Mixing the input offset currents likes this may lead to instability of the paralleled channels.
I think you need a DC block on each +IN input.
The +IN input sees an effective 1k0 + 2*[47k] as the route to ground for it's input current. That is ~95k.
The -IN input sees 22k (21k shown) as the route to an effective ground.
This will result in a big input offset voltage and to a bigger than necessary change in input offset voltage with changes in temperature.
This will lead to a variable output offset for the two paralleled amplifiers.
I suspect this proposal to suffer similar stability problems as other paralleled chipamp implementations.
thank you all for helping!
Andrew so you suggest having a DC block circuit on each +IN input (with a 47k or a 23k5 resistor for each +IN input?)
what about the -IN input? how can i prevent those instabilitys?
plus i am a little confused because this is how it apears in the orginal national schematic.
http://www.national.com/ds/LM/LM4780.pdf (page 5)
Andrew so you suggest having a DC block circuit on each +IN input (with a 47k or a 23k5 resistor for each +IN input?)
what about the -IN input? how can i prevent those instabilitys?
plus i am a little confused because this is how it apears in the orginal national schematic.
http://www.national.com/ds/LM/LM4780.pdf (page 5)
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.