The circuit can be found here:
http://www.tech-diy.com/LM4780_bridgedamp.htm
I followed it except that I used my existing power supply.
All the values I used for the amplifier are per the schematic.
http://www.tech-diy.com/LM4780_bridgedamp.htm
I followed it except that I used my existing power supply.
All the values I used for the amplifier are per the schematic.
I'm sitting listening to a 4780 driven sub at the moment, and it's fine. The 4780 chip is actually two 3866s on one die, so I suspect your implementation is at fault.
Mr Pinkmouse
I understand that the 4780 has two LM3886's in it
If I followed the schematic that I put a link to would my implementation still be at fault?
Somehow I just don't think so..
burnedfingers said:The circuit can be found here:
http://www.tech-diy.com/LM4780_bridgedamp.htm
I followed it except that I used my existing power supply.
All the values I used for the amplifier are per the schematic.
The circuit looks OK. The high-pass frequencies from the input coupling and feedback are down to a few Hz, so that cannot explain the bass problem.
How do you know you have a bass problem? If you indeed have one, how do you know it is in the amp, and not in the speaker or the room? Is the problem in both channels? Did you measure the frequency response, or maybe the gain at some spot frequencies?
Jan Didden
Hi Burned,
you have bridged two 50W into 4ohm amps to give 100W into 8ohm.
The supply is just 4 * 10mF.
That is equivalent to just +-10mF per 4ohm amplifier.
For good bass, I would recommend +-40mF/amp to +-50mF/amp, yes that's 16 to 20*10mF.
Seems like you will need to spend some money on more caps before you trash it.
Do you realise that if you had built a normal 100W into 8ohm amp the smoothing requirement would have been +-20mF/amp to +-25mF/amp?
you have bridged two 50W into 4ohm amps to give 100W into 8ohm.
The supply is just 4 * 10mF.
That is equivalent to just +-10mF per 4ohm amplifier.
For good bass, I would recommend +-40mF/amp to +-50mF/amp, yes that's 16 to 20*10mF.
Seems like you will need to spend some money on more caps before you trash it.
Do you realise that if you had built a normal 100W into 8ohm amp the smoothing requirement would have been +-20mF/amp to +-25mF/amp?
The circuit looks OK. The high-pass frequencies from the input coupling and feedback are down to a few Hz, so that cannot explain the bass problem.
How do you know you have a bass problem? If you indeed have one, how do you know it is in the amp, and not in the speaker or the room? Is the problem in both channels? Did you measure the frequency response, or maybe the gain at some spot frequencies?
How do I know that I have a bass problem. In my field I am required to be able to spot any and all problems associated with any given audio system very large or small.
You have very good questions and certainly the same that I would ask of someone.
Given the same placement of the speakers in the room the same source gear and the same signal in this case the same CD with the same track I tested 3 different amplifiers. There was clearly no lack of bass from the LM3886 chip amp, the Quicksilver mono block 8417's, or an Altec 9444.
Now, I could go further and hookup a warble generator and calibrated microphone and plot a frequency response graph like I did. I tried to send the graph but it is too big.
Are both speakers connected in-phase?
First thing I checked....Yes
Hi,
I think the lack of understanding of the current requirements of a bridged pair of amplifiers has led to the poor bass.
This in turn has prompted the poor implementation.
Burned,
go back and review what you did and why.
Then imagine starting again with a 50W into 8ohm amplifier and instead of using 4*10mF for smoothing you substitute 4*2500uF.
Would you expect a change in the bass delivered?
It's down to that factor of four increase in current that the bridged amplifier needs.
I think the lack of understanding of the current requirements of a bridged pair of amplifiers has led to the poor bass.
This in turn has prompted the poor implementation.
Burned,
go back and review what you did and why.
Then imagine starting again with a 50W into 8ohm amplifier and instead of using 4*10mF for smoothing you substitute 4*2500uF.
Would you expect a change in the bass delivered?
It's down to that factor of four increase in current that the bridged amplifier needs.
AndrewT said:Hi Burned,
you have bridged two 50W into 4ohm amps to give 100W into 8ohm.
The supply is just 4 * 10mF.
That is equivalent to just +-10mF per 4ohm amplifier.
For good bass, I would recommend +-40mF/amp to +-50mF/amp, yes that's 16 to 20*10mF.
Seems like you will need to spend some money on more caps before you trash it.
Do you realise that if you had built a normal 100W into 8ohm amp the smoothing requirement would have been +-20mF/amp to +-25mF/amp?
Andrew,
As I noted, the supply caps have nothing to do with the bass response except in the limiting case of constant full power output. This is clearly a freq response problem.
I also noted that apparently everybody here likes to chase ghosts about supply caps without really being interested in WHY the bass lacks.
Anyway, I wish you all a lot of fun chasing those ghosts (and possibly trashing a perfectly good amp).
Jan Didden
Most likely it is the input coupling -- in the bridged board I left space for a variety of coupling caps -- and in the article showed an alternate arrangement whereby two 100uF electrolytics could be back-to-backed with a polypropylene or other. The parallel board shows 1uF which has a much different f3 point. This is OK, but remember that the phase is shifting all the way out to a few hundred hertz with a 1uF input capacitor !!!
It would be helpful if you had a power vs frequency chart for your setup, and some more info on the general application. In the absence of that we are left to wonder whether there isn't something lacking in the completion of the board. The pic below shows gain and phase plots with two different coupling caps.
It would be helpful if you had a power vs frequency chart for your setup, and some more info on the general application. In the absence of that we are left to wonder whether there isn't something lacking in the completion of the board. The pic below shows gain and phase plots with two different coupling caps.
An externally hosted image should be here but it was not working when we last tested it.
Hi,
the two input filters are at 0.6Hz (260mS) and 3.4Hz (47mS).
I can't see where 5.4Hz comes from.
Could the phase difference between 260mS & 47mS be part of the problem?
I suppose it would be fairly easy to set both filters to about 100mS and listen. Change C4 from 1uF to 4.7uF and the combination of C5//Cb2&3 from 54.7uF to 22uF.
Janneman,
I don't agree.
the two input filters are at 0.6Hz (260mS) and 3.4Hz (47mS).
I can't see where 5.4Hz comes from.
Could the phase difference between 260mS & 47mS be part of the problem?
I suppose it would be fairly easy to set both filters to about 100mS and listen. Change C4 from 1uF to 4.7uF and the combination of C5//Cb2&3 from 54.7uF to 22uF.
Janneman,
I don't agree.
AndrewT said:[snip]
Janneman,
I don't agree.
Hey Andrew,
That's OK. It would be boring indeed if everyone did!
Jan Didden
Andrew,
Maybe some additional reasoning:
Suppose we have an amp with +/- 40V supplies, with 50.000uF capacitance. Suppose this amp is delivering 50W into 8 ohms at, say 20Hz. That's 28V peak output signal.
Now change the supply caps to 10.000uF. What will happen to the output signal level? Hint: nothing! Therefore, IF this amp gives less bass, the problem is freq response, also hinted at by Jack above.
Jan Didden
Maybe some additional reasoning:
Suppose we have an amp with +/- 40V supplies, with 50.000uF capacitance. Suppose this amp is delivering 50W into 8 ohms at, say 20Hz. That's 28V peak output signal.
Now change the supply caps to 10.000uF. What will happen to the output signal level? Hint: nothing! Therefore, IF this amp gives less bass, the problem is freq response, also hinted at by Jack above.
Jan Didden
It seems that everyone or just about everyone except for me feels that adding mega capacitors in the power supply is going to up the bass. Sorry, but I feel a need to cry BS on this one. Unless someone proves me completely nutty I am going to assume that 20,000 mfd should be more than enough to run the pair of bridge amps. Hell, a 9440 Altec amp doesn't have over 20,000 total capacitance in the power supply to deliver over 800 watts in mono bridge mode and it will take your head off with the low end capability.
Jack,
I will look over the board again. I have already used (2) 47uf caps in the ca2,ca3 position bypassed by ca1. I have changed this to (1) 47uf poly cap bypassed by a .1
I will yank these out and try a single 4.7 uf cap. I will yank out my 1.0uf cap in the C4 position and measure it.
I have a signal generator that is capable of frequency changes as small as .1 HZ. I will try some caps that I have measured and see if the response changes. After thinking about this all day I am wondering about the quality of my 1.0uf cap.
Jack,
I will look over the board again. I have already used (2) 47uf caps in the ca2,ca3 position bypassed by ca1. I have changed this to (1) 47uf poly cap bypassed by a .1
I will yank these out and try a single 4.7 uf cap. I will yank out my 1.0uf cap in the C4 position and measure it.
I have a signal generator that is capable of frequency changes as small as .1 HZ. I will try some caps that I have measured and see if the response changes. After thinking about this all day I am wondering about the quality of my 1.0uf cap.
feel free to send it back -- i just looked at 3 bridged boards which I had assembled and they conform pretty much to what I showed --- you can get wonky response if the power supply isn't putting out much juice -- and you can get the chip to oscillate if the heat sink isn't adequate (it's a long story...)
jackinnj said:...and you can get the chip to oscillate if the heat sink isn't adequate (it's a long story...)
Yup, been there and smelled the smoke...
5 Hz to 500 Hz -- this is at very low level with an attenuator in the path:
An externally hosted image should be here but it was not working when we last tested it.
As far as I understand..the main purpose of having a mega capacitor on the supply rails is to prevent the rails from sagging when you need a sudden burst of current from the PSU...as long as they serve that purpose ..doesn't make any difference how big a capacitor u have on the rails...at least for me it didn't..
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
- Home
- Amplifiers
- Chip Amps
- Gainclone LM4780 (poor bass ?)