Check out "another Lm1875 question" thread. I have found that using two sets of parallel lm1875 in bridge mode can give half the distortion and much greater power than when compared with the other solutions mentioned here.It even has much less distortion than a one lm1875 amp. Unfortunately I havent been able to post the diagram but, I will e-mail it to any one interested.
Jez.
Jez.
That will give you a +-33 V supply. Maybe a little higher or lower depending on the line voltage (which varies) and other things.
Two 3886's would be perfect. With two, you'll easily get 120 W power output.
As for 1875's, the supply voltage is too high. In the "Absolute Maximum Ratings" section of the data sheet, the supply voltage is spec'd at 60 V max. Remember, this in the sum of the two power supplies, which is 66 V in your case. They may tolerate the extra voltage if you reduce the load on each chip by paralleling say, four of them. Then you should be able to get quite a high power output (no more than the 3886's though). That is, if the chips survive long enough.
Ah you want to use them with a 4 ohm speaker.
In my experience a 3875 with supply voltages above 28V DC (after rectification, under load) gets pretty hot and "spike-s" way too fast driving 4 ohm speakers.. I think the 3886's behaviour is similar in this matter. Lowering the supply voltage prevents this. My 3875 is curently running on 24 VDC (agai, after rectification, under load; 18v secondaries on the toroid).
In my experience a 3875 with supply voltages above 28V DC (after rectification, under load) gets pretty hot and "spike-s" way too fast driving 4 ohm speakers.. I think the 3886's behaviour is similar in this matter. Lowering the supply voltage prevents this. My 3875 is curently running on 24 VDC (agai, after rectification, under load; 18v secondaries on the toroid).
Actually, bridging quadruples the power not doubles it as some in this thread have stated. The 4x power is the theoretical increase from ohms law and can only be approached if the amps have very low output impedance (true voltage source),can supply the necessary current without blowing up and have an equally low source resistance (i.e "stiff") power supply, but I have seen amps manage about 3.8x the power in bridge mode.
With "gainclone" type amps the problem is being able to supply the necessary current output to back up the voltage without the protection circuitry (SPike etc) kicking in.
It is because of the above that the best way to achieve really high power from chip amps is to use two sets of paralleld amps in bridge mode, this can also halve the distortion if done correctly,
however, it can be an absolute £$%^&* to get it all working right and without oscillating, be VERY carefull with grounding and power wiring regimes or it won't work.
Jez.
With "gainclone" type amps the problem is being able to supply the necessary current output to back up the voltage without the protection circuitry (SPike etc) kicking in.
It is because of the above that the best way to achieve really high power from chip amps is to use two sets of paralleld amps in bridge mode, this can also halve the distortion if done correctly,
however, it can be an absolute £$%^&* to get it all working right and without oscillating, be VERY carefull with grounding and power wiring regimes or it won't work.
Jez.
Another question about power
Hello all,
I'm planning to build a bridged parallel/bridged amplifier.
I have the following questions regarding implementation/
I read the instructions and am not sure. When I read the
National description of BPA200 it talks about it being able
to drive a 4 ohm load at 330 W. When you use the spreadsheet
and enter a 4 ohm load at 37V the spreadsheet warns about the
current limit.
Should I lower the voltage in order to be able to drive a 4 ohm load?
(I would like to keep the power as high as possible). My heatsink
is 14" x 5.25" x 3" with vertical fins (Per channel). I calculated
that thatwill be enough. I think I'll be introuble with the
insolated version (Any thoughts). Does anybody have a source
for mica insulators? (On digikey's website it's confusing).
If anybody has a part number that would be great.
The other issue is I would like to use a 2 x25V 625 VA toroidal
transformer for this configuration for both channels (Is in stock).
Is this heavy enough or cshould I go with a custom made
heavier solution. There's not much space in the cabinet to
put 2 transformers in.
Thanks all
Harry Andree
Hello all,
I'm planning to build a bridged parallel/bridged amplifier.
I have the following questions regarding implementation/
I read the instructions and am not sure. When I read the
National description of BPA200 it talks about it being able
to drive a 4 ohm load at 330 W. When you use the spreadsheet
and enter a 4 ohm load at 37V the spreadsheet warns about the
current limit.
Should I lower the voltage in order to be able to drive a 4 ohm load?
(I would like to keep the power as high as possible). My heatsink
is 14" x 5.25" x 3" with vertical fins (Per channel). I calculated
that thatwill be enough. I think I'll be introuble with the
insolated version (Any thoughts). Does anybody have a source
for mica insulators? (On digikey's website it's confusing).
If anybody has a part number that would be great.
The other issue is I would like to use a 2 x25V 625 VA toroidal
transformer for this configuration for both channels (Is in stock).
Is this heavy enough or cshould I go with a custom made
heavier solution. There's not much space in the cabinet to
put 2 transformers in.
Thanks all
Harry Andree
Re: Another question about power
Harry, may I offer you my 2 cents, being completely amateur.
I am running a 2 channel (LM3875) from one 200 VA Abbott trannie.
25 V, after rect. 37 V and 8 Amp.
The speakers I am using to burn them in are Pioneer CS-R 590 Gold series (old, 8 Ohm) 3 way 100W and the amp is mighty happy with them.
I put my LM's TF v. in one design, in one enclosure 5 by 8 by 2.5"
, (the enclosure body is about 0.5 mm thick) attached them to the aluminum enclosure, It runs slightly warm but not hot at all. It has been continueously playing for about 2 days now.
Personally I think it is a bit overkill for the gc's but I am not the pope.......
J-P
h_andree said:
Harry, may I offer you my 2 cents, being completely amateur.
I am running a 2 channel (LM3875) from one 200 VA Abbott trannie.
25 V, after rect. 37 V and 8 Amp.
The speakers I am using to burn them in are Pioneer CS-R 590 Gold series (old, 8 Ohm) 3 way 100W and the amp is mighty happy with them.
I put my LM's TF v. in one design, in one enclosure 5 by 8 by 2.5"
, (the enclosure body is about 0.5 mm thick) attached them to the aluminum enclosure, It runs slightly warm but not hot at all. It has been continueously playing for about 2 days now.
Personally I think it is a bit overkill for the gc's but I am not the pope.......
J-P
How about the load
I read you.
According to the Overture Design Guide the
temp coefficient for the T package is
2.6 C/W and 1.6 for the TF packages (At 37 V).
That's still not bad at 8 ohms. When you change the load
to 4 ohms it lists: 0.8 and -0.2 C/W respectively.
So the thermal requirements for a 4 ohm load seem
to be much worse.
My speakers currently are 4 ohms. I've decided
to go with a 4 chips parallelled design (8 chips
total per side). This should run nice and cool.
It should also give me quite a bit of output current
so that I never run into those limits.
Harry
I read you.
According to the Overture Design Guide the
temp coefficient for the T package is
2.6 C/W and 1.6 for the TF packages (At 37 V).
That's still not bad at 8 ohms. When you change the load
to 4 ohms it lists: 0.8 and -0.2 C/W respectively.
So the thermal requirements for a 4 ohm load seem
to be much worse.
My speakers currently are 4 ohms. I've decided
to go with a 4 chips parallelled design (8 chips
total per side). This should run nice and cool.
It should also give me quite a bit of output current
so that I never run into those limits.
Harry
My parallel expericence
Hi all
i have just finished my LM3886 Amp it has 4 single chip amps (front and surrounds) and a parallel amp for a 4ohm sub woofer.
the problem is the parellel amp just keeps on heating up with no input
I have built the parallel amps exactly the same as the other four which is the linkwitz design. i connected the inputs of the parallel amps together before the 1k input resistor and the outputs together after 0.1R. similar to Matttcattt's picture on page one.
now i have disconnected the outputs,and the amps are fine. they play music perfect.
there is 1.4mV DC diference between the outputs of each 0.1R.
the DC supply is +/-37V and 14100uF on each rail.
the 0.1R resistors are 5W and 5% all others are matched by me to 0.1%
please help
thanks,
Con
Hi all
i have just finished my LM3886 Amp it has 4 single chip amps (front and surrounds) and a parallel amp for a 4ohm sub woofer.
the problem is the parellel amp just keeps on heating up with no input
I have built the parallel amps exactly the same as the other four which is the linkwitz design. i connected the inputs of the parallel amps together before the 1k input resistor and the outputs together after 0.1R. similar to Matttcattt's picture on page one.
now i have disconnected the outputs,and the amps are fine. they play music perfect.
there is 1.4mV DC diference between the outputs of each 0.1R.
the DC supply is +/-37V and 14100uF on each rail.
the 0.1R resistors are 5W and 5% all others are matched by me to 0.1%
please help
thanks,
Con
Ok. There are many possible causes.
where to start?
When the amp is idle (no music) but powered on:
Measure the DC voltage across the output terminals. There should be very little voltage (100 mV or less). Any DC voltage here causes quiescent power dissipation (heating when idle).
Now measure the AC voltage or if you have a 'scope, use it. This should be 0 or flat. If it isn't, then the amp is probably oscillating. Note that even if you measure zero, the amp could still be oscillating (at to high a frequency to measure). Do the same measurement across each resistor. If using a scope, be careful that you don't short anything to ground!
Disconnect the speaker. Measure the DC voltage across each of the 0.1 Ohm output resistors. There should be very little voltage. Remember I=V/R, so even a few mV here means quite a bit of current. All current present here will be flowing back into the other amp, since there is no speaker. This is the most common source of idle power dissipation in parallel amps. One of the only cures is to use a DC servo on each amp to eliminate the DC offset. Since each amp will likely have a different output offset voltage, this voltage difference will cause current to flow from one or more of the amps into the other even when idle.
(e.g.) If you measure 30 mV across one of the R's: you have 0.030V/0.1R = 0.3 A = 300 mA of current. If your power supply is +-37 V, then just that one amp is dissipating (37V * 0.3A) = 11 W of power when idle, which is a lot. And that's from just a 30 mV difference in the offset voltage of one amp to the other(s). Many chip amps will have much more.
Is this the measurement where you measured 1.4 mV? In that case, each amp is dissipating only 0.5 W of heat as a result, which isn't much.
If you think that one or more of the amps may be oscillating, add the zobel network (small cap and resistor) to the output of each of the amps, before the 0.1 resistor. The values for these components are not critical: a 0.1 uF cap and 10 ohm R work well. You can also try adding a 50 to 100 pF cap in parallel with the feedback resistors. This reduces gain at extremely high frequencies (>>100 kHz) and can help eliminate oscillations.
In most cases, you won't know that you cured an oscillating amp except that you'll notice that it runs a whole lot cooler after making changes!
macboy thankyou so much for your post
i have it working now without heating up.
0.8mV difference in the offset voltage of one amp to the other.
AC was 0 with a DMM. i have no scope.
i couldnt measure any voltage accross the 0.1R resistors with my very cheap DMM.
i currently have no zobel on the output but will add one if it needs it .
i already had a 100 pF cap in parallel with the feedback resistor.
thanks again macboy,
Con
i have it working now without heating up.
0.8mV difference in the offset voltage of one amp to the other.
AC was 0 with a DMM. i have no scope.
i couldnt measure any voltage accross the 0.1R resistors with my very cheap DMM.
i currently have no zobel on the output but will add one if it needs it .
i already had a 100 pF cap in parallel with the feedback resistor.
thanks again macboy,
Con
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.